Moog Source analog monophonic synthesizer

moog source glamor
moog source with taurus ii
moog source

12xx
12xx-front
12xx-rear
12xx-side
Update 03-25-2012

Where's the knobs? Features Tuning your Source
Modifications
MIDI Retrofit
Technical
Sounds
Why does it sound so good?

The Moog Source (its official product name was "The Source") was designed as a programmable Minimoog, released in 1980.  It was the first product from (1970s Norlin) Moog Music to incorporate microprocessor control (the Z-80) and also one of the first synthesizers to replace the panel full of knobs with a touch membrane panel and a data wheel.  I have been a happy Source owner since 1985.  During the analog revival of 1990-2000, this instrument was largely ignored and undervalued because it lacked a panel of knobs and switches.  Fashion wise it was a bit gaudy with its brushed aluminum case framing black bordered blue membrane panels peppered with gold, light blue, and red buttons.  But the orange program buttons?  Someone at Moog R&D let their kids loose with the crayons.  Walnut trim complimented the aluminum case nicely although they did work loose as they aged.   And it was a lot lighter and more compact.

So... where's the knobs?

Above the left hand pitch bend and modulation wheels is a large data wheel for editing the sound.   To edit the sound, you selected the parameter on the right hand panel which displayed the current setting in the small LED display, and moving the data wheel changed that setting up or down.  The data wheel is pleasantly weighted - it spins freely on a ball bearing shaft, and because the weight provided inertia to the wheel you can whip it back and forth with ease.  This makes an excellent controller for filter cutoff wankers.  The data wheel is basically a primitive rotary encoder, with an obviously in-house factory designed radial transparent ring with alternating stepped black stripes interrupting the light beams on a pair of optocouplers.  Basically a quadrature position and direction encoder.  The Source is an adequate application for a data controller/parameter select interface because of the low count of patch parameters - I always felt that the same interface applied to complex polyphonic systems with high parameter count was too cumbersome and unwieldy.  Close on the heels of the Source was the ARP (later Rhodes) Chroma with a similar interface, and indeed the common complaint was too many parameters for a data controller interface (and the Chroma had a slidepot data controller not a wheel).  And any data controller is far better than a pair of up/down buttons.

Many people worry about the membrane panel wearing out.  I can tell you that I have not had a single problem with mine.  What can happen is that people push the buttons with their fingernails and those sharp nails WILL cut through the membrane surface and damage it.  So the secret is use your FINGERTIPS not your finger nails.  top

Features

While it omitted some features of the Minimoog, the presence of a microprocessor permitted some useful additions.  Two 88 note sequencers, a 24 note arpeggiator, a program sequencer, and a patch archive cassette tape interface have been added.  The sequencer is a real time (not step) sequencer that records note duration and interval between notes.  Using the "page 2" functions, you just press REC1 or REC2, start playing, press STOP then press PLAY1 or PLAY2 to play it back.  The sequencer clock is driven by the LFO rate.  According to the manual, you could use the program seqeuncer to assign different patches to any sequence event.  The arpeggiator is very easy to use - when it is activated, you simply play any pattern up to 24 notes (although if you play C-E-E-E-G-C, the arpeggiator plays back C-E-G-C and doesn't repeat the triple E pattern).  The arpeggiator is driven from the LFO so varying the LFO rate varies the rate the arpeggiated notes are played back.  Your input pattern must be such that the first and last notes are the same as this defines the repeat point, and when you press that last note the arpeggiator immediately starts playing.  While the arpeggiator is active, you can change patches and tweak the sound in real time - but if you slip and touch a key, the arpeggiator stops because it "thinks" you are entering another pattern.

The orange PROGRAM buttons give you a whopping 16 locations to store your patches (hey it wasn't bad for 1980).  The large LED displays the current program number.  The buttons also perform double duty as "page 2" - you access them by press the LEVEL 2 button on the far right.  The legends printed above the PROGRAM buttons are the "page 2" functions.

The third oscillator of the Minimoog was replaced with a simple triangle/square/S&H LFO routable to VCO and/or VCF, but the LFO would only go up to 100hz which limited the special effects that were popular on the Minimoog (you can't get clangorous audio FM on the filter or VCO cross modulation effects).  The keyboard was shortened from 44 note F-to-C to 37 note C-to-C although the keyboard mechanism is more reliable (of all my Moogs, it's the most reliable) and the left hand panel has octave buttons that raise or lower the keyboard range by an octave.  Below the octave buttons are the pitch bend and modulation wheels. 

Both oscillators sport triangle, ramp, and variable 5% to 95% pulse waveforms (the minimoog did not have variable pulse waveforms, only fixed 50% 35% and 15% pulse) and footages of 32', 16', and 8' while the detune control of VCO#2 had a two octave range pushing it into 2' territory.  Missing is the Minimoog "tri/ramp" combination waveform.  Variable pulse width is a very useful addition, but it's a shame they didn't include pulse width modulation from the LFO.  VCO#2 can be hard synced to VCO#1, and when hard sync is active the pitch bend wheel is converted to bend OSC#2 only - one of the best features as the hard sync on this beast SCREAMS.  Unfortunately you can't route the filter EG to the hard sync'd oscillator like you can on the Moog Rogue or Memorymoog.  When tweaking the detune, the data wheel has a finer resolution for easy tuning.  The beating between closely detuned VCOs has that classic Minimoog sweeping sound and the VCOs do not lock.  Both VCOs have temperature compensation for stable tuning, and my unit has been very stable all these years.  The Source includes the obligatory glide processor which is essential for that Minimoog sound.

A mixer for varying levels of both VCOs and pink noise provides the option to overdrive the filter, a desireable trait from the Minimoog.  Pink noise unfortunately isn't all that useful with the limited modulation options of the Source.  Since I relegated my unit to bass duties, I modified it so that the NOISE control actually adds audio rate FM to the filter via VCO#2 which is a very useful modulation for bass sounds.

There is a master VCO2 scale "soft-trim" in the mixer that trims the volts per octave scale for VCO#2.  If you want the detuning of all your patches to be the consistent between sessions, this is a good practice when you turn the synth on.  Allow a five minute warmup before making this adjustment.  I usually have a "tuning" patch with both oscillators tuned to known unison after a calibration.  When I call up this patch it has only one oscillator firing so I can adjust the master tune on the rear panel to a reference, then I can quickly bring in the second oscillator using the mixer and verify the scaling.  Before making the scale adjustment the mixer must have both VCOs audible.  This "soft-trim" is activated by holding down the STORE button and pressing the OSC 2 button.  Then while holding high C you adjust the data wheel until the two oscillators are in tune.  The reason you use high C is this is the range that is affected by the "soft-trim".  In my experience with my unit I have had to touch up scaling every time I turn the Source on because the internal trimpots (there are SCALE trimpots for each VCO on the PC board in addition to the master VCO2 scale "soft-trim") don't always retain their position.  If you cannot get the oscillators in unison using the master scale "soft-trim" then your Source needs to be calibrated.  Make sure your tech understands this scaling system. top

Moving on to the filter, this is the classic 24dB/oct lowpass ladder filter made famous by the Minimoog.  You can adjust cutoff frequency, resonance, EG amount (no negative sweeps), although keyboard tracking is limited to fixed none/half/full.  The filter can self-oscillate into a sine wave, and when keyboard tracking is set to full the frequency of the sine wave can track the keyboard and the data wheel has finer resolution for tuning the sine wave to pitch.  This particular filter has better resonant colors than the Minimoog.  Creamy Minimoog lead sounds and fat basses are easy to get - the bass sounds have a little more bottom end with that classic Minimoog midrange girth.  Between the colorful filter resonance and variable pulse width on the oscillators, the Source is an excellent bass synthesizer capable of a bigger variety of sounds than the Minimoog, and it serves quite well as a "poor man's Taurus" in that you can get real close (not quite) to the classic Taurus pedal preset that was popular on the Taurus I bass pedal synthesizer.  However there is no external audio input for the filter and there are some Minimoog sounds not possible on the Source.  Spoinky basses - patches with filter resonance turned up while the filter cutoff is spiked with an EG with fast decay to zero sustain - aren't quite the same on the Source because the resonance on the Minimoog filter is nonlinear.  "Nonlinear" means the resonance on a Minimoog filter does not stay constant across the full frequency cutoff range - it disappears as the filter cutoff falls to low frequencies.  So when a "spoinky" bass is dialed up on a Minimoog, the filter resonance disappears as the EQ sweeps the filter toward low cutoff - the net effect is the bass sound is full sounding.

Moog subsequently viewed this nonlinear resonance as an engineering "fault" and corrected it in new designs - including the Source.  So the resonance on the Source filter stays constant across its full cutoff frequency range and the engineers were happy.  But this was also why later Moog instruments also didn't sound like a Minimoog and the customers were not happy.  Moog went through a lot of head scratching to find out why and never realized that their "correction" was the source of the difference.

Another elusive Minimoog sound is that "growl" on certain filter cutoffs - the only keyboard in my arsenal that duplicated that "growl" was my Moog Voyager. 

The Source has a dedicated envelope generator for the filter and another for the VCA.  They are full ADSR envelope generators - no more "shared" decay/release control  like the Minimoog.  Musicians had to wait ten long years for Moog to progress to this system - the "shared" D/R was propogated through the Micromoog, Multimoog, Polymoog, Taurus I, Liberation, Prodigy, Rogue, and Taurus II.  There is a popular myth that the EGs are digital - in actuality they are analog (technical dissection of the circuit is below).  One nice addition thanks to microprocessor control is the EGs can be single or multiple trigger, the latter being essential for bass sounds.  You can also use the LFO to trigger the EGs.  The EGs on the Source have a nice fast snappy attack - the range of transient times are 3 milliseconds to 10 seconds.

To store your patch, you press the HOLD button on the left side.  This freezes the panel so no further edits can be made, and it makes the STORE button active (two decimal points will light up in the LED display).  In HOLD mode you can also do a compare by pressing any of the PROGRAM buttons - while holding the button down the patch in that position is active, when you release the button it snaps back to the edited patch waiting to be stored.  When you are ready to save the patch, hold down the STORE button then press the appropriate PROGRAM button where the patch is to be saved.  This convention prevents accidental patch overwrites from an inadvertant press of the STORE button

But don't be dismayed because it doesn't sound like a Minimoog - the Source is an excellent instrument.  Only recently is the Source getting its recognition as a respectable synthesizer - thanks to its filter - and market prices are beginning to reflect that.

Rear panel interfacing included CV input and output, and S-trigger input and output.  "S-trigger" means the EGs are triggered with a short to ground, "V-trigger" or voltage trigger means the EGs are triggered with a voltage.  Synthesizers never have a standard trigger system (and still don't), and to further confuse the market some voltage triggers use +5 volts, some use +9 volts, some use +12 volts...    If you're trying to drive the Source from a voltage trigger system like an external sequencer, you will need a V-trigger to S-trigger conversion circuit which can be found online.  Gone but not missed on the Source are the awkward cinch-jones sockets for S-trigger, now replaced by more convenient 1/4" jacks. top

Modifications and Factory Service Bulletins

I made a few modifications to my Source over the years - some are custom and some are for fixing known problems.  One annoying feature is that when you power on the Source, the OCTAVE button defaults to +1. Since I was using it as a bass synthesizer, I added a capacitor across the "OCTAVE 0" button, which changed it so it defaults to "OCTAVE 0" on power up.  This works because when the machine is off, the voltage across the cap drops to near zero essentially shorting the contacts of the button.  When the power is turned on, the voltage on the cap charges slow enough for the microprocessor to detect that the "OCTAVE 0" button contacts are closed.  Another cool modification is modifying the Taurus II controller to control the Source remotely, described in factory service bulletin #853B (scans of page 1, page 2, and page 3).  The Taurus II controller is strictly a bass pedal controller only with no internal synthesizer circuitry - it has CV and trigger outputs only.  The reason I did this was when I routed the Taurus II control outputs to the CV/trigger inputs on the rear panel of the Source, you cannot get multiple trigger and the CV input is not processed by the glide processor.  This modification corrects that.  I used the Taurus II/Source combination to great effect for many years, and Mike Rutherford of Genesis adopted the Taurus II/Source system in the last years of the band.  You can catch glimpses of the Taurus/Source in their 1992 Genesis Live video (not sure it ever got released on DVD).

There are known problems on many Sources.  One is the noisy volume control on the top left corner - it will "crackle" when you move the knob.  This pot is actually in the feedback loop of the final audio opamp, which isn't a great design because intermittent wiper resistance drop-outs due to mechanical design of the pot will cause the opamp to momentarily jump to full gain, which causes the "crackle".  I re-designed that volume circuit by soldering a fixed 4r7K resistor in the opamp feedback loop, replacing the 5K log pot with a 100K audio pot, and tapping the output of the final audio opamp to the pot with the wiper wired to the audio output jack - a simple voltage divider like everybody else!  Another personal modification of mine is filter FM by VCO#2 modification which displaces the noise source (I never considered the noise source as appropriate in a bass synth). This mod lets you vary the level of FM to filter cutoff via VCO#2, and by using the NOISE parameter in the mixer this level is programmable. This is a really cool mod that is useful for adding a little "dirt" to bass patches, for clangorous bell sounds, and for getting that elusive "Tom Sawyer" opening rezz sound.

The second common problem is the "Crazy Source" which is caused by oxidizing leads on the power supply regulators.  The factory service bulletin #827A addresses this issue and it can be found online.  Another known issue according to bulletin #862 is "several notes playing out of scale non-tempered tuning or sequencer playback rate varies" which is caused by a firmware error.  You need firmware version 3.3 or later to fix this, unfortunately you'll need a friend's Source with the correct EPROM and another friend to copy them with an EPROM burner.  The only way to check the firmware version is to open the hood, locate U23 on the digital board, and the version will be hand written on a sticker on top of the 2532 EPROM.  If your Source loses patches, check for 1) the presence of the modification specified by bulletin #863, and 2) your backup battery is good.  Batteries have about a thirty year life, I had to replace mine in 2011.  Fortunately Panasonic still makes the Lithium BRC battery, but I highly recommend ordering a battery with tabs (shown on the right in this pic) and using an external battery holder with the new battery.

There were two versions of the Source.  Later models serial #3180 and higher added the five pin Roland DIN sync interface.  If you count three 1/4" jacks and a Cassette DIN jack on the rear panel, it's the early version.  If you count five more 1/4" jacks on the right of the rear panel, it's the later version.  Moog offered an upgrade package for early units - if you see two black overlays around the sets of jacks on the rear panel, this is an early version with the upgrade added.  While some wiring changes were needed, unfortunately it also required a small PC board and firmware version 3.2 or later both of which are no longer made.  This upgrade is specified in factory service bulletin #833.

Moog designed a digital sequencer in similar case design, but it was never put in production.  The only known specimen currently resides at Audities Foundation. top

Encore Electronics MIDI Retrofit

In 1991 Encore Electronics introduced a comprehensive MIDI retrofit - I was one of the first customers to buy one.  Firmware version 4.08 added MIDI reception of note on/off, program change, pitch bend, filter cutoff, filter resonance, portamento, MIDI clock control of the arpeggiator with start/stop/continue and clock divider, MIDI to CV conversion (using the rear panel jacks), expansion of patch memory from 16 to 256, and patch dump/restore over MIDI sysex!  MIDI Bank Select is used to access the 256 patches via 16 banks of 16 PROGRAM buttons.  And you could control the Source over a four octave range from a MIDI controller.  Because of EPROM space, the sequencer and cassette interface features was omitted from the firmware to make room for the MIDI functions (not a big loss with MIDI sequencers becoming more widespread and MIDI sysex being far superior to cassette tapes).  This retrofit was done by disassembling the original Source firmware and re-writing new code for it - far superior to simple MIDI to CV converters.  Because of the stock design of the mod wheel, MIDI control was not possible.  This is another reason that I relegated my Source to bass duties.  There is also no MIDI transmission of note, program, or pitch wheel - also not a missed function as the Source isn't much of a controller anyway.

Firmware version 4.09 added MIDI sysex patch and edit buffer dump/load request commands, but at the expense of removing the arpeggiator.  Version 4.10 is the latest that fixed the problem of synth engine malfunction when spinning the data wheel. top

Technical Discussions

There is a common myth that the envelope generators in the Source are digital.  This dissection will explain that they are actually analog. 

Near the bottom of this schematic is a section labeled "Contour Generator".  We'll study the filter envelope generator as the VCA EG is identical in operation.  There are two CVs labeled FLT CNTR LEVEL (near R196) and FLT CNTR RATE (near R199). The 3080 OTA (U44) is configured as a voltage controlled resistor that charges capacitor C57 - by varying the resistance, you vary the time it takes for the cap to charge (attack time) or discharge (decay or release time). The actual envelope originates from C57 which is a true analog signal. FLT CNTR RATE varies how fast C57 is charged, while FLT CNTR LEVEL varies the final charge voltage of C57 (IE sustain level).  You can create the separate attack, decay, sustain, and release stages of a complete EG with just these two CVs.

The contour of the EG is converted to a TTL compliant signal via LM393 opamp (U41B). If you view this schematic in the section labeled OCTAVE CONTOUR & MODULATION INTERFACE, this TTL signal (labeled FILT CNTR) is monitored by the CPU via the 74LS367 (U32). This way the CPU can detect the completion of any EG stage and transition to the next one.

Very slick, compact, and elegant design. And it can be expanded beyond a four stage EG by software. I'm surprised they didn't patent this circuit, I've never seen it anywhere else.  Why didn't Moog implement voltage control of existing ADS designs?  The answer is economics.  The Source only needs two control voltages to implement a full ADSR.  Converting an old design to shoehorn CV control in place of four passive ADSR controls would be more expensive in part count.  You need a control voltage for each ADSR envelope stage - each control voltage comprises a S&H circuit and an OTA circuit.  Multiply those TIMES FOUR for a full ADSR.  Double that for separate VCF and VCA EGs.  Those parts add up quick.  Also the design used in the Source eliminated the dreaded control feedthrough problem so common with VC'd EGs.  Oberheim ran into this problem while trying to design a programmer for their polyphonic four voice SEM system and eventually had to enlist the services of the late Doug Curtis who designed a workable VC'd EG in a chip - the predecessor of the CEM3310.

The keyboard scanning system is also unique.  It is a low priority system that uses a single contact buss, but the scaling is far from 1 volts/octave.  The reason for this is that the keyboard is an input device for the processor, it does not generate the pitch control voltage directly.  Remember the sequencer and arpeggiator?  The microprocessor needs to know what keys were pressed.

The system uses a successive approximation technique to decipher what key is pressed.  To reduce cost, Moog foresook an ADC for a comparator system.  Since there is no other analog voltage that needs to be monitored by the system, the cost of an expensive (in 1980) ADC was not justified.  A comparator is a specialized opamp with two inputs that are "compared", and an output that indicates greater than or less than result.  So what the system does is generate a series of known step voltages to compare to the unknown keyboard voltage.  The comparator tells the system if the unknown keyboard voltage is between two known "steps" - greater than one, less than the other.  It is continuously scanning for a pressed key.  This continual comparision of known "steps" against an unknown voltage level is called successive approximation.

If no key is pressed, the system knows this because the lowest possible step voltage is not found.  If any key is pressed, the system finds the two known "steps" and deciphers which key is being played.  When a higher key is pressed, the system finds that one and can generate a multiple trigger if enabled.  Now because the DAC is generating the known keyboard steps in a zero to ten volt range, the keyboard scaling is set to about 3 volts/octave to avoid quantization errors - that is why it is not 1 volts/octave.  The real 1 volt/octave pitch voltage is generated by the microprocessor at the DAC output.

One advantage of a single buss keyboard is reliability.  With older designs utilizing two separate busses for trigger and pitch voltage (some ARPs used a third buss for "gate"), the closure of the busses had to be slightly offset or a "glitch" could be heard in the oscillator pitches.  The single buss system eliminates this problem.

If your data wheel only changes any setting in one direction but not the other, you have a bad optocoupler.  Installing a new one is a bit tricky.  The two optocouplers must be physically spaced such that the phase offset of the outputs signals are 50%.  Since the parts must be unsoldered to physically move them, it would be much better if we knew the correct physical location of each optocoupler.  I had to replace a bad optocoupler on my unit many years ago.  Before I removed the defective one, I simply "gauged" the gap between the existing optocouplers.  I simply took a set of precision drill bits and found the one that represented the gap - in my case I used a #55 drill (0.052 in diameter).  So when I installed the replacement part, I used that drill bit to locate the correct position of the new part - worked like a charm.

And finally, the Source has a design convention that spells reliability: the sample and hold (S&H) system is on the same board as the analog voiceboard.  The S&H system receives a multiplexed analog voltage from the DAC and generates all the control voltages for a synthesizer voice.  One major reason why hybrid analog polyphonic synthesizers such as the Memorymoog, old Oberheims (FVS, OB-X, OB-SX, OB-Xa, OB-8), and old Prophets (P-5s P-10s and T-8s) have so much trouble with reliability is that the S&H system is on a separate board from the analog voiceboard whereto a connection system distributes these voltages to the voiceboards.  Too often, the connection system is the achilles heel as the contacts oxidize and the voltages go out of whack.  By placing the S&H system on the voiceboard, connector contact count is greatly reduced and the system is more reliable.  The tradeoff is more parts count as the S&H circuits are duplicated across each voiceboard.  This is also the convention with the Rhodes Chroma (actually an ARP design before they were liquidated), which has a good record with voiceboard reliability.  top

What does it sound like?

Some bass sounds that this thing can do:

Rush "Subdivisions"







Vintage Sound - Why?

Why can't modern synthesizers or softsynths duplicate the creamy fat sounds of the Source or the Minimoog?  Some people think it's the filter, some people think it's the oscillators.  The real reason is the Minimoog's discrete VCAs and the OTAs sprinkled throughout the circuitry of the Source.  In order to implement microprocessor control in the Source, many passive components formerly used in the Minimoog (such as pots) had to be replaced with OTAs. 

An OTA - Operational Transconductance Amplifier - is a rudimental opamp with an extra input that controlled the gain of the opamp.  I use the term rudimental because the OTA can't handle large signals, has a weak current output stage, and can only perform a subset of traditional opamp functions.  OTAs do have other talents because they can be configured as a voltage controlled variable resistor (to replace a passive component) or as a VCA.  However OTAs were not high fidelity audio devices - they have a linear low distortion region only for very small signals, and if you exceeded this region your distortion would increase.  This distortion also varied by frequency.  This was the best that technology had to offer in the 1970s.  And although the Minimoog did not use OTAs, the VCA was a crude design that wasn't low distortion high fidelity either.  Back when it was designed in 1969 that was the best that technology had to offer.

Moog designed the signal path in the Source a bit hot that drove the OTAs into subtle distortion, creating that creamy fat sound that we all know and love.  OTAs are used for mixer control, in the feedback path of the filter, and in the VCA.  That's five OTAs wreaking havoc on the sound!  The accumulation of all that subtle distortion at each stage is creating that creamy fat sound.

Today high fidelity VCAs are readily available and they are easy to implement in softsynths.  What they missed out is the distortion of the classic OTA which is a direct contributor of that "vintage sound".  Modeling dynamic distortion in softsynths is a big challenge because the exact mathematical model is not easy to derive and it is a major number cruncher to implement on a microprocessor.  Many owners of Moog Voyagers have noticed that when you opened the filter all the way, it was missing that Minimoog high end "sheen".  I was able to confirm that the Minimoog VCA was the contributing factor by routing the Voyager output to the Minimoog external input which put the Voyager through the Minimoog VCA and - wala - instant Minimoog "sheen" on the Voyager.  The Voyager VCAs are too clean!

The same OTA distortion was responsible for that big Oberheim sound on their early polyphonic synthesizers.  By the time the OB-8 was designed, the OTAs were replaced with clean VCAs and they lost "something" in the sound.  OTAs contribute more to the "vintage sound" of synthesizers than designers care to admit.  When Moog Music set out to re-issue the Taurus I bass pedal synthesizer in their Taurus III, they even duplicated the dirty sounding OTA in key circuits.  Taurus enthusiasists all over the world were delighted at how close the Taurus III sounded to the original. top

The Source - a long neglected synthesizer that is finally getting its due.

What the h#ll is that thing in the first picture on this page?  That's the coupler of a railroad caboose.  Choo-choos are another hobby of mine.

contact info

Home