Yeah, a 1 pf capacitor and a 1 turn coil would resonate at a higher frequency. The long leads, the capacitance of the tunnel diode, the quality of the capacitor (which is more like an open stub transmission line), it all lowers the frequency.
@@gkdresden 1 pf Is about the lowest value capacitor you can buy. These crazy low values are relatively expensive too! It is much easier to use a small island of copper on a PCB, or stripline. Microwave electronics is more like plumbing. :)
Interesting series of clips on Tunnel diode. I believe your stability or phase noise problem is because of supply noise. Here how you can improve it : * Use a RF rated capacitor (10nF) possibly chip capacitor * Use small chock in series with your supply. * Use microstrip transmission line with stubs, instead of L1 & C1 * I guess the wires connecting your circuit to your supply act as antenna and puts noise on your oscillator. If any body else have any suggestion on how to improve the stability this oscillator, please share so we learn from each others techniques.
Probably some kind of closed type resonator (spiral, cavity), to reduce receiving radio frequency noise. Tunnel diode is known for its simple, but sensitive schematics. Old timers make VHF FM receivers containing only antenna, LC tank, diode, choke and filter capacitor.
I used to make those little twisted wire capacitors back in the 1980s. Quartz watches were quite new back then, and the cheap ones weren't too precise. Because the quartz crystals used in them were cheap and not working at exactly 32768Hz. Usually they were a little fast, up to a minute or two per day, meaning the crystal frequency was too high. The way to tune them properly would have been with a trimmer cap, but there was no space inside for that. So instead I'd twist 2 wires like you did, and solder them across the crystal. That would increase the capacitive load and decrease the frequency. The next day I'd come back and see what the drift was, trim a piece of the twisted wire if the watch was now too slow, etc.
Really interesting series, thank you! I have a Tektronix 284 Pulse Generator that uses a tunnel diode mounted in an air line that produces 70 picosecond or less rise time pulses into 50 ohms. Not bad for 1967. It was used for time domain reflectometery and for checking bandwidth of the scopes of the day among other things. Regards, David
Seeing the capacitor as you made it, being just two pieces of wire, seems like a fantastic visual representation of why/how capacitors block DC. Obviously (almost) no direct current can flow though it, as the conductors aren't even touching! Such an example should be used to teach new electronics students. Thanks for what you do.
Philbrick's old trick to make a 1 pF capacitor - yes! The tuning may depend on the voltage-dependent capacitance of either the wire insulation or the voltage dependence of the 10nF blocking capacitor.
PTFE insulation is much better there than PVC as well, which will increase the Q of the circuit, and reduce that instability. You will probably want to make the decoupling capacitor leads shorter as well
You have a bit of load on the circuit from the SA probe, and, if it's voltage sensitive, power supply ripple, or 60 Hz pickup could make it less than lovely... than unshielded "capacitor" wire might pick up static... For tuning the capacitor, you could also untwist the pair a bit... Not sure if there might be funny effects from the sharply cut wire ends, too. Just an old time FAE thinkin' out loud...
I’ve seen a similar circuit as a passive transmitter. Same as yours but no battery. You excite the circuit from an external oscillator. After excitation there is a small delay then it transmits for a moment. Used by US intel back in the day as sort of a tripwire when the device got close to a monitoring point.
I think the story goes that Mr. Esaki working for Sony in Japan invented the tunnel diode to demonstrate quantum tunneling, which is perhaps one of the weirdest phenomena in quantum mechanics. But nowehere in his wildest dreams did he imagine that GE would start to mass produce them and that in 2024 someone would produce a youtube video series on them and very creatively demonstrate how they work. I believe that initially the effect of the electrons tunneling through the barrier was too fast for the poor curve tracer and that is why there was a blank region on the screen, until it was slowed down by inductors and capacitors making it oscillate instead of tunneling almost instantaneously. That's my story and I'm sticking to it :-)
A resistor in series with the with the power supply would help protect the diode from overcurrent and make it easier to adjust the supply voltage for best setting of the forward bias current
In 1961 Jim Kyle K5JKX/4 got a 1N653 to oscillate using resistors between 30 and 150 megacycles. Jim commented anything within 12 feet of the restive circuit caused instability. He wrote it up in "Lost in a Tunnel", 73 Magazine, Jan, 1961, pp. 20-23. I've built problem circuits in circuit board boxes often as `dead bug` configurations to reduce microphonics and powered each part with its own battery power supply to eliminate paths for noise, cross talk, etc.
A twisted wire capacitor is called a 'gimmick'. You'll sometimes find them in old tube radio circuits. I built a good sounding FM radio transmitter years ago using one, taking advantage of the circuit's 'jumping around' in frequency for modulation.
if you connect the tunnel diode without inductor and pf capacitor, you will have the max oscillation frequency, limited by td internal capacitance and inductance terminals.
A gimmick cap got used in S38 A through C, msybe D receivers to make the IF amp oscillate in BFO mode. Only because 1 pf caps weten't available in 1952.
I'd be curious if a quarter wave shorted coax would oscillate too - A 6.5 inch piece of RG59 (foam) would be about right for 350 MHz (shorted at far end lol). Coax shield should go to ground side. Also need a better bypass - maybe 1000 pf chip. I had a tunnel diode back in the 1960s !😏
I would be curious to see if it's more stable with a regular ceramic cap. I think the twisted wires would have a significant inductance as well in this circuit (given the single loop of wire used as an inductor).
In 1962 TI made a gas tunnel for lo in their radar sets because of the higher power than ge, but had to withdraw them because they couldn't be used as switches
I think the early Fuzz Busters used a DSO stabilized tunnel diode oscillator as their LO. At the time tunnel diodes were a cheap way to get into the GHz.
Talking 'bout weird oscilators, i once wrote an article about reverse avalance mode oscilators. Looks pretty much like a tunneldiode osc but more in the LF range.
On the tunability of Tunnel diode When I was studying the tunnel diodes in school there was no mention of its tunability, only its negative resistance. The circuit above shows that tunnel diode are tunable, because the small signal AC model of above circuit diode's capacitor is in parallel with C1 ( Power rail is ground in AC). This makes sense because the depletion region on the diode increases, as result the capacitance decreases and frequency increases, as we increase voltage on the diode. To measure the capacitance of tunnel diode I thought to put the tunnel diode under the emitter of npn transistor put a inductor in parallel with it. The DC value of Emitter voltage is Vbase - Vbe ( Vbe ~ 0.7V). So as I increase the base voltage, so does the DC value of emitter voltage, and can measure the oscillation frequency and infer the capacitance of the diode. The capacitance of the diode should be not only the result of depletion capacitance, but also diffusion capacitance. This is interesting situation because in reverse biased diodes we have only diffusion capacitance, and in forward biased diodes we have only diffusion capacitance. But in tunnel diodes we both have depletion region and diffusion of carriers. Any idea how to measure them separately?
This was very interesting, thank you. I always learn something from your videos. It had never occurred to me to make a capacitor from a twisted pair, so I learned a second thing today. I like to try to understand how systems work, even small systems. Three gimmick capacitor questions for anyone who knows: Does the twist introduce unwanted inductance? Does the thermoplastic-looking insulation (vinyl?) absorb RF energy and thereby damp out RF oscillations? If these pose problems then can they be eliminated by using a non-twisted but closely-paralleled pair of varnish-insulated magnetic wires, perhaps adhered together with a coat of e.g. nail polish or glyptal cement or conformal coating or similar?
Your 10nf cap looks like it has a Z5U dielectric. They are sensitive to almost everything. The capacitance is voltage sensitive, they have shock/capacitance sensitivity, they are temperature sensitive, and for your circuit the frequency/capacitance/voltage sensitivity makes them mostly unusable. You need at least a good X7R or, better yet, a COG (NPO). One other thing: Is the diode in a metal package with glass fret? If so it may be light sensitive.
I believe it is called a gimmick capacitor. At least in this Era. If you got this from that early 60s book then maybe that's what it used to be called?
I think you need to shorten the wiring around the diode and the 10nF capacitor. I guess they also resonate at some frequency causing the circuit to be unstable.
When you get close to the gimmick capacitance you can fine tune the capacitance by unstwisting a little to lower it or twist a little tighter to increase it.
Frequency seems a bit low considering the circuit you've got there. Wouldn't you expect something in the GHz range? Are you sure you're not just picking up rubbish from your power supply? According to the data sheet that particular device isn't intended to be used as an oscillator.
Hi, Chad... I take it, that like me, you are not a classically trained electrical engineer. If so, I understand your frustration. I am also a self-taught hobbyist. Accordingly, I have only studied those things that interest me. This method has left huge gaps in my understanding, making it difficult to learn new things. But, I am not young (at all) and I need to only do things in an enjoyable way. So, I am not going to change my (evil and poor) ways of learning. All that rambling for several purposes... 1) Imsai Guy is a pretty good teacher. Hang in there and he will come back around to something that interests you. 2) I watch videos that I don't understand, just to get exposure to things that I do know (yet.) So, you didn't really waste your time here. You just got some exposure to things that you can add to, in the future. 3) I want to encourage you to hang in there. I don't know your skill level, but, if your are just learning, like me, it does get better. I pick devices/circuits that I think do something interesting and then I try to build them. I go to EE forums and pester the gurus with dumb questions. Eventually, I learn a thing, or two, and make new friends. I hope that helps. It is good to have met you. Cheers.
Fellow greybeard here... I have never heard that style of DIY capacitor called a "mimic", I've always known them as "gimmick" capacitors.
You are right, they are called a ‘Gimmick’.
Yeh Hams call them gimmicks
Same here.
Agreed. They are called gimmick capacitors.
sorry, brain fart, they are gimmics
Never heard of a mimic capacitor. Gimmick capacitors on the other hand I have. They were common in the vacuum tube era RF power meter circuits.
2:22 - in Germany we say "cut 10 times and still too short" (10 mal abgeschnitten und immer noch zu kurz) 😂
A universal truth!
1973 Nobel prize for Leo Esaki's demonstration of tunneling using his invention, the Esaki diode.
At 375 MHz those long leads on the bypass cap are going to matter.
And the inductance within this special capacitor as well.
Yeah, a 1 pf capacitor and a 1 turn coil would resonate at a higher frequency.
The long leads, the capacitance of the tunnel diode, the quality of the capacitor (which is more like an open stub transmission line), it all lowers the frequency.
@@Buzz_Purr and you can also buy a 1 pF capacitor. There is no need to make it by yourself.
@@gkdresden 1 pf Is about the lowest value capacitor you can buy.
These crazy low values are relatively expensive too!
It is much easier to use a small island of copper on a PCB, or stripline.
Microwave electronics is more like plumbing. :)
Interesting series of clips on Tunnel diode.
I believe your stability or phase noise problem is because of supply noise.
Here how you can improve it :
* Use a RF rated capacitor (10nF) possibly chip capacitor
* Use small chock in series with your supply.
* Use microstrip transmission line with stubs, instead of L1 & C1
* I guess the wires connecting your circuit to your supply act as antenna and puts noise on your oscillator.
If any body else have any suggestion on how to improve the stability this oscillator, please share so we learn from each others techniques.
Probably some kind of closed type resonator (spiral, cavity), to reduce receiving radio frequency noise. Tunnel diode is known for its simple, but sensitive schematics. Old timers make VHF FM receivers containing only antenna, LC tank, diode, choke and filter capacitor.
I used to make those little twisted wire capacitors back in the 1980s. Quartz watches were quite new back then, and the cheap ones weren't too precise. Because the quartz crystals used in them were cheap and not working at exactly 32768Hz. Usually they were a little fast, up to a minute or two per day, meaning the crystal frequency was too high. The way to tune them properly would have been with a trimmer cap, but there was no space inside for that. So instead I'd twist 2 wires like you did, and solder them across the crystal. That would increase the capacitive load and decrease the frequency. The next day I'd come back and see what the drift was, trim a piece of the twisted wire if the watch was now too slow, etc.
Really interesting series, thank you! I have a Tektronix 284 Pulse Generator that uses a tunnel diode mounted in an air line that produces 70 picosecond or less rise time pulses into 50 ohms. Not bad for 1967. It was used for time domain reflectometery and for checking bandwidth of the scopes of the day among other things. Regards, David
I envy you. The best I can produce with a russian diode is like 150 ps (home built reflectometer) .
Seeing the capacitor as you made it, being just two pieces of wire, seems like a fantastic visual representation of why/how capacitors block DC. Obviously (almost) no direct current can flow though it, as the conductors aren't even touching! Such an example should be used to teach new electronics students.
Thanks for what you do.
Put it in a cavity with a regulated power source and loose output coupling.
Maybe tunnelling diodes will work better with spade connectors….
🤣
Philbrick's old trick to make a 1 pF capacitor - yes!
The tuning may depend on the voltage-dependent capacitance of either the wire insulation or the voltage dependence of the 10nF blocking capacitor.
PTFE insulation is much better there than PVC as well, which will increase the Q of the circuit, and reduce that instability. You will probably want to make the decoupling capacitor leads shorter as well
You have a bit of load on the circuit from the SA probe, and, if it's voltage sensitive, power supply ripple, or 60 Hz pickup could make it less than lovely... than unshielded "capacitor" wire might pick up static... For tuning the capacitor, you could also untwist the pair a bit... Not sure if there might be funny effects from the sharply cut wire ends, too.
Just an old time FAE thinkin' out loud...
I’ve seen a similar circuit as a passive transmitter. Same as yours but no battery. You excite the circuit from an external oscillator. After excitation there is a small delay then it transmits for a moment. Used by US intel back in the day as sort of a tripwire when the device got close to a monitoring point.
I think the story goes that Mr. Esaki working for Sony in Japan invented the tunnel diode to demonstrate quantum tunneling, which is perhaps one of the weirdest phenomena in quantum mechanics. But nowehere in his wildest dreams did he imagine that GE would start to mass produce them and that in 2024 someone would produce a youtube video series on them and very creatively demonstrate how they work.
I believe that initially the effect of the electrons tunneling through the barrier was too fast for the poor curve tracer and that is why there was a blank region on the screen, until it was slowed down by inductors and capacitors making it oscillate instead of tunneling almost instantaneously.
That's my story and I'm sticking to it :-)
A resistor in series with the with the power supply would help protect the diode from overcurrent and make it easier to adjust the supply voltage for best setting of the forward bias current
In 1961 Jim Kyle K5JKX/4 got a 1N653 to oscillate using resistors between 30 and 150 megacycles. Jim commented anything within 12 feet of the restive circuit caused instability. He wrote it up in "Lost in a Tunnel", 73 Magazine, Jan, 1961, pp. 20-23.
I've built problem circuits in circuit board boxes often as `dead bug` configurations to reduce microphonics and powered each part with its own battery power supply to eliminate paths for noise, cross talk, etc.
A twisted wire capacitor is called a 'gimmick'. You'll sometimes find them in old tube radio circuits. I built a good sounding FM radio transmitter years ago using one, taking advantage of the circuit's 'jumping around' in frequency for modulation.
if you connect the tunnel diode without inductor and pf capacitor, you will have the max oscillation frequency, limited by td internal capacitance and inductance terminals.
Bob Pease would be proud!!!
A gimmick cap got used in S38 A through C, msybe D receivers to make the IF amp oscillate in BFO mode. Only because 1 pf caps weten't available in 1952.
Better way is make resonant circuit ( tank ) by cut a coaxial H1000 cable , short at one end , and conect tunnel diode at second.
Can you show on vídeo?
I guess you'd expect it to be TUNNELble oscillator hahaha
I'd be curious if a quarter wave shorted coax would oscillate too - A 6.5 inch piece of RG59 (foam) would be about right for 350 MHz (shorted at far end lol). Coax shield should go to ground side. Also need a better bypass - maybe 1000 pf chip. I had a tunnel diode back in the 1960s !😏
I would be curious to see if it's more stable with a regular ceramic cap. I think the twisted wires would have a significant inductance as well in this circuit (given the single loop of wire used as an inductor).
Tekteo ics used a lot of tunnel diodes in some scopes.
In 1962 TI made a gas tunnel for lo in their radar sets because of the higher power than ge, but had to withdraw them because they couldn't be used as switches
Super! Thank you very much!
I think the early Fuzz Busters used a DSO stabilized tunnel diode oscillator as their LO. At the time tunnel diodes were a cheap way to get into the GHz.
Talking 'bout weird oscilators, i once wrote an article about reverse avalance mode oscilators.
Looks pretty much like a tunneldiode osc but more in the LF range.
Thanks a lot! I'm YO3HST, radio-ham.
Would love to see a tunnelling diode amplifier. I've seen a schematic on Wikipedia but I never got it to work
On the tunability of Tunnel diode
When I was studying the tunnel diodes in school there was no mention of its tunability, only its negative resistance.
The circuit above shows that tunnel diode are tunable, because the small signal AC model of above circuit diode's capacitor is in parallel with C1 ( Power rail is ground in AC).
This makes sense because the depletion region on the diode increases, as result the capacitance decreases and frequency increases, as we increase voltage on the diode.
To measure the capacitance of tunnel diode I thought to put the tunnel diode under the emitter of npn transistor put a inductor in parallel with it.
The DC value of Emitter voltage is Vbase - Vbe ( Vbe ~ 0.7V). So as I increase the base voltage, so does the DC value of emitter voltage, and can measure the oscillation frequency and infer the capacitance of the diode.
The capacitance of the diode should be not only the result of depletion capacitance, but also diffusion capacitance.
This is interesting situation because in reverse biased diodes we have only diffusion capacitance, and in forward biased diodes we have only diffusion capacitance.
But in tunnel diodes we both have depletion region and diffusion of carriers.
Any idea how to measure them separately?
Thanks 👍
Please try inductor turns see if its become more stable higher or lower in freq...
and now we know why no one uses tunnel diodes anymore(😂)
PLLs made them obsolete. Great Retro Electronics Video. Now it's on to Varactors.
I've already done several videos on varactors, varicaps
@@IMSAIGuy 👍 how about IMPATT diodes? back diodes? joking aside i am curious about back diodes.
This was very interesting, thank you. I always learn something from your videos.
It had never occurred to me to make a capacitor from a twisted pair, so I learned a second thing today.
I like to try to understand how systems work, even small systems. Three gimmick capacitor questions for anyone who knows:
Does the twist introduce unwanted inductance?
Does the thermoplastic-looking insulation (vinyl?) absorb RF energy and thereby damp out RF oscillations?
If these pose problems then can they be eliminated by using a non-twisted but closely-paralleled pair of varnish-insulated magnetic wires, perhaps adhered together with a coat of e.g. nail polish or glyptal cement or conformal coating or similar?
Top video, merci. Very interesting.
Your 10nf cap looks like it has a Z5U dielectric. They are sensitive to almost everything. The capacitance is voltage sensitive, they have shock/capacitance sensitivity, they are temperature sensitive, and for your circuit the frequency/capacitance/voltage sensitivity makes them mostly unusable. You need at least a good X7R or, better yet, a COG (NPO). One other thing: Is the diode in a metal package with glass fret? If so it may be light sensitive.
What other circuits are they actually useful for? Why/what did (do?) people buy them for?
what is the dielectric for that bypass cap,...
some how I doubt there is much of any bypass at that frequency.
Well does anyone know at which frequency that LC circuit was supposed to oscillate?
I believe it is called a gimmick capacitor. At least in this Era. If you got this from that early 60s book then maybe that's what it used to be called?
Dude, your LCR is able to read femtofarad... crazy.
yes, and this one does atto farads th-cam.com/video/pTyXONAJt6k/w-d-xo.htmlsi=J90SRI27UcZLLX4-
I think you need to shorten the wiring around the diode and the 10nF capacitor. I guess they also resonate at some frequency causing the circuit to be unstable.
We called them gimmick
How would it go if you use a reverse bias diode instead of the 1pf capacitor? Please make it if it interests you.
9mm would be the inside of the socket not the outside. Cool video.
As he said, the diameter of the socket is 9mm; it's not a 9mm socket.
It would be a lot bigger than that.
gimic
When you get close to the gimmick capacitance you can fine tune the capacitance by unstwisting a little to lower it or twist a little tighter to increase it.
What's the theory of operation?
Hope this helps: www.worldradiohistory.com/BOOKSHELF-ARH/Technology/GE-Books/General%20Electric-Tunnel%20Diode%20Manual-First%20Edition-1961.pdf
Make the loop bigger, and increase the capacitor to 3pf and see if you can pick it up on an fm radio, then you will be able to hear it's modulation
Could you tell us the brand and model of your LCR meter please.
th-cam.com/video/247VKbnZ8TM/w-d-xo.html
Thanks!@@IMSAIGuy
WOW, thank's. from Bill
Frequency seems a bit low considering the circuit you've got there. Wouldn't you expect something in the GHz range? Are you sure you're not just picking up rubbish from your power supply? According to the data sheet that particular device isn't intended to be used as an oscillator.
I know absolutely nothing about this stuff. I have no idea what I'm doing here
Learning?
@@fritzkinderhoffen2369 no. I have no idea what anything is. I understood nothing
I thought it was a diode tuned LC circuit.
Watch this channel for a couple of years and it will start to sink in :-)
Hi, Chad... I take it, that like me, you are not a classically trained electrical engineer. If so, I understand your frustration. I am also a self-taught hobbyist. Accordingly, I have only studied those things that interest me. This method has left huge gaps in my understanding, making it difficult to learn new things. But, I am not young (at all) and I need to only do things in an enjoyable way. So, I am not going to change my (evil and poor) ways of learning. All that rambling for several purposes... 1) Imsai Guy is a pretty good teacher. Hang in there and he will come back around to something that interests you. 2) I watch videos that I don't understand, just to get exposure to things that I do know (yet.) So, you didn't really waste your time here. You just got some exposure to things that you can add to, in the future. 3) I want to encourage you to hang in there. I don't know your skill level, but, if your are just learning, like me, it does get better. I pick devices/circuits that I think do something interesting and then I try to build them. I go to EE forums and pester the gurus with dumb questions. Eventually, I learn a thing, or two, and make new friends. I hope that helps. It is good to have met you. Cheers.
“gimmic capacitor” I think is what you mean.
Mimic diode? No, you meant to say a gimmick capacitor.
Never seen that. It works but sure ain’t pretty.