I need some help diagnosing a very bad octave difference between OSC A and OSC B. You might already know this but on both oscillators there is an octave switch with four increments.
Here’s my problem:
If I turn both octave switches to any MATCHING position (position 1 on A and position 1 on B, or position 3 on A and position 3 on B, etc.), both oscillators should play in the same octave. But my synth doesn’t. My OSC A is a full TWO OCTAVES lower than OSC B. I can only get both octaves to match if I turn OSC A to two hole increments higher.
What would cause this oscillator to oscillate at such a slow frequency? And when I say slow, I mean slowwwwww. At the lowest octave position for OSC A it sounds just like an auditable LFO. OSC B has no issues and is working and sounds perfectly. Can anyone point me in the right direction of what could be this issue??
BACKSTORY on the synth which could give a clue:
I bought this synth a few months back extremely cheap. The person who had it before me parted it out and left it for scrap. About 40% of the synth was missing. The power supply was gone, mod & pitch wheel gone, ALL IC’s gone, voltage regulators missing, half of all the potentiometers were missing, AND someone had spilled epoxy all over the OSC A & B section on the circuit board. It was a mess to say the least! I rebuilt the entire synth and everything it works great, except for that weird octave difference.
I’m good a measuring and replacing things but I have no idea what to look for here. Like I said, OSC A works but it’s a full two octaves lower and I have no idea where to start. Any help would be VERY APPRETIATED!
Hello! Yes, both A and B have been tracking with the keyboard correctly.
Since my last post, I was able to bring OSC A’s slow LFO-like oscillations to a normal rate. Since I had to rebuild both OSC A and B sections I removed and replaced the 20 resistors in both sections. I went back and rechecked my work and found a that I had about 4 resistors in total that were the wrong spec. I then swapped them out for the correct resistors and that brought OSC A up - BUT - the two octave difference between A and B is still present. OSC A is now two octaves higher making position 2 on OSC A’s octave switch sounds like a dog whistle and position 3 so high it’s inaudible. Now that it’s two octaves higher I’ve went through about 4 times thinking that maybe, yet again, I had the a few more wrong resistors. But everything I replaced is now correct.
Being that this a 1v per octave synth, OSC A must now be receiving two more volts than OSC B is right? Reading the schematic, OSC A and B’s octave summer is a single LM348, which I’ve replaced many times thinking that this could be the cause. But still the problem exists. Any ideas on what I should check next??
I’d first remove all resistors at the voltage mixers (for both VCOs) with xception of the Offset-potentiometer (R124 and R144), trim then both VCOs to the same frequency (which will be very low), and then add one mixing resistor after the other. Start with R126 and R146 (Mastertune), then R131 and R153 (Frequency potentiometer), then octave switches (R133 and R155), then R127 and R145 (Pitchwheel), then R125 and R147 from the matrix.
Thank you very much for your help! I did as you recommended, still didn’t have any significant change with the high octave OSC A. OSC B on the other hand would gradually increase from sounding like an audible LFO to normal oscillations with adding the resistors back one by one, just like you said.
At this point I was completely lost, so I ended up removing and replacing all resistors in the OSC A section … again. Then I stumbled upon something very weird. With replacing the resistors on the entire OSC A section resistors, in the process I removed R123 (220k) that links to the +15v power rail. Once removed OSC A would suddenly drop in pitch down to a LFO-like oscillation. Even with all remaining resistors installed, if I left out R123, the pitch would never change from this steady slow oscillation. And when I would install this final 220k resistor, I was right back to my 2 octave higher pitch. So, I then decided to install a 680k resistor as an experiment. Once the 680k resistor was in, OSC A was roughly 2 octaves lower! Played around with this a bit and finally landed on a 300k resistor to get me exactly in tune with OSC B!
I’m no technician so I haven’t the slightest idea why replacing with a higher ohm resistor worked but, with fingers crossed, I hope this sounds like an ok fix??
A higher value resistor will bring down the bias to the sum pin on the 3340. I can see a direct path through R123, R132, and R135 from +15v to ground. It may be working, but it’s definitely not to spec. Have you checked the voltages on R120, OSC A offset? It should be adjustable between +15v and -5V, and I would think that adjustment might bring OSC A to the correct range.
I don’t have either of my Pro-One’s anymore, so I can’t check anything locally unless my Behringer is comparable, and I’m looking at two copies of the schematics. There are differences between my two schematics, and between Osc A and B that are confusing. R123 is a 220K resistor, and the OSC B equivalent is R1212, which is either 820K or 1M depending on which copy I look at.
I’m going to have to locate my spec sheets for the 3340 to see what’s going on with this, and I might have to draw up some new schematics to make it more readable.
I think I may have found something. Pin 13 is labelled as “LFM,” and there isn’t much detail on it in the datasheets other than to identify it as a possible place to modulate frequency linearly.
According to an article on cabintechglobal, this pin is one of the four building blocks that affect 1v/Oct tuning. Specifically, they describe pin 13 as a reference current for the initial pitch of the oscillator at 0v.
I would look into R136, R137, and C106. They should match R158, R159, and C114.
Also, take a look at C105; this is the VCO’s timing capacitor. The schematics specify that this should be a 1000pF polystyrene capacitor.
Hello! Thanks again for your help here. I checked the voltages at R120 and got a good +15v and -5v reading. R136, R137 and C106 values are identical to what I replaced OSC B’s corresponding section. R136 and R158 are both 470 ohm, R137 and R159 are both 2.2 meg, and C106 & C114 are both 1000pF polystyrene capacitors. I had also swapped these poly caps for ceramic caps earlier in my troubleshooting thinking maybe something weird was occurring at this point on the board. Still getting the +2 octave difference if I leave the original 220K resistor in at R123.
It is odd that OSC B’s matching position for R123, being R1212, is an 820K resistor while everything else on the schematic seemingly matches.
Okay, you see where I’m thinking. Is something wrong with one of the parts? Sounds like both 3340’s are okay. That at least tells us that the two circuits are different in some other way. If all of the components are okay, we’re left looking at something else on the board that causes a change in the circuit. It could be a short, or some of that epoxy is still causing some trouble.
Note that the value of R123 is all over the place, and Sequential used a 357K 1% on the Prophet 5. Because of this, I’m thinking that your 300k resistor is an ok fix if it tracks well. It’s biasing the node to the lower level.
This is that article from Cabintech that I refrenced. It talks about tuning the AS3340, but it should still be applicable to the CEM3340:
Thank you for sending that article over, that was a very interesting read. What’s odd is that I see the 357k resistor you had mentioned on the P5, and to make matters a bit more confusing the Prophet T8 uses a 294k resistor there as well! Very weird.
I’ve pretty much thrown everything that I can at this point. I’ve swapped IC’s, replaced/swapped offset pots, scale pots and frequency pots. Changed out all capacitors. In the beginning I even removed all of OSC A & B components and removed all the epoxy off the board. With everything removed I then checked for shorts and for possibly hidden resistance with every trace to and from the OSC sections. I think at this point I could probably work on this thing with my eyes closed! Perhaps the issues I’m having now is why so much was missing on this synth to begin with??
Either way, it seems to be working perfectly with my 300k resistor swap for R123. I’ve stress tested everything for at least the last 72 hours and it’s holding up well. The calibration is holding tight and it genuinely sounds and looks great. I’m probably going to throw in the proverbial towel and chalk this up as “good enough?”.
I just want to let you know I’m very appreciative of your help on this. Without your suggestion of removing and replacing the resistors responsible for voltage mixing I don’t think I would have stumbled on what I did with R123. I’d like to, if you’re willing, have a pizza sent to you to show my appreciation. Completely on me. If that sounds ok, just PM me and we’ll iron out the details.
I’m glad you tried all of that with the board. I did suspect a short or residual epoxy might be causing trouble, but I think that might cause a larger issue. I’m wondering if this issue is the reason that he decided to part it out.
The different values used in the different designs has me wondering too. I wish that I still had one of my Pro-One’s so that I could compare. That 300k resistor must be compensating for something else. I hate to modify the circuit from spec, but I can’t argue with the performance.
The thanks go to fanwander. It was his suggestion to remove the resistors in the summing circuitry. That was a very good way to isolate the 3340 from the other circuits.
Since you’re in there, are you tempted to do any modifications to soup it up any? There’s an unused waveform on OSC A just going to waste…
Okay, I am looking through a copy of the technical manual for the Pro-One, and it has a clearer copy of the schematic along with a detailed description of the circuit surrounding the oscillators. I’m hoping to learn more about the design.
The clearer schematic is definitely helpful. I’m able to see that what I thought before was a 47M resistor is actually a 4.7M resistor.
I believe that the main reason for a different value for R1212 is that OSC B is designed to be used as a LFO when S116 is closed. If R1212 was a lesser value, the LFO would not go as low. R139 is open ended in normal use, so it’s value should float higher.
Looking backwards, the OSC A octave switch is the only input to OSC A that is not shared by OSC B provided that the mod wheel is not switched in. It provides a bias to the oscillator from 1v to 3v in 1v increments.
These two areas are the only differences between the two oscillators at the input stage. For one of the oscillators to be off by 2 octaves, we must be looking at an incorrect part that is causing a 2v difference at pin 15.