Boss CE-3 repair...out of ideas

[rockhorst]

So I've got this Boss CE-3 in for repair and here's where I'm at:

- Clean signal in bypass and when on
- No effect/chorus signal on the output
- Audio probe shows signal on the input of MN3207 but none on outputs 1 and 2 of the chip
- LFO works (checked with a resistor + LED on the output of the LFO)
- Verified that MN3207 and MN3102 work as expected in another chorus I had lying around
- Replaced Q9 and Q10
- No discontinuities found between output of LFO and MN3207.

I did notice that if I max the bias trimmer there's a very weak, gated signal out of the effect output (which carries only the time shifted signal), but I think that's an artifact from the clean signal path leaking into the effect path.

I'm out of ideas. Any suggestions?

Here's a schematic:

[robthequiet]

How about voltages on IC3 and IC4? Maybe something in the circuit is not providing the right voltage to get the BBD to kick on? I believe there is a ratio, like 14/15ths of Vdd that exists as Vgg from the 3102 to the 3207? Need to double-check my recollection.

[reddesert]

Do you have an oscilloscope / can you check the clock signal between the MN chips?

[rockhorst]

I haven't got a scope at home but I could borrow one from work next tuesday. Hoping to solve it before that though.

Voltage measurements:

MN3102
Pin 1: 8.2
Pin 2: 4.4
Pin 3: 0.6
Pin 4: 4.4
Pin 5: 0.9 (slightly varying)
Pin 6: 8.2
Pin 7: 3 (slightly varying)
Pin 8: 8.2

MN3207
Pin 1: 1.3mV
Pin 2: 4.6
Pin 3: 12 mV
Pin 4: 7.9
Pin 5: 8.4
Pin 6: 4.4
Pin 7: 3.5
Pin 8: 3.5

I do have to add that I fiddled with the bias control and set it back to where I *think* it was when I first opened the unit.

[robthequiet]

Vgg should be 14/15 of Vdd -- pin 4 on the 3207 should be the same as pin 8 on the 3102. I don't know if there is a tolerance for this but it has to do with biasing every other BBD cell in the 3207 from what I'm reading. May be a good idea to double check the readings. Pin 3 on the 3102 also seems to be floating up a bit, where it should be ground.

Datasheets:

http://www.experimentalistsanonymous.com/diy/Datasheets/MN3102.pdf
http://www.experimentalistsanonymous.com/diy/Datasheets/MN3207.pdf

[rockhorst]

There's a silicon diode connected to pin 3 of the 3102 (as per schematic) which would explain a 0.6V drop to ground. So that checks out.

[robthequiet]

There's a silicon diode connected to pin 3 of the 3102 (as per schematic) which would explain a 0.6V drop to ground. So that checks out.

Ah, yes, makes sense.

It would be nice to be able to scope the clock pulses. If the chips work in another box it suggests not a chip problem, obviously. How about the speed and depth pots? If they were noisy it might annoy the 3102. Other than component or PCB trace issues, or maybe noisy power supply, there are not that many things left to check.

[rockhorst]

I'll go over the traces once more. I've tried different power supplies, makes no difference. Could you tell me exactly what to probe with the scope (pins) and what to look for?

[robthequiet]

IC4 MN3102 pin 4 to IC3 MN3207 pin 2
IC4 MN3102 pin 2 to IC3 MN3207 pin 6

If you look at the schematic, they have the pinout names for the chips. "CP_" stands for Clock Pulse[1 or 2]. On either line, you should see a squarish wave.

Checking for continuity on those two lines is one thing you could try now.

The lines run out of phase, while one pumps up, the other pumps down.

The other issue with the clock is the oscillator control circuit which runs into the 3102 on pins 5 and 7. I'm not sure how to interpret the voltages there from the datasheet, probably requires further study. But seeing the clock pulse would tell you whether or not that is working. If the chips work OK in another circuit, it suggests that the problem might be in the IC5, Q10, Q9 chain.

[rockhorst]

Sorry to bring this back after a two week break, but I only just had the time to check the ICs with a scope. I indeed get a square wave on pins 4 and 2 of the MN3102 relative to ground (with continuity to the MN3207). When I measure pin 2 relative to pin 4 I get a square wave of double amplitude which confirms that they're out of phase. Still no chorus effect though

Also, I noticed that I get a very very weak (and gated) signal out of the chorus only output when cranking the level on the practice amp I use for testing. So there is something there, but not a real lot. Sounds like a transistor that's not really turning 'on'. Which one(s) would be the most likely to check out?

[robthequiet]

IC1, Q6, Q8 would be good to get voltages on.

The other line of thought is the setting of VR3 -- there may be a very narrow range where it biases the audio path components.

[reddesert]

To return to your original post, now that we think the BBD is getting proper clock signal: the audio probe showed it gets input, but no signal on the output pins. Perhaps it is a very faint signal and you didn't pick it up with the audio probe, but it is audible through the cranked amp.  I don't know how loud it should be after going through the BBD, but presumably more than near-inaudible.  Audio probing and voltages of the section from the output through Q8, Q4, and IC1 could help.

[rockhorst]

I probed the output of IC3 again and it now makes a sort of motorboating noise (say 'drrrrrrrrrr' out loud and you've got the idea). It's not affected by the rate/depth controls.

IC1 readings seem normal:
Pin 1: 2.1 V
Pin 2: 2.1 V
Pin 3: 2.1 V
Pin 4: 0V
Pin 5: 2.1 V
Pin 6: 2.1 V
Pin 7: 2.1 V
Pin 8: 8V with battery, 7V with 1Spot

The 2.1 V can be varied with the bias trimmer. I noticed that if I turn the trimmer in the opposite direction I get the dry signal out of the chorus output. I replaced the electrolytics surrounding the BBD but to no avail.

Also, this thing reacts differently on different power supplies. Using a Pedaltrain Volta gives a very dim, pulsing LED, a battery gets a hefty LED signal and using a 1Spot seems to be ok (but not as good bright as with a battery). This is checks out with a lot of comments online though.

[robthequiet]

By any chance, do you think you could get the voltages on Q6 and Q8? Also, can VR3 get the bias voltage up to 4.5 V, or approximately 1/2 of Vcc?

The motorboating suggests something out of alignment. Is the signal still good at the input of IC3?

[rockhorst]

Fixed! seems I missed a trace lift at the base of Q6. Thanks for pointing me there At first it didn't seem to fix anything, but then I turned the bias pot and got a chorus effect. Is 4.5V the ideal bias point for this?

[EBK]

If 4.5V was the definitive answer, then there wouldn't be a pot there, right? 

[rockhorst]

Good point

[robthequiet]

If you ever get a chance to put a scope on it, check out the waveforms at the emitter of Q8. If possible, look at the waveform as you adjust VR3 and you'll be able to set the bias point close enough so that you then use your ears to find the best setting. Congrats! Enjoy. 

[Fender3D]

If you have a 'scope then ears are optional...

[DrAlx]

I always set bias by ear. Find the bias levels that give distortion and chose a level midway between the two.