Transistor Matching

[@jj74qformerlyjailbreak3]

Thank you. A year ago my idea of matching was by package, and numbers on the device. Now I enjoy making them line up and earn their numbers. Some gripe and moan about it, others just do their job, then there are those you lead to water, and they use the bathroom instead. Great video. None of my time is waisted. Even when I thought it was. God Bless.

[@cr0wland]

Good video AO! This is something a lot of newcomers struggle with (as did I).

[@zeiram8230]

I just went through a bag of 557s looking for matches the other day. I had pretty good luck.

[@ke9tv]

The legendary Bob Pease called breadboards 'the white slab of trouble.' I'm not sure I'd go to quite this much work. I don't know how often I need a 10 ppm match. (I usually use emitter resistors, which hides some of the mismatch.) The last batches of DMMT5401's and DMMT5551's that I got in all seem to have a Vbe match within about 300 microvolts (The data sheet says 1 mV typical, 2mV max, so maybe I got lucky, or maybe their production process is better than they claim.) Having both transistors in one package is convenient. I solder it into a breakout board, epoxy a tiny RTD on the back of the package, and glue on a piece of blue foam. An opamp and a couple of resistors, and I have a temperature compensated matched pair suitable for a 1000:1 log converter. For currrent mirrors and current references, I really like the REF200. Searching for transistor matches for that application is worth a lot more than ten bucks of my time.

[@kellerdev2905]

So cool, I imagined you have to match everytime you need a matched pair but with this technique you can do it once and have alot of matched pairs

[@Tripmaster2006]

thx for the explanation...Im new to this and overly pedant over some instructions (which just make me uncomfortable to try), so its good to see, that this more easygoing approach works :)

[@frankstein7481]

There is a guy called elektrouwe, he invented the yatma. 8-channel Transistor matcher. Fairly Simple circuit. Highly functional.

[@rty1955]

I restore 2" quad broadcast video tape machines. I use a curve tracer to match transistors

[@mickeygreenbaum9037]

This looks great! Just what I was looking for to construct a phaser guitar effect pedal which needs 4 matched transistors. Would I be able to use a 9 volt battery though, instead of the 12 volt power supply that you are using? If so, would I still need a "ground" connection on the circuit board? How would I go about that?

[@analog_output]

I'm not an expert on this but I should think you could use not one but two 9 V batteries. Connect the + terminal of one to the - terminal of the other, and that's where the ground connection goes. Then the other + terminal goes to the 3.9k resistors (where I show +12V) and the other - terminal goes to the 100k resistors (where I show -12V). In essence what you have there is a dual ±9 V supply. It might be necessary to scale the resistors down accordingly, to something like 3k and 68k. Try it anyway, see how it goes.

[@mickeygreenbaum9037]

@@analog_output Thanks for the quick reply. I'd have never come up with that myself. I'll definitely be trying it out. I was surprised that there didn't seem to be much information out there showing an accurate, yet simple way to match transistors. But your way appears to do the trick while being quite easy to follow, and you posted it just in time too (i.e., just ten days ago). Thanks for taking the time and effort to do it!

[@cobar5342]

Great video. Perhaps you could go into the board you made in more detail to allow your viewers more benefit

[@iblesbosuok]

Magnificent

[@cskeet1320]

Great video, thanks. Do I just reverse the polarity for testing pnp transistors?

[@analog_output]

Now there's a coincidence for the ages. I just this morning matched some PNP transistors for the first time and had to look that up. The circuit's very similar to the NPN one, but with the diode and the voltage connections reversed. See Ian Fritz's article (link in the description).

[@woosix7735]

in the pdf he sais that for building OTAs they need to be matched to less than 50uV. How would you mesure that? multimeters don't usually go below the millivolt range so I don't know how to do it.

[@RNMSC]

It might take a while to work it out, but this looks like a good use for an op-amp as a comparator, or simply use 2 op amps with matched resistor pairs to get higher values. Say a 100 ohm, and 1 k ohm for the inverting to get a 10 factor increase, turning the provided 50uV to a .5 mV difference, which should be visible on most reasonably good multimeters. (When I say matched resistor pairs, each op-amp is going to need one of each of the two resistances you select, and you want to make sure that the resistance of each value is a very close match. That can be difficult on budget multimeters. Part of that problem is that the more variance across the test equipment you're working with, the more inaccurate your results are going to be. You might use trimmer pots, with a common source voltage to tune the output of the op-amp circuits so that you can get them to a common value, or tune one so that the difference in output between the two is 0, then switch to testing transistors with that setup.)