MS20 Inspired VCO WIP

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

So I first want to say hello to everyone, I'm a newb here.
This seems like the best place on the web to share in my new hobby with like minded folk.

I should make it clear that I am very new to electronics and synth diy in general (just over a year since I started learning) so please bear that in mind when reviewing this work.

I thought I would share my current project with you bunch, and try and get some useful feedback in the process.

As the title says I have been trying to build a MS20 style VCO, with 1v/oct. control, better waveshapers, and soft syncing. Not a task many seem to have taken on. there is one video on youtube of a MS10 VCO clone, and Jurgen Haible made a linear control VCO around the MS20 core.

So far I have a working core, producing a very crisp downwards saw tooth waveform.
Also soft syncing is tested and working (sounds awesome to my ears).
1v/oct control is working (as best as it can without matched trannys), but the high notes need some form of compensation as they are falling flat.

So the main area I think I need to work on is improving the stability, I'm going to place an order for some CA3046's tomorrow. I think they should go a long way towards improving the stability of this VCO, compared to discrete unmatched trannys. Also I can build the entire core and the expo converter from a single CA3046 Smile

Since this is my first VCO that I have designed to some extent, I would really appreciate any comments about my design.
I can post photos from my oscilloscope/breadboard if someone wants to see the waveform/etc.

The attached schematic shows (hopefully Wink

exactly what I have on my breadboard.

XMS20_VCO.JPG

Description:

Filesize:

141.54 KB

Viewed:

2502 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

v-un-v · Posted: Thu Feb 17, 2011 1:57 pm Post subject:

Hi Pantheon!

welcome

to electro-music.com! Very Happy

Here's probably not the best sub-forum to post, but I'll leave your post here for now. Smile

I was never a big fan of the VCO on the MS20, I found it do thin and 'Casio'-like. But I was a big fan of the external signal processor, which I still find hard to understand why no one has ever done a decent clone of it (with minor improvements, to keep it kosher Smile

).

Tom

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

Hi:
I wanted to add a second VCO to my MS-10 for some time, thanks.

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

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

Yes, i know, I think connecting the CV adder directly to C.C.C will do the trick.

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

There is one thing from the original schematics that still puzzles me, perhaps some kind soul here could help me understand its purpose. I have attached this section of the MS50 schematic, and highlighted the section in a red box.
This same structure appears in the MS series of VCOs and even Jurgen Haibles VCO, so i'm guessing it may be important.

On another note, the MS50 schematic shows very clearly v/oct control of this oscillator, and even has a temperature compensation scheme detailed.
So I have ordered CA3046's and the correct Tempco resistor, so I think I will use the MS50 expo converter now.

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

Ahh, yes, you right, I forgot the current sink Sad

I will check the MS-50 schematics, thank you.

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

Just a small update to say that I have now changed the core of the VCO to use a CA3046 transistor array. This has vastly improved the stability, and increased the frequency range a lot.

I also have been playing around with the spare transistor in the CA3046, using it to generate sync pulses from an input wave. This works great, however the hard sync method I got working is very unusual. It sounds amazing though, I'll post some recordings and oscilloscope pictures soon.

All that is left to do is to try out the wave shapers, and the sub-oscillator.
The sub-oscillator will produce a sub-octave saw wave, using a transistor bistable multi-vibrator and a mixer. I have only simulated these elements for now, so i'm keen to hear how they sound.

Once I'm happy that this VCO is working correctly I'll post up the final schematic and a PCB layout for home etching.

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

Great!!! Thanks.

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

Hi,

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

so I thought I might as well post some current info from the project here for posterity. I still have to juggle with the ca3046 transistors a bit after reading that one pin (the susbtrate pin) must be pulled lower than the rest. I havent mustered up the will power yet to change all this on the breadboard.
So instead I have decided to model the core of this oscillator in LTSpice to get a better understanding of it. I have attached some of the data from the simulation:

[edit]
I also wanted to mention that this same vco core is extreemly simmilar to that used in the Polyvoks synth. It was quite surprising when I first saw that.
[/edit]

vco_data.JPG

Description:

Filesize:

102.91 KB

Viewed:

638 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

vco_core.JPG

Description:

Filesize:

40.42 KB

Viewed:

692 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

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

Hi,

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

I have attached my layout for a prototype of this VCO.
Im just about to etch this and check it works 100%, I'll post my findings once its running (hopefully). I have left the sync off from this board, as there were a few kinks I havent been able to work out yet.

[edit 1]
its etched and drilled now, im just stuffing it with all the parts.
[/edit 1]

[edit 2]
pcb is mostly stuffed, but I need to buy a small T0-92 7805 tomorrow, as well as a 10k trimpot. The orientation of the TO-92 componants on the silk screen is correct.
[/edit 2]

MS50VCO.zip

Description:

PCB artwork for the prototype MS50VCO

Download

Filename:

MS50VCO.zip

Filesize:

43.51 KB

Downloaded:

1019 Time(s)

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

Nice one!

Let us know when your proto is alive and running.

THX!

Greg

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

Great work!!!

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

ITS ALIVE!!!

So building went without a hitch so far, and the initial tests seem encouraging.
All the waveforms look very clean, the triangle in fact looks perfect.

I'm about to connect it to my Midi-CV converter and tune it in.

So notes so far:
The saw tooth can easily become loaded down (when I connected it to my little speakers), i'll have to investigate that some more.
And the triangle wave has a lower amplitude than the saw and pulse.
I'll fix these issues.

One thing im not sure of is the standard way to output a waveform.
Should it be decoupled with some caps? Or do i just leave them floating above ground or whatever? Its not been so clear from the schematics I have looked at.

I'll try and find some batteries and take photos from my oscilliscope shortly. Maby even some sound recordings when I fix the loading issue.

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

So I just hooked It up to my Midi-CV converter and..... It totaly works Smile

The pitch was realy easy to tune in and my ears tell me that I have about 6 good octaves before it flats out.

The triangle wave sounds deep and thick! The saw wave is realy fat and has loads of bite to it! I havent tested the pulse width yet, but that will happen shortly.

I'll also try to get some AA batteries for my camera...
I'll post up a schematic fairly soon, once i've tested and tweaked everything...

FYI: I havent used any precision resistors or anything, just E6 series mostly.
So I dont know how each VCO will vary.

[edit]
ok i've tested the exp cross modulation and it sounds increadible!
I'm too excited about this guys Smile

[/edit]

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

Sorry for repeated postings... Embarassed

Here are a few pics of the prototype...

Pulse will come as soon as I find my 68k resistors Smile

Oh and you might spot that the TempCo hasnt been pushed onto the CA3046 on my board yet.

DSCF3260.JPG

Description:

Filesize:

562.31 KB

Viewed:

528 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

DSCF3258.JPG

Description:

Filesize:

248.73 KB

Viewed:

482 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

DSCF3259.JPG

Description:

Filesize:

195.74 KB

Viewed:

443 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

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

Great work. I let go of my MS20 a few years ago, and always regretted it. I'm looking forward to attaching these VCOs to my etched MS20 LPF + HPF, and listening for that familiar tone. I'll just have to wire together the simple VCA from the schematics.

It might be nice to have two of these VCOs on one board with the CMOS gate "ring mod" and transistor noise source.

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

hey eric,

That sounds like a great idea. I think im also going to make a dual version for myself. That ring mod idea sounds great; I had planned to use a CGS real ringmod myself, but probably on a sepourate card.

Since im still a newby with synth circuits i'll post up my PCBExpress layout file to help others that may need to tweak it to be more compatable with their current setup.

I need a little clarification with two points:

- Is there a standard voltage peak-to-peak for VCO outputs?
currently the saw is 3v p-p, and triangle is 2v p-p, IIRC.
I think i'll make them all a nice 5v p-p for the final version, unless someone suggests otherwise.

- Should they all be centered around ground at the output?
currently the saw is centered, but triangle and pulse are not.
Is a simple cap decoupler sufficient?
Some circuits like the ms50 seem to use resistors to push the signal towards ground.

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

Hi:
There are various standards in synths, I think that every manufacturer have his own,
but in modular world the most common for VCOs are 10Vpp (+5 to -5V) centered
around 0V.
And using a capacitor to remove DC is not a good idea, as this will acta as a highpass filter, will deform the waveforms a little.
The most common way is to inject a DC to cancel the offset of the waveform,
usually using a trimmer and a opamp as a summer. You need an oscilloscope to calibrate it.

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

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

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

Hi,

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

Strange, somehow I didnt notice the MS20 had a ring mod. It seems very interesting though, so I'll have a look into it. If its simple enough, then I dont see why I cant do a dual vco + this ring mod Smile

I also just wanted to post to say that I have revised the waveshapers so that there are not decoupling caps in the signal line and each output is +5v to -5v.

I'm going to do a layout for this new version in the coming days. In the meen time i'll research this RingMod. If you guys have any good info on it then it could be a big help to post/link it Smile

[edit]
ok a quick google around shows that its probably a realy simple circuit. I'll definately try and make one. I'll proceed to make my new design first to be sure that the vco is working correctly. After that i'll put together the dual version + ringmod I think.
[/edit]

[edit2]
I threw together a quick spice simulation of a MS20 ringmod, and all seems well with it Smile

Output is also +5v/-5v to scale well with the output of the vcos. The new VCO design has an opamp going spare, so it should in theory be easy to add this on.
[/edit2]

[edit3]
I might as well upload my current design for the VCO, it should be attached. A few things arent detailed in, but you will get the general idea.
[/edit3]

RingMod.JPG

Description:

MS20 Ring Mod

Filesize:

30.15 KB

Viewed:

817 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

MS_Series_VCO.pdf

Description:

MS_Series_VCO - Revision 2

Download

Filename:

MS_Series_VCO.pdf

Filesize:

19.75 KB

Downloaded:

1099 Time(s)