Author |
Message |
Jaba
Joined: Feb 27, 2009 Posts: 48 Location: Genova, Italy
|
Posted: Fri May 27, 2011 7:42 am Post subject:
VCO questions Subject description: questions about Yusynth VCO's details |
|
|
Hi,
I already built several Yusynth modules: VCO, Noise-S/H, Dual Gated Slew, CV Standard, some ADSRs, ARP4072 VCF, Minimoog VCF, EMS diode VCF, Dual Balanced Modulator, MinMax and a couple mixers, and I am very happy with them all, so happy that I planned to make another VCO and a Wavefolder
(Thank you Yves!)
while I was debugging my second Yusynth VCO, two questions arose:
1) it seems to me that the PCB does not correspond to the schematic in the hi freq compensation: R53 goes to the wrong side of R15, directly to the emitters of U4 instead of going to pin7 of U3b
did anyone notice ?
2) Yves' page recommends to replace R15 with a 100K if a SSM2210 is used to replace the LM394; would someone please explain the reason for this ? I tried to compare the datasheets, but still I don't understand.
my other Yusynth VCO, built before I saw that note, still has R15=10K with a SSM2210 and it works fine;
it seems that in many other similar designs (CGS, MFOS, ...) the LM394 is replaced by SSM2210 (or 2SC1583, or 2x 2N3904) without any resistors change, would someone please confirm ?
(in the .gif layout, R15 looks like a 51K, but with a LM394)
and if I use a 2SC1583, or two matched 2N3904s, should I change R15 ?
I am also considering replacing R53 with a 1M trimmer and a 47K resistor, as suggested in Mariano's PDF ("C.- HIFREQ TRACK mod"), but I would prefer to solve issue 1 before
many thanks to you all
Paolo(Jaba) |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Sat May 28, 2011 9:15 am Post subject:
|
|
|
1) How long since you checked the VCO page ? This discrepancy has been corrected more than a year ago.
2) The reason for the change of this resistor when using a SSM2210 is that if you have a look (with an oscilloscope) to the SAWTOOTH wave you will notice that for frequencies above 1kHz the slope of this one becomes less and less linear (it becomes curved) as the frequency increases. Using a 100k resistor instead of a 10K linearizes the slope. This way the wave shapes behave and sound like they should above 1kHz. _________________ Yves |
|
Back to top
|
|
|
Jaba
Joined: Feb 27, 2009 Posts: 48 Location: Genova, Italy
|
Posted: Mon May 30, 2011 9:03 am Post subject:
|
|
|
Hi Yves,
thank you for answering my questions
1) you are right, my copy of the schematic is old; but I believed that the old schematic was right, and the PCB was wrong.
You changed the schematic instead, and this confuses me.
I tried comparing it with other VCOs (MFOS VCO, Ian Fritz's Sawtooth VCO, Thomas Henry's X4046 VCO and XR2206 VCO, René Schmitz's VCO4), in all of them the HiFreq compensation goes from the base of a transistor in the matched pair directly to the output of the opamp, not to the coupled emitters. There are other differences, such as bigger timing capacitors, but the structure seems repeating. Even in Mariano's modifications PDF he seems to follow it.
It probably does not make much of a difference, but I would like to understand.
2) ok, I do not have the technical skill to understand this, but I believe you.
On my old Yusynth VCO, with SSM2210 and R15=10K, the sawtooth wave starts "bending" at about 8kHz, so I would not care much; do you think it may affect the tracking or tuning too ? |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Mon May 30, 2011 9:21 am Post subject:
|
|
|
There is no problem in trying to understand, I may say it's a good practice indeed !
1) Both will work but do not require the same resistor value, as a matter of fact injecting the correction current at the point I put it avoids that it ibeing sensed by the feedback loop of the servo which would then tend to anihilate the compensation (and don't want this do we ?). The fact that I am using a smaller cap is done in purpose, it solves some of the problem of the high tracking which is due to the deviation of the Ic from an ideal curve (Ic vs Vbe) for high currents. By using a smaller cap, I can reach higher frequencies without going in that bent part of the curve and therefore I have less problems of high tracking.
2) no this won't affect the tracking and tuning, it just impacts on the wave shape at high frequency that's why I recommand using a 100K in this case. Therefore if you're satisfied by your VCO as it is there is no need to replace the resistor. _________________ Yves |
|
Back to top
|
|
|
Jaba
Joined: Feb 27, 2009 Posts: 48 Location: Genova, Italy
|
Posted: Wed Jun 01, 2011 3:33 am Post subject:
|
|
|
Ok, thank you very much for your explanation
I think I'll leave the first VCO as it is.
In the new one, I put a 100K for R15 and we'll see after calibration
some other minor mods:
100k pots for Lin FM In, EXP FM In and PWM In, for a higher input impedance;
a switchable series cap for AC coupled Lin FM (directly behind the panel);
make C9 switchable for HardSync/SoftSync.
... and a lot more fun is coming ! (two great VCOs is better than one)
next step: Yusynth version of the Haible Wavefolder
thank you Yves!
Jaba(Paolo) |
|
Back to top
|
|
|
Jaba
Joined: Feb 27, 2009 Posts: 48 Location: Genova, Italy
|
Posted: Tue Jul 05, 2011 6:04 am Post subject:
|
|
|
while doing some experiments with ringmodulated sine and triangle outputs from the same VCO, I found that the small nonlinearity in the saw waveform causes a subtle asymmetry in the triangle wave, which can sometimes be annoying.
In the end, I'll put the 100k at R15 in the old VCO too, just to improve the waveform linearity and symmetry.
thank you Yves! |
|
Back to top
|
|
|
|