\\[First published in the September 27, 1990 issue\\]
Recently I was invited to a meeting to see the results of a new high-performance, low-noise transistor project. I looked at the technical report. The new transistors were indeed quieter than the old ones. In fact, they were two to four orders of magnitude quieter than the conventional ones. I was suspicious. What was the test method? Oh, here is the test circuit (see the figure).
Now, I asked, were the betas high or low? I was told they were pretty low, but they can be brought up later. I explained that when the betas get low, if theyâ€™re not very well matched, the test circuitâ€™s output can pegâ€”right at the + or â€“ rail. Then, of course, the apparent output noise gets rather small. (Ohhhhh!)
When this circuit is running okay, the current noise is amplified by a big resistance: 1 MÎ©Â X (N+1), where N is the closed loop gain, (RF/R1). So, the output will include: (1+ noise of Q1A) + (I-noise of Q1B) X 1000 MÎ©. However, the offset current (I+) â€“ (I-) will also be magnified by 1000 MÎ©. So even 9 nA of offset current will cause the op ampâ€™s output to try to go to +9 V. If the power supply wonâ€™t let the output get to a fair balance, the output will peg. Naturally the output becomes very quietâ€”the circuit has stopped amplifying the noise.
I also pointed out that ideally the circuit could measure the base-current noise of the transistor, or device under test (DUT). But the layout of the circuit is quite critical. Just 1 pF of capacitance (CF) from the output to the base of Q1B will cause a lag in the response: 1000 MÎ©Â X 1 pF = 1 ms, so the noise will roll off above 160 Hz. You can make a layout with less than 0.1 pF, but you have to think about it and engineer it.
When we checked the test box, it was laid out very neatly: The output wire was bused alongside the summing-point (base of Q1B) wire, and the bandwidth was indeed less than 100 Hz. In a future column, I will talk about what a picofarad looks like and the harm it can do to you. So, even if the output wasnâ€™t pegged and you looked for the noise at 1 or 10 kHz, this test circuit would give an answer thatâ€™s considerably quieter than the theoretical minimum for the transistor. Now hereâ€™s a good place for a sanity check.
Itâ€™s not impossible to measure the noise of a transistorâ€™s base circuit, but you must have a suitable circuit. I wrote a paper back in 1968, and as I look at it today, the only things that changed are the names of the op amps. You canâ€™t buy any of those old discrete transistor, potted module op amps any more, but the testing approaches are just as valid. Maybe Iâ€™ll write an updated version. In this case, the problem was that a 1-MÎ©resistor and a noise gain of 100 or 1000 (provided by RF and R1) wasnâ€™t a good idea. The stray capacitances and the noise of the 1 MÎ©Â are detrimental to accuracy. Itâ€™s better to use a real 100-MÎ©Â or 1000-MÎ©Â resistor.
In fact, a 20-MÎ©Â or 5-MÎ©Â resistor is justified because it will still give plenty of signal-to-noise ratio, and a lot more bandwidth. More on how to do this in the next issue.
All for now. / Comments invited! / RAP / Robert A. Pease / Engineer