Transistor Tester Current Levels

[Stocker]

I am about 1/2 through with building my transistor tester. When done, it will be able to show simultaneous readings of Vbe, Ib and Ic. Current capability for Ic will be up to several amperes, hopefully 10 or more amps. This is fine for power transistors but small signal has got me thinking, now that I have come to building that part.

Currently, I can set this thing up to read as high as 1.9999A Ib. (never anticipate reaching that though! ). Is this resolution enough for the first-watt-best testing of small signal transistors? 0.1mA is not much, but then how much current is going to the base of the first differential pair in an amplifier at ***LOW*** listening volume? I can set this thing to measure down to single microamperes but that seems a bit much, and would involve another switch on an increasingly complicated control panel.

What *minimum* level of base current should I consider when testing transistors for audio equipment?




p.s., when this is done, stand by for some DIY amplification!

[AndrewT]

Hi,
it depends on what level of resolution you want at the lower Ib currents.
If you test a signal transistor at Ie = 1mA and gain is say 800 then Ib = 1.25uA. I think with very low Ie and high gain you might just get down to Ib min > 400nA. When matching transistors, is two significant figures close enough for you? or do you think you would ever need three significant figues?
So your answer is either 100nA or 10nA for a resolution. Absolute accuracy here is not imperative but repeatable resolution is important.

[Christer]

Originally posted by Stocker
... but then how much current is going to the base of the first differential pair in an amplifier at ***LOW*** listening volume?

This makes me wonder if you are going to test AC parameters or DC parameters? The signal currents (AC) will vary with listening lever and could be very very low, but the DC bias currents won't change with listening lever. Andrews answer applies to the DC case. For AC, I don't think it is useful to test extremely low values.

[Stocker]

Christer: The design is not quite finalized (obviously). It could be AC but I am thinking two (or more) transistors that have similar properties at several levels of DC will have similar properites with AC as well.

AndrewT: Wow that's pretty low current. If I gave it a full-scale range of 1.9999mA, the resolution is 100nA... I think that should be good enough for me. Unfortunately it means minor revisions to the plans but it should be doable. Actually, it means a pretty simple variation now that I think of it, but this chassis is getting tighter and tighter inside. I really want to keep it small but function may have to dictate form for once.

Thanks for the responses guys. I am trying to get this done right once and never again. I was working on it during slow times at work... the fact that it's been over a year tells you how much slow time we've had at work... So now it's rapidly approaching the #1 spot on the project list at home and I am trying to nail down the final design.

So that means now is the time to throw open the door to suggestions. What features would your dream/ideal transistor tester have?

[pinkmouse]

Originally posted by Stocker
... What features would your dream/ideal transistor tester have?

Thermostatically heated heatsink block to test devices at temperature. Label printer. Fully automated.

Well you did ask!

[Christer]

The de Luxe version would be to control it from the PC so you can use the PC as a curve tracer. Then there are plenty of things you could do in software, like extracting derived parameters, matching on arbitrary parameters etc.

[Stocker]

No really, swing for the fences guys. Solid-gold wires anyone?

Okay, I did ask.

The thermostatically controlled heatsink... I can do that with parts on the shelf already. Mount a screw clamp and it's done. I'll have to think about that some. It gets to 75deg. C in the current configuration, but I am pretty sure I can adjust it up or down a bit. Variable temperature is likely not going to happen in this implementation.

I was doing some calculations last night... for 100nA, I'm looking at some pretty huge gain on the voltage to get to the full-scale for my meter... Time fore multi-stage low noise low-offset buffering action I think.

I toyed with a peak/hold circuit setup, but that is unlikely to make it in there unless I start matching a buncho transistors and find it would alleviate a huge hassle. I can't see it being worth that much effort to remember 2 or 3 numbers long enough to write them down (yet).

For a *manual* trasistor tester, any other ideas?

[AndrewT]

Hi,
I hope you are not talking about amplifying up to give a full scale reading of 100nA.
My suggestion was for a resolution of 100nA or 10nA. This will be the last digit on your readout.
I recall your scale was 1.9999mA =1999.9uA. the last digit is already 100nA but I followed up by saying that it should be repeatable.

[Stocker]

CTM yes I was not thinking straight when I was thinking of the x1BILLION multiplication! please forgive the 2AM maths! We are still on the same page after all, even if half-asleep while reading it...

I did a lot of calculating and came to the conclusion that such a level of resolution would actually be seen by the amplifier's input and I'll want to test for it in such a case.

MSD 1mV is LSD 100nV... much less amplification required... fortunately... That gets me from top to bottom in two ranges with a little overlap, which is nice.

I am considering, leaning toward and liking the maths for nice round numbers for my resistances which will limit max. current to 60mA Ic... does anyone think I should go higher than this?

Oh yes, AndrewT, I just realized that I didn't answer one of your questions... I think I would like 2 significant digits of resolution. Wait, that gets me back to 10nA LSD on my readout.

OK guys give me another day or two to figure out how to do the current reading.

p.s. I did some looking around the web and 'transistor testers' are mostly a completely different feature set to this one... I think I should start calling it a transistor matcher, since that's what I'll be doing with it anyhow...

[jackinnj]

Anyone who purchases a PMD1208LS from Measurement Computing ($99 I think) can have the code (written in Softwire with Visual Basic. Net) for this transistor tester -- the analog outs will source 15ma Max so there is a bit of jiggering for larger devices -- of course you can use almost any of the MCC USB devices:



i could interface an RS232 label printer -- but then would have to write code for all the bells and whistles.

the Measurement Computing data acquisition card could also be used in the bench power supply which I described a few weeks ago. All you have to do is load a different program for the device.