Tunnel diode polarity

[FlyingHacker]

Anyone know how to identify the polarity of one of the germanium tunnel diode. One side has a red dot. Is that red as is positive/anode? Or not?

Thanks.

[Alex Eisenhut]

What package is it in? By the red dot, I assume it's the tiny black blob on two beam leads? Is it a single dot or a bunch of dots?

[FlyingHacker]

Black blob with two flat leads that are brass in color.

One side has a single red dot.

Are these particularly sensitive? At first it was testing OK on a Keithley in diode mode. But then later it did not, as though perhaps I ruined it by testing it.

[rfeecs]

Are these particularly sensitive?

Yes.  Extremely ESD sensitive.

With 500mV breakdown voltage, a multimeter in diode mode applying 2V could possibly kill it.

[Alex Eisenhut]

Black blob with two flat leads that are brass in color.

One side has a single red dot.

Are these particularly sensitive? At first it was testing OK on a Keithley in diode mode. But then later it did not, as though perhaps I ruined it by testing it.

You probably did. They are meant to be used reverse polarity; in the normal-diode-forward they are mostly a dead short. Too much current will kill these. I don't know what "OK" would mean to a diode tester testing a tunnel diode. What did you see and why would you think it was OK? Tunnel diodes make fantastic oscillators, who knows what the Keithley saw.

The dot usually indicated the cathode, but since these diodes were mostly used in-house by companies like Tektronix and HP, there might be some oddly-marked diodes out there. Tek would paint the part number in color code on one of the leads, this was the cathode.

The usual way I test these is with a function generator and a resistor and a scope; since you rarely find diodes with a Ip greater than 10mA, this works fine. Limit the current to about 12mA and use a triangle wave. You'll see the snap when the diode shifts state.

[FlyingHacker]

These are some tunnel diodes from Russia. I am trying to repair a pulse generator from Systron Donner. It turns out one of the guys that designed the original circuit was on the Tek scopes mailing list, and he said that most any tunnel diode should work.

What is OK? I don't know. It did test differently in the two different polarities. However, later, but before I installed it, it did not. So I think I fried it in testing it. I do have a couple more.

You are saying to limit it to 12mA by selection of the resistor connected in series with it coming off the function generator? My function generators cannot limit current on their own. Perhaps a more detailed test setup explanation would help this hobbyist 

Thanks.

[Alex Eisenhut]

Basically the idea is that a generator with a 50 ohm output will output 20mA into a short with only 1V, so best to limit that with an extra resistor in series.
I have an old Wavetek that can power the universe from its output, I put a 1K resistor in series with the reversed diode, and this gives me plenty of room to dial in accurately.
It's also not that hard to do the same test with a multimeter; put the diode in reverse, dial in the current with a power supply that has fine control, and watch the voltage across the diode, when it snaps over to the valley voltage, the current you read just before was the Ip.

[edavid]

You probably did. They are meant to be used reverse polarity; in the normal-diode-forward they are mostly a dead short.

That is backward.  Tunnel diodes are used in the forward direction (where they look like a normal diode except for the famous negative resistance region), and in the reverse direction they start conducting very close to 0V.  When used in the reverse direction, they are called "back diodes", even though it's the same device.

I don't see any reason why a DMM diode test function would hurt a tunnel diode, unless it were a very tiny low current device (in which case it would be useless in a pulse generator).  It's more likely that 30-40 year old surplus Russian electronic parts are poorly tested/flaky/unreliable.

Also, tunnel diodes are not particularly ESD sensitive, since they have such low conduction voltages in both directions.

[Alex Eisenhut]

You probably did. They are meant to be used reverse polarity; in the normal-diode-forward they are mostly a dead short.

That is backward.  Tunnel diodes are used in the forward direction (where they look like a normal diode except for the famous negative resistance region), and in the reverse direction they start conducting very close to 0V.  When used in the reverse direction, they are called "back diodes", even though it's the same device.

I don't see any reason why a DMM diode test function would hurt a tunnel diode, unless it were a very tiny low current device (in which case it would be useless in a pulse generator).  It's more likely that 30-40 year old surplus Russian electronic parts are poorly tested/flaky/unreliable.

Also, tunnel diodes are not particularly ESD sensitive, since they have such low conduction voltages in both directions.

Ah yeah, you're right... I don't know why I thought they are used backwards! But a normal diode tester can harm them because they simply can't handle a lot of current. Especially the small blob ones.
I think I wanted to say that backwards they are dead shorts. 

[edavid]

But a normal diode tester can harm them because they simply can't handle a lot of current. Especially the small blob ones.

I just tested the 3 closest DMMs, and the diode test current was 0.5-1.0mA 

[Alex Eisenhut]

I guess the warnings from old databooks and service manuals stuck with me. Probably they were talking about VTVMs and analog meters, they didn't have diode test so you'd use the resistance scale. Those can pump high currents.

[rfeecs]

Well the only common application left for germanium tunnel diodes is in zero bias microwave detectors.  And these have such tiny junctions that if you breathe on them, they blow.  So extremely sensitive.

But 1960's era devices are probably another story.  It looks like there were some pretty beefy devices made with peak currents in the hundreds of milliamps.

[Alex Eisenhut]

Well the only common application left for germanium tunnel diodes is in zero bias microwave detectors.  And these have such tiny junctions that if you breathe on them, they blow.  So extremely sensitive.

But 1960's era devices are probably another story.  It looks like there were some pretty beefy devices made with peak currents in the hundreds of milliamps.

I have never seen diodes with more than 10mA Ip. I know there is a 100mA diode in one of my pulsers but since repair is impossible, it's one of the few things I've never taken apart. And according to my databooks, they made tunnel diodes into the dozens of *amps*. I don't know what for.

The zero-bias detectors are "back diodes". *those* are the ones you use backwards! 

[edavid]

Looking at GE datasheets, it appears that the absolute maximum ratings are 5-10X the specified peak current.

For example, a 1N2929 or 1N3712 1mA tunnel diode has a max forward current spec of 5mA, and reverse current of 10mA.

OP never said what Russian device he was using, but if it's a low current device, yes, it would be a good idea to be careful when testing...

[FlyingHacker]

The particular diodes are labelled as:

GI304A Ge Amplifier Tunnel Diode military USSR

The seller is no longer there, and I paid less than this, but here is a similar posting:

http://www.ebay.com/itm/GI304A-Ge-Amplifier-Tunnel-Diode-military-Lot-of-4-pcs-/131743984112?hash=item1eac8da9f0

Looks like this one shows the polarity in one of the pictures.

What I was trying to do was to use a sharp pair of side cutters to trim the blades down, and then put them into through holes. This worked fine, other than the fact that the diode doesn't work now ;-)

I am going to test another one of these out.

[jimmc]

Looks like RCA made a 200Amp ( that's AMP not milliamp) GaAs Tunnel diode type 40079!
see  https://archive.org/details/RcaTunnelDiodeManual p147

Jim

[FlyingHacker]

So I hooked another one of these tunnel diodes up to both a scope and an analog ammeter. I could see the tunnel effect on the analog meter, and the oscillation pattern on the scope.

The odd thing is that the diode that I removed from the pulse generator seems to work as well. Although the one I removed stop oscillating if I turn the voltage up a little. I will try the new one, but perhaps my troubleshooting was incorrect and the tunnel diode is not the culprit. Though the service manual had some pretty good steps to isolate it.

[mmagin]

I know in my Tek 284 http://w140.com/tekwiki/wiki/284 there's a front panel trimmer related to the TD bias which needs occasional adjustment.  I seem to recall when I was going through the adjustment procedures there were a couple other things that needed adjustment.

I had to replace a dead transistor somewhere in the lower frequency parts of the circuit -- but it was just a boring old 2n3904, not some unobtainium!

[Alex Eisenhut]

So I hooked another one of these tunnel diodes up to both a scope and an analog ammeter. I could see the tunnel effect on the analog meter, and the oscillation pattern on the scope.

The odd thing is that the diode that I removed from the pulse generator seems to work as well. Although the one I removed stop oscillating if I turn the voltage up a little. I will try the new one, but perhaps my troubleshooting was incorrect and the tunnel diode is not the culprit. Though the service manual had some pretty good steps to isolate it.

Depends on the passive components around it. Carbon composition resistors drift a lot in value (20%!) and can change the polarization (load line) of the diode.

Perhaps the troubleshooting part of the manual assumes the resistor values would stay the same.

See any of these in your instrument?



be suspicious. Try to measure them out of circuit.

[FlyingHacker]

It turned out to be another issue.... Not the tunnel diode. I was not reading the waveforms properly because of a conformal coating. Once I got past that I found a transitor with a loose base pin. It's loose at the transistor itself. So I need to get a replacement. But I got it working by wiggling it a little.

It has the new tunnel diode in there now, though.

I am all too familiar with carbon comps. This guy doesn't seem to have them, though.

[orolo]

I'd suggest testing your diode with a gentle ramp generator and an oscilloscope. That worked for me with several exotic ebay tunnel diodes. In any case, I have attached the typical curve for your diode: ??304 (GI304) is just the military rated version of 1?304 (1I304), it seems to be rated for a peak current of 5mA, and it certainly is not a backwards diode.

Also, check this great video from w2aew, who is a contributor here: https://youtu.be/PuG8CCUbg58

Edit: cyrillic letters seem to fail in this forum, or is it my browser?

[FlyingHacker]

I referenced W2AEW's video to do my testing. His stuff is great!

[starphot]

  Good video.  I got into electronics in 1972 when I joined the US Navy. There was a semiconductor curve tracer in my shop and I put a tunnel diode in it and observed the unique curve. There were articles available in the electronics mags in the 1972-1975 time frame for the tunnel diode circuits. I built an FM broadcast band micro-transmitter operated by 2 coin cells, microphone, tunnel diode and very little parts.

Joe