902A Bandpass Filter Adapter (Clone)

[CZ Rider]

A pair of auctions for two original circa 1965 R.A.Moog modules. (Not mine.)
The one is the 902A Bandpss Filter Adapter, the other an early 907 Bandpass Filter Bank
Moog 902A Banpass Filter Adapter Ebay Auction
Moog 907 Bandpass Filter Bank Ebay Auction
These modules look very early with the audio portion using RCA type phono jacks. Eric Siday's early system also had these type jacks. Interesting to get a look at what was behind the panel on these.
The 902A has additional wiring and would be used in conjunction with a 902 VCA module. There was an early AES writeup by Bob Moog explaining the patch to make the 902 VCA into a bandpass filter. Would like to study the schematic on this one as it would probably be easy to clone. Just an audio in/out with a 4 position rotary to select a range with an additional on/off rotary. The audio went to/from the adjoined 902 VCA. Interesting!

That 907 is a configuration I have never seen. One of the first ones for sure.

[JohnLRice]

Wow, thanks for posting. I've never seen those before!

[thealien666]

They're listed as pre-production prototypes circa 1965.
Pretty rare, and interesting to see, indeed. Thanks for the link.

[BrianK]

They're listed as pre-production prototypes circa 1965.
Pretty rare, and interesting to see, indeed. Thanks for the link.

I see people in guitar and other vintage worlds naming things "prototype" and pre-production, without knowing what they're talking about. My friend was researching VOX amps for his (great) book, and anyone having something unusual or early, they called it "a prototype". In this case, these are indeed production models - made to go into an early production system. Early production, yes, but anything made after Herb Deutsch's system and sold would be (by its very nature) production models. I have seen early versions of Moog modules, some stamped as belonging to the factory, and some truly being test-bed production models, but it's usually quite easy to see.

These are indeed VERY early, and fascinating to see. Obviously, the Bandwidth control is something you'd not see on a later 907 or 914. I wonder if the 902 is just used as make-up gain for the filter module?

These early modules are also skinnier than the modules made later. Quite possibly they are from the first half of 1965, and only a few systems were sold then, although they did a lot of business selling just a few modules at a time...

[thealien666]

The fact that they are made of a breadboard pcb would tend to indicate pre-production and possibly even prototyping. A production pcb would certainly contain copper traces, even in those early days of pcbs nearing the end of the lug mounted electronics era.

[CZ Rider]

I wonder if the 902 is just used as make-up gain for the filter module?

The 902A is more interesting than that. This is a helper module much like the 904C filter coupler, and does nothing without being connected to a 902 VCA.
This is a module listed in the R.A.Moog 1965 catalog that I have never seen a photo of or any info on. Using the auction photos, one can figure out what this did and how it did it. Looking at the back the multi-pin connector and harness is still with this module. The other end would have connected to a powered 902 Voltage Controlled Amplifier. So this 902A is getting power from the 902 along with sending audio in and out of the 902 VCA. Just like the rotary switch on the 904C coupler, that on/off rotary switch would interconnect the 902A to the 902 forming a bandpass filter. The connected 902 VCA would operate as a normal VCA with the switch in the off position. Further controls on the 902A are the four bandpass filter ranges to select. Switching the 902A on, the lower audio in/out of the 902A would be the Bandpass Filter in/out with the adjoining 902 handling the voltage control functions.
Bob Moog described this making a 902 VCA into a bandpass filter in his July 1965 AES article "Voltage-Controlled Electronic Music Modules". The math is a little over my head, but by combining a "high input impedance buffer" and a "high output impedance buffer" through the VCA with a R/C circuit in the feedback loop, a voltage controlled bandpass filter is formed.
The two pages explaining this can be found on the Moog Archives site:

http://moogarchives.com/aes05.htm
http://moogarchives.com/aes06.htm
(The one block diagram should be center block VCA and not VCO?)

Looking at that 902A circuit board, there are two mirrored circuits. The mirrored input and output buffers routed through the 902VCA and the rotary selected cap and resistor pairs in the feedback loop for different ranges. Really simple to construct if I had all the values and transistor types. I guess the real mystery is what did it sound like? Bob claims high Q values were possible with this filter. But Bob had dropped this module from the 1967 catalog. Might not have been that great, or the 904 filter combo made this obsolete. In the full 1965 article Bob never mentions the 904 ladder filter. Probably still under wraps applying for the patent?
Here is the first page of seven that Bob Moog presented on October 14, 1964 to the AES from Roger Luther's site:
http://moogarchives.com/aes01.htm
Interesting read! Describes the 901 oscillator, 902 VCA, that bandpass filter patch and a few other early insights.

[CZ Rider]

The fact that they are made of a breadboard pcb would tend to indicate pre-production and possibly even prototyping. A production pcb would certainly contain copper traces, even in those early days of pcbs nearing the end of the lug mounted electronics era.

Not necessarily. Those early modules were production, but not yet mass produced. I think every 901C I have seen were on Vector Plugboard. They only made about 15 units, so they never got to the mass produced etched board stage, but they made them over a three year period. I also have an early 950 keyboard here with the same plugboard. I suppose in the early going they were still making small revision changes on some modules. Some production modules never made it from plugboard to mass produced eched board. But the 901C was a production module and many new release models sold, started out on plugboard. That August 1967 Dolenz system shown recently here has both a 914 and the 901C on plugboard along with the CP mixers. And the circuit wasn't shown, but that newly released 911A Trigger Delay module was likely on plugboard too.

Now that 907 bandpass filter bank might be a one off or the first. That has no coils, just caps and resistors, and only six bands with bandwidth and amplitude controls. Nothing like the production 907. And like that 902A, I wonder what it sounded like? Probably not as sculpted as the production 907.

[BrianK]

I see it ended early, and not too high a price. Someone got ahold of the seller to prevent a bidding war. Wonder if we'll know where it goes eventually.

[unfiltered37]

The 907 looks like a simple fixed 4 band eq with no frequency control. Looks like the peaking baxandall type.

[CZ Rider]

I see it ended early, and not too high a price.

Yeah, I noticed that today.
I wasn't as interested in owning an original as I am in putting together one to try it out.
I can almost figure it all out, but can't seem to find a ground on that circuit?
There are only 6 wires comming in. In the AES paper, Bob describes the 902 VCA as a four terminal "black box". The 902 is a differential VCA with both positive and negative inputs and outputs. And that 902A looks to have two pairs of audio/signal going in and out. There is a black/grey pair comming from the front panel switch, so a pair of signals leaving the 902A via that switch. Then there are a pair of yellow and orange wires returning the signal. That only leaves two wires to power the module?


Assuming the power color code is the same on that 907 where red is +12, and blue is -6, but there is no seventh green wire. The whole circuit does not connect to ground anywhere on the front panel either? Never studied a Moog modular circuit that does not use ground somewhere? Perhaps I'm missing something. I would still like to try this one out though. There is chassis ground for the audio between the 902 and 902A.

To study the circuit I was able to use those auction photos and superimpose the inverted board back on top of the front to see where everything goes. This is an old Photoshop 5 PSD file about 11 MEGS, where I can bring in the top or bottom view or blend them together.
Low rez capture of a 50/50 blend so you can see both sides of the board. Cool huh!


That board has 5 seperate sections I can identify. At the top a green wire carrying the audio output with a small circuit of 2 resistors,1 cap,1 transistor. To the left of that a brown wire with the audio input. Small circuit with a resistor and disk cap. From there it travels the left side of the mirrored "high input/output impedance buffer" circuits. The third section is those cap pairs selected via the 4 position rotary. Connected to each side of the mirrored buffers via both sides of the vertical 470K resistors. Leaves the left side via a white/blue and white/orange twisted pair of wires to the rotary. From the rotary to the 4 selected pairs of each large cap returning to the right side buffer via both sides of another vertical 470K resistor. The fourth and fifth sections are those mirrored buffers. These have grey and black outputs going to the on/off switch, with a grey and black twisted pair leaving the switch and routed to the 6 pin socket. (Would guess this is the +/- inputs to the 902 VCA). Looks like the signal returns via a yellow and orange. Both wires go to opposite sides of the buffers. (Guessing this is the return form the 902 +/- outputs? Unless this is where the missing ground is, but I don't think so?). Power looks to be +12 on the red wire and -6 on the blue. (Same color code as that 907, missing the green ground?)
The mirrored "high input/output impedance buffers" use only a handfull of parts. Right and left sides identical parts with slightly different wire routing. Each side had three transistors with a matched pair of NPN's (2N3392?) and a PNP (2N 4058?). 7 resistors and 3 small caps each side. Haven't yet identified all the parts though. Looks like 1% resistors, four on each side labled 1503, might be 150K? The 2 small caps look to be .005uF with a slightly larger unidentified one at the top. (.01uF?) Not really sure about any value of those small yellow axial caps.

Still have not identified the parts on the small output circuit. Two 1% resistors labled 3320F (332ohm?) and 1001F (1K?), a yellow cap a little larger than those .005uf caps. I can make out "058" on the transistor so it is a 2N4058 on the top circuit.

If anyone who knows these old Moog designs wants to help guess the parts I can't identify it would be greatly appreciated. Would like to try this one out, as I have a 902 to patch this circuit into. Might be interesting to hear the other Moog filter no one knows much about.

I know I would have fun with a different and perhaps strange sounding Moog filter!

EDIT: Had to edit the first wording. The input audio/signal disc cap connects to the left side of the buffer at the yellow cap, same connection as the black wire going to the panel mounted on/off rotary. Looked like it connected below to those two 1K resistors, but that is where the -6 volt supply is comming in. Would have let out the magic smoke. A closer look revealed where the other leg of that disk went. Luckly the auction had a close up from a different angle of the disk.
Also those matched NPN's might be 2N2926's like the ones used in the 984 mixer circuit. Looks like I can make out a 26 on the one. I asked the seller if they can help identify. Added identified parts to pic above.

2nd EDIT: Added a drawn plugboard schematic with color to help folow the internal signal flow. Other than the chassis ground from the audio inputs, there seems to be no ground in the circuit.
Any one ever see something like this before? The 902A is sort of a sub circuit to the 902 to make it a bandpass filter.

Fairly confident I have the parts right.

[CZ Rider]

Finally got around to putting this all together. First time attempt to clone a Moog module from just a few photos. I revised the schematic above to the latest version. The circuit does not use ground but gets it's voltage from the +12 and -6 supplies.
This looks to be the same as the original modular Moog voltage controlled filter. Probably in very early units 1964-65, before the well known ladder filter. Taking a second look at the original first unit, I noticed what looks like a 904A filter might actually be the 902 VCA. The controls are gain and a knob below with exp/lin. This is probably a 2-pos. rotary switch and the exact same set of controls found on the 902 VCA. Not typical filter controls. To the left a 1U module labled "904 Bandpass Filter". This would most likely be the combination of the later labled "902A Bandpass Filter Adaptor" and the Moog 902 VCA to form this early filter.
Can see the 2U wide VCA that almost resembles the 904A Low Pass Filter with a dymo label.


So all I have to go on is a photo of an original early 902A from an Ebay auction.
After carefully drawing out the components and wiring I started with a blank modular Moog frame. I used the same Mallory 3 pole 4 position rotary for the range select and same Centralab 2 pole 2 position rotary for the on/off switch and use standard Switchcraft 112 jacks. (the old original had RCA phono type jacks.)


Using the drawing I begin by placing in the Vector pins in the pattern for component placement.


Once the pins are in place the components can be soldered in.


My parts are a bit larger than the original. The precision 1% resistors are quite fat and I could only find 200 volt El-menco filter caps so those are a bit larger too. Was not able to find the 2uF El-mencos at a reasonable cost, so instead went with the Moog method of paralleling two 1uF caps together to get 2uf. So I have one extra cap for each set.
A tight fit with my wider parts, but it all wired up.


The clone parts side, next to the 902A photo, can see my extra filter cap and wider 1% resistors.


Front panel is close enough?


And the rear circuit wiring.


It looks like the original, but does it work?
And the million dollar question...What does it sound like?

Well it works and is quite unlike any filter I have ever used. Not sure if I have it exactly right because I do not have an original or any reference to compare it to. But it works as a filter, so it must be close? I subbed many parts experimenting and there is a feedback oscillation I was able to eliminate, but it would no longer sound like a filter, just a VCA. So this oscillation seems to be part of the design. Almost like a Minimoog filter at the brink of oscillation where the oscillation effects the tone. It is a constant tone though and not a wide adjustable range, but is seems to go away when the is a signal present. I discovered when processing audio, it helped to modulate via the CV input with a similar tuned source. This colored this oscillation and made it change with the note like a Minimoog filter would when set to track the keyboard.
The sound is raspy and has a grit to it. Similar to a violin bow scraping or throaty/buzzy sound from a brass instrument. Really unique the way the oscillation tone interacts with the audio signal.

Uploaded a few sond samples to give an idea what this filter sounds like.
First, what I would guess is a typical filtered 902A tone. It gets a different tone at higher registers and is somewhat smoother then the lower registers.
MP3 of typical tone

Can hear that constant tone and the raspy quality of the filter.
Tried using white noise and the oscillation can not be heard. Actually gets a pretty good hi- hat type sound
MP3 of white noise through 902A

Doesn't sound too bad and probably the best hi-hat I ever got from the Moog.
But these are the clean sounds it can do. The 902A gets this awesome dirt and grit when pushed in a different direction.
Here is a filter sequenced type patch. Sounds nasty, almost evil sounding.
MP3 of 902A filter sequenced

The resonance is somewhat a constant tone on the last clip. this is where I noticed if I modulated the CV input with an audio rate tuned oscillator, that static sounding resonance would now more closely follow the note played. Here is more of a lead tone with the modulated filter similar to using osc. 3 on the minimoog modulating the filter at a tuned audio rate.
MP3 902A lead with FX

Sounds bad ass! Getting the tracking modulation makes it less static and more like a traditional filter.
So this morning I was trying to get something completly different from the filter, when this scrapping violin sound comes out. The grit and dirt from the feedback oscillation turns into the "zing" sound similar to a bow scrapping.
MP3 902a bow sound

If this filter is broken, I will leave it that way. Some really strange tones this is producing. That type grit and dirt is missing from many electronic patches I get from the Moog. This is just the 902a in the samples and naturally in a modular system can be recombined with any other tone. But that was the best electric violin I ever got out of the Moog. Did this before and had to multi track a few passes for the strings and another few passes for the bow sound.
The best way to show the filter effects is to blend in each tone on the next sample. First just the feedback oscillation. A tiny bit of sequenced CV to give a slight difference between notes. Then I blend in the modulation oscillator that is in tune at audio rate patched into the 902 CV input but not the audio input on the 902A. Can hear how the oscillation morphs into the "zing" part of the bow sound as the oscillation fades away. (Could just use this method and skip the audio input.) Then I bring in the audio from two 901's mixed together in a CP3 mixer. Gets a bit gritty bacause the filter is being modulated at audio rate. Have to watch the levels but it sounds unlike anything else in that Moog system.
MP3 902a feedback sound blended in

This filter is wild!
So why did Bob and company discontinue this? Unknown. Perhaps because of the unstable nature of it. It is quite radical and not smooth like the ladder filter. But I like it and can see where this unstable filter will be usefull for adding that grit to patches. Like all Bob Moog designed circuits, this has a lot of depth and many uses.
So has anyone heard the original Moog filter much? I think so. Eric Siday more than likely used this filter on his early 1964/65 sound logos. They were all over radio and TV. One of the most memorable, the Screen Gems "S from Hell" sure sounds like that same filter and was produced at that same 1964/65 time period. I was always fascinated by that sound, and it has that same electronic yet gritty tone. As a kid I thought it sounded very futuristic.

https://www.youtube.com/watch?v=N6j8EhsJrIA

So this was a fun project to try. Might be the only working 902A? Might be broken and way off the original, don't know. The most important part, it sounds unique and bad ass!
I'll put it next to the 904 filter combo. Module on the top right.

[EricK]

For whatever reason those pictures don't display and the auction ended.

Did anyone snag the pics?
CZ, so is your goal to reproduce exactly as possible every Moog module? The clones you have done look fantastic! You put in a shitload amount of research and work into that. How long would you say it takes to fabricate a module like that?

[CZ Rider]

Hey EricK, I was able to snag the photos from that January auction. Only a few pics from each auction. My goal was to get a few modules I did not have to add to my system. When I first found that original 1969 IP it had a large 8U blank and 4U blank. So I had 12U of spaces to fill.
This is what it looked like when I first found it.

The system just turned 45 years old with build date Aug 11 1969.


I was able to find a few circa 1969 modules and added an extra 911 envelope generator, and a 911A dual trigger delay. Also a 912 envelope follower, and was lucky to find the 904B high pass filter and 904C filter coupler to complete the filter set, all circa 1969 original modules. The only original module the system was missing that the bigger systems had was the 984 mixer, so I cloned the smaller 982 and just added a ring modulator and now the 902A bandpass filter. So it kind of makes sense to add those different modules, and I had to make them myself.



Interestingly, a second look at that odd 907 filter bank, I just might have to try and build one of those next. Or at least experiment with a single stage of that filter. All six stages look exactly the same with different filter caps. Could make just one or two and add a switch to get the different cap values.
Here are the pics from the 907 auction.









Enough info to clone this rare 907 one too.
Have the pics of the 902A, but most were posted next to my copy above.

[EricK]

Thanks, I surely didn't see the date stamp on the op.

It's a nice thing having the cash to jump on the lucky finds. Not that I'm ever in that position but I assume it is nice.