So I’ve run into this issue with the CEM3340 that I’ve never seen before. I’ve built this them half a dozen times on breadboards for different playing around projects. I went to put it together again since I got some more rev g chips in and wanted to build a few to play together with sync and such.
However, every time I build it now (I’ve repeated the full build several times, based on the LMNC schematic because it’s one of the simplest with fewest components to keep it simple while figuring this issue out) I end up with it stuck at a very high note, somewhere I’d say around 7-9kHz. The tune knob and scale adjust work, but it stays in that very high pitch range. Tuning up it quickly disappears into inaudibly high frequencies.
I’ve never run into this before, and nothing I do seems to be able to bring it down to the range I’m expecting. I am not doing anything weird or subbing any components. All of my values match (and have been tested, including the caps). I’ve checked over all voltages and they are nominal (around -7v-ish on VEE, +12 on VCC, etc), all the passives check out. The IC is decoupled properly.
Oddly enough, if I am checking resistance across the scale 1/2 circuit the oscillator seems to settle on a much lower range which leads me to believe there may be some current screwery going on where the Multimeter is introducing a new variable, but I can’t reproduce it without.
My power supply is perfectly fine (and relatively clean to boot) and other modules have no issues.
I can reproduce this issue across different 3340 chips.
It’s driving me nuts, I can’t figure out what could possibly be causing this - especially since everything seems nominal/expected. There may be some weird thing I’m doing or some known cause of this, so I’m hoping someone here may have dealt with and solved this before to fill in any blanks.
Thanks, all.
3 Likes
You need quite a bit of CV to hit the audible range, so if you land on 7-9 kHz odds are there’s some problem with the oscillator section. I’d start by checking the oscillator cap (pin 11 to ground), then the pins next to it (pin 12 goes directly to ground, pin 13 to +12V via 1.5M, and pin 14 to ground via 1.8k).
If that all checks out, turn it on and measure the CV components, on the other side of the summing resistors (or across them). See here:
5 Likes
Timing cap is nominal.
13 is nominal as well. 12 and 14 (yes via 1.8k) have a dedicated ground each (same ground plane, just their own sinks to try and rule out ground issues). These are all what I’ve been checking frequently and redoing from scratch to make sure they are as expected.
I’m seeing about 6v coming out of the tune pot, but only about 1mV on the other side of the summing resistors (from the pull down resistor/cap, no?)
And oddly, as I mentioned, if I bridge 1 and 3 with a MM checking resistance (of the trim pot) it dramatically drops the pitch to almost what I would expect. Just an aside.
Pin 15 is a virtual ground, and will be held at 0V by the input opamp. You need to measure all CVs on the other side of the 100k summing resistors (or you can measure the voltages across them).
Pin 1 is part of the tempco circuitry. You should have ~30k from pin 1 to 3, and 5.6k between 2 and 3 (with the circuit turned off).
2 Likes
It’s generally not useful to try to measure resistances in a powered circuit. Any voltages present may affect the measurement, and the voltage generated by the MM may affect the circuit behavior.
3 Likes
fredrik:
I only have the tune knob CV, no other CVs are in use (broken connection on CV in jack). Still ~6-12v coming out of the pot from bottom to top (no ground lead) when tested against the other side of the summing resistor.
Resistance across 1 and 3 is ~30k, adjustable obviously from scale trim pot.
2 and 3 are also nominal as expected.
yes, resistance wasn’t being tested when powered on, it was a fluke that I happened to power it on with a lead still attached and noticed the newly introduced behavior.
3 Likes
I identified my problem, and it ended up being really silly.
I had my CEM3340’s in IC headers to protect their pins (these 3340s seem to bend way easier than my other ICs) and even tho the header was sitting flush it must not have been making a good connection within the breadboard.
Pulling one of the 3340s out and placing it directly in the breadboard appears to have solved this issue.
Thanks for the input everyone who commented.
7 Likes
Ive been there. Gosh when its something like this i have mixed feelings. Its like: “yay, nothing horrible”, but also: “dangit, that shouldnt have happend”. End of the day, at least it works right? 
3 Likes