I also thought that was a typo. Haha
Ok. So I should defiantly see voltage on pin 15? I’m not getting anything there. I threw it 4v from my bench top supply with moderate success. The tuner kicked on, but it was unresponsive to knob inputs. I think with that knowledge and the druid article you linked me, I should be able to trace this sucker. Thanks again!
Pin 15 is control voltage input. It’s not that you should find voltage there, [Edit: Sorry, you should! ] but Jos was suggesting you put voltage there. I think the frequency should be audible with 0V there but in case it’s sub-audible, putting a few volts there should bring it up enough (1 octave per volt) to make it audible.
Again, ignore the tuner! It’s not going to do anything useful until you have a good audio signal.
*Edit edit to delete: * Edit to add: As @fredrik says below, there should be three voltage sources feeding into pin 15, from the rotary switch, the centre note pot, and the fine tune. So when you say
that’s very perplexing. Are you literally measuring on pin 15, or on the pad it’s supposed to connect to? Try both. Could be a bad solder joint right there. Or possibly whatever you’re using as your ground reference is actually not connected to ground. If you see no voltage at the pad and you’re sure of your ground, then the problem is upstream — but with three things feeding into pin 15, there’s not much left for it to be.
Do you see some voltage in all three places – on the center wire of the rotary switch (if the switch isn’t turned all the way counterclockwise) and on the wipers of those two pots?
1 Like
That’s what the coarse tune (center note) does already, though, so you could debug things the other way around – measure the voltage at pin 15 at the potentiometer output, on the potentiometer side of the 100k, and tweak the center frequency trimmer to see if it changes.
If you cannot get a tone out of the chip with the available knobs/trimmers (there’s three of them on board that feeds directly into the CV input, before you even plug anything in), something’s broken. See here for suggested trimming order.
EDIT: Pin 15 is a summing node that’s held at 0 V by the input stage, so the input voltages are on the opposite side of the 100k resistors.
2 Likes
Oh, ok. I threw 4v like I said, and listened with the probe and got a hum that I’m not sure is the wave I’m looking for or just noise on the line. I plan on sitting in a corner with some coffee and the 3340 data sheet tomorrow and see what I can get my brain to process.
So to be clear, you’re saying I probably won’t see any voltage on the cvinputs line, but that doesn’t mean there isn’t signal on it? Damn. I thought I’d pinned it down.
Also, I know I should be ignoring the tuner, I just get excited when it does something.
1 Like
No, it’s a summing node, and I shouldn’t post when I’m low on caffeine 
You need to measure the voltage on the other side of the 100k resistors to see the voltages going into the node.
So I’d suggest:
- With power off, make sure that you have 100k resistance from the centre note trimmer output to pin 15. Check the same for fine tune, the octave selector, and CV in.
- Power on, and measure the voltages on the centre note trimmer and the others, on the trimmer/knob side of the 100k.
- Pin 15 will see the sum of these voltages (but your multimeter will read near zero on the pin itself, since it’s a summing node and the input opamp pulls it down to a virtual ground).
- Sweep the centre note trimmer across the range, check that the voltage at the trimmer changes, and listen for an audible tone.
Alternatively, use Jos suggestion with a separate voltage into CV in, after verifying that you have 100k continuity to pin 15. Pin 15 will see the sum of the ones you measured in step 2 and your additional CV.
5 Likes
Oh right. I shouldn’t post at midnight.
3 Likes
If none of @analogoutput 's or @fredrik 's or my suggestions lead you to a working situation, it may be that you are dealing with a broken chip. As you have already tried 2 others one may think that is not very likely, but it is difficult for anybody to see what else you have been doing that could have caused them to get fried.
However all is not lost.
To test whether the above is the case I’d suggest you build a test circuit as shown in the CEM3340 VCO (Voltage Controlled Oscillator) designs article. But keep it to a bare minimum. Do this on a bread board or solder a minimal version on a strip board and do your tests on that.
This eliminates any errors there may be in your PCB build that you have been unable to find. Of course you should tred carefully here as you may introduce errors when building this. So DO NOT COPY the PCB wiring, but follow the schematic in the article. As you obviously have been learning how to diagnose things now for a few days, I wouldn’t worry about making mistakes. In case of a strip board use a socket for the IC so that you can exchange it easily.
1 Like
One more thing to note is that with the typical resistor/capacitor values, the 3340 is calibrated to behave nicely from ~1 Hz all the way to 20 kHz and beyond, and you have to feed in quite a bit of voltage to get an audible signal.
Assuming nominal resistor/capacitor values, the 10k tracking pot set to the middle position, and that I didn’t completely screw up my algebra, the CV for a given frequency is:
cv = 1.463 * ln(freq) + 0.239
so for e.g. 440 Hz you need to feed in just over 9 V in total (coarse tune + CV).
The coarse tune on its own gives you a range of just under 1 Hz to 3100 Hz:
(note that the above equation isn’t exactly 1 V/octave, that’s what the tracking pot is there for)
3 Likes
Good of you to find out. My suggestion was just a guess to work from. So 4 Volts would result in a really low frequency, possibly not audible but maybe visible on the loudspeaker.
2 Likes
Hi everyone!
An update:
I broke down and bought a new pcb and a posh soldering iron. I reused the voltage regulators, the op amps, the 7 segment, and the 3340 from the last attempt.
After about 2 hours and with great trepidation, I turned on the power supply and…
No Problems. Worked perfectly. Actually, the bottom led on the 7 segment has blown, but I’ll live.
So, I’m going to use the working one to troubleshoot the last one. Hopefully, I’ll find the bit that’s f**ked. If I can manage to track it down, I’ll let you all know.
Thanks again for all the input. Now I’m off to the filter.
6 Likes
be careful if a led burned out, there may be a little problem somewhere.
have you tested it with a multimeter ?
1 Like
I have. I just didn’t want to wait for a new one. Hopefully I’ll never need to know the difference between an A and a B, or an E and a F.
I’ll just stick to playing in the key of 9
5 Likes
excellent, well done for keeping at it,
I would highly recomend looking at the schematic on the website as it’s broken down into blocks. Just be aware that the LABLES are all shown as input flow which can be confusing.
Good that the 3340 is ok too , not expensive in the real world but as synth components go its top of the £££
Rob
How sure are you of this voltage/frequency relationship? I ask because 440 Hz at 9 V seems awfully low to me. I don’t see anything about frequency range in either the CEM3340 or AS3340 datasheets (added: oh, there it is, never mind) and it depends on the capacitor at pin 11 (and resistance at pin 14?), but the Electrosmith ES 3340 datasheet gives a frequency range of 0.05 to 80000 Hz for control voltages from -8 to 12 V, and that implies 10 kHz for a 9 V CV.
Calculations based on the CEM3340 datasheet with Sam’s simple VCO component values, plus a sanity check on an AS3340 build using those values. Haven’t measured on the tuner VCO, but I think that’s using the same component values for the important parts.
1 Like
Hmm. Interesting. Wonder why it’s as low as it is, if your control voltage range is 0 to 10 V that’s about 0.86 Hz to 880 Hz. The Electrosmith range would be 20 to 20 kHz which seems more reasonable albeit high … a piccolo for comparison can go to C8 which is 4186 Hz, more than 4 octaves higher than that 880 Hz limit. If I ever build one of the LMNC VCOs I’ll want to think about that.
Added: OK, found what I was overlooking in the datasheet:
and the 1222 does use 1 nF (1000 pF). But it only gets to 880 Hz at 10V CV compared to the 10 kHz the datasheet talks about. I need to think about this more.
1 Like
The CEM3340 sample circuit has a 360k pullup to +15 V, which adds the equivalent of 15×100k/360k = 4.17 V to the CV, or around 15 Hz. Sam’s circuits have tuning knobs instead.
EDIT: Checking my notes, the frequency depends on Vcc as well, so wouldn’t surprise me if it’s closer to 20 Hz for a 15 V system.
4 Likes
Sharing some things I found when troubleshooting a build, just in case it’s of value for others…
The advice above on the CV at pin 15 is good - an alternative I found was removing the 3340 chip and measuring the voltage on the IC socket. This should change as you alter the tuning post/trimmers or add a CV input. Setting it at around 3.5V (without a CV input) produced a good starting point for debugging once you put the chip back in as this should provide a sensible output.
The other thing I found is that an incorrect value of R5 produces strange output on the 3340 - too low a value and the CV on pin 15 goes positive (surprising), and all three outputs go to a steady negative voltage (this measured using a scope, so the probes may have had an impact).
2 Likes
Even simpler and cheaper would be to use a crystal earpiece to listen to the circuit.