VCO - weird readings...

[MFOSser]

I've built a VCO and wall wart power supply.

At the negative pins on the chips I'm getting -12.04V but the positive pins read -2.73V. That's a minus value for the positive pins.

The power supply is producing +11.73V and -12.04V. The wall wart itself is a 12V AC 300mA supply.

Many of the resistors no longer have their original values. The 1M ones now only read 100K or so.

The LM13700 got extremely hot and is now completely cold.

The track from 12V to R65 was virtually non-existent so I've repaired that with some wire. Maybe the lack of a decent track altered the resistance of the circuit.

I have blown something or other but have no idea where to start. as I probably have more than one problem.

[andrewF]

At the negative pins on the chips I'm getting -12.04V but the positive pins read -2.73V. That's a minus value for the positive pins.

if the chips are in sockets, pull them out, turn the power back on and check the voltages on the power pins.

possibly a short - little solder bridge or the +12V is not getting to places it is meant to.


The power supply is producing +11.73V and -12.04V. The wall wart itself is a 12V AC 300mA supply.

should be fine

Many of the resistors no longer have their original values. The 1M ones now only read 100K or so.

thats (usually) normal, once resistors are in the circuit some of them will be in parallel so you will see lower readings.

The LM13700 got extremely hot and is now completely cold.

hope you have some spares, don't put any in until you are sure the power problems are sorted

The track from 12V to R65 was virtually non-existent so I've repaired that with some wire. Maybe the lack of a decent track altered the resistance of the circuit.

It doesn't need to be very big, there is not much current going thru. Is this PCB purchased from Ray or a home etch?

I have blown something or other but have no idea where to start. as I probably have more than one problem.

yep, there are few people who have built VCOs to have them work perfectly from 1st start-up.
Check with your multimeter (on continuity 'beep' mode) to make sure there is no short between +V and ground (also between -V & Gnd and +V &-V)
have a good look for solder bridges
(as i wrote b4) with chips out, turn on power, check you are getting +V and -V to the appropriate power pins.
If you are getting +V to some pins and not others, trace the +V track and try to find the point where it goes wrong, maybe a bad solder joint or broken trace

if none of this helps, take some close-up photos of both sides of the board and post them.

have fun

[MFOSser]

Thanks for the check list.

The -2.3V on the +ve pins was without the chips in. I only get +11.73 on the 12V to R65 track. After R64 and nowhere else on the board (self-etched) do I find anything above 11V. I can find -12.04C at many places but no +11.73V anywhere but at the input.

I have done a continuity check between +ve, -ve and ground with no bleeps.

For today, I will take a rest and do something else.

I will start again tomorrow.

[andrewF]

did you etch both top and bottom traces?
If so, some components (pin 8 of U6 for example) need to be soldered on both sides of the board as the +V trace jumps sides a few times.

[MFOSser]

Reading back what we had written made me realise that the +ve problem was close to the +12V input. I had forgotten to solder through the via near the 12V input.

I get +12V on the chips now.

However, no wave detected on the sinewave wire and a flat line on pin 1 of the OP275 with my scope.

Is there anywhere else I can test for a waveform before I start buying a new chipset?

[jordroid]

If the core is oscillating you will see a 0 volt to +2.2 volts (or thereabouts) ramp wave at the output or U2-A. The output of U1-A will be the product of the summed control voltage, so will normally be flat (DC), the flat line should move up and down when you twist the coarse and fine tuners.

If you don't see the output of U1-A moving up and down with the tuning pots, and you don't see a ramp at the output of U2-A then the core is not oscillating, or oscillating much too slowly (this happened to me with my first VCOs). If that is the case there are not many things to go wrong and we should be able to get you through it pretty easily.

If you do have a ramp at the output of U2-A, the next thing i would look for is a triangle (or triangle-ish, depending on where the trimmers are) at the output of U4-C. Then a pulse at the output of U7, and a sine(ish) at pin 8 of U5-A.

I have been assuming that you are using the latest schematic for Ray's VCO, if that is wrong then a bunch of that may be incorrect.

Let us know how you are getting on

[MFOSser]

Thanks for the help.

I had another go but flat lines from all the pins you mentioned.

Pin 1 on OP275 is at -0.01V

No deflection on the scope when I turn the pots.

[jordroid]

I would start with the oscillator core and then move on from there. Page 1 of the schematic is the entire ramp core oscillator, just two dual opamps, two transistors and a fet.

If you have the same chips in there from when the power was wonky there is a possibility that they are buzzed, if you have two cheap dual opamps and two general purpose NPN transistors you might want to pop them in until the danger of smoking them has passed. U1 should be a good opamp for proper tracking, but it will still oscillate with a TL072 or whatever for testing, likewise if you have a socket for U3 you can stick two NPN transistors in there for the time being, it would be a shame to toast OP275s and SSM2210s.

There isn't that much going on on page 1 of the schematic, and if you have one it might be a good idea to breadboard just that part of the circuit until you have it working, and then figure out what is different between the working breadboard and the non-working pcb.

If not then just quadruple check everything on page 1, you might want to print out a copy of the schematic and ring out every connection with your meter set to beep mode, highlighting the schematic with a marker as you confirm continuity.

I also seem to remember some recommended mods to the high frequency compensation circuit (in an old EM thread) that cause the core to stall out at certain settings if the HF trimmer was dialed for too much compensation (mine did exhibit that behavior). If that was suspected you could lift one end of D1 or R9 to remove the high frequency compensation from the circuit, but that is not the first thing i would suspect.

I don't know exactly where you are at or what you have done so i'll just toss out some of the first things i would check for-

D8 pointing the right way? All points that should be grounded or have a voltage reference correct? Do you see an adjustable voltage at the wiper of R2 and R3? Absolutely sure that your scope is adjusted and working properly? Polystyrene caps are pretty easy to cook(if that is what you used), could C6 be fried? Q1 oriented correctly? U3 (or two NPNs) oriented correctly? Is R16 really 100 OHMs?

[MFOSser]

Just to check...

This monster is C6. I hope it is a .0047uF silver mica cap.

---

cap.JPG
Description:	C6 silver mica cap
Filesize:	643.03 KB
Viewed:	286 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[MFOSser]

I've been over the oscillator part of the circuit with the voltmeter and continuity check.

All places on the diagram which are supposed to be +12V, -12V or GND are exactly that.

All outputs from the 3 chips deflect when I twiddle the pots.

The only anomaly I can see is that trimmer R16 shows continuity on any combination of the three pins when I test it. Doesn't that mean that there is an internal short in the trimmer as there is no short on the tracks or between the solder blobs.

I ask afore I desolder it.

But would that make a difference? Wouldn't it just mean that the circuit at that point has zero resistance or does the circuit at that point need some sort of resistance to get it oscillating?

[jordroid]

C6 looks to be a mica cap of the correct value. I don't think those big mica caps are sensitive to heat like the polystyrene ones that often get used are, you are probably fine there.

R16 is only 100 ohms, which your meter may read as a short in beep mode, could be normal. You could try measuring the resistance between pins, i don't *think* the rest of the circuit will throw off this reading too much, but i could be wrong... If you check it and the resistance is something around 100 ohms between the outer lugs, and the readings from 2 to 1 and 2 to 3 add up to about 100 ohms it's probably fine.

The function of R16 is to scale the summed and attenuated CV from U1-A to precisely the amount needed for 1v/octave response. I *think* if everything else were in order it would still oscillate if R16 was shorted.

Just to be sure, are the same chips in there that were in it when the 13700 got hot?

With your scope probe clipped to the output of U2-A (pin 1) and the coarse and fine tuners roughly in the middle of their range, sit and watch it for a while (30 seconds or so should be plenty), does anything happen on the screen? Like a slow ramping up, or a series of ticks every few seconds?

If all designated points are getting the correct voltages, likely candidates are missed or cold solder/solder bridge, incorrect value resistor, fried component (opamps, transistor pair, fet), backwards diode... from the artwork on Ray's site it looks like there are some tricky pads beneath sockets on the top side, are you sure you got all those?

It may be time for clear pictures of the board like andrewF suggested, it might help.

You will get it

[MFOSser]

VOTRIM and HFTRIM are okay and I now have them at the centre of their ranges.

Yes, I have the same chips in their sockets. All chips (and diodes) are as per circuit diagrams.

Now, my oscilloscope. It's probably not suited to the task. Assuming that it works at all. It's a 1967 valve operated Heathkit. I have connected it to an audio source and it looks fine. I am trying to get xoscope to work on my PC but with no luck.

However, I am beginning to get a little ripple from U2-A (pin 1).

I'd like to connect this to an amplifier and maybe just listening to something would give me an indication that the board is working.

Just to check. I don't have everything wired up. Just the pots, CV1, Sine and a secondary ground for readings. At present, I connect an AA battery to CV1 and ground to get my little ripple. Could I get a sound out of it if I connect Sine and ground to an amplifier?

[jordroid]

Sounds like you are making some progress, and that sounds like a cool old scope, as long as it works.

The ripple at the output may be proper oscillation, hard to tell from here with the available information. If it looks like a ramp and gets more frequent as you increase the voltage at the inverting input of U1-A then it's what you want to see. If it starts oscillating when hooked up to the battery, but not when you turn up the coarse tuner then you know something is miswired or wrong about the tuning pots, but the core is working.

Hooking it up to an amp and listening to it directly is not a bad idea, it might give you some good information. You can also rig up a quick audio probe to listen to various parts of the circuit. Ray shows this in the synth DIY section under "hot tips" and then "guitar amp signal tracer"

Once again, if it was powered on incorrectly for long enough to cook the 13700, it is fairly likely that the other chips were damaged as well. Don't tear your hair out with troubleshooting until you have removed fried chips from the list of likely suspects.

[MFOSser]

I've managed to put Visual Analyser on an old Windows laptop and am using that instead of my oscilloscope.

Here is a screen grab from analysing U2-A (pin 1).

There is a ramp, of sorts, but it is neither constant nor steady. It breaks up and changes in character all the time. Through the laptop speakers there is a rapid pulse.

Fiddling with the pots can make the wave increase, decrease or breakup.

What sort of voltage range do I expect to get? I barely get a volt out of pin 1? And that is with a CV1 input of 1 rechargeable AA battery.

I can't find any faults in the resistors. All have been checked. Capacitors, diodes, tracks. Everything but the chips.

[ignore the bitmap download, it is just a copy of the jpg]

---

grab.jpg
Description:
Filesize:	168.31 KB
Viewed:	263 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

---

untitled.bmp
Description:		Download (listen)
Filename:	untitled.bmp
Filesize:	2.25 MB
Downloaded:	284 Time(s)

---

[jordroid]

Waves on a screen and sounds coming out are a good sign, it is sort of working, just not great. There can't be much in the way of it working properly at this point.

U1-A is an inverting summer with a gain of -0.02 for anything tied to one of the 100k input resistors, so if you have 1 volt at the input you should see -20mV at the output, so it is normal to not see a large voltage swing there.

Trying new chips is probably going to be the most informative thing to do at this point. Subbing two NPN transistors for the SSM2210 (if that is what you used) would not be a bad idea either, i had an issue with a VCO running much too slowly (which it seems like yours is doing as well based on the "rapid pulse" description of the sound) that i tracked down to weirdness in the expo converter (i used NPNs with a weird pinout and got it wrong, but the symptoms were similar).

cheers, it seems like you are pretty close now

[MFOSser]

Thanks for the help. I have another set of chips in the post.

I intended to build a second VCO so a second set was always going to be needed.

[MFOSser]

Well, this is where I am up to.

I did a little chip swapping after checking the board to make sure there were no errors.

It looks as there is nothing wrong with the oscillator on page 1, I think. The raw.jpg picture below is the raw output.

I can get something out of Triangle and Ramp but the triangle is nothing like any triangle I ever saw. You can see it in the tri.jpg picture below.

I can get no signal at all out of Sine or Square.

Hopefully, this will help someone pin the problem down.

I twiddle away at all the trimmers but I don't see any change in the waveforms.

The trimmers are from Murata - http://www.rapidonline.com/sku.aspx?tier1=Electronic+Components&tier2=Resistors+%26+Potentiometer&tier3=Preset+Potentiometers&tier4=%26frac14%3bin+Multiturn+cermet+potentiometers+-+Murata&moduleno65271&catref=68-1100

It is plain to see on the pictures that the basic oscillator is quite regular, although it does throw a mental quite often.

The "page 2" circuit doesn't seem to do anything other than turn raw into pink noise.

Anyone want to buy an expensive pink noise generator?

The only thing out of the ordinary that I can think of is the power supply with the 11.75V +ve and the -12.03V -ve. I might fiddle with the transistors on that board to get the +ve over 12V.

Other than that... I'll buy an analogue off eBay.

---

raw.JPG
Description:
Filesize:	149.01 KB
Viewed:	258 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

---

tri.JPG
Description:
Filesize:	143.48 KB
Viewed:	245 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[mark_olson]

I skimmed through this thread and, maybe I missed it, but, how does your VCO actually sound?

[MFOSser]

The raw output from pin 1 U2-A makes a sort of "duh duh duh duh" sound.

The ramp and triangle outputs make a similar sound. Nothing from sine and square.

I have ground and whichever output connected to my PC's sound card for input to the PC's oscilloscope software.

There is not enough signal to register on my old valve oscilloscope. I have to use the zoom option and the input level control to get the screen grabs below. The sound is barely audible when I feed it into an amplifier.

It is mentioned on MFOS VCO page to have your probe on 10X but there is no signal from anywhere if I do that so I have it on 1X.

[mark_olson]

Oh.
I guess I was hoping it was more of a wave shaping issue but evidently not.
I am not good at troubleshooting - too easily frustrated.
Mark

[jordroid]

Rat's. I don't blame you if you want a break from it, when you come back to it i would be happy to help again if i can. It's hard to do this across the internet though.

I think my best suggestion was to breadboard just page 1 of the schematic and compare readings once you have it working. It's a proven circuit and yours is sort of working, so it is probably something minor, just hard to find.

cheers,

jordan