[sdiy] Synthex Oscillator

[rsdio at audiobanshee.com]

More notes on the Synthex digital-controlled current sink.

I'm sure folks who looked at the schematic noticed that the resistors are not exactly twice the value of each smaller neighbor.

I filled out a quick spreadsheet and learned that the 7.15 kΩ resistor is 4.67% low, the 470 Ω resistor is 2.08% low, and the 14.7 Ω resistor is 2% low. Meanwhile, the 62 Ω resistor is 3.33% high and the 1 MΩ resistor is 4.17% high. I picked the 30 kΩ, 120 kΩ, and 240 kΩ resistor values as the reference.

I graphed the weights resulting from binary combinations, and noticed that there are a couple of points in the 256-code range where the resistance actually drops when the code increases, simply because the resistor values don't add up in a way that produces a monotonically increasing value. Of course, this simple analysis is beside the point, because the resistors are combined in parallel. This is the light gray line on the attached image.

I then graphed the voltage divider output that would be produced by the parallel combinations of resistors across the full 256-code range, assuming that there is an ideal op-amp voltage follower. In other words, this ignores the significant current flowing out of the capacitor, but I still thought it would be interesting to see how well this circuit would perform if it actually were a Voltage output DAC rather than a digitally-controlled current sink. This is the blue line in the image.

Here's the graph, although I'm not sure whether the mailing list will strip the attachment.



Brian Willoughby

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20170710/4d58e235/attachment.htm>