This page is NOT to provide the reader with the theory of electronics and how transistor work in a circuit. It is to provide those interested analog music synthesizers and appreciation of tools and building techniques to ensure success in DIY projects. Some additional background info:
- Transistor Matching Here
- Ian Fritz Matching Here
- Transistor Matching Here
- MFOS Transistor Matching Here
Transistor Selection & Test Tools
You can buy cheap component testers on eBay, or various hobby shops. The work for “testing” and even identifying components, but they are really not accurate for measuring Hfe, or Vbe of a transistor. I don’t even think they are accurate for measuring capacitors either (there are some VCF circuits that require matched diodes and capacitors).
If you really get into electronics DIY and analog synthesizers, you should probably put some of the outlay into tools instead of more synths. Yea, I know, tools are… well just not as much fun. One tool I found incredibly valuable is the Peak DCA pro transistor tester. It is very accurate, and even has the capability of providing curve tracing when connected to a PC. I use this tool for testing transistors and measuring Hfe. This tool is also great for testing and identifying various transistors and JFets, prior to installing them into a project.
For complimentary pairs, i measure the Hfe of each transistor by type. all my NPN, then PNP, etc. If I need an NPN/PNP pair, I will take 25 of each, from 3 or 4 different lots or manufacturers. Say we require a 2n3906/2n3904 pair. I will take 100 2n3904 (25 from 4 lots) and 100 2N3906 (25 from 4 lots), and I may find 2 pairs that have the same Hfe. I have in the past required several hundred to get a couple pairs. Small signal transistors can have a wide Hfe specification.
Transistor Matching Tools and Techniques
What does it mean to “match transistors”? In a very general sense, many VCOs use a pair of transistors to provide a logarithmic response so that each increase in voltage will double the frequency output. Basically, given an input voltage, your VCO should increase some proportion exactly. This requires that the transistors have minimal differences in specifications between them, so as voltage input increases, then the circuit provides the exact, proportional output in response. This is how an analog VCO can stay in scale tune as the voltage increases by pressing the keys going up a synthesizer keyboard.
Specifically, transistors must be identical in the voltage across the Base and Emitter (Vbe) for like pairs (NPN/NPN or PNP/PNP), and identical in current gain (Hfe) for complimentary pairs (NPN/PNP). Other individual transistors may need to meet minimum current gain (Hfe) specifications for the circuit to work optimally. You can measure them with many transistor testing tools, but in order to really do this accurately, you need to ensure the voltage, current and temperature are stable, before taking a reading. Two methods commonly used is by measuring, and then labelling each transistor individually for it’s Vbe. The other is by placing two transistors into a circuit and testing as a pair, with an indicator letting you know they are a match or not.
On the MFOS site, Ray Wilson provides a series of circuit designs for matching and measuring transistors. In the pair of schematics above, note that the left side measures NPN, while the right side circuit is to be used for PNP. What I did, was build a single circuit that combined both tests into one. I did this by noting that the circuits are identical, except for the yellow highlighted areas in the schematic on the left. With the assistance of a switch, and ZIF socket, this tool was easy to build and use.
Basically, I run a heater set at 110 to 115 degrees F, and blow the warm air over the transistors waiting to be tested and the test rig. Then, I simply write down the value displayed on paper or tape, and then tack or place the transistor to that spot on foam or paper. You have to let the transistor sit for a good two minutes and allow the Vbe to settle.
You can buy a PCB and build your own transistor matching tool from Synthcube. This method is very simple as you do not need to measure all your transistor, just enough to get the number of pairs needed. This method, you place two transistors and then measure the difference between Vbe. If you measure +/- 1mV or better, zero ‘0’ on the multimeter, you are gold! Same process with the fan and/or heater. Also note, NEVER touch the transistors with your hands prior to measuring, it can effect the readings. Use tweezers to place them into the sockets. I pre-bend the legs and per-fit, before hand. Then place them off to the side to sit under the fan/heater for about 20 min prior to testing.
I actually use both methods. I use the Ray Wilson circuit first at 110 degrees F, and get several pairs, then I double check each pair with the Ian Fritz method after. Either method with work for matching, I just hate taking apart a matched pair due to the inability of a VCO to track properly. By double checking, I can usually find the error in my circuit without resorting to replacing the pair.
In some circuits, VCOs, VCFs the use of JFets are very common. Some designs will utilize a pair in the design. For VCOs, it can be critical the the JFets turn off at the same time. To turn the JFET off, the gate must me made negative with respect to the source. Vgs(off), or the “pinch-off voltage”, is when the current in the drain-source falls to zero and the JFET is fully off.
The Myriad JFet testing tool is perfect for this task. I use the same process as with matched pairs, measure and document a pile, and make matched pairs.
The article linked above is perfect for testing and selecting op amps for critical functional specs. Op amp operational specifications generally come in ranges, with the average performance in the middle somewhere. In many cases you want to select an op amp, based on the lowest Input Offset Voltage you can find (of one particular type). I take a couple dozen, say LM301s and measure for this by using a tools I built (based on the article in the link).
Because the offset voltage is very, very small and not measurable on most multimeters, this design uses a second op amp to boost the DUT output by 1000 so it is measurable. Then you divide the reading by 1000, and get the value in microvolts. Below, I am using the test rig to measure the Input Offset Voltage of LM201 Op Amps.