PLEASE HELP! My preassambled XoX is broken!

[mario1089]

Just got my xoxbox from an australien company preassambled. But it is broken: Only kinda pulse peak on every note, sounds very much like a "zero"-decay-no-sustain-sounds. I can hear some very subtle remainings of the oscillator in the very back of the noise it is producing, also accent (and resonance a slight bit) changes that pulse-peak sound. All other filter related knobs have no effect at all. All digital functions (LED, Sequencer etc) are 100% ok, "only" sound module f** up.

This morning, when I got it out of the box it did not work, but after 30 minutes it suddenly started to work ok, and I tested my new firmware with new smart randomizer/pattern creator unit and tweaked it a little and everything was just fine, but then later it stopped doing real notes again, and then never got back to work. tried for hours now. so sad!

I checked with a different power supply (500mA - yes, AC-AC), but no difference.

I`m 100% sure its just one part of the analoge circuit that`s not really healthy, but, damn.. which? where to look? I have very old oscilloscope (maybe still working, not sure) and normal meters here.

Please help me, I was so enthusiastic with that firmware modifications I`m doing and now this damn ..

Thx,
Mario

[MMM]

I'd suggest tracing the signal flow from the oscillator through the filter.

While I don't want to question your judgement on the "very subtle remainings of the oscillator" I would like to make sure that it is really there. When the oscillator fails or produces a very low frequency the symptoms are not unlike the ones you describe.

So first I would check the output of the oscillator. You can do this with the oscilloscope or with a cheap pair of headphones; the procedure is described in the building manual. If you do the headphone route you can connect ground of the headphone jack to the metal tab of the 7805 in the IOboard (or any other ground plane) and then touch the tip of the headphone jack to the middle pin of the waveform switch.

[mario1089]

Thanx!

I checked it! oscillator definitevely works well. what now?

[MMM]

Now we're going to check the envelopes.

Please measure (voltage against ground) D34, D35, D36 and D37 when playing a note. Do the voltages change and if yes, how?

[mario1089]

did not know which side of the diodes, so here is both. Tried to get it right, but not exactly sure..

------ no notes ------------ | ---- notes playing ---------
marked side -- other side -- marked side -- other side
d34: 0V || -5V || 0 - 5V || -3 - +5.5V
d35: 0V || -5V || -1 - +5.5VV || -4 - +5.5V
d36: -4.4V || -5V || -4.4V || -5V
d37: -2.3V || -5V || 0 - 3V || -5V

[mario1089]

Also I tested the filter output at R90 like said in the building guide.
I get a clean dirac pulse there, one per note. That "remainings of the oscillator" are missing at R90, also much less noise than at Mix-Out. If I think about this (the mix out sound compared to the filter out), it shows that the VCA is clearly open the way it should, and it gets a reasonable env voltage. I`m not sure about the env signals at the diodes, but my feeling is they are ok, at least doing something looking ok. So I think I have a filter problem! Where can I check the filter input signal and how can I check parts of the filter?

[MMM]

Also I tested the filter output at R90 like said in the building guide.
I get a clean dirac pulse there, one per note. That "remainings of the oscillator" are missing at R90, also much less noise than at Mix-Out. Where can I check the filter input and other parts of the filter?

R90 has nothing to do with the filter; this is the basically output of the DAC for the oscillator CV. The builder's guide suggests attaching 3v there (to simulate a proper CV and hence put the oscillator in audible range); this isn't needed for your since I believe the oscillator is "playable" already?
Test point for the filter is VR4.

As far as the envelopes go: I have some suspicions here (especially the values for D36 look wrong); however since the envelopes are a changing voltage (and not static) they are a bit diffcult to debug without an oscilloscope. I'll try to think of something, hang on.

[mario1089]

[url][/url]Yes, osc is playable.
I ll check VR4. And have another look at D36.

D36 is def. constant -4.9 / -4.4 V while slowly triggering random notes, and also if not triggering notes.

Assuming VR4 is one of the filters pots, I can say I dont get anything "audible" on any of the cutoff and resonance poti pins. (I have 6+3 pins here on the back, that I think relate to the pots, but not sure)

[mario1089]

for anyone reading this, here the facts until now:

1) osc is fine, playable
2) VCA and VCA-env very likely to be fine, as on the mix out I get all the noise and some oscillator leakage (from where?) at note rate and in the right length and shape.
3) As far as I understand, from the original osc signal only a dirac pulse (maybe DC going on and off) is left before going into the VCA.

[MMM]

The values for D34 to D37 are definitely all far off. Either the meter has a problem (then your envelopes actually could function) or the meter is correct - then your envelopes will definitely not be ok.

Make sure that the meter is set to "DC" and that you obtain ground at a safe point. As an example, the middle pin of J4 is a good ground point.

Then let's start with just D34; this is the basic Gate signal that finally also goes to the Gate output. Starting here has the advantage that we are basically dealing with just two different DC; the Gate has just two states (off or on)

The part of the diode with the stripe is called the cathode, the other (unmarked) side the anode. Both sides should be roughly at 0 volts when not playing a note and at approximately 12 volts (or a bit less) when playing a note.

If this is not the case please measure the middle pin of Q37 without playing a note and playing a note.

[MMM]

Assuming VR4 is one of the filters pots, I can say I dont get anything "audible" on any of the cutoff and resonance poti pins. (I have 6+3 pins here on the back, that I think relate to the pots, but not sure)

Please check again; it's the 6-pin potentiometer and the pin closest to C13.

[mario1089]

I was using the inner-most pin of the 7805 as ground reference, that ok?
If it is, I`m very sure all voltages I reported while no notes are triggered are kinda right (does this already tell something?) while the ones while triggering notes are very uncertain, as the meter is too slow. I will try to find some of that clamps to be able to trigger notes manually, to get better results.

I could try to setup my old oscilloscope if we get to a point where this is helpful.

The meter is ok.

I tested all pins of VR4, with the same setup as the osc - there was no audible (periodic) signals. The setup was an audio cable mix-out to external audio mixer to have the mixer at ground, then another audio cables plus-pin to the mixer to touch the pins I wanted to listen the signal. that worked very well for the osc. the mixer I think has high impedance so I thought thats a reasonable setup to track audio signals on the board.

I`m in the office now, so I will be testing things, especially the diode you wanted me to start with, again a little later, in the evening (now is 11 am my time).

[MMM]

Further thoughts:

If I see this correctly from your measuring results ALL four diodes are at -5v when no note is played, correct? This is insofar not plausible as the whole circuit doesn't use any negative supply. Since all four diodes are actually supposed to be a 0v without a note being played I suspect that you've obtained the ground for the meter at a point where there are actually +5v; this of course would result in such a behaviour.

Please check the voltages at J4 on the mainboard. Check both against the middle pin of J4 and the metal tab of the 7805.

[mario1089]

Yeah, I was digging into the schematics the whole morning ( i learned about these things but its 15 years ago in which I only did software), and I also ended up doubting my results:) I think now with better understanding I ll repeat all measurement carefully, to see if these gates work. But from both my VR4 checks and from how the audio sounds at the very end, I still think the VCA and the gates work, but there`s no output from the filter section - Is it possible/useful/safe to try to find the audio signal somewhere inside the filter module, say at Q12/Q21/Q22 ??

[MMM]

Is it possible/useful/safe to try to find the audio signal somewhere inside the filter module, say at Q12/Q21/Q22 ??

Not without an oscilloscope. The voltages inside the filter are very small (in the millivolt range) but ride on a large DC offset. So the only way to really get some useful results there is by using an AC coupled oscilloscope.