This guide about transistor matching is specifically about matching a pair of transistors for use in a log converter of a voltage controlled oscillator of a synthesiser. You may also need them in voltage controlled filter designs.
The idea is to find two transistors whose base–emitter voltage or Vbe is within a couple of millivolts of each other, under a constant current.
There are other applications where this type of matching is needed.
The more accurate the match, the better tuning of the voltage controlled oscillator across a larger range. It will also make the setting up of the VCO easier. If the transistors are badly matched you may struggle to get any form of consistency when it comes to trying to calibrate them.
Traditionally the way to do it was to use a stable current source and measure and record the Vbe of each transistor and record the results. You then simply selected two transistors with similar readings.
The problem with this method was that the voltage being measured was around 0.6V so you needed something very accurate to be able to do it properly. Another problem was one of temperature variation. Changes in the temperature of the transistor altered the voltage being recorded. If you touch the transistor with your fingers you can see the difference it makes to the readings. The values still change for up to a minute or two after you’ve touched them as they are still cooling and changes in the room temperature and draughts also throw the results out!
I was about to give up on being able to match transistors myself when I stumbled upon an article by Ian Fritz on the internet. He comes up with a brilliantly simple idea and circuit. As he points out in the article it’s actually the difference in the Vbe not the actual values that we are interested in. Measuring the difference is a lot easier. You can do it with a digital multimeter instead of an expensive digital voltmeter.
As you are using two transistors at the same time, if the ambient temperature changes it will be changing for both so wont mess up the results.
Ian goes into more detail in his article and you can read it here. The circuit is shown below for NPN and PNP transistors.
Transistor matching NPN
Transistor matching PNP
As you can probably guess because of the quite tiny nature of the voltage measurements, the two resistors need to very accurately matched.
I built the circuit to combine both PNP and NPN transistors by wiring in a switch. All it needs to do is swap the polarity of the supply voltage around and swap the polarity of the diode’s connections. I also included a trimmer instead of one of the resistors to get the values exact. You should build the circuit to get the two transistors physically next to each other so they will be at a similar temperature. I used a DIL socket for the transistors.
Plug two transistors in and you will get a reading on your DVM. Make sure you’ve set it to measure the smallest voltage possible. You’re looking for a reading of less than 2 millivolts. The smaller the better. If you 2 mV or less those transistors are pretty evenly matched and would be suitable for the application of a log converter in a voltage controlled oscillator or similar.
As Ian recommends, you can keep one transistor in place and keep swapping the second. If it doesn’t match just keep the ones that don’t match together. For example all the ones that read 3 millivolt together, 4 millivolt together and so on. Those should be a good match. You can try them together just to confirm this.
As you can see in the picture above the two BC337 transistors on test are within 1mV and so are a good match. It took a few minutes for it to settle at this reading and if you touch either transistor for a second or two you will see the voltage shoot up until it settles down again.
The transistors that you are testing should be kept together so they are e at the same temperature and if possible use plastic tweezers to handle them so your fingers don’t warm them. You can see that just by touching the transistors how much it affects the readings.
I didn’t get on too well inserting and removing the transistors with tweezers so I used my fingers but then let them remain on test for a few minutes. You can tell when your warming effect wears off as you will see the readings stabilise. It’s not a bad idea to leave them on test for a few minutes anyway just to confirm the results stay stable.
As you can see by using Ian’s circuit, transistor matching is within the realms of the electronic hobbyist and you won’t have to go out and buy any more expensive equipment.