FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 

SVCO HFT calibration
MUFF WIGGLER Forum Index -> Oakley Sound Systems  
Author SVCO HFT calibration
Just finished calibrating a recently built SVCO-B. It is my third SVCO. It is calibrated beautifully and tracks accurately across the whole keyboard, but, and I am being pedantic, I can't quite get the HFT interval because I run out of trim pot adjustment when I'm within a whisker of achieving it. E.g with an A at 7040Hz the smallest interval I can get to for an octave above is 14106 and the 10k trim pot is at one end of it's travel (the same place it was set to provide a zero ohms reading prior to calibrating the SCL trimmer). Given the direction of travel I have observed it suggests I need a lower value than 0 ohms to get to the correct interval.

Whilst there is no perceivable audio issue (I did say I was being pedantic) I am wondering what might be causing this anomaly as I was expecting to end up with a successful setting somewhere in the middle of the trimmer range.

I think I may have also observed the same issue on my other SVCOs but would need to check.

I am using 1% metal film transistors throughout.

Any thoughts/feedback would be appreciated.
My first thoughts would be that the SCL trimming is not quite right. For it to be running sharp at 14kHz would normally suggest it to be overcompensated at say 7kHz. However, there may be other issues at work - accuracy of the log conformity of the NPN pair (in U6) or switching speed within the VCO core (U8 & Q3). I've not heard of any issues before though - Paul Darlow aka Krisp1 builds a fair few of these things and he normally flags up issues before the rest of us get to hear of them.

But if tracking is good at lower frequencies and you really need that level of tracking at 14kHz then you could try lowering R44 to 3k3. R44 actually compensates for the fixed switch time of the FET and comparator but maybe in your build they are running faster than I've anticipated.

Let us know how it goes.

Thanks for the feedback Tony, I also wondered whether it might be a really small calibration error in the output from the midiDac, i.e a tiny voltage increase in the output going up the keyboard. I will re-check the midiDac calibration at some point in the near future, I have a bit of a cabinet/module re-shuffle planned soonish.
I have now re-calibrated my midiDAC, it was a couple of 100ths of a volt out over the full 10V range so only a small adjustment required. Having adjusted that I re-did the SVCO-B. After some very careful tweaking I get results that provide accurate tracking across the whole keyboard including up to 10V. i.e my measured frequency at 5V is 261.6hz, at 1V it is 16.35hz, at 7V it is 1046hz, at 10V it is 8370hz. So basically it appears I can set the tracking across the whole keyboard without having to use the HFT circuitry, so I've left it alone! I haven't lowered R44 to 3k3 as there seems little point given the accuracy of the tracking achieved.

I'm using a K30A GR91 FET from Banzai, I used the same FET in my other two SVCOs in which I seem to remember I noticed the same issue (I'll check). Could these be the components running faster than anticipated?

As an aside, neither of my keyboards will generate an A an octave above 7040hz which is A10 midi note 117, if they could it would be a midi note above 127 so I'm not quite sure how you achieve the specified test range of 7040 - 14080hz, not that it matters.
midiDAC scaling should not affect the HFT settings on any VCO. This is because any errors in the output scaling of the midiDAC will be compensated for by the scaling (V/OCT) trims on the VCOs. HFT only acts on the compensated signal, so if midiDAC scaling matches the VCO scaling then HFT never gets to see an uncorrected midiDAC output.

Small drifts over ten octaves are expected and what one day will track exactly, will be off the next because of temperature differences. The temp co resistors can only compensate for the changes in the exponential convertors only so much. I generally expect tracking to be good over seven octaves at around +/-2 cents but this will be worse if the temperature changes in your system and room are large.

If the tracking appears good then that should be fine to leave HFT where it is. HFT compensates for the lack of accuracy in the expo convertor at high frequencies. R44 compensates for the lack accuracy due to the flyback time of the sawtooth core. The flyback being that part of the sawtooth with the steepest slope - it takes fixed finite time to go from a low state to a high state. HFT and R44 both compensate by forcing the VCO to work faster than it would normally do but work in different ways and affect frequency differently.

Normally HFT can't compensate very well for the flyback time, nor can R44 compensate for errors in the expo. However, on their own they can sometimes be good enough.

I think that R44 may be working harder than it needs to and could be reduced. But if it's working you may want to leave it be.

I'm not quite sure how you achieve the specified test range of 7040 - 14080hz...

This can't be done solely with the action of KeyCV. You need to use the frequency control or octave switch on the front of the SVCO.

I have now checked the setup on my SVCO-A, this also exhibited the HFT symptoms described earlier. I have now changed R44 to 3k3 as suggested and this has resulted in the correct calibration settings being achieved. With the HFT circuitry turned off the higher frequency notes are slightly flat and can then be 'sharpened' to the correct value with the HFT trimmer. A tiny amount of further tweaking to the Scale trimmer and job done.

I have to say I am impressed at the accuracy of the keyboard tracking on all of the Oakley VCOs. I also have two of the 2U VCOs and although they use different circuitry I also checked the HFT trimmer behaviour on these, they were performing as expected.

Many thanks Tony for spot on advice, I will update my two SVCO-Bs in the same way although there is a little more disassembly of the module required to get at R44 on these boards.

Perhaps an appropriate update to the builders guide for anyone else who comes across the same issue, all three of my SVCOs exhibited the same feature so it must be down to my component choice/sourcing?
Thank you for the update and for looking into this.

I've got an SVCO in bits at the moment as at some point I am going to look into using different switching FETs as the venerable 2SK30A-GR may get harder to find in the future. I'll also look into the flyback compensation via R44. I have not heard any one else with the issue but that is not to say it doesn't exist and I'm interested to learn more about any possible problem. Thankfully, it is an easy fix if it does prove to be more commonplace.

I'll put it on the list of things to do. It'll be a little while off as I've got a heap of things to do at the moment. Two new DIY projects are coming shortly, the analogue delay and the MOSPhaser. More on those as soon as can.

MUFF WIGGLER Forum Index -> Oakley Sound Systems  
Page 1 of 1
Powered by phpBB © phpBB Group