AoE x-Chapters, High-Speed op-amps section, DRAFT

[Klaus Kragelund]

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:

On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.

An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.

At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps 😊

Cheers

Klaus

--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[John Larkin]

On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
<klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?




--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Lasse Langwadt Christense]

onsdag den 17. juli 2019 kl. 22.16.23 UTC+2 skrev Klaus Kragelund:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps 😊

imagine fixing it, here's a reel of caps and soldering iron, see you in a week

 

[Joerg]

On 2019-07-17 14:21, John Larkin wrote:

On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?

Earl Muntz is said to have walked around the R&D lab with a pair of wire
cutters in his pocket. Snip .. out went a cap. "See? Nothing happened!
You are wasting money". Other times .. snip .. phseeeooouu .. PHUT ...
"Well, you better put that one back in".

--
Regards, Joerg

http://www.analogconsultants.com/

 

[John Larkin]

On Wed, 17 Jul 2019 14:42:02 -0700, Joerg <news@analogconsultants.com>
wrote:

On 2019-07-17 14:21, John Larkin wrote:
On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?


Earl Muntz is said to have walked around the R&D lab with a pair of wire
cutters in his pocket. Snip .. out went a cap. "See? Nothing happened!
You are wasting money". Other times .. snip .. phseeeooouu .. PHUT ...
"Well, you better put that one back in".

The Muntz TVs only worked close to TV stations. They didn't have much
gain. Not enough tubes in the IF strip.

He snipped screen bypass caps, which didn't help the gain situation
either.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Lasse Langwadt Christense]

onsdag den 17. juli 2019 kl. 23.42.00 UTC+2 skrev Joerg:

On 2019-07-17 14:21, John Larkin wrote:
On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?


Earl Muntz is said to have walked around the R&D lab with a pair of wire
cutters in his pocket. Snip .. out went a cap. "See? Nothing happened!
You are wasting money". Other times .. snip .. phseeeooouu .. PHUT ...
"Well, you better put that one back in".

https://en.wikipedia.org/wiki/Muntzing

 

[Chris Jones]

On 17/07/2019 03:19, John Larkin wrote:

On Tue, 16 Jul 2019 10:08:25 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-16 08:12, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.

An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


But it always makes for a good scare story, just like a lot of the
theories about global w... ...... bit my tongue here

I've seen absurd papers about bypassing... authored by capacitor
makers. There is one so co-authored paper, with someone from Xilinx,
that recommends three caps for *each* power ball on an FPGA. Hundreds
of caps per chip.

Modern FPGAs and CPUs have on-chip capacitance for the fast stuff.
They don't need heroic bypassing.

Even on radio chips I used to put them, because off-chip capacitors are
not much use above 1GHz in a package with bondwires. Also, I had to put
some resistors of various values in series with some fraction of the
on-chip decoupling capacitors, because otherwise there would be ringing
on the on-chip supply voltage due to resonance with the bondwire inductance.

 

[whit3rd]

On Wednesday, July 17, 2019 at 2:21:12 PM UTC-7, John Larkin wrote:

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?

Depends. I had an early Goldstar (now LG) microwave that gained time when used;
found that they had a board of TTL for the clock, with no bypass capacitors
at all. Two little 0.01 uF ceramics, and it kept correct time for a decade.

 

On Wed, 17 Jul 2019 14:42:02 -0700, Joerg <news@analogconsultants.com>
wrote:

On 2019-07-17 14:21, John Larkin wrote:
On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?


Earl Muntz is said to have walked around the R&D lab with a pair of wire
cutters in his pocket. Snip .. out went a cap. "See? Nothing happened!
You are wasting money". Other times .. snip .. phseeeooouu .. PHUT ...
"Well, you better put that one back in".

The assumption, of course, is that every TV manufactured would be
identical to this one. ...and stay that way for its lifetime.

 

[Joerg]

On 2019-07-17 17:57, whit3rd wrote:

On Wednesday, July 17, 2019 at 2:21:12 PM UTC-7, John Larkin wrote:

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?

Depends. I had an early Goldstar (now LG) microwave that gained time when used;
found that they had a board of TTL for the clock, with no bypass capacitors
at all. Two little 0.01 uF ceramics, and it kept correct time for a decade.

There would have been a much cheaper solution. They could have declared
this a feature and sold it as having a time machine function. "Our
microwave catapults you into the future!".

--
Regards, Joerg

http://www.analogconsultants.com/

 

[Joerg]

On 2019-07-17 19:22, krw@notreal.com wrote:

On Wed, 17 Jul 2019 14:42:02 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-17 14:21, John Larkin wrote:
On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?


Earl Muntz is said to have walked around the R&D lab with a pair of wire
cutters in his pocket. Snip .. out went a cap. "See? Nothing happened!
You are wasting money". Other times .. snip .. phseeeooouu .. PHUT ...
"Well, you better put that one back in".

The assumption, of course, is that every TV manufactured would be
identical to this one. ...and stay that way for its lifetime.

Not for life, only until its warranty is up

--
Regards, Joerg

http://www.analogconsultants.com/

 

On Thu, 18 Jul 2019 06:25:21 -0700, Joerg <news@analogconsultants.com>
wrote:

On 2019-07-17 19:22, krw@notreal.com wrote:
On Wed, 17 Jul 2019 14:42:02 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-17 14:21, John Larkin wrote:
On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?


Earl Muntz is said to have walked around the R&D lab with a pair of wire
cutters in his pocket. Snip .. out went a cap. "See? Nothing happened!
You are wasting money". Other times .. snip .. phseeeooouu .. PHUT ...
"Well, you better put that one back in".

The assumption, of course, is that every TV manufactured would be
identical to this one. ...and stay that way for its lifetime.


Not for life, only until its warranty is up

Well, I did say *its* lifetime.

 

[Winfield Hill]

Joerg wrote...

Shoddy products usually still have a lot of life left
in them but the performance becomes so bad and the
potential repairs so expensive that people give them
up prematurely. Cars of certain brands are a classic
example.

Often a sign of shoddy engineering. We're taught that
designs should not ask for more out of a part than
can be delivered at end of life. For example, BJT beta.
Good designs should perform acceptably with low betas.


--
Thanks,
- Win

 

[Joerg]

On 2019-07-18 19:12, krw@notreal.com wrote:

On Thu, 18 Jul 2019 06:25:21 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-17 19:22, krw@notreal.com wrote:
On Wed, 17 Jul 2019 14:42:02 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-17 14:21, John Larkin wrote:
On Wed, 17 Jul 2019 13:16:18 -0700 (PDT), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Tuesday, July 16, 2019 at 5:13:01 PM UTC+2, John Larkin wrote:
On Tue, 16 Jul 2019 07:29:59 -0700, Joerg <news@analogconsultants.com
wrote:

On 2019-07-15 20:06, Clifford Heath wrote:
On 16/7/19 12:21 pm, John Larkin wrote:
My experience with opamps rectifying RF on PC boards is that there are
usually several narrow resonances where it's very sensitive, in the
low 100s of MHz.

One of my (ex) competitors NMR temperature controllers, with a
thermocouple sensor, could be hard shut down with a GR signal
generator, from clear across the room.



An RF design friend of mine has had issues with parallel capacitors
(like 0.1uF/100pF) adjacent on a supply line. He's had quite high-Q
resonance between the 100pF and the parasitic inductance at between 400
and 900MHz. Traps, and he's not a "young player".


IME such small bypass cap values make no sense at all. Definitely not in
the days of MLCC. 0.01uF to 0.1uf in 0603 is fine. It's the layout that
counts, plus the proper use of large planes as "free" bypass capacitors.
This is part of my daily bread, helping engineers route and place vias
properly. "But that's less than 1/10th of an inch!" ... "Yes, but it
does matter here".

I use 1 uF 0805 caps by default. Scattered around a power plane, close
to a ground plane, they work fine.

The oft-repeated "analysis" of paralleled, stepped-value bypass caps,
is nonsense. Lots of signal-integrity lore is nonsense.


At a earlier employment we had a large board with 1000+ decoupling caps and an error in the production meant none was mounted

The board had a good stack up , so it worked without all those caps ?

Cheers

Klaus

I met one guy who worked for Lockheed (I think) who did big TTL logic
boards with no bypass caps. All his stuff worked.

An FPGA does deserve enough bulk C so the supply doesn't droop much
when clocks start or stop or something, amps of Vcc-core step maybe.

I don't think I've ever made a board that had too few bypass caps. How
many million caps have I wasted?


Earl Muntz is said to have walked around the R&D lab with a pair of wire
cutters in his pocket. Snip .. out went a cap. "See? Nothing happened!
You are wasting money". Other times .. snip .. phseeeooouu .. PHUT ...
"Well, you better put that one back in".

The assumption, of course, is that every TV manufactured would be
identical to this one. ...and stay that way for its lifetime.


Not for life, only until its warranty is up

Well, I did say *its* lifetime.

Shoddy products usually still have a lot of life left in them but the
performance becomes so bad and the potential repairs so expensive that
people give them up prematurely. Cars of certain brands are a classic
example.

--
Regards, Joerg

http://www.analogconsultants.com/

 

[Jan Panteltje]

On a sunny day (19 Jul 2019 17:03:28 -0700) it happened Winfield Hill
<winfieldhill@yahoo.com> wrote in <qgtlog06uv@drn.newsguy.com>:

Joerg wrote...

Shoddy products usually still have a lot of life left
in them but the performance becomes so bad and the
potential repairs so expensive that people give them
up prematurely. Cars of certain brands are a classic
example.

Often a sign of shoddy engineering. We're taught that
designs should not ask for more out of a part than
can be delivered at end of life. For example, BJT beta.
Good designs should perform acceptably with low betas.

That is a bit over over-simplification.

In the tube days you would design keeping in mind 100% or so degeneration of the thing.

Semiconductors OTOH do not really degrade that much (unless some .. uses the be zener with kA ;-0 ).
so no problem relying in some circuits of a high beta.
Simple BC547 etc have an A, B, C suffix and a minimum beta specified for each.

If you are against using that, then why all the stories about the evolution of MOSFETs?

In MY opinion (open fire if you feel the need) if I need some part I use google for the specs I want.
I could not care less what's in those things as long as it meets those specs.
Same with processors and more complicated chips.

The story about IR is interesting though, deceasing capacitances over time too.
I use IRLZ34N a lot. for low voltage stuff.

 

[Joerg]

On 2019-07-20 00:29, Jan Panteltje wrote:

On a sunny day (19 Jul 2019 17:03:28 -0700) it happened Winfield Hill
winfieldhill@yahoo.com> wrote in <qgtlog06uv@drn.newsguy.com>:

Joerg wrote...

Shoddy products usually still have a lot of life left
in them but the performance becomes so bad and the
potential repairs so expensive that people give them
up prematurely. Cars of certain brands are a classic
example.

Often a sign of shoddy engineering. We're taught that
designs should not ask for more out of a part than
can be delivered at end of life. For example, BJT beta.
Good designs should perform acceptably with low betas.

It can have many causes. In the case of a car company I have inside info
that a main factor was quality control and here mainly incoming
inspection. Or to a large extent a lack thereof.

That is a bit over over-simplification.

In the tube days you would design keeping in mind 100% or so degeneration of the thing.

Not so much for the large tube in the HV and horizontal deflection
section. On color TVs those had to be replaced quite regularly. In
Europe that was the PL509 and later PL519.

Semiconductors OTOH do not really degrade that much (unless some .. uses the be zener with kA ;-0 ).
so no problem relying in some circuits of a high beta.

They do if treated harshly.

http://datasheet.octopart.com/PMGD290UCEA-NXP-Semiconductors-datasheet-20734454.pdf

Quote page 5 "If the transient temperature resulting from an avalanche
event, as illustrated in Figure 6, rises beyond a recommended rated
value, the device risks being degraded".

Simple BC547 etc have an A, B, C suffix and a minimum beta specified for each.

If you are against using that, then why all the stories about the evolution of MOSFETs?

In MY opinion (open fire if you feel the need) if I need some part I use google for the specs I want.
I could not care less what's in those things as long as it meets those specs.
Same with processors and more complicated chips.

Like the exorbitant wattage rating of some FETs? Good luck.

[...]

--
Regards, Joerg

http://www.analogconsultants.com/

 

[amdx]

On 7/16/2019 3:27 PM, Lasse Langwadt Christensen wrote:

tirsdag den 16. juli 2019 kl. 22.01.45 UTC+2 skrev Phil Hobbs:
On 7/16/19 11:26 AM, John Larkin wrote:
On Tue, 16 Jul 2019 14:18:41 GMT, Jan Panteltje
pNaOnStPeAlMtje@yahoo.com> wrote:

On a sunny day (Tue, 16 Jul 2019 06:56:47 -0700) it happened John Larkin
jjlarkin@highlandtechnology.com> wrote in
eqlrielfeeeovifsku3olimru2635svnld@4ax.com>:

On Tue, 16 Jul 2019 06:53:17 GMT, Jan Panteltje
pNaOnStPeAlMtje@yahoo.com> wrote:

On a sunny day (Mon, 15 Jul 2019 16:41:44 -0700) it happened John Larkin
jjlarkin@highland_snip_technology.com> wrote in
0e3qie9ivv3io0dcsf0uhihf61m4lbo028@4ax.com>:

When I used to work on Judah Street, I drove to work over a road on
Twin Peaks, just under the Sutro tower. My old ratty Fiesta had a
cheap add-on radio. I could hear buzzing from the speakers with the
radio turned off.

People live up there.

It is common, my Sony alarm clock FM radio also makes noise when my cellphone interacts with the tower,
even with that radio off, so do my PC speakers.



Here's the street. That's 22 megawatts right above those houses.

https://www.dropbox.com/s/4fzsmipz0rar4ke/Panoramic_Drive.jpg?raw=1

Mega Watt? not kW?

Megawatts.

What frequency is that?

Everything. On a spectrum analyzer, the AM, FM, and TV bands are about
full. And the tower is used for police and fire comms too. It
overlooks the entire city, and then some.


Just one nickel-plated connector could cause some havoc.


https://youtu.be/lMuJKsUjD_o


Sounds like some voice audio at :56 and 1:05.

Mikek

 

[whit3rd]

On Friday, July 19, 2019 at 5:03:39 PM UTC-7, Winfield Hill wrote:

... We're taught that
designs should not ask for more out of a part than
can be delivered at end of life. For example, BJT beta.
Good designs should perform acceptably with low betas.

A related concern is the longevity of LEDs for lighting; while selected
LEDs have high efficiency (140 lm/w), the run-of-the-mill is more like 90 lm/w
and will degrade with time. The better manufacturers that give lamp
specs (Philips) are promising 75 lm/w and years of service life.

In the old days (probably for chemistry reasons) the LED recommendation
was to expect 30% per year degradation; I've seen optointerrupters
become inoperable with just a few months of use, that were assembled
in circuits with factory-adjusted trimmers. That wasn't good engineering, IMHO.

 

[Winfield Hill]

whit3rd wrote...

The better manufacturers that give lamp specs (Philips)
are promising 75 lm/w and years of service life.

Yes, I bought into this line of thinking.
I purchased about six of the more-expensive,
higher-quality Philips LED bulbs. Four of
the six have failed, many after only a few
hundred hours, and none anywhere near their
claimed lifetime. They were installed in
dropped-down-from-the-ceiling fixtures,
with lots of room and generous airflow.
No autopsies were performed.


--
Thanks,
- Win

 

[gray_wolf]

On 7/21/2019 9:03 AM, Winfield Hill wrote:

whit3rd wrote...

The better manufacturers that give lamp specs (Philips)
are promising 75 lm/w and years of service life.

Yes, I bought into this line of thinking.
I purchased about six of the more-expensive,
higher-quality Philips LED bulbs. Four of
the six have failed, many after only a few
hundred hours, and none anywhere near their
claimed lifetime. They were installed in
dropped-down-from-the-ceiling fixtures,
with lots of room and generous airflow.
No autopsies were performed.

Does anyone what know the common failure modes are?