White Noise above certain cutoff points on SMR4 mkII board

Hi, this is my first post and the Shruthi-1 is my first electronics project, so right off the bat I’d like to apologise for any ignorance on my part… Anyway, to my question…

Having ‘finished’ building it, everything appeared to work fine, tested all the voltage points successfully, all LEDs, the LCD screen, pots, buttons, MIDI I/O, Audio I/O, no problem. But, when I came to tune the filter I could not make it self-oscillate. I then discovered that the filter breaks up into white noise above a certain cutoff point. The exact point at which this happens varies according to resonance and midi note value, but for instance, playing C3, with resonance at 63, the noise appears at cutoff 111, and above. In the preceding 25 cutoff values I find a kind of high pitched ring mod sound that gets slightly louder before finally breaking into noise at 111. I guess it goes without saying, but to confirm, yes, both oscillators are off, I’m pretty confident I’m only hearing the filter board. Does anyone have any pointers as to what I may’ve done wrong?

Thanks.

This behaviour is quite normal, every Filter breaks somewhere with to high resonance values. Either you have swapped some Resistors or you have a very unlucky summation of tolerances, id look for the first and check the Resistors around IC 4, namely R21 (18K) R20 (2.2k) C17 (100N) Q4 3906 and further around IC3 R18 (1K) R17(220R) R16 (10K) R19 (10k) C15 (220N)

If you wonder why i am so smart, check the schematics on the lower left theres the connector from the digital board J4 , just follow the flow of the RES (onance) . . . .

This is normal behaviour.

One thing that defines the Shruthi-1 “gritty” sound is the use of 1-bit PWM instead of DACs. The drawback of this technique (without which the Shruthi-1 would more expensive and would offer less synthesis options) is that the audio signal coming from the oscillators has a strong frequency component at about 40kHz. This is usually not a problem because it’s not audible, and because it’s going to be killed by the filter anyway. However, when the resonance is high and the cutoff is high (outside the audio range), the intermodulation between this 40kHz tone and the high-pitched sine wave emitted by the filter in self-oscillation can get into the audio range (remember the product to sum trig identities?)

  • Suggestion 1: don’t bother with it.
  • Suggestion 2: increase C13 (and lose in brightness / punchiness).

I’m still surprised that you can get this tone without getting the filter to self oscillate. What do you have when cutoff is set to a middle range value (say 60) and resonance to 63; with all oscillators set to none. Just silence?

Thanks for the replies. I’ve checked around IC4 and IC3, and all seems correct. One of the 10k resistors around IC3 is measuring around 5k, but presumably that’s something to do with it being in parallel with something else? They’re all correctly colour-labelled anyway.

I’ve turned the v/oct tuner anticlockwise until the break-up point disappears beyond cutoff 127, and frankly it generally sounds really good, so I’m tempted to leave it there, but am still concerned that I can’t get the filter to self-oscillate.

At res 63 / cutoff 60, I get a tiny - almost inaudible - noise and a similarly quiet tone that remains the same regardless of what midi note I send. It’s more as though a gate has opened and I’m hearing interference, rather than anything the filter is doing.

Locate R21, this is the one that goes into IC4 pin 2. Try using a 15k here (or strap a 100k in parallel to it). Does it give a self-oscillation? If not there’s probably a problem with the filter itself - I suspect that one of the stages is not working, so not bringing enough phase shift for self-oscillation to occur.