MS20 Inspired VCO WIP

brother303 · Joined: Nov 02, 2010 Posts: 139 Location: ruhr-area/germany

Hi,

brother303 · Joined: Nov 02, 2010 Posts: 139 Location: ruhr-area/germany

Hi Pantheon,

have built the vca.

The cv side seems to work out correctly,but I have problems in the audio-path.

I tried attenuating the input signal (Wiard Anti-osc/Hertz Donut/Arp-VCO) but it always sounds very distorted and the high freqs are cut off massively. Confused

Have to go thru the circuit once again...

What is the correct mounting-direction of c2 and c4?

- C2 : "+" > right
- C4 : "+" > left

Can you confirm this?

Ah,and please,could you run a saw wave thru your Spice simulation to see if the highs are cut off?

Many thanks and best regards (as always... Wink

)
Greg

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

Ah sorry to hear its not working correctly,
I'll get back to you with the details in a couple of minutes.

remork · Joined: Aug 02, 2009 Posts: 25 Location: bhellgium

wow, first time reading through this topic. great work so far!
got so much catching up to do..

nothing to contribute to the tech discussion, but i can tell you that screenprinting an already drilled plate is not gonna work as you'd probably tear up the screen. not too many printers would be happy about that.
undrilled plates would work (if the edges are rounded/deburred!), but preparing a screen for a single print is a pain in the *ss.
larger numbers are definitely more interesting when screenprinting, since preparations can take a couple of hours before you can actually start printing.
..i'd go with the decals if i were you Wink

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

I had used the vca provided by eric as a basis, so you can look at that for the most part: http://urekarm.tripod.com/synth/jx_vca.pdf.

Yes your cap orientation seems correct.
I have tried a few things in the simulation and yes I also have found that it was cutting off the high-frequencies. Here is what I found:

If you look at Erics VCA, in the feedback loop of the first opamp, there is a [4n7 and 100k] from the output to the negative input. I think this is there to compensate for the High-frequency loss. I stupidly removed that from my schematic for some reason, but when I added it to my simulation the highs arnt as rolled off. You could try adding them in and see if that corrects the problem. I cant think right now why you would be getting such distortion Confused

You can try pulling lower Transistor out entirely, the one with its collector at ground. That would be a little more true to the original MS20 vca.

Let me know if that EQ fix works and then we can see how to tackle the distortion Smile

Thanks for you input Remork, yes it sounds like I should give screen printing a miss for this pannel. I got a little excited and couldnt help myself but to start working on the pannel. I will definately look into transfers for this.

Im looking forward to getting this board etched tommorow.
Im going to stick two VCO's onto one PCB I think, and the ring mod with the octave/finetune curcuits for each VCO on another, making just two PCBs for two vcos, much easier to mount.

ericcoleridge · Joined: Jan 16, 2007 Posts: 889 Location: NYC

Sebo · Joined: Apr 27, 2007 Posts: 564 Location: Argentina

Some more links:

Early MS-20 VCF (non OTA) with PCB:
http://www.aleph.co.jp/~takeda/radio/MS20clone2E.html

Late MS-20 VCF (OTA) with PCB (in japanese):
http://www.aleph.co.jp/~takeda/radio/MS20clone.html

Also there was a MS-20 VCA schematic there, but I can't find it now, I think the VCA here is based on it:
http://www.aleph.co.jp/~takeda/radio/OneBoardFarm.html
The PCB for this VCA can be taken from monosynth PCB.

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

I might have found the schematic you were looking for:

http://www.aleph.co.jp/~takeda/radio/harvest/img/schem/MS20CloneSet-scheme.gif
and
http://www.aleph.co.jp/~takeda/radio/harvest/img/schem/MS20-VCA-schem.gif

The structure looks the same as the one eric linked to.
Hmm, I'm going to breadboard this after the VCO's are complete and see if it works.

[edit]
I just tested the aleph MS20 vca in spice and it seems to perform much better.
Input was a 1khz saw tooth, and there is only 3db of frequency loss at 10khz, where as the other one was about 6db iirc. Is this how the actual MS20 vca operates or is something amiss here :S
[/edit]

ericcoleridge · Joined: Jan 16, 2007 Posts: 889 Location: NYC

As I remember it, it was difficult to discern the behavior of the VCA on the MS20, cause it's always hardwired to ENV Gen 2, and can't be disconnected. There's an "initial cv" input without an attentuator. As I remember, it's kind of a crappy VCA.

brother303 · Joined: Nov 02, 2010 Posts: 139 Location: ruhr-area/germany

Hi Eric,

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

brother303 · Joined: Nov 02, 2010 Posts: 139 Location: ruhr-area/germany

Hi Pantheon,

vca is working and yes,it`s not a "high-end" sounding one. But that´s MS20,isn`t it? Wink

The distortion-bug was gone after resoldering a wire,my fault... Embarassed

Any idea how to improve the high freqs at 10k and above? I added the 4n7 and a 100k resistor as you suggested and the behavior was much better. Perhaps increasing the value of the 4n7 gives less high cut?

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

Great news 303, do you have any idea if it has "the ms20 sound"?
I cant think right now of a good way to improve the eq, i'll have a closer look at the ms20 vca from the service manual and see if i notice anything.

Did you try to remove the lower transistor, to see what effect that would have on the sound?

ericcoleridge · Joined: Jan 16, 2007 Posts: 889 Location: NYC

Sebo · Joined: Apr 27, 2007 Posts: 564 Location: Argentina

Well, I have a Korg MS-10, if you want to run some tests on it, let me know what you want.
It has an audio in, I can do some spectrum tests...

brother303 · Joined: Nov 02, 2010 Posts: 139 Location: ruhr-area/germany

Hi all,

found some time today to play around with the vca (wiring up a Klee is an absolutely time-killer...).

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

Sebo · Joined: Apr 27, 2007 Posts: 564 Location: Argentina

Hi:
I'm a little busy now, I will try to record some audio trought the MS-10 as soon as I can.

brother303 · Joined: Nov 02, 2010 Posts: 139 Location: ruhr-area/germany

Hello,

Sebo · Joined: Apr 27, 2007 Posts: 564 Location: Argentina

The audio pass through the VCF, but if I can make me enough time I will try to dissasemble the synth and inject the signal directly to the VCA.

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

Sebo · Joined: Apr 27, 2007 Posts: 564 Location: Argentina

Great work!!!

brother303 · Joined: Nov 02, 2010 Posts: 139 Location: ruhr-area/germany

Hi,

Yummy! Very Happy

Hope,troubleshooting will be successful. Thumbs up!

@Sebo: Don`t kill your MS... Let us pray

Pantheon · Joined: Feb 17, 2011 Posts: 36 Location: Edinburgh

ok, so i've tested both sides now and they both seem to work almost identicaly which is quite a good sign.

Things of note:
- The triangle symmetry looks nice.
- The gain trim actualy works, it just doesnt have a great amount of range, but its sufficient.
- There is a very slight amount of wobble on the output of each vco (maby 100mv or less), that seems to increases with the presence of my finger by the CA3046 chip. I've had this same wobble on each version of the vco, and I havent been able to get rid of it so far. Does anyone think this will be a problem? Its not audible. I think the best I can hope for is that someone else who makes this vco might be able to figure out whats causing it. It might be present on the MS20 for all I know though.
- The saw still has to be biased a little negative for a good triangle symetry, leading to both the triangle and the saw being offset by around -500mv. I know this can be improved by changing a resistor in the triangle circuit, but its ok for my needs, so I dont think i'll change it, sry Embarassed

. I'll point out the resistor when I release the plans.
- Not much else to tell other than I think it sounds realy nice.

So the final thing that I need to do is to etch and test the octave and ring mod board, which I will do tommorow. Assuming that works, I'll post a PDF with the PCB for toner transfer and a small guide on how to tune it in and wire it.

Remember this will be no high end vco, I just want to emphasise that. Though hopefully once it starts getting built by people, some improvements will be suggested. For now, expect something very analoge, warm and with plenty of colour.

abovenyquist · Joined: Aug 31, 2009 Posts: 55 Location: Atlanta

Dear Pantheon and other electro-music folks,

My apologies if I'm posting this in the wrong spot; I confess I don't drop in here too often. (I hang out a lot on the SDIY mailing list, but sometimes I find the many forums and topics on EM overwhelming.)

Anyway, I'm teaching my "Electronics for Music Synthesis" class again this semester:

http://users.ece.gatech.edu/~lanterma/ems11

And my students are working on their final projects. One of my suggested project ideas, which a team of two students picked, was to build your VCO (or the original Korg MS-20) sawtooth core. They're having a very difficult time, and I was hoping folks here might be able to offer some help.

They started with this schematic:

http://electro-music.com/forum/phpbb-files/xms20_vco_150.jpg

They added the JFET between the expo converter and the 10K resistor in the sawtooth core. I think you asked what this JFET was in another post; some SDIY list folks figured out the JFET is a "common gate" amplifier acting as a *current* buffer (instead of the usual voltage buffer).

After this didn't go, we figured out that the expo converter needed a reference current. So they attached a resistor (megaohm range I think) between the negative terminal of the op amp in the bottom left corner and the +15 volt rail.

They did get a sawtooth oscillation going - but it's quite high frequency (20 kHz to 40 kHz range or something like that), and very tiny (200 mV pk-to-ok or something like that). I'm wondering if these issues are related - i.e. if the "comparator" is basically tripping at the wrong point - tripping earlier than it should be - hence making the wave amplitude smaller than it should be and also have a lower period than than it should be. There's also a big DC offset that varies with the "initial tune" pot (I don't know if that's expected or not.)

In the original diagram listed above, the leftmost part of the sawtooth core has 1K and 4K7 resistors. In the SPICE diagram, these are instead 22K and 220K resistors. They've tried both; same problem, high frequency wave, tiny amplitude. They also noticed the 3906 emitter is attached to the +5 supply in the original schematic, but the 3906 collector is attached to the +5 supply in the SPICE schematic; again they've tried it both ways.

I should mention they were using a CA3083 for the NPNs, using the extra-matched pair in the expo converter, and a 2N3096 for the PNP. They used a 2N5457s for the current buffer JFET. They have the substrate pin of the CA3083 tied to the -15 V supply.

Any advice that might help towards getting this working would is greatly appreciated. Smile

I should also mention that they've also tried to get the original Korg MS-20 VCO design working too; alas, what they have on the breadboard is not oscillating *at all*, but that's another story.

Along those lines, I noticed the Pantheon and the MS-20 VCO designs are quite different; I must confess I don't have a very good understanding of either design.

Thanks!

- Aaron