Author Topic: Matching transistors room temprature  (Read 1443 times)

0 Members and 1 Guest are viewing this topic.

Offline Jan AudioTopic starter

  • Frequent Contributor
  • **
  • Posts: 820
  • Country: nl
Matching transistors room temprature
« on: May 07, 2019, 10:37:25 am »
So i am matching transistors again, this is difficult to do.
My room is steady temprature without heating 17,5 degrees.
With heating on to 20 degrees the temprature is not stable enough, the readings vary to much.
Is it bad to measure at cold temprature ?

i readed in a lab with climate control is 21 to 25 degrees ?
Should i wait for a warmer day ?

Some more info how i measure :
cost me 1 breadboard + cables, some flexible styrofoam to place around the row transistors.
The cables are to have my hand away from the transistors, it changes the reading when to close.

So how do hobbyists fix this without climate controlled lab ?
 

Offline jeremy

  • Super Contributor
  • ***
  • Posts: 1079
  • Country: au
Re: Matching transistors room temprature
« Reply #1 on: May 07, 2019, 11:04:30 am »
If you are going to run the transistors at 17.5 C, then it is probably best to match them also at 17.5 C. It doesn't really matter at what temperature you match them as long as they are also used as close as possible to that same temperature (or have a well characterised temperature coefficient). It's roughly the same problem with precision mechanical measurements (fancy CNC machines, optics, etc); the absolute temperature is not so important itself, it's more important that the temperature is stable.

Laboratory climate control is not the same in all contexts, but (for example) the US NIST Normal Temperature and Pressure (NTP) is 20 C. This is more out of necessity than anything else; you just have to pick a standard and stick with it, and 20 C is relatively easy to achieve in most parts of the world (and I suppose comfortable enough for humans).

You can find a list of common standards here: https://en.wikipedia.org/wiki/Standard_conditions_for_temperature_and_pressure
 

Offline Jan AudioTopic starter

  • Frequent Contributor
  • **
  • Posts: 820
  • Country: nl
Re: Matching transistors room temprature
« Reply #2 on: May 07, 2019, 01:51:13 pm »
Thanks.
Wont the transistors matched at 17,5c be matched at the temprature inside the enclosure ?
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23099
  • Country: gb
Re: Matching transistors room temprature
« Reply #3 on: May 07, 2019, 02:27:23 pm »
I just don't do it. If I need a matched pair I usually bodge a CA3046 / LM194/394 array in somehow.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Matching transistors room temprature
« Reply #4 on: May 07, 2019, 08:21:01 pm »
What characteristics are you matching?  I usually settle for simple Vbe measurements to get within a couple millivolts and it does not take long enough for the temperature to change significantly.  For transistors in the same lot, Vbe correlates well with other characteristics.

A more complex method is to make a delta Vbe measurement to get the temperature of the transistor junction while forcing it to a specific temperature and then making the measurement of whatever is being matched.
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Matching transistors room temprature
« Reply #5 on: May 07, 2019, 08:34:08 pm »
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7678
  • Country: ca
Re: Matching transistors room temprature
« Reply #6 on: May 07, 2019, 09:06:12 pm »
Are you matching beta or VBE?
Touching the transistors to put them in the breadboard then takes time for them to cool off.
Using a high test current heats them up, this is the most frustrating because readings are always moving.
Add some capacitance like 0.1uF to prevent RF pickup from getting amplified, which also makes readings wander.
A box to prevent convection currents, drafts, heating system, sunlight, IR etc. helps.

I don't see a point to super-matching (<2%) because your end circuit will have thermal gradients and they'll end up running at different temperatures anyway. Like in an audio power amplifier. You can apply heat-shrink to a diff. pair though.
 

Offline Jan AudioTopic starter

  • Frequent Contributor
  • **
  • Posts: 820
  • Country: nl
Re: Matching transistors room temprature
« Reply #7 on: May 08, 2019, 02:13:55 pm »
I am measuring for HFE for exponential converters.

Those 1 part measuring devices i think are no good.
Like you say you cant toutch them.

I add a whole row in heat isolating material and measure from a distance.

I dont know why the readings change, i will try a 100n cap, it wont work, the readings change over time, you have to measure really fast before it changes again.
 

Offline Jan AudioTopic starter

  • Frequent Contributor
  • **
  • Posts: 820
  • Country: nl
Re: Matching transistors room temprature
« Reply #8 on: May 08, 2019, 02:17:55 pm »
I don't see a point to super-matching (<2%) because your end circuit will have thermal gradients and they'll end up running at different temperatures anyway. Like in an audio power amplifier. You can apply heat-shrink to a diff. pair though.

I use heat paste from computer, 2 transistors + tempco resistor together.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7678
  • Country: ca
Re: Matching transistors room temprature
« Reply #9 on: May 08, 2019, 06:03:01 pm »
Dr. Leach simple blurb on matching diff. amp transistors for his power amplifiers. http://leachlegacy.ece.gatech.edu/lowtim/part2.html
I add a 0.1uF E-B capacitor to shunt RF if the test fixture has long leads. You can also put a scope on the transistor under test to see if anything odd is happening.
Cheap DMM's apply maybe 3V to measure hFE so that is not so great if you have high VCE expected.

If you use multiple power supplies (one for base drive, another for collector, or bi-polar), be careful of the sequence to turn them on/off so you do not reverse or damage a transistor.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf