moog (modular) mixer circuit

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: Joe., lisa, luketeaford, Kent

User avatar
wavehead
Super Deluxe Wiggler
Posts: 1235
Joined: Sat Nov 14, 2009 7:58 pm

moog (modular) mixer circuit

Post by wavehead » Thu May 16, 2013 8:00 pm

has anyone etched or produced PCBs for the mixer from the moog modular? basically something like the STG .MIX, but would be cool to get that same saturation/distortion with more channels and maybe some other features. could easily see having 4 of these in my setup.

User avatar
revmutt
Knutty For Knobs
Posts: 1040
Joined: Sat Dec 06, 2008 4:24 pm
Location: NYC

Post by revmutt » Thu May 16, 2013 9:02 pm

I'm curious which mixer you are talking about, the CP mixer or the Matrix Mixer?
Tommy DOG
Punk Rock Star, gambler and stunt double

User avatar
wavehead
Super Deluxe Wiggler
Posts: 1235
Joined: Sat Nov 14, 2009 7:58 pm

Post by wavehead » Thu May 16, 2013 9:24 pm

CP3

User avatar
CZ Rider
Veteran Wiggler
Posts: 509
Joined: Sun Jun 14, 2009 11:19 pm
Location: The edge of space

Post by CZ Rider » Thu May 16, 2013 10:30 pm

Those circuits are so basic even Moog made them on Vectorboard on the early ones.
Image
The later units had a pre etched board, but they were very small. They were usually mounted on a fabricated L bracket that went behind one of the mixer pots.
This one is an updated CP mixer from a model 10/12/15. The circuit had minor resistor change and an added trimmer for DC balance.
Image
My experience with the actual CP3 is they do not work well with 10 Volt P-P oscillators, and begin to distort almost immediatly. The CP3 design is not only a mixer but a 2X amplifier on each channel. So those 22K input resistors would need to be matched for a hotter output oscillator. The Moog oscillators these are designed to mix are 1.2 Volt P-P. Would need a decade box or some math to figure out the best value so the distortion was in a more manageable range. I have seen those CP3 mixers boards used as single channel amplifiers on custom Moog modules too.
I never bothered to make etched boards for the few I made and just went with the original old Moog way a making them using Vectorboard and those pins.
Here is a very similar circuit to the CP3 on Vectorboard, 1969 old school style. It is actually a circuit of the 901C Output Stage, with the added square wave cilpper circuit. But you can recognize the typical layout of two matched transistor pairs at the top of the Vectorboard. Bottom board is a Moog trigger converter.
Image
A pre etched board would make it easier.
Also experimenting with the 984 type clone of the 982 mixer. Have all the parts and schematic, just have to breadboard and test. Always wanted to see/hear what those mixers sounded like too. They also boost 2X and have bass and treble controls. They are intended as final audio mixers/routers and won't mix DC signals like the CP3.

User avatar
wavehead
Super Deluxe Wiggler
Posts: 1235
Joined: Sat Nov 14, 2009 7:58 pm

Post by wavehead » Fri May 17, 2013 4:32 pm

Also experimenting with the 984 type clone of the 982 mixer. Have all the parts and schematic, just have to breadboard and test. Always wanted to see/hear what those mixers sounded like too. They also boost 2X and have bass and treble controls. They are intended as final audio mixers/routers and won't mix DC signals like the CP3.
this sounds super interesting.... the only reason i am specifically interested in these circuits is the distortion they create for waveform mixing before filters etc., but using the 984 as a final output could be really cool as well.

i know conversion for the moog power specifications is also necessary on these (at least if it's going to run at 15V or 12V).

User avatar
sinemod
Wiggling with Experience
Posts: 306
Joined: Thu Jan 15, 2009 11:16 am
Location: Montreal QC

Post by sinemod » Mon May 20, 2013 9:54 am

I woud like to see a simple 12 v version
(Eurorack)

I have the stg version and i realy like using it pre filter

À schematics would do for me

JohnLRice

Post by JohnLRice » Mon May 20, 2013 10:24 am

The schematics are in the Moog Modular Service Manual, easily found thanks to Kevin Lightener! :tu: :hail:

http://www.synthfool.com/docs/Moog/modu ... Manual.pdf

User avatar
Dubka
Wiggling with Experience
Posts: 366
Joined: Sun Nov 27, 2011 11:39 am
Location: London, UK

Post by Dubka » Tue May 21, 2013 7:20 am

Does anyone know how to change the appropriate values for 10v pp?

Also, what's the best way to derive -6v from my -12v line?

User avatar
Monobass
thonk.co.uk
Posts: 8760
Joined: Fri May 29, 2009 5:39 pm
Location: Brighton, UK

Post by Monobass » Tue May 21, 2013 7:37 am

It doesn't saturate or distort per se, it just clips.

I really like my STG Mix.

[video][/video]
[video][/video]
Thonk - CLICK HERE - Modular Synth DIY + Eurorack Accessories Store

JohnLRice

Post by JohnLRice » Tue May 21, 2013 10:36 am

Dubka wrote: Also, what's the best way to derive -6v from my -12v line?
I needed -6v for one module so I made this little adapter board using a LM7906 regulator:
http://www.fairchildsemi.com/ds/LM/LM7905.pdf

Image

User avatar
Dubka
Wiggling with Experience
Posts: 366
Joined: Sun Nov 27, 2011 11:39 am
Location: London, UK

Post by Dubka » Tue May 21, 2013 10:52 am

So, are they 2.2uf and 1uf Tants just like in the pdf?

JohnLRice

Post by JohnLRice » Tue May 21, 2013 11:36 am

Dubka wrote:So, are they 2.2uf and 1uf Tants just like in the pdf?
I'm sorry, I can't recall. I may have used the specified values but . . . . :hmm:


Check what the notes say though and maybe just go with the electrolytic?:
For value given, capacitor must be solid tantalum. 25μF aluminum electrolytic may be substituted.

User avatar
Dubka
Wiggling with Experience
Posts: 366
Joined: Sun Nov 27, 2011 11:39 am
Location: London, UK

Post by Dubka » Tue May 21, 2013 2:20 pm

Cheers John :tu:

goom
Ultra Wiggler
Posts: 808
Joined: Sun Jun 07, 2009 8:03 pm

Post by goom » Wed May 22, 2013 12:38 am

CZ Rider - thanks for the pics and info. I tried looking for a pic of the CP3 front panel (just curious), but only found the CP3A. Do you have a pic of the front panel by chance?

Is your 2nd pic of the pre-etched CP mixer PCB an actual CP3 (not a CP3A)? It's very interesting to see pics of these. Thanks for that!

I'm going to try my hand at breadboarding the CP3, but I'm not sure how to select the input resistors for use with 10V Pk-Pk input signals. I guess the input pots would have to be increased to a similar value as the input resistors? If worse came to worse, I'll just experiment with different values to see what happens.

A different approach that I was thinking about is keeping the input pots/resistors the same, but adding a voltage divider after them. I wonder if that would work to scale the inputs down to a useable level. On the output side, I'd like to add an amp circuit to boost it back to 10V Pk-Pk. I wish I were a bit smarter with all this. I'm not sure of what the best implementation would be.

User avatar
Dubka
Wiggling with Experience
Posts: 366
Joined: Sun Nov 27, 2011 11:39 am
Location: London, UK

Post by Dubka » Wed May 22, 2013 3:54 am

Someone here MUST know.... they always do...

User avatar
daverj
Vintage Video Wiggler
Posts: 8514
Joined: Fri Jun 19, 2009 11:09 am

Post by daverj » Wed May 22, 2013 2:12 pm

goom wrote:I tried looking for a pic of the CP3 front panel (just curious), but only found the CP3A.
These match the CP3 schematic in the service manual John linked to:

Image

Note the switch for the "Click Filter".

The CP3A replaces the switch with a 4-way multiple.

Image
CZ Rider wrote:The Moog oscillators these are designed to mix are 1.2 Volt P-P.
Are you sure that's not 1.2v RMS? (typical studio line levels are +4dBu which is 1.22v RMS, or just under 3.5v p-p)

Based on the schematic in that service manual it looks like the gain of the mixer is just a tiny bit higher than x2 (about x2.14). So I would expect clipping to happen with somewhere around a +/-5v to +/-6v output. So a +/-2.5v to +/-3v input (or multiple smaller inputs).

So if my guesses are right, you could change the input resistors to 47K and make it a unity gain mixer with clipping a bit above the normal +/-5v levels. Or change the 22Ks to 68K and add op amps on the output with a bit of gain so the clipping happened higher up.

The CP3A is all op amps, and also has the gain set to exactly x2. It won't have the clipping that the CP3 has.

goom
Ultra Wiggler
Posts: 808
Joined: Sun Jun 07, 2009 8:03 pm

Post by goom » Wed May 22, 2013 8:06 pm

Thanks for the panel pics and info, Dave. It's interesting that the panel doesn't indicate the mixer model number (no silkscreen). The three CP3 panel images look to be identical, other than the illuminated switch colors.

I just got through looking up how to calculate gain in a transistor circuit. My head hurts! There are so many explanations that I came across, some with a lot of calulations.

For the CP3 circuit, is the 47K acting as a feedback resistor (as in an op amp circuit)? Using the 22K as the input resistance, and calculating the gain (as is done with opamp circuits), I see that would be a gain 2.14. Is that how gain is calculated for this circuit?

I would definitely like to keep the distortion characteristics towards the higher end of the input level pot settings (using +/-5V input signals).

Anyway, I plan to order the parts and try my luck. :hyper:

User avatar
CZ Rider
Veteran Wiggler
Posts: 509
Joined: Sun Jun 14, 2009 11:19 pm
Location: The edge of space

Post by CZ Rider » Thu May 23, 2013 1:27 am

daverj wrote: Are you sure that's not 1.2v RMS? (typical studio line levels are +4dBu which is 1.22v RMS, or just under 3.5v p-p).
Aprox. 1.2 Volt P-P. The Moog manuals list the 901B as "0.5 Volts RMS (-4db)".
And the 921B is listed as "-4dBm {approximately 1.3 volts peak to peak}".
Have to remember the 901's and discrete mixers are 1960's designs. The nominal outputs stated vary on each waveform. The square has the highest energy, followed by the sine, triangle then the sawtooth. The saw having the least energy output, but to the ears sounds louder than the sine/tri.

The pic of the circuit board on the Voyager is one from a Moog 15. The mixer on the model 15 seems to be a later revision of the original CP3 mixer. There are a few changes of values and most notably the extra trimmer for the DC balance.
Here are links to the two different versions to compare the minor differences.
CP3 MIXER SCHEMATIC
MODEL 15 MIXER SCHEMATIC
I am not certain when this revision took place, or what models they were definitely on. My 1969 has the ones without the extra trimmer. But I have seen several old original CP mixers with the two trimmers. All of the later Norlin era model 35 and 55 consoles with the 930 power supply had the "improved" CP3A. This had a +/- 15 volt supply and was based on the 1458 op-amp. A very different mixer.

Some pics of the 1969 CP3 mixer board. Small board on the right, attached via a fabricated L bracket. The bracket held in place by one of the mixer pots with two 4-40 machine screws holding the circuit board.
Image


A close up of the board. Just two matched pairs of 4058's and 3392's, with a few resistors.
Image

In this pic you can see the tropical fish cap for the click filter. And buried under the wire the .0033uF cap. The CP3 will break into oscillation at full gain without this cap.
Image

Another angle showing the cap. Oddly the two +/- jacks are missing the small .01 caps?
The hardware on the Moog for the CP3 mixer has four 25K AB type J pots, with a double ganged 10K attenuator after the circuit. So any clipping/distortions are adjusted via the four mixer pots. The double ganged pot just attenuates both the positive/negative final outputs simultaneously.
Image

Distortion and clipping is one of those areas of personal taste. I prefer the soft clipping/distortions I get when mixing four 901 waveforms together. Too much and it hard clips and begins to sound like mud. But just the right ammount, when the oscillators are slowly beating, and I hear this very alive sound.
So what happens to a stock CP3 mixer when a typical 10 Volt P-P oscillator is patched into it?
Here is a scope shot of a Roland 100M triangle wave into one input turned up to 10 looks like.
Image

The waveform clips hard, very squared. To get the Roland not to clip the mixer dial needs to be set about 3.5. This is just a single waveform in one mixer input. Mixing two or more would require further dialing down the mixer levels to avoid total hard clipping.
Here is the Roland sawtooth. This wave is positive biased. The scope shot is from the negative output of the CP3.
Image

Not much saw left, and this is just a single waveform. So some type of adjustment needs to be made to make this compatable with hotter oscillators. And most importantly, to be able to dial in the soft clipping/distortion with a workable range when mixing waveforms. Right now the workable range on the mixer with four hot waveforms being mixed is from about 0 to 1.
Using the lower output 901 oscillators, this workable range is around 8 on all four dials giving a clean output. Any bit above the 8 mark and you begin to get slight distortions in the waves, depending on the type wave.
Here is a Moog 901 saw wave output via the CP3 positive output. For this demonstration, I have the saw output split to all four inputs.
The saw before any distortions.

Image

Nice clean straight line without distortions. As I begin to turn up the gain, the line starts to bend toward the negative side. (little blurry)
Image

And when pushed further will hard clip on the negative side.
Image
The fun part here, is that is a scope shot of the positive output. The negative output shows a rounded clip where the hard line was. I prefer the sound on the negative output. And this negative output is needed if you want to use feedback loop trick from a 904A lowpass filter. Due to the 904A inverts the input. The positive side patched into a 904A will subtract if looped back into the CP3. Notice how curved the top of the waveform is. This hard clip inverted to a soft clip happens all over the Moog system when signals are inverted. Possibly due to the +12/-6 power supply.
Image

This rounding at the edge of clipping is the sound I think gives a special tone. Especially with the moving of oscillators against each other transitioning from that soft clip to a ripping hard clip. Thats my take anyway?

Any info on a good resistor change for the hotter oscillators, please post. I would not mind building a CP3 designed to handle all those other oscillators I have here. I may have to experiment with this myself.

Panel shot of the 1969 CP3.
Image

User avatar
daverj
Vintage Video Wiggler
Posts: 8514
Joined: Fri Jun 19, 2009 11:09 am

Post by daverj » Thu May 23, 2013 3:07 am

goom wrote:For the CP3 circuit, is the 47K acting as a feedback resistor (as in an op amp circuit)? Using the 22K as the input resistance, and calculating the gain (as is done with opamp circuits), I see that would be a gain 2.14. Is that how gain is calculated for this circuit?
Yes. With the two differential amplifier pairs the overall gain of the amp is quite high. So the input and feedback resistors determine the final gain.
CZ Rider wrote:Aprox. 1.2 Volt P-P. The Moog manuals list the 901B as "0.5 Volts RMS (-4db)".
OK, so that's +/-0.6v, which is 8.3 times smaller than +/-5v. So changing the 22K resistors to about 180K should compensate for larger inputs. The outputs will still be the same size.

User avatar
Dubka
Wiggling with Experience
Posts: 366
Joined: Sun Nov 27, 2011 11:39 am
Location: London, UK

Post by Dubka » Thu May 23, 2013 5:24 am

if the 22k changed to 180k, would the pots still be 25k?

:hmm:

goom
Ultra Wiggler
Posts: 808
Joined: Sun Jun 07, 2009 8:03 pm

Post by goom » Thu May 23, 2013 7:35 am

CZ - thank you for the fantastic insight to the CP3, and also for taking the time to take scope pics of the various patching examples. Do you know what type of caps the click filter and .0033uF are? I think I spec'd out a poly cap for the .0033uF, and an electrolytic for the click cap.

Thank you also to daverj for the info about the circuit gain. I'm my search for a simple gain calculation, I was amazed at how many formulas there were for the different transistor circuits.

I hope to order the parts tonight.

User avatar
CZ Rider
Veteran Wiggler
Posts: 509
Joined: Sun Jun 14, 2009 11:19 pm
Location: The edge of space

Post by CZ Rider » Thu May 23, 2013 8:37 am

goom wrote: Do you know what type of caps the click filter and .0033uF are? I think I spec'd out a poly cap for the .0033uF, and an electrolytic for the click cap.
Not sure just how much that click filter contributes to the sound. When that switch is engaged, it cuts off a little high end. I have never read any documentation on what this was designed to do. And when switched out, is completly out of the circuit. Was showing the pic because those two parts are not located directly on the circuit board. Those Mullard type vintage caps can be found if you want to build it like the original.
Ebay 1uF Mullard tropical fish caps

goom
Ultra Wiggler
Posts: 808
Joined: Sun Jun 07, 2009 8:03 pm

Post by goom » Thu May 23, 2013 10:28 pm

Thanks. I thought I remember reading that the click filter was to reduce the DC thump when using fast attack or release settings. I could very easily be wrong though.

User avatar
Dubka
Wiggling with Experience
Posts: 366
Joined: Sun Nov 27, 2011 11:39 am
Location: London, UK

Post by Dubka » Fri May 24, 2013 4:29 am

You'll have to excuse my lack of knowledge but - do any of the values change if I only have 3 inputs?
AND how important are the input pot values in mixer circuits?

User avatar
daverj
Vintage Video Wiggler
Posts: 8514
Joined: Fri Jun 19, 2009 11:09 am

Post by daverj » Fri May 24, 2013 3:41 pm

Dubka wrote:You'll have to excuse my lack of knowledge but - do any of the values change if I only have 3 inputs?
No.

Dubka wrote:AND how important are the input pot values in mixer circuits?
There are two things to consider:

1 - The load seen by whatever module you plug in is determined by the value of the pot, the setting of the pot, and the input resistor to the mixer. That load will always be the value of the pot or lower. (it is lower as you turn up the pot and the input resistor to the mixer is put in parallel with the pot value).

So with a 25K pot and a 22k input resistor the load seen by whatever you plug in is 25k when the pot is all the way down and is 11.7k when the pot is all the way up (25k in parallel with 22k). With the pot in the middle it's somewhere between those two values.

So a larger pot and a larger input resistor presents less of a load to whatever module you plug in. The larger the input resistor value is relative to the pot value, the less variation in load there is as you turn the pot.

2 - The value of the pots combine with the value of the input resistors in a way that changes the linearity of the pot. Any value pot will work (no signal when all the way down, full signal when all the way up). but the higher the value of the pot relative to the value of the input resistor, the less linear the pot becomes.

Here are three graphs showing the linearity of the pot as you rotate it. Each graph shows 4 different sized pots for the same size input resistor. I made graphs for 10k, 25k, 50k, and 100k pots vs 22k, 100k, and 180k input resistors. Clearly the larger the input resistor, the less of an issue the pot value is.

Image

Image

Image

Post Reply

Return to “Music Tech DIY”