Moog Modular Clone - DIY

[analogmonster]

Hi folks,

as already announced in June I've started my new modular project - a Moog modular clone. My goal is to build a System 55 like modular, but as I don't know how much time it will take to complete a full system I started with a clone of a 904B high pass filter. I don't have such a module in the first modular system I built, the Formant, so if I decide to stop that new modular madness I can use that one in my first modular
So this is my first clone of a Moog module, and perhaps it would have been better TO CHOOSE ANOTHER ONE as a first try, because to build the 904B filter was tricky
3 different schemos, two of them buggy, no setup hints (I asked this forum last year, but no one could help me). But after 3 months (!) I got it, my clone works satisfactorily.
And here you can find pics and a detailed technical description:
http://www.analog-monster.de/mmt904b_en.html
I really hope that I will proceed faster with the next modules, otherwise I will finish this system when I'm a grandpa.
As discussed in June I've chosen Schaeffer Apparatebau in Germany for the front panel. I did not enclose the board into an aluminium sandwitch tin like Moog did, and I implemented the Moog port (connection to the Moog modular bus) with a double row multi-pin connector for ribbon cable use instead of an edge connector.
I am still waiting for the cabinet I've ordered at a local carpenter, so I can't show system pics now. Nevertheless I continue with the PSU to be autarkic when my cabinet arrives.
Any comments or any questions? Don't hesitate to ask I would be happy to share my experiences with others...

Cheers

Carsten

[MC]

Nice work. If you stumbled on the 904B, wait until you clone the 921.

[Benjamin AM]

Great job! Any chance that you will share the board art? I would love to add this to my modular system.

[analogmonster]

Nice work. If you stumbled on the 904B, wait until you clone the 921.

Oh yes, thanks. I've already had some looks at the schemo. I thought about replacing the saw core by a microcontroler solution, but that would'nt be very true to original, would it? But as I made good experience using LTSpice I will try to clone the 921 close to the original some day.
But first I will try the 901 oscillator. Yes, I know, the 921 is the better one, but as a lot of artists worked with the 901, I will have a look at it.

[analogmonster]

Great job! Any chance that you will share the board art? I would love to add this to my modular system.

Thanks for the compliment. Just send an email to my address on my homepage, and I will send a .pdf to you.

[Kevin Lightner]

But first I will try the 901 oscillator.

You'll probably be using different equipment than me*, but making 901s track well can be very difficult.
Not to mention there must be at least 3 or 4 different designs.
It's very similar to a Paia VCO anyway, which were also very hard to keep in tune.


*

[analogmonster]

You'll probably be using different equipment than me*, but making 901s track well can be very difficult.
Not to mention there must be at least 3 or 4 different designs.
It's very similar to a Paia VCO anyway, which were also very hard to keep in tune.

Yes, I've heard this from different sides. But I will try it anyway, although I don't have any special equipment. Just oscilloscopes, a multimeter and my ears I've found 2 schematic drawings yet, one in my service manual, and one on the Moogarchive pages. So if you build 901 module sets: which schema do you use, and what is your CA3019 source?

[Kevin Lightner]

I have some 3019s, but haven't a clue where to find any more nowadays.
My price is very high on them too.
You might be able to find a substitute diode array that would work just as well though.
The schemos I use are for whatever 901 I'm working on at the time.
I think there's 18 of them here now and not all are the same.
But if you're building something, it's best to use the latest schemos possible. (and don't always assume they're all error-free.)
I think the 901 schemos published are correct, but there's still a wide degree of variance possible, especially with the transistors.

Their main problems are no temperature compensation, need for selected caps for octave switching (difficult to do, even when paralleling caps) and making them match other vcos in the same linear way.

901s are oscs where you might get octaves 3 or 4 scaled fine on the same note across that range, but some notes in between can be slightly off. Normally once you scale a common osc to its highest and lowest points, all the notes in between fall perfectly into place.
Not so on 901s.
So whereas a normal VCO could be described as a straight line, 901 response can be more likened to a jagged or even curvy line.

If you DO make any 901s, make a set at the same time and match every component across all pairs or sets.
This will help them track against each other closer and more reliably. Use good caps too. Polystyrenes if you can.
I recommend sockets on the 2N4058 transistors and the UJT.
Using a multimeter to match or measure transistors can be a drag, but it is possible.
Substituting trannies until you find some that work well together is possible, but hit or miss.
You might even find a curve tracer for use with your scope that can display visually how a transistor under test will perform.
But alas, you'd need two curve tracers and a double trace scope if matching them via this method... or maybe a camera.

It is not difficult to get them to oscillate tho usually.
You could prototype a sounding osc in about 15 minutes on a breadboard.
They probably have the fewest components needed of any VCO to create their core oscillation. (thank their UJT transistor.)
Most everything else on them is for tracking, octave switching and waveform production.
Waveforms on them can be amazingly accurate too, with their sines being their weakest accomplishment but often still quite pure.

Hope something here helps.
They ARE fun to play around with in a hobby manner, but very difficult to make sets of them that perform well musically.
Open a window and let a draft in and they might sail away in pitch.

Good luck!
K

[analogmonster]

Hi Kevin,

thanks for all the information. This will surely help. I think I will try to build one bank first, one oscullator controller and three oscillators, to see how it works and whether I get along with it.

I have some 3019s, but haven't a clue where to find any more nowadays.
My price is very high on them too.
You might be able to find a substitute diode array that would work just as well though.

I might ask you for selling 2 or 3 of them when I am at this point.

The schemos I use are for whatever 901 I'm working on at the time.
I think there's 18 of them here now and not all are the same.
But if you're building something, it's best to use the latest schemos possible. (and don't always assume they're all error-free.)
I think the 901 schemos published are correct, but there's still a wide degree of variance possible, especially with the transistors.

Ok, this is what I've learned from building the 904-B, too. but 18 or even more varieties is much more than I expected

Their main problems are no temperature compensation, need for selected caps for octave switching (difficult to do, even when paralleling caps) and making them match other vcos in the same linear way.

Perhaps I should use adjustable capacities like in radio circuits?

Thanks for all the hints!

Carsten

[MC]

Perhaps I should use adjustable capacities like in radio circuits?

They're pf range, way too small for this application

As for CA3019s, they can be scavenged from Oberheim OB-SX preset synths or dead OB-Xs. They're on the voicecard motherboard.

[Benjamin AM]

This seems to be a good deal: http://cgi.ebay.com/RCA-CA3019-NOS-TO-1 ... 7909453322

[Kevin Lightner]

Some great posts here by others. Thanks for helping out.

If those 3019s on Ebay are good (and I see no reason why they're not), get them.
I can't even come close to $9.00 ea.

MC is right. Adjustable trims are far too low in capacitance for this job.
They'd be useless for anything but the most high range pitch adjustments possible and even then won't be of all that much use.

I also didn't mean to imply there are 18 different revs of the 901s.
Perhaps there's 3-6 of them with various changes.
I already have scans of several different ones from the component side, but never scanned their PCBs.
I can if I get a chance.

Your choice of 1 A and 3 B's is a wise one. That's what Moog did and what I'd do too.
You can also expand on the basic 901 design by adding a pulse width CV input on the A module.
Moog didn't include this (only a knob), but it can make a big difference in sound and capability.
You may also decide to externalize the low end tracking trim on the 901A.
It's the most often needed adjustment and this is what Paul Beaver and CalArts did with their 901s.

Also, don't forget that these things want not only the standard Moog +12 and -6 volt power, but also a -10 volt supply.
It doesn't need to supply hardly any current, but is needed to make them work at all.
An LM7910 is fine usually.

[analogmonster]

If those 3019s on Ebay are good (and I see no reason why they're not), get them.
I can't even come close to $9.00 ea.

Unfortunately I was too late.

Your choice of 1 A and 3 B's is a wise one. That's what Moog did and what I'd do too.
You can also expand on the basic 901 design by adding a pulse width CV input on the A module.
Moog didn't include this (only a knob), but it can make a big difference in sound and capability.

Yes, I already thought about that. I use CV driven PWM often in my other modular.

You may also decide to externalize the low end tracking trim on the 901A.
It's the most often needed adjustment and this is what Paul Beaver and CalArts did with their 901s.

Also a good hint.

Also, don't forget that these things want not only the standard Moog +12 and -6 volt power, but also a -10 volt supply.
It doesn't need to supply hardly any current, but is needed to make them work at all.
An LM7910 is fine usually.

The PSU of my clone will deliver +/-15V. The modules will get own voltage regulators for +12, -6 and whatever other voltages. I already did this on my T904-B, see also
http://www.analog-monster.de/mmt904b_en.html

But there is a point I don't understand: There is a connection point "B" marked in the 901-A - schematic (controller). This point is directly connected to the +12V supply. The 901-B -schematic (oscillator) refers to this point as source for the oscillator current sink (R2 R3 Q7...) So what is the difference between Point "B" and +12V supply?

Thanks

Carsten

[davemoog]

How is your 901 project going?

I built a few a couple of years ago and they work quite well. Curious about the Beaver low freq tracking mod. Mine track quite well over at least 3.5 to 4 octaves. Outside that range the bottem few notes are miles out but the high end freq not so bad. I found that it helps to use the Fixed CV to balance against CV coming in and keep the oscillator in its sweet tuning spot.(ie higher keyboard voices and lower Fixed CV) I also added pulse width modulation which sounds great. Oscillators also sound good tuned very low for rhythmic pulses.

http://minimalist.davidleith.com/DeitzModSynth4.html
http://minimalist.davidleith.com/DeitzModSynth6.html

Cheers dave

[analogmonster]

Hi Dave,

as you might have seen I want to build a complete System 55 clone. See also

http://www.analog-monster.de/mmclone_en.html

I am still on the filters. I am currently building the 904A lowpass filter, then I want to continue with the filter coupler and the 914 fixed filter bank. So it will take some time until I continue with the 901, but all information is very important for me, because someday I will need it.

I've been on your site. Impressive. I will contact you directly because I have some questions.

Cheers

Carsten