Build Thread:MS76

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
That's right! Everything is pointing to the problem being around the ratio board/switch.

I've slightly lost track of who's had which issues over the last little while, but am I right in thinking that your problems went away when you touched T7? That transistor is right next to that 6-pin header on the main PCB, so maybe something mechanical is at play, a broken trace or somesuch.

Does the problem go away (with the ratio board connected) if you put a small amount of pressure on the connector housing? Is there good continuity from the connector pins to the PCB pads, and to the other ends of the traces that lead from the connector?
 
I've slightly lost track of who's had which issues over the last little while, but am I right in thinking that your problems went away when you touched T7? That transistor is right next to that 6-pin header on the main PCB, so maybe something mechanical is at play, a broken trace or somesuch.

Does the problem go away (with the ratio board connected) if you put a small amount of pressure on the connector housing? Is there good continuity from the connector pins to the PCB pads, and to the other ends of the traces that lead from the connector?

Hi there, thanks for the suggestions! I've tried a lot of physical manipulation of the header connector, and it doesn't produce even a flicker of movement on the meter. If I poke T7 with something insulated, is has no effect, regardless of how firmly I touch it. The curing effect of touching it only occurs with a naked finger.

If you didn't catch my post on the last page, I also created a dummy 6-pin connector with the resistances needed for 2:1 ratio hard-wired. When I plug this in, the comp works perfectly.
 
Hi there, thanks for the suggestions! I've tried a lot of physical manipulation of the header connector, and it doesn't produce even a flicker of movement on the meter. If I poke T7 with something insulated, is has no effect, regardless of how firmly I touch it. The curing effect of touching it only occurs with a naked finger.

If you didn't catch my post on the last page, I also created a dummy 6-pin connector with the resistances needed for 2:1 ratio hard-wired. When I plug this in, the comp works perfectly.


- It works on the other ratios (Means connections should be fine)
- You measured the equivalent resistance on the switch itself (not a wrong resistor, not an internal short in the switch etc, since that would put off the measurement)

So in essense, all checks out, but with the dummy, it works, and with the switch, it doesnt.....

Long shot, totally ignoring that it cant really be the problem with reference to the above

1. Check if the color sequence on the molex is wrong (seen this on rare occasions), make sure 1 goes to 1, 2 to 2 etc.

Gustav
 
- It works on the other ratios (Means connections should be fine)
- You measured the equivalent resistance on the switch itself (not a wrong resistor, not an internal short in the switch etc, since that would put off the measurement)

So in essense, all checks out, but with the dummy, it works, and with the switch, it doesnt.....

Long shot, totally ignoring that it cant really be the problem with reference to the above

1. Check if the color sequence on the molex is wrong (seen this on rare occasions), make sure 1 goes to 1, 2 to 2 etc.

Gustav

Hi Gustav, the sequence is definitely correct. When I was setting up my dummy connector, I was checking all the measurements from dummy to the actual ratio connector, so any situation whereby a pin was not in the right place would have immediately presented a non-matching reading.



I have another issue that we discussed a bit last year, perhaps we could try and figure this one out first?

When I turn my output pot to the last 1/8 turn or so, I get a loud squealing sort of distortion. At absolute maximum it almost fizzles out to static, like radio static. This happens with or without any input signal. It causes the meter to bury to the right (in VU out mode) even though the actual level of the distortion from the output of the comp is at around -10dB digital.

I suspected the output pots initially - one of them has always been a bit jumpy from the very start. But I tried disconnecting the out pots, and replacing with a short between pin 2 & pin 3 of the connector and 100K between pin 1 & pin 2. This simulates a 100k pot fully open, and I still get the interference/distortion.

The right channel is worse than the left. The left channel still buries the meter (VU out) but the static is quite low level. Here's the strange bit though.. as I turn up the right channel out pot, the static starts off loud in the right channel, then moves over to the left channel as well.

The comp is in split mode whilst this is happening. It feels like an oscillation to me, especially since it happens without any input connected to the comp. I used an audio probe to check for the distortion in various places, check out the attached pic. Good means no distortion, or clean test tone if supplying input. Bad is where I can hear the static obvs.

I swapped T3 as a blind effort, but that had no effect. I checked all of the resistors around that zone and they are correct. Also I don't hear the distortion when the blend pot is fully to dry, only when there is wet in the signal path.

-Ross
 

Attachments

  • MS76 one channel OUT POT NOISE.png
    MS76 one channel OUT POT NOISE.png
    299 KB · Views: 17
Hi Gustav, the sequence is definitely correct. When I was setting up my dummy connector, I was checking all the measurements from dummy to the actual ratio connector, so any situation whereby a pin was not in the right place would have immediately presented a non-matching reading.



I have another issue that we discussed a bit last year, perhaps we could try and figure this one out first?

When I turn my output pot to the last 1/8 turn or so, I get a loud squealing sort of distortion. At absolute maximum it almost fizzles out to static, like radio static. This happens with or without any input signal. It causes the meter to bury to the right (in VU out mode) even though the actual level of the distortion from the output of the comp is at around -10dB digital.

I suspected the output pots initially - one of them has always been a bit jumpy from the very start. But I tried disconnecting the out pots, and replacing with a short between pin 2 & pin 3 of the connector and 100K between pin 1 & pin 2. This simulates a 100k pot fully open, and I still get the interference/distortion.

The right channel is worse than the left. The left channel still buries the meter (VU out) but the static is quite low level. Here's the strange bit though.. as I turn up the right channel out pot, the static starts off loud in the right channel, then moves over to the left channel as well.

The comp is in split mode whilst this is happening. It feels like an oscillation to me, especially since it happens without any input connected to the comp. I used an audio probe to check for the distortion in various places, check out the attached pic. Good means no distortion, or clean test tone if supplying input. Bad is where I can hear the static obvs.

I swapped T3 as a blind effort, but that had no effect. I checked all of the resistors around that zone and they are correct. Also I don't hear the distortion when the blend pot is fully to dry, only when there is wet in the signal path.

-Ross

If you are osciallating on the mix circuit, you can probably kill it with a cap across the feedback, but I am thinking, something is unstable.

Check for ripple/measure AC on all all voltages.

Gustav
 
If you are osciallating on the mix circuit, you can probably kill it with a cap across the feedback, but I am thinking, something is unstable.

Check for ripple/measure AC on all all voltages.

Gustav

Hi Gustav, I'm getting some odd readings here, using a UT61E auto-ranging digital meter:

-15V supply: 0V AC (actually starts at 0.3V but drops towards zero over 20 seconds or so)
+15V supply: Meter jumps around, 3V, 160V, then OL and emits an overload bleep, as though mains voltage is connected
+30V supply: Same as +15V supply
-10V supply: 1V AC
+12V supply: Same as +15V supply

Not sure what the jumping around and bleeping means on the positive DC supplies, something not right in the power supply? All the voltages measure the correct DC.

The 1V AC in the -10 supply seems troublesome, does this suggest something being unstable?

-Ross
 
Hi Gustav, I'm getting some odd readings here, using a UT61E auto-ranging digital meter:

-15V supply: 0V AC (actually starts at 0.3V but drops towards zero over 20 seconds or so)
+15V supply: Meter jumps around, 3V, 160V, then OL and emits an overload bleep, as though mains voltage is connected
+30V supply: Same as +15V supply
-10V supply: 1V AC
+12V supply: Same as +15V supply

Not sure what the jumping around and bleeping means on the positive DC supplies, something not right in the power supply? All the voltages measure the correct DC.

The 1V AC in the -10 supply seems troublesome, does this suggest something being unstable?

-Ross

Can you try this just on the PSU, not attached to the main board, measuring from 0 on the send to respective DC supply lines? (or is that what you already did?)

Gustav
 
Can you try this just on the PSU, not attached to the main board, measuring from 0 on the send to respective DC supply lines? (or is that what you already did?)

Gustav

Those readings were with the PSU connected to the main board. I think my meter is reacting strangely in auto-ranging VAC mode though. I have just tried again using the low voltage mV AC mode, with the PSU wires disconnected, which has given me some clearer readings:

-15V: 15mV AC
+15V: 21mV AC
+30V: 15mV AC
-10V: 15mV AC
+12V: 21mV AC

-Ross
 
Those readings were with the PSU connected to the main board. I think my meter is reacting strangely in auto-ranging VAC mode though. I have just tried again using the low voltage mV AC mode, with the PSU wires disconnected, which has given me some clearer readings:

-15V: 15mV AC
+15V: 21mV AC
+30V: 15mV AC
-10V: 15mV AC
+12V: 21mV AC

-Ross

Ok, nothing out of the ordinary there.

I dont have a unit here, its just been so long at this point, since I built one.

Anyone here have a unit for 1:1 measurement comparisons and checks?

Gustav
 
If you are osciallating on the mix circuit, you can probably kill it with a cap across the feedback, but I am thinking, something is unstable.

Check for ripple/measure AC on all all voltages.

Gustav

Could you expand on the cap across the feedback Gustav? I tried a 100pF ceramic across pins 1 & 2 of the NE5532, is this the right place? It didn't stop the oscillation, I also tried across pins 6 & 7.

Ross
 
Question for clarification.

Your wording makes me interpret it, like the needle shows does not stay on 0, if you dont hold it.

so to be clear. Does the needle rest on 0VU in GR mode after initital calibration , but then move left when you switch ratios, even with no signal coming in?


Gustav

Yes, it calibrates correctly at 20:1 but moves far to the left at 2:1 or 4:1. I believe this happens both with/without signal coming in.

When I touch the transistor, the needles moves to 0VU in GRmode on both 2:1 and 4:1.

I'm out of town for the next couple days but I can check again when I get back.

Could you expand on the cap across the feedback Gustav? I tried a 100pF ceramic across pins 1 & 2 of the NE5532, is this the right place? It didn't stop the oscillation, I also tried across pins 6 & 7.

Ross

I have oscillation on mine as well at T7 when I put my probe on the top of the transistor (weird).

My tech suggested the 100pF mod a while ago but I have yet to try it as I need to order the cap. This is where he suggested:

100pf cap.png

Also, I tried adding the same value resistors in parallel for R74 and R73. That helped the meter move a little in the right direction, but still pretty far away from 0vu (approximately from -7vu without, to -5). When I did this, I forgot to check if the oscillation changed so it hard to tell if it made the correct improvement... still need to check that.

I also feel silly modding a kit that never worked correct initially, or is somehow unstable. Like fixing a problem that shouldn't be present to begin with. Is this normal?
 
I dont really have a direction for further remote trouble shooting on this, really wish I did.

So far, I managed to realise, you are getting meter movement with no signal, and that it works with a dummy.

I hope someone with a working unit can chime in, or maybe I can find one locally (Ive dropped off a few at friends places over the years).

And yes, I would not start doing modifications to the basic design to get elements in place, when theres an overall issue.

Gustav
 
I have oscillation on mine as well at T7 when I put my probe on the top of the transistor (weird).

My tech suggested the 100pF mod a while ago but I have yet to try it as I need to order the cap. This is where he suggested:

View attachment 92788

Also, I tried adding the same value resistors in parallel for R74 and R73. That helped the meter move a little in the right direction, but still pretty far away from 0vu (approximately from -7vu without, to -5). When I did this, I forgot to check if the oscillation changed so it hard to tell if it made the correct improvement... still need to check that.

I also feel silly modding a kit that never worked correct initially, or is somehow unstable. Like fixing a problem that shouldn't be present to begin with. Is this normal?

Hi James

Well this suggestion cured my output oscillation! I added a 100pF cap across the equivalent pins of T5 (in the signal preamp section) and now I can crank the output to full.

As for our other problem regarding 2:1 and 4:1, I came up with my own fix in the last couple of days. I've spent a lot of time referencing the original 1176 schematic over the last year whilst trying to get this comp working, and it stuck out to me that the signal path to and from the ratio board uses shielded connections on the original units. I thought I would give it a try on my MS76, and it worked! 2:1 and 4:1 now resting at 0VU GR with no input, and compressing as expected with input signal.

The shielded wire replaced the connections to pins 1 and 4 on the ratio connector. The shields are connected to ground at one end only to avoid any loops. I used pin 6 on the ratio connector.

I have no idea why my unit would need this mod, since so many of these comps are out there working perfectly without it. Same goes for the 100pF over T5. I think I'm going to settle though, I've been troubleshooting it for over a year. It sounds like a 1176, get it in the rack!

I tried adding a 100pF across R73 as your tech suggested, just to see what effect it had on the comp. It made the response of the meter sluggish, and seemed to throw the calibration out of whack. Will be interested to see if it works for you James.
 
I dont really have a direction for further remote trouble shooting on this, really wish I did.

So far, I managed to realise, you are getting meter movement with no signal, and that it works with a dummy.

I hope someone with a working unit can chime in, or maybe I can find one locally (Ive dropped off a few at friends places over the years).

And yes, I would not start doing modifications to the basic design to get elements in place, when theres an overall issue.

Gustav

my 2:1 ratio dummy connector possibly worked because it was very short - just resistors poking out of a 6-pin connector. Now I have shielded wires 1 and 4 on my actual ratio board connections, it behaves the same as the dummy connector. Perhaps I have unusual levels of radio interference in my area, something like that?

Talking of mods Gustav, can you suggest any other values for the sidechain high-pass filter components? I like this feature a lot, but it feels like the filter freq is a bit high for my taste.

Ross
 
my 2:1 ratio dummy connector possibly worked because it was very short - just resistors poking out of a 6-pin connector. Now I have shielded wires 1 and 4 on my actual ratio board connections, it behaves the same as the dummy connector. Perhaps I have unusual levels of radio interference in my area, something like that?

Talking of mods Gustav, can you suggest any other values for the sidechain high-pass filter components? I like this feature a lot, but it feels like the filter freq is a bit high for my taste.

Ross

I have no idea why this would be different on different units. I dont think Ive built a unit without cutting the 6 pin connector to size and twisting it, so maybe thats a variable - but I also dont think Ive ever seen this error over the years, and if that was the variable, its odd it hasnt shown up before/someone else had it.

To lower the HP, lower the capacitor value, simple as that :)

Gustav
 
I've made some progress on mine but am still having the weird oscillation issues. The shielded wire with a short length got the ratio board working correctly- thanks for that everybody! That was acting erratic before.

I tried the 100pf cap across T5 which seemed to help but the oscillation is still there. I tried increasing that to 220pf which may have helped a little bit more. Here's the symptoms:

-Only present on wet signal. Not present at all with dry signal/ratio board removed
-More noticeable on right channel, more noticeable with MS mode than with split
-Ratio selector changes frequency of oscillation pitch, making it highest at 2 and lowest at 20. ALL mode has a similar pitch to 2. This happens with compression in bypass as well.
-The VU in/out/GR switch changes the pitch as well. The VU in/out have a similar lower pitch while the GR has a higher pitch
-It seems to go away when compressing, i.e. if the attack is slow it takes longer to return. If the attack is fast it goes away then comes back quickly.
-There also seems to be some lag when switching between modes- I'm not sure if this is a symptom or normal behavior but it's like it takes a couple seconds to adjust between compression/bypass and different ratios.

I've ruled out power supply and toroid proximity by physically moving them with no effect. I've ruled out the input/output pots by jumping the headers. I'm a little lost at this point on where to go next. I've been wondering if something with R17 on the ratio board would change things? It seems like the oscillation is starting somewhere on the ratio board or before the ratio board. Would increasing the resistance here or adding a cap here be a reasonable thing to try?
 
Also a weird one- I adjusted q-bias on the right channel and got a smoking resistor (R2) in the power supply. I replaced it and R1 with 12ohm resistors since it was what I had around. R1 was fine but I thought I'd keep them even. It was ok with appropriate voltages when not connected to the main board but started smoking again once I connected. I measured the voltages and the -12v and -15v were much lower. Removing the board again the voltages were fine, with the burnt resistor.

Looking at the schematic the -12v feeds into q-bias. I'm not sure where to go on this one but wondering if the voltage being -12v instead of -10v could have affected this? I've looked around for anything else like a piece of metal that may have fallen in and shorted something. Obviously it's happening somewhere on the main board. I'm probably going to try swapping out the 7912 with a 7910 to be "safe"... Also not sure if somehow this is related to the oscillation problems. Not ruling out the possibility of something silly that was my fault like a tiny piece of metal crossing two paths.
 
Just a little update- mine’s working great now too. The burning resistor turned out to be a shared connection between pins 5 and 6 or 4 in a ratio board out that was creating creating some bad loop back to the power supply board. I replaced the 7912 with a 7910 which likely didn’t matter, added the 100pf capacitor as noted above and had shielded wired on the ratio board. Calibrated on ratio 20 and it went off without a hitch! Many thanks everybody!
 
Hi all,
A week ago I had some time to visit my friend James (above) who was having issues with his build similar to others here. The problem turned out to be oscillation on the -12V line which was confirmed with a scope. The datasheet of these regulators does state that a minimum output capacitance is needed for stability and this design only has 0.1uF on each line. Clearly the other voltages are fine but for whatever reason this one has instability without the extra capacitance. The fix is to switch out the 0.1uf cap on the 7912 output with a 10uF aluminum or tantalum electrolytic. Remember since we are dealing with a negative voltage rail here, the positive side of the capacitor should go to ground, and the negative to the output of the 7912 regulator.

I hope this helps anyone else with this specific issue.


Screenshot 2022-08-05 222258.png
 
Hi all,
A week ago I had some time to visit my friend James (above) who was having issues with his build similar to others here. The problem turned out to be oscillation on the -12V line which was confirmed with a scope. The datasheet of these regulators does state that a minimum output capacitance is needed for stability and this design only has 0.1uF on each line. Clearly the other voltages are fine but for whatever reason this one has instability without the extra capacitance. The fix is to switch out the 0.1uf cap on the 7912 output with a 10uF aluminum or tantalum electrolytic. Remember since we are dealing with a negative voltage rail here, the positive side of the capacitor should go to ground, and the negative to the output of the 7912 regulator.

I hope this helps anyone else with this specific issue.

Thanks for posting this info! My unit was compressing pretty much correctly after I shielded all of the ratio board wires, but the meter was still not behaving 100% as it should. Changing the attack and release controls was causing movement in the meter in GR mode, without any input signal. I swapped out the capacitor as you recommended, and my compressor is finally in full working order!

I need to learn some oscilloscope skills, it's sobering to think that one single component swap could have saved me a year of troubleshooting..
 
Back
Top