Testing transistors for gain/leakage, help?

JVanDe7 · Post by **JVanDe7** » 15 Mar 2012, 14:07

Finally bought one of these Peak Atlas doo-dads. I think it'll really enhance my builds because I'll be able to select the right trannies for the job and take notes to remind me which gain ranges I like in which circuits, etc. Thanks again for the help here fellas.

PokeyPete · Post by **PokeyPete** » 16 Mar 2012, 01:34

JVanDe7 wrote:Finally bought one of these Peak Atlas doo-dads. I think it'll really enhance my builds because I'll be able to select the right trannies for the job and take notes to remind me which gain ranges I like in which circuits, etc. Thanks again for the help here fellas.

I'm sure that Peak Atlas unit will be a handy tool for you, very accurate for silicon transistors; but, I'm not convinced that it will
prove accurate for germanium transistors. I don't own one so I admit that I'm speaking from ignorance (please inform me if you
know I'm in error), but I think your gain readings will NOT take leakage into consideration. I'm thinking that what you will be
shown as gain will actually be "gain + leakage" and will read higher than the true gain of the transistor. I posed this question on
another thread here about this meter hoping a Peak Atlas owner would enlighten me, but I don't remember getting a reply one
way or another. Now that you have one, maybe you will run tests using the meter readings, Small Bear method, and the Geofex
method and report your findings here. I hope you do!

LucifersTrip · Post by **LucifersTrip** » 16 Mar 2012, 04:31

I actually did a bunch of tests when I bought one...here are some results:
https://www.diystompboxes.com/smfforum/ ... ic=91115.0

bottom line for me is that it's excellent for modern silicon, pretty bad for vintage silicon, gives similar results to R.G's method for germanium hfe (it does report leakage separately) and is excellent for lug identification

phibes · Post by **phibes** » 16 Mar 2012, 04:49

The peak accounts for leakage. I still don't know why so many people hate on such an awesome product.

mictester · Post by **mictester** » 16 Mar 2012, 06:16

PokeyPete wrote:
JVanDe7 wrote:Finally bought one of these Peak Atlas doo-dads. I think it'll really enhance my builds because I'll be able to select the right trannies for the job and take notes to remind me which gain ranges I like in which circuits, etc. Thanks again for the help here fellas.
I'm sure that Peak Atlas unit will be a handy tool for you, very accurate for silicon transistors; but, I'm not convinced that it will
prove accurate for germanium transistors. I don't own one so I admit that I'm speaking from ignorance (please inform me if you
know I'm in error), but I think your gain readings will NOT take leakage into consideration. I'm thinking that what you will be
shown as gain will actually be "gain + leakage" and will read higher than the true gain of the transistor. I posed this question on
another thread here about this meter hoping a Peak Atlas owner would enlighten me, but I don't remember getting a reply one
way or another. Now that you have one, maybe you will run tests using the meter readings, Small Bear method, and the Geofex
method and report your findings here. I hope you do!

I have a Peak Atlas DCA55. This is my second one - a colleague borrowed and didn't return the first one. It's superb with ANY type of transistor, diode, FET, thyristor or triac. It gives accurate indications of hfe and leakage and identifies the leads of unknown or unmarked devices.

Any time I've used it with Germanium transistors (or diodes) it's worked fine. In time saved, it's a real money-maker!

mictester · Post by **mictester** » 16 Mar 2012, 06:19

phibes wrote:The peak accounts for leakage. I still don't know why so many people hate on such an awesome product.

It's probably because they either don't understand it, or they suffer from "not invented here" syndrome.

It never ceases to amaze me that Americans insist on "American Made" when the foreign products are so much better (Detroit NEVER made a car half as good as any BMW or Toyota!).

mictester · Post by **mictester** » 16 Mar 2012, 06:21

LucifersTrip wrote:I actually did a bunch of tests when I bought one...here are some results:
https://www.diystompboxes.com/smfforum/ ... ic=91115.0

bottom line for me is that it's excellent for modern silicon, pretty bad for vintage silicon, gives similar results to R.G's method for germanium hfe (it does report leakage separately) and is excellent for lug identification

In my extensive experience, it works really well on ALL discrete semiconductors!

PokeyPete · Post by **PokeyPete** » 16 Mar 2012, 06:39

phibes wrote:The peak accounts for leakage. I still don't know why so many people hate on such an awesome product.

Come on!!! I'm not "hating on it". I consider it a tool. And a tool will serve you best when you understand
its limitations. I was simply inquiring about those limitations.

mictester, thank you for clearing that up for me. In the advertisements that I read quite awhile back, leakage
wasn't covered to my satisfaction and I wasn't aware of how it dealt with it (if at all).

LucifersTrip, your results are interesting and surprising. They are not at all what I would expect.

PokeyPete · Post by **PokeyPete** » 16 Mar 2012, 07:17

mictester wrote:It's probably because they either don't understand it, or they suffer from "not invented here" syndrome.

BULL!!!

mictester wrote:It never ceases to amaze me that Americans insist on "American Made" when the foreign products are so much better (Detroit NEVER made a car half as good as any BMW or Toyota!).

Damn, mictester, you're full of ..it today. There was a time when the U.S. had a booming industrial base, and most imports
were cheap junk. The quality of import cars have only been acceptable in the past 2 or 3 decades. Before that, not so much!
I remember friends having cute little Triumphs and MG Midgets....cute, but stayed in the shop way too much for enjoyment.
No one I know kept them for very long, just too unreliable.

LucifersTrip · Post by **LucifersTrip** » 16 Mar 2012, 08:47

mictester wrote:
LucifersTrip wrote:I actually did a bunch of tests when I bought one...here are some results:
https://www.diystompboxes.com/smfforum/ ... ic=91115.0

bottom line for me is that it's excellent for modern silicon, pretty bad for vintage silicon, gives similar results to R.G's method for germanium hfe (it does report leakage separately) and is excellent for lug identification
In my extensive experience, it works really well on ALL discrete semiconductors!

I tested a sh*toload of vintage silcon (early to mid 60's with gains of 15+) and on the failures, the Peak reported mainly "darlington", some "common diode network", a few "digital transistor", and one "germanium".

These were the most recent failures: 2N3640, 2N2184, 2N2182, 2N493B, 2N2274, 2N2183, RCA34594, 2N2185, 2N2646, 2N2187

I tested at least 2 of each with the same results and also tested the transistors in circuits to make sure they weren't bad.

I bought it mainly for ge's and lug ID, so no big deal...

btw, another good thing about it...it shows forward voltage of diodes (or transistor junctions) and ID's cathode/anode of diodes.

JVanDe7 · Post by **JVanDe7** » 16 Mar 2012, 12:35

Well I had no idea this would cause such a ruckus... heh.
For $70 it's basically a steal, I think. I plan to use it for gain testing both Ge and Si transistors and it allows me to skip the datasheet lookup for pinout info. Well worth the money right there.
Plus instead of buying a pairs/sets of transistors from SmallBear or similar, I can get a larger qty of rough sorted and then hunt for the hFe that I want for the build in question. I could even test the leftovers and sell those on ebay perhaps? Lots of possibilities.

Sure I could have just built a little circuit and done the math and all that but this is way easier and it also tells me E-C-B

If the device was $500 or something I could understand some complaining but for $70 whats not to like?

mojah63 · Post by **mojah63** » 16 Mar 2012, 18:01

LucifersTrip wrote:I actually did a bunch of tests when I bought one...here are some results:
https://www.diystompboxes.com/smfforum/ ... ic=91115.0

bottom line for me is that it's excellent for modern silicon, pretty bad for vintage silicon, gives similar results to R.G's method for germanium hfe (it does report leakage separately) and is excellent for lug identification

Yeah I read your thread and I've been wondering lately about HFE gains people publish on projects. I went back to RG Keen's method using a traceable precision voltmeter and resistance bridge to set up the circuit for exactly 9v and resistors for exactly the published values . I got sick of having 3 meters give me 3 different results for the same transistor... I did find as you that RG's method seems to result in lower HFE readings than any transistor testing meter I've used for silicon transistors...

phibes · Post by **phibes** » 16 Mar 2012, 18:54

PokeyPete wrote:
phibes wrote:The peak accounts for leakage. I still don't know why so many people hate on such an awesome product.
Come on!!! I'm not "hating on it". I consider it a tool. And a tool will serve you best when you understand
its limitations. I was simply inquiring about those limitations.

mictester, thank you for clearing that up for me. In the advertisements that I read quite awhile back, leakage
wasn't covered to my satisfaction and I wasn't aware of how it dealt with it (if at all).

LucifersTrip, your results are interesting and surprising. They are not at all what I would expect.

No worries Pete, I wasn't referring to you on that. I've just noticed that in other posts people still seem to be skeptical of it when there a lot of praising posts. I'm sure there will be more threads in the future.

As far as the errors go, the only time I've gotten em has been with bad transistors. I've never got a bad reading on silicons though.

mictester · Post by **mictester** » 17 Mar 2012, 10:36

LucifersTrip wrote:
mictester wrote:
LucifersTrip wrote:I actually did a bunch of tests when I bought one...here are some results:
https://www.diystompboxes.com/smfforum/ ... ic=91115.0

bottom line for me is that it's excellent for modern silicon, pretty bad for vintage silicon, gives similar results to R.G's method for germanium hfe (it does report leakage separately) and is excellent for lug identification
In my extensive experience, it works really well on ALL discrete semiconductors!
I tested a sh*toload of vintage silcon (early to mid 60's with gains of 15+) and on the failures, the Peak reported mainly "darlington", some "common diode network", a few "digital transistor", and one "germanium".

These were the most recent failures: 2N3640, 2N2184, 2N2182, 2N493B, 2N2274, 2N2183, RCA34594, 2N2185, 2N2646, 2N2187

I tested at least 2 of each with the same results and also tested the transistors in circuits to make sure they weren't bad.

I bought it mainly for ge's and lug ID, so no big deal...

btw, another good thing about it...it shows forward voltage of diodes (or transistor junctions) and ID's cathode/anode of diodes.

I've just tried mine with 2N3640 and 2N2182 and a bunch of other miscellaneous devices - all correctly detected, and I compared the hfe with another (more expensive) meter and the results were almost identical. Perhaps yours is broken?

By the way: the "not invented here" syndrome isn't exclusively American - we have similar over here in Europe.

LucifersTrip · Post by **LucifersTrip** » 17 Mar 2012, 22:42

mictester wrote:
I've just tried mine with 2N3640 and 2N2182 and a bunch of other miscellaneous devices - all correctly detected, and I compared the hfe with another (more expensive) meter and the results were almost identical. Perhaps yours is broken?

nope, same result on my buddy's Peak...you're most likely testing more modern versions of the same transistors....try testing a bunch of late 50's and early 60's grown junction, alloy junction, "bullet" types, etc...

I just re-tested a bag of NOS bullet type 2182's and 2184's. All the hfe's are pretty accurate (20-30), but the Peak showed em all as Darlington

phibes · Post by **phibes** » 18 Mar 2012, 04:07

Yeah all those names looked pretty gnarly. There's gotta be something in the transistors specs that makes the Peak freak.

PokeyPete · Post by **PokeyPete** » 18 Mar 2012, 08:35

Dont know, just thinking.....we know that Darlingtons have two PN junctions between base and emitter.
Perhaps the Peak is reading a higher voltage drop than the normal .6 or .7........perhaps it is reading a
volt or more. You could have it read just the base and emitter leads and give you a diode reading and
see what is says.

LucifersTrip · Post by **LucifersTrip** » 19 Mar 2012, 06:31

PokeyPete wrote:Dont know, just thinking.....we know that Darlingtons have two PN junctions between base and emitter.
Perhaps the Peak is reading a higher voltage drop than the normal .6 or .7........perhaps it is reading a
volt or more. You could have it read just the base and emitter leads and give you a diode reading and
see what is says.

that's seems to be what's happening...just checked. the Peak is reading a fwd voltage of > 1.25v for the lot. It actually read the some NOS gold legged 2N2183's as "digital transistor", with fwd voltage of > 2v

mictester · Post by **mictester** » 19 Mar 2012, 15:12

LucifersTrip wrote:The Peak is reading a fwd voltage of > 1.25v for the lot. It actually read the some NOS gold legged 2N2183's as "digital transistor", with fwd voltage of > 2v

That's weird! Mine worked OK with the two 2N2183s that I had in the junk box. It also correctly identified the 2SC2368s that I tried it with - these have a potential divider resistor network from the base and emitter, so my old transistor tester just saw them as base - emitter leaky! The Peak got the hfe quite close (it displayed 120, I measured - the hard way - 114).

For the price, it's an amazing piece of kit. I've also got their LCR meter

phibes · Post by **phibes** » 19 Mar 2012, 19:21

How is the LCR? That's been on my list of junk to buy for a long time...