Testing Ge Transistors

[soggybag]

I have a small stash of Ge transistors that I would like to test for gain. I built a small PCB with the testing circuits from GEOFEX. I have a few questions about the process.

I'm using a regulated power supply, it's an old Heath Kit thing, and a Radio Shack Digital Multimeter.

I notice when I put the transistors in and measure them they imediately show a high leakage. That is, the first measurement usually shows close to 1v. Over time this goes down. The longer I wait the more it seems to drop. After a minute or more what started as 1.21V is now .64v. If I wait a little longer it seems to drop more. Allof the transistors seem to do this.

Should I wait a while before measuring the devices to let them "settle"? Or should I take the first measurement?

Has anyone else used this system to measure Ge transistors, give me some advice to getting good readings.

[Fuzzy-Train]

I noticed that as well when testing some of my germs. Are you touching the metal can? Cause the heat from your hands will make the V higher... I would just try and only touch the leads so you're not sitting there waiting for it to go down. I still noticed that it goes down slowly even when I didn't heat it up with my hands.

But yeah, I would also like to know how long you should wait.

[soggybag]

I'm not touching the transistors when they are in the tester. But Picking them up and putting them in the tester requires some contact, which probably changes the temperature a bit.

[soggybag]

It seems there can be a large change over time. The meter might show 1.2V when the transistor is initially put in the socket. After 5 to 10 minute it might show .75v.

[soggybag]

What's a good range of leakage for Ge transistors that will be useable for Fuzz Face, Range Master and Tonebenders?

I tested 20 transistors and came up with a wide range of leakage and gain. Using R.G.s method they range in leakage from .28 to 1.71. What' s the cutoff for leakage for useful devices? 1V or greater?

The gains range from 35 to 128

[soulsonic]

I've used ones with leakage as high as close to .7mA, but anything 1mA or higher is probably too much to deal with. I know the "magic number" many people shoot for is .3 or less, but I've used ones around .5 without any noticeable problems.

Those gain ranges seem pretty normal to me and will likely work in most circuits. I'd go with the lower gain ones for Rangemasters and usually the higher gain ones in Fuzz Faces and Tonebenders - though sometimes lower gain ones can sound really good in a fuzz (like around 60 to 70). I say build them with sockets so you can try swapping.

[soggybag]

Let me know if I'm understanding the R.G. test system correctly. With the switch open the voltage at the collector is the leakage. There is no conversion needed. I ask because, leakage is always discussed in mA, but the tester is read in V.

[Fuzzy-Train]

The way I understood it is:

Switch off is just voltage. You take the voltage then divide by 2.47 which will give you your uA of leakage. Then you take your 2nd reading with the switch engaged and you subtract your leakage from your 2nd reading and multiply by 100 which will give you your true gain.

Here's an example from my batch:

1st reading switch off: .52v

.52 / 2.472 = .210uA of leakage

2nd reading switch engaged: 1.07v

1.07 - .21 = .086 * 100  = 86Hfe

I think that's how it's done but I'm not sure.

[soggybag]

Thanks, the divide by 2.47 part makes the batch I have not look as bad as they did at first.

[soggybag]

One more question. Do you subtract the leakage voltage from the gain voltage before calculating the gain?

For example I have a transistor that shows .61V leakage, and 1.89V gain. Do I calculate gain as

leakage 25uA (100*.61/2.472)
gain less leakage 103 ((100*(1.89V-.61V))-25uA)

[jschwalls]

I found a multi-meter that actually measure the HFE of PNP and NPN trannys.... very easy just insert the transistors and after a few seconds it gives you the reading... cool that if you touch the transistor while testing the HFE changes as well...

[Dragonfly]

I found a multi-meter that actually measure the HFE of PNP and NPN trannys.... very easy just insert the transistors and after a few seconds it gives you the reading... cool that if you touch the transistor while testing the HFE changes as well...

Lots of DMM's will test transistors,  but theyre really only accurate for silicon devices. With germaniums, they won't read correctly due to leakage.

I believe that some of the "Peak" brand testers will read leakage and gain on germaniums....of you could just use the method outlined on www.geofex.com in the "technology of the fuzz face" article.

[Nico666]

The way I understood it is:

Switch off is just voltage. You take the voltage then divide by 2.47 which will give you your uA of leakage. Then you take your 2nd reading with the switch engaged and you subtract your leakage from your 2nd reading and multiply by 100 which will give you your true gain.

Here's an example from my batch:

1st reading switch off: .52v

.52 / 2.472 = .210uA of leakage

2nd reading switch engaged: 1.07v

1.07 - .21 = .086 * 100  = 86Hfe

I think that's how it's done but I'm not sure.

I think you're right for the leakage. For the gain, you should substract the 1st reading from the 2nd (that way you substarct V from V and not mA from V) and then multiply by 100. SO, in your example, you would get:

1.07-.52= .55*100= 55Hfe

At least that is what I got from th RG article.

[Fuzzy-Train]

^ Yeah, sorry I was wrong. Looks like my batch is perfect for Rangemasters or low gain FF's.

So soggybag the example you put up would be 128hfe true gain.

.61v / 2.472 = .25uA (rounded to nearest)

1.89v - .61v = 1.28v * 100 = 128hfe

[soggybag]

Thanks for the replies. This is starting to make sense.

What's the limit on leakage to make a device unsuitable for use? 100uA and greater?

[Fuzzy-Train]

I've used ones with leakage as high as close to .7mA, but anything 1mA or higher is probably too much to deal with. I know the "magic number" many people shoot for is .3 or less, but I've used ones around .5 without any noticeable problems.

There ya go.