I managed to tune the VCO but I still have some questions.
I didn’t quite understand what this instruction is about: “Set the pot labelled “2” to its minimal position.” I found no noticeable effect on tuning. I tried it both with the pot in neutral position (0 detune) and all the way to left (-12). If I tune it at -12 it seemed to be a little bit off when I return to neutral, not really sure, so eventually I tuned it in its neutral (0 detune) position.
I did notice something odd when tuning. If you play C3 and then C4, the tuning of C4 will be slightly different from when you first play C5 and then C4. This goes for every other pitch too. So depending on if your going up or down in octaves the tuning will be slightly different. The tuner I use feels very accurate (guitar toolkit for iphone), at least it is showing consistent and stable measurements every time. The difference in tune between going up or down is always 2 to 3 cents.
Eventually I managed to tune the range C1 - C4 with deviations between -8 and +8 cents. That seemed to be the best I could do. It made me wonder, how many bits is the midi to cv conversion?
I was able to achieve +/- 1 cent within a three octave range - a fourth, higher octave goes off by an additional 3 to 5 cents
12 bits is plenty for 5v/oct - we’re tuning to compensate for the non-linearity of the VCO tracking
the Intellijel Dixie and Rubicon designs are supposedly some of the most precise VCOs currently available with a claimed accuracy of .1% over 8 octaves (also 2164 based) and they actually have two trimmers for v/oct adjustment, one is specifically to compensate for non-linear tracking behavior at high frequencies
I have found that the design of an analog v/oct front-end to either a digital or analog oscillator is a good bit trickier than it initially appears - which is why the auto tuning system on the Anushri is so convenient even if you’ve carefully tuned your oscillator - no need to stop at 4 octaves of good tracking if you use the software tuning - although its usage reduces integration with other modular gear.
@defenestration: The Rubicon looks interesting I didn’t know it. What are you using to measure pitch? Any tips on how you got it so close?
I was just using a basic guitar tuner, with the anushri signal running directly into it (using the mic on a cheap guitar tuner is a nightmare imo) - I was also monitoring by ear as well, cheap tuners can easily throw spurious readings. Also, after adjusting to get it pretty close I will let it sit for a while and do its thing, and then come back and check again (I run in a MIDI sequence from my computer). I found it was helpful to have the MIDI sequence both slowly step through octaves, as well as arpeggios and the occasional scale (only looked at tuner and made adjustments when using octaves though).
One interesting thing about the Anushri VCO design is that there is no tempco? Didn’t realize this until I was putting it together and I’m curious about the design decision, as I know essentially nothing about analog VCO design.
the tempco is done by the v2164
the design is derived from Osamu Hoshuyama’s SSM2164 log converter ; see http://www5b.biglobe.ne.jp/~houshu/synth/
I think it was discussed on the forum already, search for “hoshuyama” you should find the topic easily.
FYI I have discussed this a bit with David Dixon and the Dixie uses the same circuit + an extra classic expo converter (not sure how it is stabilized) that feeds an extra bit of current into the VCO core to compensate for high frequency detuning. The reason for the Dixie / Rubicon extra stability is the use of a triangle core design, which eludes the cap discharge time problem.
@defenestration. I don’t have a guitar tuner which is that accurate. Mine’s a Korg pitchblack but its not nearly as accurate as my iphone app.
What do you guys think about my observation that going down my reading of the octave tuning was always 2-3 cents off compared to going up? Has anyone tried it out / able to reproduce? Just curious. My tuner displays this behavior consistently and always.
for tuning, you may want to use very simply your soundcard as an oscilloscope.
If you’re under windows, you may want to check this software
And if you’re using Linux, fmit is pretty nice.
I don’t think the problem is my tuner. I will try an alternative but I doubt it will make a difference. I’m using a Mac btw.
thanks for the scope link, MicMicMan, now i don’t have to get up from my computer at all anymore
I don’t know how you get such nice results, I can’t seem to tune Anushri properly, the best I get is a 15 cent difference on each octave, so 4 octaves up it’s almost a semitone off, is there a specific patchsetting for tuning? I set the Sub volume to 0, filter is open, no resonance, I have a decent guitar tuner…
Is there a way to proceed? First offset, then v/oct? Note the AutoTune is working fine…
I do it the other way round… First V/oct, playing octaves from around 60 Hz to 1kHz and checking that the notes are spread one octave apart, then offset.
Thanks, I tried with v/oct first, the best I get is when the trimmer is turned to the max, and I still have roughly a 10 cent deviation for each octave, then the offset doesn’t seems to make it any better, once that is done shall I set the offset to a specific note, or is it only influencing the precision of the tracking?
You should set V/Oct so that the interval (rather than the note) is followed (C2, C3, C4 yields 3 notes one octave apart) ; then offset so that the note is followed (C2, C3, C4 yields the actual note).
That’s what I did, I wonder if I could shift the range of the v/oct trim a bit up by changing a resistor? The interval tracking got better until I reached the max value of the trim…
You can increase / decrease R61 to shift the range of the trimmer down / up. But this might also mean you have a mistake somewhere else…
@smugrik: I was confused at first because I was turning the pot in the opposite direction. Sounds stupid but maybe give it a try. Counter clockwise make the intervals larger I think. At first I also had the pot turned all the way clockwise and didn’t get it to tune properly.
@0x80 yes, Counter clockwise makes interval larger, but even when I’m full clockwise, interval is still larger than the Octave on My Anushri…