Rybow
Well-known member
I made good use of adding resistors in series while I was testing FET's to get non standard values. Adding resistors in parallel works well too. I just happened to have lots of smaller values. I tried a trimmer as well, and it worked well enough, but it was just too fiddily for me.
I started off by measuring the drain voltage on each FET with a 10k source R. One thing I noticed, and I am not sure if this is true of all 2n3819's, but the higher the Drain voltage at 10K, the lower the source voltage with drain adjusted to 10V. My measurements of my 10 FET's all point to this correlation. My best FET 10.5v/2.1v @6.81k (fine tuning needed on the R) gave 19v with the 10k source R. My "worst" FET gave me just over 4v @10k source R, and over 4v @ source with drain adjusted to 10v. This gives more credence to the optimal 3.9k source R. So a quick and dirty way to sort a hundred FET's would be to pick a static source R value, measure the drain voltages, and pick the FET's with higher drain voltages. Then adjust as normal. Not sure if I have enough data to come up with a ratio.
The rotary switch plus trimmer looks like a nice macro/micro way of doing it. To tell you the truth, I really didn't mind swapping resistors in and out while adding more in series to get the bias adjusted. Turning off the phantom inbetween swaps of course. I sort of like the idea of getting a trimmer set to where the bias is right, and then just popping that trimmer right into the circuit. Then you can make more adjustments by ear if needed.
Edit: Just picked up an XR 2206 chip to build a tone generator. I'll it up to do a 1khz sine wave initially. Probably do this after the mic is constructed with a 50pf cap in place of the capsule for testing noise and the like. Then I can just plug it into a preamp with some oscilloscope software running to check out whats going on.
I started off by measuring the drain voltage on each FET with a 10k source R. One thing I noticed, and I am not sure if this is true of all 2n3819's, but the higher the Drain voltage at 10K, the lower the source voltage with drain adjusted to 10V. My measurements of my 10 FET's all point to this correlation. My best FET 10.5v/2.1v @6.81k (fine tuning needed on the R) gave 19v with the 10k source R. My "worst" FET gave me just over 4v @10k source R, and over 4v @ source with drain adjusted to 10v. This gives more credence to the optimal 3.9k source R. So a quick and dirty way to sort a hundred FET's would be to pick a static source R value, measure the drain voltages, and pick the FET's with higher drain voltages. Then adjust as normal. Not sure if I have enough data to come up with a ratio.
The rotary switch plus trimmer looks like a nice macro/micro way of doing it. To tell you the truth, I really didn't mind swapping resistors in and out while adding more in series to get the bias adjusted. Turning off the phantom inbetween swaps of course. I sort of like the idea of getting a trimmer set to where the bias is right, and then just popping that trimmer right into the circuit. Then you can make more adjustments by ear if needed.
Edit: Just picked up an XR 2206 chip to build a tone generator. I'll it up to do a 1khz sine wave initially. Probably do this after the mic is constructed with a 50pf cap in place of the capsule for testing noise and the like. Then I can just plug it into a preamp with some oscilloscope software running to check out whats going on.