Low voltage negative resistance oscillator design, lessons l

B

bitrex

Guest
Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble
 
On 8/8/19 3:24 AM, bitrex wrote:
Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble
Click to expand...

You can build an okay oscillator using piglet/Win Hill's topology from
the AoE x-chapters, I got an LC working down to about 1.8 volts with
that one with a fairly low Q coil using the negative R to cancel it. a
wide tuning range is impossible though because the slope of the negative
resistance changes with load and there's not enough headroom to easily
buffer it to isolate it from the tank, there's almost always too much
gain or too little and...bla bla bla world of pain, larry.
 
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
<user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble
Click to expand...

You did not try a tunnel diode?
 
On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?
Click to expand...

I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for microwave
frequency oscillators and such? I'm not going that fast.
 
On a sunny day (Thu, 8 Aug 2019 12:52:35 -0400) it happened bitrex
<user@example.net> wrote in <nrY2F.153633$lH5.25796@fx05.iad>:

On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for microwave
frequency oscillators and such? I'm not going that fast.
Click to expand...

Few MHz works OK too, were used in the FM modulator of some Ampex quadruplex video recorders,
say 5 MHz to 7 MHz sweep or there about.

ebay has plenty:
https://www.ebay.com/itm/254034572875

AFAIK the frequency low side does not have a limit.
https://en.wikipedia.org/wiki/Tunnel_diode

Not sure that will work in your circuit,
what exactly are you trying to accomplish with that circuit?
 
On a sunny day (Thu, 8 Aug 2019 14:26:44 -0400) it happened bitrex
<user@example.net> wrote in <FPZ2F.33617$V64.2871@fx19.iad>:

On 8/8/19 1:14 PM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 12:52:35 -0400) it happened bitrex
user@example.net> wrote in <nrY2F.153633$lH5.25796@fx05.iad>:

On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for microwave
frequency oscillators and such? I'm not going that fast.

Few MHz works OK too, were used in the FM modulator of some Ampex quadruplex video recorders,
say 5 MHz to 7 MHz sweep or there about.

ebay has plenty:
https://www.ebay.com/itm/254034572875

AFAIK the frequency low side does not have a limit.
https://en.wikipedia.org/wiki/Tunnel_diode

Not sure that will work in your circuit,
what exactly are you trying to accomplish with that circuit?


Increasing the Q factor of a pathologically low-Q coil, like a Q of say
2 at 3 MHz, so you can build a tank-type oscillator from it at a lower
frequency than up where the L has a good Q which might be 50 or 100MHz.
at low voltage, say 2 volts.

A cheap AD8515 works down to 1.8 volts and with 3 resistors in the
negative-resistance configuration, even at 2 v it has enough oomph and
bandwidth to make even a crap L ring nicely in a tank circuit like:

https://imgur.com/a/WZqskyw
Click to expand...

OK, got the idea.
If you look at the wikiedia link I gave, enlarge the picture of the Tek trace
https://en.wikipedia.org/wiki/Tunnel_diode#/media/File:I-V_curve_of_10mA_germanium_tunnel_diode..jpg

then you see the middle of the negative resistance area is around 325 mV for
a Ge tunnel diode, ebay has those too it seems:
https://www.ebay.com/itm/202519120776

But if your solution works for you OK.
 
On 8/8/2019 1:12 PM, Tom Gardner wrote:
On 08/08/19 17:52, bitrex wrote:
On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and
are
seriously fussy with respect to how you load them and have lousy
tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for
microwave frequency oscillators and such? I'm not going that fast.

The trigger circuit in Tek 4x5 era scopes was famously good. My
350MHz Tek 485 reliably triggers on a 1.1GHz signal. I suspect
it would go higher, but the amplitude is then only 0.2 divisions.

Step generators, of course; 50ps or less 0.4V into 50ohms is
achievable.
Click to expand...

Not that it has application for you, but others might like to see
an oscillator that works at 7mv. I thought it was neat.

> http://www.dicks-website.eu/fetosc/enindex.htm

Mikek
PS.
I've posted it before.
 
On 8/8/19 1:14 PM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 12:52:35 -0400) it happened bitrex
user@example.net> wrote in <nrY2F.153633$lH5.25796@fx05.iad>:

On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for microwave
frequency oscillators and such? I'm not going that fast.

Few MHz works OK too, were used in the FM modulator of some Ampex quadruplex video recorders,
say 5 MHz to 7 MHz sweep or there about.

ebay has plenty:
https://www.ebay.com/itm/254034572875

AFAIK the frequency low side does not have a limit.
https://en.wikipedia.org/wiki/Tunnel_diode

Not sure that will work in your circuit,
what exactly are you trying to accomplish with that circuit?
Click to expand...

Increasing the Q factor of a pathologically low-Q coil, like a Q of say
2 at 3 MHz, so you can build a tank-type oscillator from it at a lower
frequency than up where the L has a good Q which might be 50 or 100MHz.
at low voltage, say 2 volts.

A cheap AD8515 works down to 1.8 volts and with 3 resistors in the
negative-resistance configuration, even at 2 v it has enough oomph and
bandwidth to make even a crap L ring nicely in a tank circuit like:

<https://imgur.com/a/WZqskyw>
 
On 08/08/19 17:52, bitrex wrote:
On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world, much less
used one for anything! aren't they usually used for microwave frequency
oscillators and such? I'm not going that fast.
Click to expand...

The trigger circuit in Tek 4x5 era scopes was famously good. My
350MHz Tek 485 reliably triggers on a 1.1GHz signal. I suspect
it would go higher, but the amplitude is then only 0.2 divisions.

Step generators, of course; 50ps or less 0.4V into 50ohms is
achievable.
 
Forgot a link:
http://panteltje.com/pub//lighting_a_LED_with_a_candle_setup_IMG_3607.GIF
 
On a sunny day (Thu, 8 Aug 2019 13:40:08 -0500) it happened amdx
<nojunk@knology.net> wrote in <qihqa3$9g8$1@dont-email.me>:

On 8/8/2019 1:12 PM, Tom Gardner wrote:
On 08/08/19 17:52, bitrex wrote:
On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and
are
seriously fussy with respect to how you load them and have lousy
tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for
microwave frequency oscillators and such? I'm not going that fast.

The trigger circuit in Tek 4x5 era scopes was famously good. My
350MHz Tek 485 reliably triggers on a 1.1GHz signal. I suspect
it would go higher, but the amplitude is then only 0.2 divisions.

Step generators, of course; 50ps or less 0.4V into 50ohms is
achievable.

Not that it has application for you, but others might like to see
an oscillator that works at 7mv. I thought it was neat.

http://www.dicks-website.eu/fetosc/enindex.htm

Mikek
PS.
I've posted it before.
Click to expand...

Yes, JFET oscilators work from almost zero volts upward, done that:
http://panteltje.com/pub/lighting_a_LED_with_a_candle_IMG_3604.GIF
http://panteltje.com/pub/lighting_a_LED_with_a_candle_IMG_3604.GIF

The negative gate voltage will limit the amplitude / gain,
enough to drive a LED.


 
On 8/8/19 2:26 PM, bitrex wrote:
On 8/8/19 1:14 PM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 12:52:35 -0400) it happened bitrex
user@example.net> wrote in <nrY2F.153633$lH5.25796@fx05.iad>:

On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance
topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts
and are
seriously fussy with respect to how you load them and have lousy
tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for microwave
frequency oscillators and such? I'm not going that fast.

Few MHz works OK too, were used in the FM modulator of some Ampex
quadruplex video recorders,
say 5 MHz to 7 MHz sweep or there about.

ebay has plenty:
  https://www.ebay.com/itm/254034572875

AFAIK the frequency low side does not have a limit.
   https://en.wikipedia.org/wiki/Tunnel_diode

Not sure that will work in your circuit,
what exactly are you trying to accomplish with that circuit?


Increasing the Q factor of a pathologically low-Q coil, like a Q of say
2 at 3 MHz, so you can build a tank-type oscillator from it at a lower
frequency than up where the L has a good Q which might be 50 or 100MHz.
at low voltage, say 2 volts.

A cheap AD8515 works down to 1.8 volts and with 3 resistors in the
negative-resistance configuration, even at 2 v it has enough oomph and
bandwidth to make even a crap L ring nicely in a tank circuit like:

https://imgur.com/a/WZqskyw


Unfortunately negative resistors don't have imaginary noise. ;)
Click to expand...

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
 
On 8/8/19 3:59 PM, Phil Hobbs wrote:
On 8/8/19 2:26 PM, bitrex wrote:
On 8/8/19 1:14 PM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 12:52:35 -0400) it happened bitrex
user@example.net> wrote in <nrY2F.153633$lH5.25796@fx05.iad>:

On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance
topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts
and are
seriously fussy with respect to how you load them and have lousy
tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for microwave
frequency oscillators and such? I'm not going that fast.

Few MHz works OK too, were used in the FM modulator of some Ampex
quadruplex video recorders,
say 5 MHz to 7 MHz sweep or there about.

ebay has plenty:
  https://www.ebay.com/itm/254034572875

AFAIK the frequency low side does not have a limit.
   https://en.wikipedia.org/wiki/Tunnel_diode

Not sure that will work in your circuit,
what exactly are you trying to accomplish with that circuit?


Increasing the Q factor of a pathologically low-Q coil, like a Q of
say 2 at 3 MHz, so you can build a tank-type oscillator from it at a
lower frequency than up where the L has a good Q which might be 50 or
100MHz. at low voltage, say 2 volts.

A cheap AD8515 works down to 1.8 volts and with 3 resistors in the
negative-resistance configuration, even at 2 v it has enough oomph and
bandwidth to make even a crap L ring nicely in a tank circuit like:

https://imgur.com/a/WZqskyw


Unfortunately negative resistors don't have imaginary noise. ;)

Cheers

Phil Hobbs
Click to expand...

well clearly you just need to couple them to a "negative heater" and
cool them down!
 
On 8/8/19 4:39 PM, bitrex wrote:
On 8/8/19 3:59 PM, Phil Hobbs wrote:
On 8/8/19 2:26 PM, bitrex wrote:
On 8/8/19 1:14 PM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 12:52:35 -0400) it happened bitrex
user@example.net> wrote in <nrY2F.153633$lH5.25796@fx05.iad>:

On 8/8/19 4:29 AM, Jan Panteltje wrote:
On a sunny day (Thu, 8 Aug 2019 03:24:11 -0400) it happened bitrex
user@example.net> wrote in <w6Q2F.342353$5i.35921@fx46.iad>:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance
topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts
and are
seriously fussy with respect to how you load them and have lousy
tempcos
down there.

get an AD8515 for 45 cents and save yourself the trouble

You did not try a tunnel diode?


I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for
microwave
frequency oscillators and such? I'm not going that fast.

Few MHz works OK too, were used in the FM modulator of some Ampex
quadruplex video recorders,
say 5 MHz to 7 MHz sweep or there about.

ebay has plenty:
  https://www.ebay.com/itm/254034572875

AFAIK the frequency low side does not have a limit.
   https://en.wikipedia.org/wiki/Tunnel_diode

Not sure that will work in your circuit,
what exactly are you trying to accomplish with that circuit?


Increasing the Q factor of a pathologically low-Q coil, like a Q of
say 2 at 3 MHz, so you can build a tank-type oscillator from it at a
lower frequency than up where the L has a good Q which might be 50 or
100MHz. at low voltage, say 2 volts.

A cheap AD8515 works down to 1.8 volts and with 3 resistors in the
negative-resistance configuration, even at 2 v it has enough oomph
and bandwidth to make even a crap L ring nicely in a tank circuit like:

https://imgur.com/a/WZqskyw


Unfortunately negative resistors don't have imaginary noise. ;)

Cheers

Phil Hobbs


well clearly you just need to couple them to a "negative heater" and
cool them down!
Click to expand...

Sorry, I mean un-heat them.
 
On 8/8/19 2:53 PM, Jan Panteltje wrote:
Forgot a link:
http://panteltje.com/pub//lighting_a_LED_with_a_candle_setup_IMG_3607.GIF
Click to expand...

Ah, like the solar-powered flashlight!
 
On a sunny day (Thu, 8 Aug 2019 16:39:41 -0400) it happened bitrex
<user@example.net> wrote in <hM%2F.160565$rc7.116492@fx07.iad>:

Unfortunately negative resistors don't have imaginary noise. ;)

Cheers

Phil Hobbs


well clearly you just need to couple them to a "negative heater" and
cool them down!


Sorry, I mean un-heat them.
Click to expand...

I stil wonder, from the room temperature super conductor postings recently
if one can use ultrasound to slow the motion of atoms in a crystal down
like we use laser cooling to do it with a gas.
I want a Usenet patent for that,

I should try it and measure it but am no physicists, limited lab space and equipment, lame excuse I know.

so then a resistor element with piezo cooler... noise reduction, probably gives you the piezo signal
hey, maybe we can filter that out, oh wait
LOL
 
bitrex wrote:

Don't bother with discretes if you want a negative resistor when you
have low psu headroom, all the discrete negative resistance topologies,
lambda diodes, foldback limiters, etc. are bad at a couple volts and are
seriously fussy with respect to how you load them and have lousy tempcos
down there.
Click to expand...

A properly designed JFET-based oscillator can go as low as 25mv of
supply voltage. A random one should have no problems at 0.5V, BTDT.

Best reards, Piotr
 
bitrex wrote:

I don't think I've ever even _seen_ a tunnel diode in the real world,
much less used one for anything! aren't they usually used for microwave
frequency oscillators and such? I'm not going that fast.
Click to expand...

Give them a bulky tank circuit and they will go as low as 1MHz.
But what are you tryig to achieve?

Best regards, Piotr
 
Jan Panteltje wrote:

I stil wonder, from the room temperature super conductor postings recently
if one can use ultrasound to slow the motion of atoms in a crystal down
like we use laser cooling to do it with a gas.
I want a Usenet patent for that,
Click to expand...

I wonder if there is a loose analogy between the evanescent component of
an EM field that allows you to break the Abbe resolution limit and
something resulting from a specifically designed material structure
that would allow you to go well below the classic Boltzmann noise limit.

Classical physics doesn't allow you to build this gizmo:

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

so maybe a noiseless system is possible as well?

Best regards, Piotr
 
piglet wrote:

> There isn't a lower limit, tunnel diodes can do audio too.

Undoubtedly correct, but IMHO this is pretty pointless given the
required LC physical size. RC wins there.

Best regards, Piotr
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top