Its theory of operation escapes me. I know there is a negative resistance part of its IE curve, but doesn't that just make it into a current source through that range ?
I read they can make an oscillator out of it or a UHF amp, and Tektronix used them in their trigger circuit in scopes. (those did seem to be really good triggering circuits, would lock onto anything)
Is there a place with some explanations and sample circuits with their theory of operation ?
Thanks. I gave it a quick lookover. Saved it and will study it later.
Are they still being manufactured?
I fooled with some gunn diodes. Built little cavity resonators out of
3/4" copper pipe caps, and had a blast setting off radar detectors.
Not 'current source' (that's HIGH impedance, not negative). It makes it a gain element (in series with a positive resistor, it makes an attenuator with AC gain according to the formula).
Attenuation_factor = Z_1/(Z_1 + Z_tunnel) >1 when Z_tunnel is negative
The reason diodes don't forward-conduct until you get some voltage across 'em is that there's an insulating 'depletion region' that gets thin when you forward-bias. But, a bit of tunneling (Zener-style conduction) happens before the forward-bias, ONLY if there's a no-energy-change path (the P valence and N conduction bands have to have energy ranges that overlaps).
Forward bias removes that overlap by shifting P versus N voltage bias, reducing the tunneling.
Leo Esaki (Nobel prize, 1973) figured it out. Good treatment here:
Real (normal, non-tunneling) diode conduction happens at higher voltages.
Back diodes are still being manufactured, as zero-bias RF detectors. They are low-peak-current tunnel diodes. I think back diodes are/were the only germanium devices manufactured with a photolithograpic process.
Somebody was still making the old GE TDs a few years back, frightfully expensive. The fab process was insane.
Ge is a pain to do litho on because its oxide is water-soluble, so it's hard to control noise and leakage due to surface states at the semiconductor-insulator interface. (Diamond would be just that much worse, of course. ) :)
And the junction capacitance was equally insane, like 200 pF. The really fast ones had peak currents of about 200 mA. I have a bag full of 1-mA ones, which aren't that interesting. I got a dozen or so Russian InGaAs TDs off eBay awhile back but haven't tested them yet.
Cheers
Phil Hobbs
Some of the old GE and RCA parts had insane Ip/Cj ratios, enough to give 25 ps risetimes.
This still works, last time I checked:
and this is the TDR step generator from an old HP sampler:
For a long time, a tunnel diode was the fastest thing around.
I have a lot of 5 mA axial GE parts. There was a bin at HalTek full of them, 10 cents each.
I doubt so, they were expensive and had a very narrow market.
Some eastern EU users, mainly form Russia and Ukraine sell cheap NOS USSR made germanium tunnel diodes on Ebay. (example:
I used to design digital stuff with TDs.
My 1964 Allied catalog shows a 1N914 at $1.89, and a 1N3716 (5 mA Ip tunnel diode) for $2.55.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required