Analog Synth LFO troubleshooting

Hello all,

I have been working on my Sound Labs Mini Synth from Music From Outer Space and have come across some problems.  I am on the LFO and something is going a little weird and I can't figure out what.

Here is the schematic:

http://musicfromouterspace.com/analogsynth/SOUNDLABMINISYNTH/images/synthkit_lfo_noise_power_001.gif

I have everything wired up except for C13 and S17, which I don't NEED in order for this to work (at least I don't think.  i am short an spst switch so can't add that yet).  What is happening is it isn't oscillating.  It is outputting (at pin 7 of IC7-B) a DC voltage of either + or - 10.7V depending on the position of the Ramp/Saw switch.  At pin 1 of IC7-A, the triangle output, it is giving me about 6.4V.  I'm getting a small (2-300mV) signal at the input of IC7-A, it looks like noise on my scope, but there IS a pattern to it.  It is a really noisy waveform that isn't any of the shapes its supposed to be, nor is it in the LFO frequency range, its much higher.  The frequency pot doesn't seem to change a thing. 

What is stumping me is that there is no true input to this module.  It IS the source.  I can follow a signal through opamps when they have an input and output, but in this config, I'm having trouble understanding what is supposed to happen.  These are supposed to oscillate, and they aren't.

I posted over at electro-music.com and I just don't get the response I get here, so I'm throwing it out to you guys, to break up your recording electronics monotony! haha.  Shift into synth mode!

What op-amp are you using?

Have you bypassed the PSU pins at the op-amp to 0V with 10uF electros and 100n ceramics? Sounds like HF instability- the HF oscillation is "swamping" the LF signal.

Have you used a screened cable from the PCB to the pots and switches? Check all resistor values too!

Input is it's own output, sent to integrator!

Mark

The opamps are LF444's. 

The power supplies to the opamps are bypassed near the IC.

The design of this synth is pretty much made to be easy to operate, so the designer's build along with many others don't include shielded cable or any other type of further protection like that.  They just seem to work.  The result of his design isn't exactly pretty, but it seems to do what it says it does. 

This integrator you speak of, he mentioned that on the page that the schematic was posted.  Looks like my next area of study.  From his description, he says that the output of one opamp charges a cap and when it gets to a certain value, it flips the other opamp, and vice versa, triangle wave oscillation (switched to saw or ramp via the 2 diodes).  Mine only seems to stick in one or the other state.  It somehow isn't being told to flip status.  So maybe i get the first half of the first cycle, and then it sticks there, and the square wave just sits at either its positive or negative value.  I will do some resistor value checking and integrator research, and see where that gets me.  COULD this in fact require that other cap and switch for the lo range that I haven't added yet?  It just doesn't seem like that would make a difference.

Thanks for the help, I feel I'm getting closer, but just need more knowledge of oscillators.

Did it ever work? Is this a new build or a repair? Have you added solder-bridges or missing wires?

Is the opamp socketed? TL072 ought to work for test.

Voltage at pin 2 ought to be "zero". The real input is the voltage at the R90 wiper. Is this some adjustable fraction of the pin 7 voltage?

No it never has worked.  The main board has been sitting in my closet for a few years, and now I'm starting to wire it up.  I checked all wiring against the schematic and every offboard connection is correct.  All resistor values are correct, no solder bridges or anything like that.  Like I said, the only thing missing is C13 and S17 because I don't have either right now.  That just means that it should be operating in its higher range, with no choice to switch to the lower range. 

The op amp is socketed, but the trouble is that I don't have any replacement chips on hand.  I have to go buy some, which won't happen for a while.  Just FYI, it is a quad opamp where the other 2 are being used for the noise source, which is shown on that schematic page as well.  The voltage at pin 2 is in the hundreds of millivolts, and when i scope it out, it looks like an odd waveform (it does cycle though, not just random) with a bunch of noise clouding it up.  Frequency is way higher than an LFO should be, I can get specific measures later.

After some research into integrators, I now understand how this is supposed to work.  The trouble is that it isn't receiving any sort of square wave at the integrators input to flip the direction of voltage change.  It seems like it begins the cycle, and once it hits saturation, it sticks there.  How would this circuit make the output of the 2nd opamp a square wave?  Basically how would it change its output to the opposite voltage to flip the integrator?  The output of the integrator doesn't change.  Could it be a bad cap for C14?  Its not giving the integrator anything to charge, therefore there isn't any change in voltage, just DC being amplified by the 2nd opamp into greater DC?  I guess I will see if I have a decent replacement for the cap and try that.

Thanks for the further help though.  I will continue my search.

Maybe this page's description will help. I'm still trying to wrap my head around it for now.

http://musicfromouterspace.com/analogsynth/NewAugustRampLFO.html

The real input is the voltage at the R90 wiper. Is this some adjustable fraction of the pin 7 voltage?

OK, did some further testing.  This is really stumping me.  I tested the DC voltage at the output of the first opamp, and it was going from 0 in the neg direction, just very slowly.  Then as I was testing some other voltages, I came back and tested it again, and it was stuck at -11V.  So I went to test the input to that opamp, and I was getting some very odd readings on BOTH pots.  The amount pot, a simple level control, would read -12v on the one side, and when i turned the pot so the wiper was closest to that point, the wiper read POSITIVE 10V.  How is it that when those 2 points are essentially shorted together, I'm getting such drastic readings?  I couldn't even register the value of the pots on my meter.  I tested the other oscillators pots (same value, 100k), they registered just fine, i turn the pot, the value changes.  The LFO's pots however would just read Over Limit and wouldn't change when I turned them.  So needless to say, I usually got 0v on the input to the integrator.  And this time when I scoped the input to the integrator, I got a different waveform.  It was a series of pulses, but instead of a square "release" it was curve down, like a reverb tail shape.  And the cycle was consistent.  It didn't seem random.  I couldn't get my scope to track it that well, but when switched to trigger from "line" they stayed relatively still.  It was clean this time too, no noise to cloud it up.  Just for further info, I'm using +/-12V instead of +/-9v, since he originally planned to run off of 2 9v batteries, but I have a power supply for it.  Right now, I'm using a powered breadboard PSU.  His PCB design is a little confusing, he doesn't include a good amount of components onboard, so you have to wire them up OFFboard, which is a pain in the ass, so I chose to use a little protoboard bought from a local electronic supply store to run wire from the main board to the protoboard, through whatever components and out to the controls.  I first thought it was a wiring mistake due to all this confusing routing, but after looking at it more, my wiring is correct.  I ran the 2 leads out to where S17 will eventually go, and when I shorted those 2, nothing changed.  That would essentially short the charging cap of the integrator, not exactly sure what is SUPPOSED to happen then, but like i said, nothing changed.  This is starting to annoy me... ha.  I just want a working synth, that is all.

Does anyone have any idea as to what the devil is going on?  I unfortunately don't have any spare parts lying around to go swapping things out, so I guess that could always be an issue....

phishman13 said:

I couldn't even register the value of the pots on my meter.  I tested the other oscillators pots (same value, 100k), they registered just fine, i turn the pot, the value changes.  The LFO's pots however would just read Over Limit and wouldn't change when I turned them. 

Click to expand...

Hmm,

How did you measure these- across the main resistor (usually the outer two tags) and then from one side to the wiper? Make sure that these pots are working okay- this may be part of the problem! The Rate pot allows the feedback from the schmitt trigger (the 2nd positive-feedback op-amp) back into the integrator ready for the next cycle. (the schmitt trigger makes a clean square signal from varying ramp from the integrator, and allows a little hysteresis)

If the wiper is o/c, then there will be no feedback (except via the ramp/sawtooth diodes) and these will only conduct when the feedback volatge reaches the diode trigger points (i.e. forward or reverse bias ~0.7V depending on setting of switch...) and the integrator will just "drift" as it picks up charge at its (relatively high-Z) input terminal.

Mark

First of all, thanks to those who have continued to explain this circuit further.  I am really starting to understand what might be wrong.

It really does sound like something is wrong with the pot.  My only thing is that I have used 2 different pots in the amount and freq spot.  same value, just linear for freq and log for amount.  I have the exact same pots used in the VCO with no problems, and the only ones like these that are reading odd are in the LFO.  So it just doesn't seem like 2 different pots in the same module would go bad the same way.  Basically meaning that there might be a deeper problem.  Now let me say this, I remember testing the opamps supply voltage before putting the chips in, which was a couple years ago at this point (MUCH less experience and had basically NO idea what I was doing).  I did NOT remove them while wiring up this circuit recently, and was doing some soldering very near the opamp.  I was quick and accurate, but maybe this is an issue?  It sucks I don't have a replacement chip right now.  I guess I could swap the VCO chips into its place since they are the same.  I guess I will try that and see if the results follow the chip.  I have just never seen a situation where a pot, even connected in circuit, has read such a weird value.  Could a bad chip somehow damage a pot? 

Well,

Most pots have very low power ratings- due to their construction they are unable to withstand a lot of current flowing through the track, or through the wiper and track.

This may not seem to be an issue with a 100k pot, but imagine what happens if, for example, in this circuit the wiper of the freq. pot became very positive- e.g. became connected directly to the +Ve or -Ve 12V supply, without any other series resistance. This could happen if your 1M resistor was of the wrong value, or from a short between the +Ve rail and another component leg. The lower ("earthy") end of the pot connects to 0V. As you turn the pot anti-clockwise, there will come a point where the resistance between the wiper and ground becomes such that the current flowing increases (as the resistance decreases) to a point where it permanently damages the pot track (usually a small whisp of smoke and/or a hiss sound as it vapourises the carbon!) I've actually set a pot on fire like this when I was a kid...  

This damage will show up as an o/c (open circuit) or infinity-ohms on a meter, or else the pot value/law will be damaged.

Check your pots with your ohmmeter, and check between the track and wiper too- from both ends. It may read fine from wiper to one side of the track, as often the "break" in the track is at the very end of the track where the carbon meets the metal contact.

Same with the output level pot. If you shorted the wiper to ground accidentally, and adjusted the pot to the far end of travel (either CW or CCW) it would hit the -Ve rail, or maybe the op-amp output hitting a rail, and cause a larger-than-average current to flow, and damage the pot track.

I'd definitely trust my scope more than my DMM checking voltages in this circuit. If there is any HF instabilty, which can arise from a damaged or mis-connected op-amp, then your DMM would just resolve the voltage into a bizarrely fluctuating DC voltage.

Turn the time base right up to see if there is any weirdness going on, or wave an am radio about over the circuit! A fuzzy scope trace at LF also indicates HF oscillation- as you open out the timebase (to scan faster) you can see the waveform "break out" from the fuzzy trace.

Mark

Just to rule out the pots,

Have you got any 47k resistors lying around?

If you twist two together, and then use the twisted-together ends as the wiper, and the two free ends as the "input" and "earthy" side of the pots, you can get the circuit working with a pots-set-to-centre-position setup (the fixed resistors simulating a 100k Lin pot set at 12 o'clock)

You could use other values too- 68k/33k etc- obviously each value set simulates a different pot setting!

May help with troubleshooting though...

Mark

Great idea! I have a bunch of spare resistors laying around.  I'm gonna try this this week and report back.

Try removing R94 and let pin3 of R90 float.
If no change, short pin 1 of R91 to pin 1 of R90.
If it works then, R90 is open.
I built a similar circuit for my car last year and just used
a  10K resistor in series with a 100K pot wired as a rheostat
(one end and slider only) to make a 10K to 110Kvariable
resistor. I did not use the equivalent of R94 and its connection
to ground.
HTH