Suppress transients on old burglar alarm

[Terry Pinnell]

My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK

 

[Phil Hobbs]

On 06/04/2018 05:55 AM, Terry Pinnell wrote:

My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK

I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
https://hobbs-eo.com

 

[Terry Pinnell]

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK

 

[Phil Hobbs]

On 06/04/2018 12:05 PM, Terry Pinnell wrote:

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.

If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Cheers

Phil Hobbs







--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

 

[Terry Pinnell]

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK

Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?

If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Terry, East Grinstead, UK

 

[Phil Hobbs]

On 06/05/2018 01:52 AM, Terry Pinnell wrote:

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.

I was looking at the schematic you linked. It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?

I hope not!

If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

 

[Terry Pinnell]

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/2018 01:52 AM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.


I was looking at the schematic you linked.

But that was not the Autona schematic!

From my original post:

"Anyone have or know a source of the circuit diagram for the Autona 2250
Control Unit?"

"I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

You're undoubtedly right about the poor design (as I said, I'm a
hobbyist) but I certainly knew how it was to be used!

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?


I hope not!


If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?

The cause was down to my oversight. I'd forgotten that one of my nine
switches was a PIR motion detector in the lounge. It needs a 12V supply
and I was testing with all power removed. With power applied, the entire
loop is now closed, so that side-issue is resolved.

In the absence of any external filtering suggestions (for the AUTONA
circuit!) I dismantled the unit to access the PCB. I removed all the
caps and checked their capacitance. All were roughly OK so I replaced
them. In the morning I'll reassemble and test, but with no optimism.

In your initial reply you said "I think you'd be much better off
re-capping it," and I'm now wondering if I've interpreted that
correctly. Should I have literally replaced all of them regardless,
despite the fact that all showed correct values?

Terry, East Grinstead, UK

 

[Phil Hobbs]

On 06/05/18 12:40, Terry Pinnell wrote:

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/2018 01:52 AM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.


I was looking at the schematic you linked.

But that was not the Autona schematic!

From my original post:

"Anyone have or know a source of the circuit diagram for the Autona 2250
Control Unit?"

"I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

You're undoubtedly right about the poor design (as I said, I'm a
hobbyist) but I certainly knew how it was to be used!

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?


I hope not!


If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?


The cause was down to my oversight. I'd forgotten that one of my nine
switches was a PIR motion detector in the lounge. It needs a 12V supply
and I was testing with all power removed. With power applied, the entire
loop is now closed, so that side-issue is resolved.

In the absence of any external filtering suggestions (for the AUTONA
circuit!) I dismantled the unit to access the PCB. I removed all the
caps and checked their capacitance. All were roughly OK so I replaced
them. In the morning I'll reassemble and test, but with no optimism.

In your initial reply you said "I think you'd be much better off
re-capping it," and I'm now wondering if I've interpreted that
correctly. Should I have literally replaced all of them regardless,
despite the fact that all showed correct values?

Terry, East Grinstead, UK

In the modern era (say since 1975) it's really only wet electrolytics
that have the problem. It's not the capacitance that's usually the
issue, but the effective series resistance (ESR). As caps dry out, the
ESR goes up, so an ESR tester is the thing.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Terry Pinnell]

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/18 12:40, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/2018 01:52 AM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.


I was looking at the schematic you linked.

But that was not the Autona schematic!

From my original post:

"Anyone have or know a source of the circuit diagram for the Autona 2250
Control Unit?"

"I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

You're undoubtedly right about the poor design (as I said, I'm a
hobbyist) but I certainly knew how it was to be used!

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?


I hope not!


If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?


The cause was down to my oversight. I'd forgotten that one of my nine
switches was a PIR motion detector in the lounge. It needs a 12V supply
and I was testing with all power removed. With power applied, the entire
loop is now closed, so that side-issue is resolved.

In the absence of any external filtering suggestions (for the AUTONA
circuit!) I dismantled the unit to access the PCB. I removed all the
caps and checked their capacitance. All were roughly OK so I replaced
them. In the morning I'll reassemble and test, but with no optimism.

In your initial reply you said "I think you'd be much better off
re-capping it," and I'm now wondering if I've interpreted that
correctly. Should I have literally replaced all of them regardless,
despite the fact that all showed correct values?

Terry, East Grinstead, UK

In the modern era (say since 1975) it's really only wet electrolytics
that have the problem. It's not the capacitance that's usually the
issue, but the effective series resistance (ESR). As caps dry out, the
ESR goes up, so an ESR tester is the thing.

Cheers

Phil Hobbs

Thanks, understood.

I reckon it was a waste of effort removing and replacing five caps
yesterday. I'd missed Cursitor Doom's reply to stevwolf18 in the
'Capacitor Tester' thread: "If whatever meter you have won't measure
ESR, you're wasting your time." And also not spotted that his tester is
an Atlas ESR https://www.peakelec.co.uk/acatalog/jz_esr70.html
not an Atlas LCR like mine. I'm reluctant to spend on an ESR tester.
Seems unlikely I'd use it a lot.

Is there any other way I can test whether one of the five caps (1uF, 3 x
47uF, 2200uF) is the cause of the spurious alarm activation? BTW, I see
no 'bulging' in any of them.

Returning to Q3 of my original three questions, from the little you know
about the Autona unit can you suggest any practical advice like:

"Try a filter like this...across the mains."
"Try a filter like this ... across the transformer secondary."
"Try an RC filter like this ... across the 12V DC supply, or like this
in series with it..."
"Try X ohms in series with input to connector #13 and Y uf to any 0V
connector". (#13/#14 are the 'loop' connections, as shown in my original
illustration.)

IOW, anything practical I can try.

Terry, East Grinstead, UK

 

[jfeng@my-deja.com]

On Wednesday, June 6, 2018 at 9:38:26 AM UTC-7, Terry Pinnell wrote:

Is there any other way I can test whether one of the five caps (1uF, 3 x
47uF, 2200uF) is the cause of the spurious alarm activation? BTW, I see
no 'bulging' in any of them.
Here in Silicon Valley, the total price for all of those would be less than US$2 for 50 WVDC at a retail hobbyist store (and even less for lower voltages). If you do not already have them in your junque box, and because this is a one-time event, I would just replace them all. Even at hobbyist rates, your time is probably worth more than that.

 

[Phil Hobbs]

On 06/06/2018 12:38 PM, Terry Pinnell wrote:

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/18 12:40, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/2018 01:52 AM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.


I was looking at the schematic you linked.

But that was not the Autona schematic!

From my original post:

"Anyone have or know a source of the circuit diagram for the Autona 2250
Control Unit?"

"I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

You're undoubtedly right about the poor design (as I said, I'm a
hobbyist) but I certainly knew how it was to be used!

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?


I hope not!


If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?


The cause was down to my oversight. I'd forgotten that one of my nine
switches was a PIR motion detector in the lounge. It needs a 12V supply
and I was testing with all power removed. With power applied, the entire
loop is now closed, so that side-issue is resolved.

In the absence of any external filtering suggestions (for the AUTONA
circuit!) I dismantled the unit to access the PCB. I removed all the
caps and checked their capacitance. All were roughly OK so I replaced
them. In the morning I'll reassemble and test, but with no optimism.

In your initial reply you said "I think you'd be much better off
re-capping it," and I'm now wondering if I've interpreted that
correctly. Should I have literally replaced all of them regardless,
despite the fact that all showed correct values?

Terry, East Grinstead, UK

In the modern era (say since 1975) it's really only wet electrolytics
that have the problem. It's not the capacitance that's usually the
issue, but the effective series resistance (ESR). As caps dry out, the
ESR goes up, so an ESR tester is the thing.

Cheers

Phil Hobbs

Thanks, understood.

I reckon it was a waste of effort removing and replacing five caps
yesterday. I'd missed Cursitor Doom's reply to stevwolf18 in the
'Capacitor Tester' thread: "If whatever meter you have won't measure
ESR, you're wasting your time." And also not spotted that his tester is
an Atlas ESR https://www.peakelec.co.uk/acatalog/jz_esr70.html
not an Atlas LCR like mine. I'm reluctant to spend on an ESR tester.
Seems unlikely I'd use it a lot.

Is there any other way I can test whether one of the five caps (1uF, 3 x
47uF, 2200uF) is the cause of the spurious alarm activation? BTW, I see
no 'bulging' in any of them.

Returning to Q3 of my original three questions, from the little you know
about the Autona unit can you suggest any practical advice like:

"Try a filter like this...across the mains."
"Try a filter like this ... across the transformer secondary."
"Try an RC filter like this ... across the 12V DC supply, or like this
in series with it..."
"Try X ohms in series with input to connector #13 and Y uf to any 0V
connector". (#13/#14 are the 'loop' connections, as shown in my original
illustration.)

IOW, anything practical I can try.

Terry, East Grinstead, UK

LoopA 0-----RRRR-----*---------> Terminal block 13 or 15 or 17
510 |
CCC
CCC 100 nF
|
LoopB 0--------------*---------> Terminal block 14 or 16 or 18

The 510-ohm resistor will let the cap supply 24 mA briefly, which is
probably enough to wet the contacts but certainly not enough to burn
them. The peak power dissipation will be about 1-1/4 watts, but it's
only for 50 microseconds, so a half-watt resistor should be fine.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

 

[Terry Pinnell]

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/06/2018 12:38 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/18 12:40, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/2018 01:52 AM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.


I was looking at the schematic you linked.

But that was not the Autona schematic!

From my original post:

"Anyone have or know a source of the circuit diagram for the Autona 2250
Control Unit?"

"I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

You're undoubtedly right about the poor design (as I said, I'm a
hobbyist) but I certainly knew how it was to be used!

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?


I hope not!


If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?


The cause was down to my oversight. I'd forgotten that one of my nine
switches was a PIR motion detector in the lounge. It needs a 12V supply
and I was testing with all power removed. With power applied, the entire
loop is now closed, so that side-issue is resolved.

In the absence of any external filtering suggestions (for the AUTONA
circuit!) I dismantled the unit to access the PCB. I removed all the
caps and checked their capacitance. All were roughly OK so I replaced
them. In the morning I'll reassemble and test, but with no optimism.

In your initial reply you said "I think you'd be much better off
re-capping it," and I'm now wondering if I've interpreted that
correctly. Should I have literally replaced all of them regardless,
despite the fact that all showed correct values?

Terry, East Grinstead, UK

In the modern era (say since 1975) it's really only wet electrolytics
that have the problem. It's not the capacitance that's usually the
issue, but the effective series resistance (ESR). As caps dry out, the
ESR goes up, so an ESR tester is the thing.

Cheers

Phil Hobbs

Thanks, understood.

I reckon it was a waste of effort removing and replacing five caps
yesterday. I'd missed Cursitor Doom's reply to stevwolf18 in the
'Capacitor Tester' thread: "If whatever meter you have won't measure
ESR, you're wasting your time." And also not spotted that his tester is
an Atlas ESR https://www.peakelec.co.uk/acatalog/jz_esr70.html
not an Atlas LCR like mine. I'm reluctant to spend on an ESR tester.
Seems unlikely I'd use it a lot.

Is there any other way I can test whether one of the five caps (1uF, 3 x
47uF, 2200uF) is the cause of the spurious alarm activation? BTW, I see
no 'bulging' in any of them.

Returning to Q3 of my original three questions, from the little you know
about the Autona unit can you suggest any practical advice like:

"Try a filter like this...across the mains."
"Try a filter like this ... across the transformer secondary."
"Try an RC filter like this ... across the 12V DC supply, or like this
in series with it..."
"Try X ohms in series with input to connector #13 and Y uf to any 0V
connector". (#13/#14 are the 'loop' connections, as shown in my original
illustration.)

IOW, anything practical I can try.

Terry, East Grinstead, UK


LoopA 0-----RRRR-----*---------> Terminal block 13 or 15 or 17
510 |
CCC
CCC 100 nF
|
LoopB 0--------------*---------> Terminal block 14 or 16 or 18

The 510-ohm resistor will let the cap supply 24 mA briefly, which is
probably enough to wet the contacts but certainly not enough to burn
them. The peak power dissipation will be about 1-1/4 watts, but it's
only for 50 microseconds, so a half-watt resistor should be fine.

Cheers

Phil Hobbs

Many thanks, hope to try that later today.

Terry, East Grinstead, UK

 

[Terry Pinnell]

Terry Pinnell <me@somewhere.invalid> wrote:

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/06/2018 12:38 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/18 12:40, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/2018 01:52 AM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.


I was looking at the schematic you linked.

But that was not the Autona schematic!

From my original post:

"Anyone have or know a source of the circuit diagram for the Autona 2250
Control Unit?"

"I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

You're undoubtedly right about the poor design (as I said, I'm a
hobbyist) but I certainly knew how it was to be used!

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?


I hope not!


If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?


The cause was down to my oversight. I'd forgotten that one of my nine
switches was a PIR motion detector in the lounge. It needs a 12V supply
and I was testing with all power removed. With power applied, the entire
loop is now closed, so that side-issue is resolved.

In the absence of any external filtering suggestions (for the AUTONA
circuit!) I dismantled the unit to access the PCB. I removed all the
caps and checked their capacitance. All were roughly OK so I replaced
them. In the morning I'll reassemble and test, but with no optimism.

In your initial reply you said "I think you'd be much better off
re-capping it," and I'm now wondering if I've interpreted that
correctly. Should I have literally replaced all of them regardless,
despite the fact that all showed correct values?

Terry, East Grinstead, UK

In the modern era (say since 1975) it's really only wet electrolytics
that have the problem. It's not the capacitance that's usually the
issue, but the effective series resistance (ESR). As caps dry out, the
ESR goes up, so an ESR tester is the thing.

Cheers

Phil Hobbs

Thanks, understood.

I reckon it was a waste of effort removing and replacing five caps
yesterday. I'd missed Cursitor Doom's reply to stevwolf18 in the
'Capacitor Tester' thread: "If whatever meter you have won't measure
ESR, you're wasting your time." And also not spotted that his tester is
an Atlas ESR https://www.peakelec.co.uk/acatalog/jz_esr70.html
not an Atlas LCR like mine. I'm reluctant to spend on an ESR tester.
Seems unlikely I'd use it a lot.

Is there any other way I can test whether one of the five caps (1uF, 3 x
47uF, 2200uF) is the cause of the spurious alarm activation? BTW, I see
no 'bulging' in any of them.

Returning to Q3 of my original three questions, from the little you know
about the Autona unit can you suggest any practical advice like:

"Try a filter like this...across the mains."
"Try a filter like this ... across the transformer secondary."
"Try an RC filter like this ... across the 12V DC supply, or like this
in series with it..."
"Try X ohms in series with input to connector #13 and Y uf to any 0V
connector". (#13/#14 are the 'loop' connections, as shown in my original
illustration.)

IOW, anything practical I can try.

Terry, East Grinstead, UK


LoopA 0-----RRRR-----*---------> Terminal block 13 or 15 or 17
510 |
CCC
CCC 100 nF
|
LoopB 0--------------*---------> Terminal block 14 or 16 or 18

The 510-ohm resistor will let the cap supply 24 mA briefly, which is
probably enough to wet the contacts but certainly not enough to burn
them. The peak power dissipation will be about 1-1/4 watts, but it's
only for 50 microseconds, so a half-watt resistor should be fine.

Cheers

Phil Hobbs

Many thanks, hope to try that later today.

Terry, East Grinstead, UK

Before trying your filter I re-assembled in the slight hope that my
removal/replacement (and touching up a couple of joint that looked
mildly suspicious) might have fixed the spurious triggering. I tried a
few dozen times, switching the cupboard lamp on/off as described
earlier, and so far so good!

Left the temporary sounder in place to try over a few successive nights
and will report back on final outcome.

Thanks for your patient help, which is much appreciated.

Terry, East Grinstead, UK

 

[Phil Hobbs]

On 06/07/2018 07:03 AM, Terry Pinnell wrote:

Terry Pinnell <me@somewhere.invalid> wrote:

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/06/2018 12:38 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/18 12:40, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/05/2018 01:52 AM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 12:05 PM, Terry Pinnell wrote:
Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

On 06/04/2018 05:55 AM, Terry Pinnell wrote:
My 20 year old burglar alarm control has developed a fault that I've
spent the last couple of days trying to fix. Before I ditch it and go
through the hassle of installing a new one I'm seeking help here please.

Q1: An extreme longshot as I researched in 2012 but the company had
long folded, but... Anyone have or know a source of the circuit diagram
for the Autona 2250 Control Unit?

FWIW, screenshot inside the control unit's case
https://www.dropbox.com/s/4xd19m9cvdqfpr9/Autona2250-1.jpg?dl=0

Q2: A useful second best would be *any* circuit of a bog-standard,
commercial, wired unit designed for a series of normally closed
microswitches.

I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It's being triggered spuriously. Two successive nights, 01:05 and 12:35.
During late stages of trouble-shooting (i.e. after checking all
switches) I found that I could trigger it by switching on light of the
cupboard in which unit is mounted, powered from same 240V source.

I suspect a failed capacitor. But I don't relish dismantling it to get
at the two PCBs and even if I did I doubt I'd find it.

Q3: Using the access I have to the power supplies (240 V AC, 15 V AC,
13V DC) and the pairs of wires from the string of N/C microswitches,
what transient suppression should I try adding at one or more of these
external points please, and in what priority?


Terry, East Grinstead, UK


I think you'd be much better off re-capping it, but you could try a
Tripp-lite Isobar. They're my standard power strip, and the 8-outlet
ones have four filter stages, so that each pair of outlets is isolated
from the others as well as from the line.

If the problem is high-frequency crud on the line, that will probably
help, but for low frequency stuff you'd need a ferroresonant transformer.

On the other hand, it might easily be getting in on the sensing loop, in
which case a 10 nF cap right at the terminals might be the ticket. Make
sure to keep the leads super short, like zero millimetres short.

Cheers

Phil Hobbs

Thanks Phil. The Tripp-lite Isobar would be impractical in my case. Not
only is the mains supply to the alarm unit wired directly from the wall
light switch but the under stairs cupboard is jam-packed with stuff.

I'll try the small caps idea. As I've said I don't intend dismantling,
it will have to be connected at the external connections. Without the
circuit diagram I'm handicapped. but my loop of N/C switches are wired
to #13 and #14 in my diagram. So are you suggesting a single cap to
those? Obviously the lead will then be at least 5 mm in that case.

Would a simple filter on one or more of the three power supplies I
described also be helpful?

Terry, East Grinstead, UK


Thanks, much appreciate your sticking with me. However I'm finding much
of your reply rather hard going. I'm just a shed hobbyist, and a lapsed
one at that. (I finished my last project about 20 years ago. I'm into
video-making now) That's mainly why I posted here rather than in S.E.D,
my usual group in the past!

Which circuit are you referring to there and in the following please?

Using megohm pulldowns in an alarm circuit is mental. You can't wet the
switch contacts, and any water that gets in is liable to make the whole
thing useless by shorting out one or more of the switches.


I was looking at the schematic you linked.

But that was not the Autona schematic!

From my original post:

"Anyone have or know a source of the circuit diagram for the Autona 2250
Control Unit?"

"I made this one myself about 35 years ago and *suspect* that it will be
broadly similar, but I'd like to eliminate guesswork:
https://www.dropbox.com/s/cxl6zarvmqsnkem/Burglar%20Alarm-ex%20Leigh.gif?dl=0

It has the loop running from
+12 to a CMOS logic input, with gigantic pulldown resistors, 12M in
parallel with 4.7M = 3.7M in one case. The logic threshold can be
anywhere from about 1/3 to 2/3 of the supply rail, so if there's a
microamp of leakage in the loop with a switch open, the alarm may not
go off, and if there's 2 microamps, it definitely won't go off. That's
pretty poor design, given that the alarm maker has no idea how it's
going to be wired up or how the cable is going to be routed. Basements
get wet, for instance.

You're undoubtedly right about the poor design (as I said, I'm a
hobbyist) but I certainly knew how it was to be used!

What are your assumptions about the (to me) largely unknown Autona
circuit? As mentioned, I'm guessing that it's similar to my own home
made alarm. But you sound as if you're confident of that?


I hope not!


If the input structure is unprotected 12V CMOS with megohm pulldowns,
the 10-nf cap should help, and 100-nf might be better. You've got at
least 4V of noise immunity in that circuit, so it's unlikely to be RF,
in which case 5 mm leads won't make a difference.

With a biggish cap across the loop terminals, any capacitive or
inductive transient will pretty well go away. Another possibility is
that the line transient is making the +12V rail sag by a lot if it hits
at the wrong time, and when it comes back up the supply pin responds a
bit faster than the loop, leading to a transient LOW input. The cap
would actually make that worse, but it's a long shot--metal gate CMOS is
pretty slow, and the transient LOW would only last for 100 ns or so.

So the cap ought to fix it, I think. Don't use too big a capacitance,
because you don't want to damage the switch contacts. You might even
want to put 1k or so in series with the loop, with the cap right at the
terminals. That'll let the cap wet the contacts without burning them,
so the cap can be as big as you like.

Anyway, unfortunately I've had to suspend work on fixing the noise
problem. Annoyingly, although it checked out OK two days ago, I'm now
*not* getting close to zero resistance with my DMM on #13 and #14. So
there's a break somewhere in my 9-switch N/C loop. I wired this up two
decades ago and regrettably my documentation on wiring was hasty and not
done with maintenance in mind! So fixing that issue is now priority.

Any tips from anyone on how to accelerate that please?

Try working all the switches several times. What resistance are you
getting?


The cause was down to my oversight. I'd forgotten that one of my nine
switches was a PIR motion detector in the lounge. It needs a 12V supply
and I was testing with all power removed. With power applied, the entire
loop is now closed, so that side-issue is resolved.

In the absence of any external filtering suggestions (for the AUTONA
circuit!) I dismantled the unit to access the PCB. I removed all the
caps and checked their capacitance. All were roughly OK so I replaced
them. In the morning I'll reassemble and test, but with no optimism.

In your initial reply you said "I think you'd be much better off
re-capping it," and I'm now wondering if I've interpreted that
correctly. Should I have literally replaced all of them regardless,
despite the fact that all showed correct values?

Terry, East Grinstead, UK

In the modern era (say since 1975) it's really only wet electrolytics
that have the problem. It's not the capacitance that's usually the
issue, but the effective series resistance (ESR). As caps dry out, the
ESR goes up, so an ESR tester is the thing.

Cheers

Phil Hobbs

Thanks, understood.

I reckon it was a waste of effort removing and replacing five caps
yesterday. I'd missed Cursitor Doom's reply to stevwolf18 in the
'Capacitor Tester' thread: "If whatever meter you have won't measure
ESR, you're wasting your time." And also not spotted that his tester is
an Atlas ESR https://www.peakelec.co.uk/acatalog/jz_esr70.html
not an Atlas LCR like mine. I'm reluctant to spend on an ESR tester.
Seems unlikely I'd use it a lot.

Is there any other way I can test whether one of the five caps (1uF, 3 x
47uF, 2200uF) is the cause of the spurious alarm activation? BTW, I see
no 'bulging' in any of them.

Returning to Q3 of my original three questions, from the little you know
about the Autona unit can you suggest any practical advice like:

"Try a filter like this...across the mains."
"Try a filter like this ... across the transformer secondary."
"Try an RC filter like this ... across the 12V DC supply, or like this
in series with it..."
"Try X ohms in series with input to connector #13 and Y uf to any 0V
connector". (#13/#14 are the 'loop' connections, as shown in my original
illustration.)

IOW, anything practical I can try.

Terry, East Grinstead, UK


LoopA 0-----RRRR-----*---------> Terminal block 13 or 15 or 17
510 |
CCC
CCC 100 nF
|
LoopB 0--------------*---------> Terminal block 14 or 16 or 18

The 510-ohm resistor will let the cap supply 24 mA briefly, which is
probably enough to wet the contacts but certainly not enough to burn
them. The peak power dissipation will be about 1-1/4 watts, but it's
only for 50 microseconds, so a half-watt resistor should be fine.

Cheers

Phil Hobbs

Many thanks, hope to try that later today.

Terry, East Grinstead, UK

Before trying your filter I re-assembled in the slight hope that my
removal/replacement (and touching up a couple of joint that looked
mildly suspicious) might have fixed the spurious triggering. I tried a
few dozen times, switching the cupboard lamp on/off as described
earlier, and so far so good!

Left the temporary sounder in place to try over a few successive nights
and will report back on final outcome.

Thanks for your patient help, which is much appreciated.

Terry, East Grinstead, UK

No worries. Now move somewhere where there are fewer burglars.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

 

[Terry Pinnell]

Terry Pinnell <me@somewhere.invalid> wrote:

Before trying your filter I re-assembled in the slight hope that my
removal/replacement (and touching up a couple of joint that looked
mildly suspicious) might have fixed the spurious triggering. I tried a
few dozen times, switching the cupboard lamp on/off as described
earlier, and so far so good!

Left the temporary sounder in place to try over a few successive nights
and will report back on final outcome.


Thanks for your patient help, which is much appreciated.

***** That last status report was prematurely optimistic. Turned out to
be two faulty microswitches, both intermittent. One was a rusted (and
greatly weakened) magnet, the other a sticking shed window. And both at
the far end of the loop, so many hours before I found them.

Terry, East Grinstead, UK

 

[Phil Hobbs]

On 06/19/2018 02:24 AM, Terry Pinnell wrote:

Terry Pinnell <me@somewhere.invalid> wrote:

Before trying your filter I re-assembled in the slight hope that my
removal/replacement (and touching up a couple of joint that looked
mildly suspicious) might have fixed the spurious triggering. I tried a
few dozen times, switching the cupboard lamp on/off as described
earlier, and so far so good!

Left the temporary sounder in place to try over a few successive nights
and will report back on final outcome.


Thanks for your patient help, which is much appreciated.


***** That last status report was prematurely optimistic. Turned out to
be two faulty microswitches, both intermittent. One was a rusted (and
greatly weakened) magnet, the other a sticking shed window. And both at
the far end of the loop, so many hours before I found them.

Terry, East Grinstead, UK

It's always something.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net