1

Is there anywhere I can download some current-voltage data for a bipolar junction transistor? (It doesn't matter that much which one.)

To explain further, for a simple BJT circuit we have the following variables. (Assuming NPN.)

  • Base-Emitter voltage bias
  • Base-Emitter current
  • Collector-Emitter voltage bias
  • Collector-Emitter current
  • Transistor HFE (beta) which is a function of temperature
  • Transistor temperature, which other variables such as the Collector-Emitter current, might have a second order effect from

I don't know if these are all independent - I believe they all are. Some might become dependent on each other if a complete circuit is drawn including an impedance on the base and collector (or emitter) nodes.

Are there any example devices which have (presumably large) tables of data available for download for measurements of these parameters?

  • Why do I want this?

I am interested in writing some simulations involving BJTs which requires more information than can be obtained from the typical transistor equations.

In other words, all of the non-linearities are important, and the behaviour in regions outside of as well as including the active region are also important. If some data source only has data for the active region, this is probably good enough.

user3728501
  • 227
  • 4
  • 12
  • 1
    From the manufacturer possibly. – Solar Mike Sep 26 '22 at 19:05
  • 1
    You could find a curve tracer and gather the data yourself, if you have the budget. You could probably find a Tek 370A for a few thousand bucks these days; it can save collected data to a 3¼" floppy disc, or send it over GPIB to anything that can communicate that way. – Hearth Sep 26 '22 at 19:14
  • @Hearth Out of budget unfortunatly - if nothing else I can get something very approximate with a couple of multi meters and enough time ... but there are automated ways of doing these things. Doing it by hand isn't really practical. – user3728501 Sep 26 '22 at 19:25
  • >Transistor HFE (beta) which is a function of temperature< ... not only. It depends of Early voltage, of Ic current .... Some characteristics may only be known with pulsed currents ... And some simulators use ten's of parameters ... More then "typical" equations. – Antonio51 Sep 26 '22 at 19:28
  • Can you use spice simulated data until you find better? – jonathanjo Sep 26 '22 at 19:59
  • 3
    @user3728501: [Have you got an Arduino?](https://josepheoff.github.io/posts/iv-1-toc) – JRE Sep 26 '22 at 20:09
  • Hfe varies so much it is an useless parameter. Under well-defined measurement conditions, manufacturing tolerance might be 100 to 300 for a given transistor model. – Justme Sep 26 '22 at 20:09
  • @JRE sadly you don't have any example data in your gitbub! – jonathanjo Sep 26 '22 at 21:18
  • Academic resources are likely the place to go. I don't have any cites unfortunately, but looking into where the SPICE teams got their data from might provide some clues. And Gummel and Poon, etc. Likely you'll just arrive at the latter model anyway; or other familiar models for MOSFETs. – Tim Williams Sep 26 '22 at 21:58
  • @user3728501 There's the [Mextram model](https://www.eng.auburn.edu/~niuguof/mextram/_downloads/MextramDefinition.pdf). It's pretty much the state of the art, these days. The [newest stuff](https://etd.auburn.edu/handle/10415/8210) won't be out for a bit, yet. And as others have pointed out, the model is just a model. Reality is quite variable, part to part, family to family, etc. So if you are just simulating, that's one thing. If you are designing for repeatable construction, that's another. – jonk Sep 26 '22 at 23:56
  • @JRE Arduinos have a 10 bit dac, which when noise is taken into account is probably reliable to about 8 bits. It would require a lot of work to build a completely custom Arduino based solution which was accurate to 10 bits... And it would have to be tested for linearity etc... It isn't a suitable solution. The non-linearities in a transistor in the active region (although they obviously differ by device) are likely to be much smaller than 1 part in 256, probably smaller than 1 part in 1000. So the Arduino just wouldn't work. – user3728501 Sep 27 '22 at 11:29
  • @user3728501: [Have you had a look at it?](https://electronics.stackexchange.com/questions/537899/what-is-this-effect-i-am-seeing-in-a-current-voltage-plot-of-a-2n2222-npn-transi) – JRE Sep 27 '22 at 12:04
  • @JRE I have used them before, which is how I know 8 bit performance is optimistic – user3728501 Sep 27 '22 at 14:19
  • Have you looked at the results? – JRE Sep 27 '22 at 14:20
  • @JRE I don't really wish to become drawn into what is essentially an argument between yourself and someone else on this site - and I have no involvement with either parties, however I am not particularly impressed with what you have presented. In particular I happen to know that the ADCs take a reference voltage, which you don't have in your design, and without which you do not have a stable reference. This is perhaps why your results appear to have lower levels of noise than I would otherwise expect. Did you make any measurements of noise on the supply? (cont...) – user3728501 Sep 27 '22 at 14:28
  • (... cont) This is really just not a good setup for measurement. There are no independent measurements of the voltages... It means that if your supply voltage is moving all over the place (which it will because the load is not constant) you will not likely see anything in the results, because the setup is not capable of measuring that... Regardless this is really a tangent to the question I asked. No where here did I ask "how can I measure X". I asked "where can I obtain measurements someone else has made of X". – user3728501 Sep 27 '22 at 14:32
  • If that isn't clear enough... When looking for small scale nonlinear effects, you cannot use a measurement device which is likely to be more non-linear than the effect to be measured, and with a higher level of noise such that any non-linear effects are obscured. There also has to be some way to calibrate the device. Even relatively low cost multimeters are calibrated in some way at the factory. – user3728501 Sep 27 '22 at 14:35

1 Answers1

1

I am interested in writing some simulations involving BJTs which requires more information than can be obtained from the typical transistor equations.

Get a simulator (which uses dozens of parameters to model a transistor) and do some simple graphs like this one (for a 2N7002 MOSFET that I had to hand): -

enter image description here

Given that you are trying to make a homespun simulation tool, you might as well develop it along the recognized lines used by several vendors.

Micro-cap is free BTW and comes with hundreds of transistor models. If you get used to it you'll find a ton of things you haven't even thought of yet in your quest to develop a simulator tool. Alternatively you may decide that there's no point re-inventing the wheel.

All micro-cap graphs available with numerical output of course (same for LTSpice too I believe).

Andy aka
  • 434,556
  • 28
  • 351
  • 777