Minimoog model D differences

[CZ Rider]

Had a little time to get some scope shots of the different Minimoog oscillator board waveforms. I only have the CA3046 type used stock on almost all from below serial 10175, and the uA726 type used on all above serial 10175. Only the first few hundred or so Mini's had the original discrete oscillator board, and I did not have one of these to test. Over the years many Mini's were fitted with the latter "stabilized" uA726 type boards.
So, how different are these two oscillators? I was quite surprised at what I found, and never noticed untill I looked at them side by side under the scope.
Both oscillator waveforms measured straight from the oscillator output before the 25K mixer potentiometer, same "C" note on both.


From those scope shots, the CA3046 oscillator has a slightly hotter output. The CA3046 sawtooth is a little bent in the middle where the uA726 is more of a straight line. The surprise was the difference in the rectangle waves. While the square wave is about the same with the CA3046 having slightly hotter output, the rectangular waves are opposite. The exact width is a bit different too with the uA726 having a narrower pulse. Modulating with oscillator 3 the difference is easy to hear. Modulate the oscillators using a LO setting from the pulse waveform the CA3046 will go high longer, and the uA726 stays low longer. Filter sweeps with slow pulse modulaton would be different too. Owned these two Minimoogs for 30+ years and never noticed this easy to spot difference. Always a learning process!

The final output from the two Mini's pulse looks different too, and the signal from the CA3046 Mini is a bit hotter.


The filter differences seem a bit deeper. Not only were the first few hundred Mini's the only ones with fully matched pairs of ladder transistors, the capacitor brand/selection varied from year to year. From what I have gathered from photos and schematics of Mini filter boards: Most all Minimoogs have the bottom pair in the ladder matched wiith the remaining pairs somehow selected within a certain range. The caps used varied from the first thousand or so units using those Mallory "tropical fish" type. Some of the early '72 minis having those white type. And the remainimg majority of the Williamsburg era using these brown type in the ladder. (Haven't identified the brand for those white and brown type yet. Anyone?) At some time, perhaps with the Buffalo era, Mini's began using those plastic box type caps in the ladder.
The first 1000 Mini's filter boards had a few different resistor values, with the output to the VCA hotter (through 47K resistor) and the final buffer output lower (pair of 27ohm resistors). After about serial 2000 all Mini's had a higher resistor value feeding the VCA (82K) and lowered values on the final buffer (pair of 8.2ohm resistors). The very early Mini filter boards had a pair of 220K resistors feeding the filter buffer, (R20, R24 tag on newer schematics). And early on was changed to 100K for R20. Unknown why this was changed, but almost all Mini's have the 100K, 220K pair ? A few other minor changes like a range trimmer added and different resonance feedback resistor, but only the very first filter boards had those differences.
A few photos of the Mini ladder from various years.


Not sure what the impact of the various type caps have on the sound quallity? But this would be an interesting project, to construct a Mini VCF/VCA clone with selectable sets of capacators like the Moog 904A did. Could not only have matched sets of the various .068uF caps, but even a set or two of higher values like the 904A did. A six position-four pole rotary would do the job.

Something like this:


I'm tempted to swap in a set of matched "tropical fish" .068uF caps for those square plastic box caps in my 101XX Mini. And at the same time swap in a tightly matched set of TIS97 transistors, so I can measure the installed "selected" ones. And to see if I can actually hear any sonic difference. But the clone with a cap selector sounds like it may be more fun. Could have a 1971, 1972, 1974 and 1977 setting along with a deeper set of .1uF.
Have a matched set of caps ready and a couple hundred TIS97's to match.

Could be a fun project!

Anyone have Mini ladder photos with serial info to add ?

[thealien666]

Interesting. But isn't the CA3046 only used as current mirror for generating the exponential current to drive the oscillator frequency ? If so, it shouldn't affect the sound at all. Only the actual oscillator circuit design itself and the components tolerances would account for audible difference in timbre (and visual differences on the scope). BTW, the fact that the square and pulse waves are inverted is also probably a result of the new circuit design of the more recent version of oscillator circuit design, and not the fact that it uses the uA726 instead of the CA3046. I might be wrong. I'm not an electronics engineer.

Anyway, here's my #6909 Minimoog D board #4 from 1975.



EDIT Those Japanese electrolytic caps were all checked for their ESR, and all show better-than-new values, even after 39 years ! That's why I didn't replaced them. Now, that's high quality for ya !

[CZ Rider]

Interesting. But isn't the CA3046 only used as current mirror for generating the exponential current to drive the oscillator frequency ?

Not sure what else to call/identify the common oscillator board ? Many call this the early board and is confusing because the discrete R.A.Moog is really the early one. To add to the confusion many of the CA3046 boards had R.A.Moog etched in the copper. My 1973 serial 39XX had that marking on the oscillator board. I have seen someone offering a rare R.A.Moog oscillator board, thinking it was somehow worth more, but it was just an early CA3046 common board. Not sure what else to refer to the three different oscillator versions to avoid confusion? Discrete board, CA3046 board and the uA726 board.

Any idea what brand those brown caps are ? They seem to be the most common on mid production Mini's. The only difference I can find on 2000 to 13000 serial Mini filter boards are the caps used.

[thealien666]

Silly me.
I didn't realize that you were using the names of those ICs just as a means to differentiate between those oscillator boards...

As for those caps, I don't know. I guess Kevin Lightner might be able to identify them, but unfortunately he doesn't seem to come here anymore...

One thing surprises me about them though, the markings on some of them state 50 volts maximum operating voltage when only a few volts actually run thru them (if I'm not mistaken)... Talk about overkill !

[MC]

Those brown caps look like ceramics but are early mylars. You would never use a ceramic cap for RC circuits like the VCF, they are not suitable for them.

The white caps are polyester and are a telltale sign of Norlin era units.

Those caps are normally available in high voltages and do not affect the sound.

I bought a surplus discrete oscillator board from Kevin, haven't gotten around to fixing it yet.

It is very common for square waves to not be square. The pulse width was set by fixed resistors and they would drift with age. Two of my VCOs do that, and I actually find that to be a feature.

The ua726 board has opamps in the audio path. The only opamps on the 3046 board are the CV summing and cap reset control circuit, not in audio path. The discrete VCO board is completely discrete.

The ramp waveform on the 3046 board has some curve to it, whereas the ua726 ramp is nice and linear.

Interestingly, the Voyager waveforms emulate the "curvey" waveshapes of the early VCOs.

Also the filter boards on early Minimoogs did not have a scale trimpot.

[megavoice]

Might be a little help for the identification of those white foil-caps:
I've been told yesterday they can be found in many, many Italian products. Probably anybody knows more about this and let get some other ones forward.

But I've also been told there are modern ones available of very high quality and to buy, but not matched, unfortunately. The brand is called WIMA.
Other brands that had been of high quality are "Plystrol" or "Tyroflex", but they're also unfortunately out of production for a longer time now.....................

[BrianK]

Interesting results. But maybe (if there was a way) to test between boards of the SAME type, to keep it scientific. And make sure those boards behave the same, THEN we can more safely compare different board eras as "typical" for this era...?

As I used to say when people asked "does it sound as good as a Minimoog" (about anything), I would replay "ALL minimoogs sounds different" - and you're finding some of the specific reasons why this is so. Good stuff...!

[megavoice]

Interesting results............

As I used to say when people asked "does it sound as good as a Minimoog" (about anything), I would replay "ALL minimoogs sounds different" - and you're finding some of the specific reasons why this is so. Good stuff...!

All items are developed and fabricated by human beings. There is NO voodoo anywhere.........neither in the items nor in our heads........
Till now I believed only in two persons in my life: first in Father Christmas till I was 6, and later up to now in my tech............

[thealien666]

From the stand point of electronic design, it all boils down to : components tolerances and inevitable analog signals variations.

Whenever there is analog signals to be generated, and/or processed, there will be differences between them. Like the differences between two Minimoog D of the same year, even only one serial number appart.

[megavoice]

.........Whenever there is analog signals to be generated, and/or processed, there will be differences between them. Like the differences between two Minimoog D of the same year, even only one serial number appart.

Oh yeah, but this is an explenation based on science..................

[thealien666]

Oh yeah, but this is an explanation based on science..................

Is there anything else ???

[megavoice]

L

Oh yeah, but this is an explanation based on science..................

Is there anything else ???

Yes, look back in the years to lots of "connoisseurs":............"All Minimoogs sound different, but I don't know why".....................

[CZ Rider]

Looking at the differences in board no.4 right now, with the filter and dual VCA's.
Although there were always minor differences between Mini's manufactured at the same time. I did notice more of a drastic sonic difference when comparing a newly made Buffalo Mini back in late 1977/ early 1978 when I tried one. This was before the stable oscillator board was offered and had the same oscillator type as the earlier 37Xx I had. And my Mini was only a few years old while the Buffalo Mini was new. The new one just sounded so much brighter/harsh compared to the mellower/softer sounding Mini I was used to hearing. If the circuits are exact in both, the component sources might be part of the difference in sound. Have noticed the different type caps used in the ladder filter and plan to try a few different types to see what effect it has on the tone.
Still did not identify the manufacturer of those brown caps. These seem to be used on many of the 2000 thru 6000 serial Mini's. Might be difficult to source those. There is a triangle on them, but most of my surplus connections have not turned up anything similar looking.


The type caps used may or may not have any effect on tone, but I plan to make a clone with the various types switch selectable just to try. So far I have a set of the original Mullard "Tropical Fish" type. A type never used in Mini filters but used in the many early Moog modular filters made by EL Menco. (Brown glossy looking.) And a set of those box type metal film like the latter Buffalo Mini's had. Also plan on switching in a set of both Mullard and El Menco's with a higher value, (.22uF/.27uF or higher 1.0uF) just like the Moog 904A did. Have to experiment. There are seven main caps in the audio chain, the four .068uF in the ladder, two .22uF feeding the filter buffer, and the .33uF entering the VCA.
Trying to match them by uF with a close set of four.
A few Mini cap types and a 5 position, 6 pole rotary to select the different sets.



I redrew the filter/VCA/mixer schematic outlining the differences between the early and later revisions. Some changes were minor like the electrolytics were changed, probably due to availability. The early Moogs used both 80uF and 200uF that today are difficult to find and changed to a more standard 100uF and 220uF. The other changes in the VCA gain might be more of a sonic difference. Making a clone, it would be easy to have a two pole, three position rotary to try each resistor combo in the dual VCA's.
Here is a schematic with the differences I found. (A little large JPG.)


So those are the minor differences in the circuits. The change of R20 from 220K to 100K was done early and looks like only the first 50 or so had that combo. While the VCA resistor change was the first 1000 or so units. (Before ser. 2000.) All the early Mini's used those Tropical Fish caps and some time after 2000 began using those brown type most Mini's had untill they switched over to the box type. Not sure when they made the switch to the box type? (Most of the parts on my 10,4XX like the transistors are stamped with "Thialand" from the Norlin era.)

Matched sets of transistors are another possibility. Moog would change the tolerances used after Bob was no longer in control of such matters. (1972-on) Perhaps cost/time cutting measures. There should be no doubt, if one was making 5 filter/VCA boards. The ones that had all matched parts should all sound consistent. While a batch of 5 boards without carefully selecting components might have inconsistencies.

The Moog sheet on matching transistors seem to have an error. While the common schematic for the NPN is correct, the lesser used PNP circuit is missing the 10K resistor.
Here is a fixed schematic with pinouts of the common Mini/Modular transistors.
(Handy printout!)


Would not make sense to make a Mini filter clone without matching the ladder transistors and all the other mirrored pairs. Also would not be that difficult to have a rotary selector on the filter to select the different type caps, and rotary on the VCA to select the different gain combos. I'll post any results when I get this together. Still sourcing parts.

[thealien666]

Unless the comparison between Minis is done scientifically and rigorously, by using the exact same audio monitoring equipment, side-by-side simultaneously, any subjective reference of the type "it seemed brighter, or softer, warmer or colder", or any other qualifier of that type is pretty much meaningless. Especially when we all know how memory is something that is easily distorted with time.

Another thing I've learned over the years is, to never rely solely on what you see on an oscilloscope to evaluate differences in sound or timbre. Our ears, and mostly our brain, is an extremely sensitive equipment when it comes to variations in harmonic content and waveforms, often barely visible on a scope, if not invisible at all. Just the variable curving of the sawtooth ramp in low frequencies can have a dramatic effect on the resulting timbre, however small that curve is...

[CZ Rider]

I have two Minimoogs side by side. The 104XX, next to the 41XX. The 104XX sounds brighter/buzzier. It also has the newer oscillator card. I am exploring why there are these "not so subtle" differences. I could swap filter boards, but building a clone with a rotary switch is an easier way to A/B the difference. I changed the resistors on one board to match the original spec. and it did overdrive differently. An A/B/C switch would give the best of all comparison for that particular VCA gain change. Could change the drive in the middle of a riff.

Having a 10% pulse wave and 90% pulse wave is radically different when used as a modulation source. I am confident, anyone knowing what to listen for could pass the blindfold test and identify what oscillator board given the choice between the common second version and the later third version. Just by listening to them modulate with the pulse wave. Unknown what the rare first version discrete board has. This was discovered using an oscillicope and posted earlier. Untill looking at the waveforms, I too thought they were close. The oscillators are not the same. And I can't get any more scientific than showing the scope difference where the one is 10% while the other is 90% pulse waves.

Another sonic difference I notice between oscillators is phase cancellation. Oscillators like the Moog 901 have very shallow cancellations between two closely tuned sawtooth oscillators compared to other types. Listening to just sawtooth waves from various oscillators, and they all sound somewhat similar. Getting two of the same waveform closely tuned side by side reveals these cancellations, and not all oscillators handle this the same.