Build Thread:MS76

[c1184701]

Change on touch is not unusual, but if it actually works on touch, I would heavily suspect you are making an unconnected ground connection by doing so - especially since its both channels falling perfectly into place, and you measured all resistors going in (going by memero of what you wrote here).

Did you ever share pics of your build here?

Gustav

Hi Gustav, I tested your theory last October by checking extensively for any missing ground connections, but I wasn't able to find any. My XLR pin 1 has continuity with chassis.

Perhaps we could try and deduce the missing connection, if there is one, using the file I've attached. For both James and me, touching the top of the first BC107 in the GR control amp causes the comp to work normally at the lower ratios (instead of needle-burying into heavy compression regardless of input level).

The body of a BC107 has continuity with the collector, so we are touching the junction between the first and second BC107, the 1M and the 47K. The voltage here is supposed to be 15.1V, mine reads 14.9V. This voltage does not change when I touch with my finger - maybe a few millivolts, but that's all. I've marked the spot on the schemo with a red arrow.

My touchy finger has nearly infinite resistance to chassis ground, so I'm struggling to understand what the effect it is having on the the BC107. It can't be shorting it to ground. Perhaps the extremely high resistance of my touch is working in parallel with R73 (1M) and R74 (470K), which fall between the point I'm touching and ground. Adding my finger, assuming some high resistance, would have the effect of lowering the values of R73 and R74 would it not?

Ross

 

[Gustav]

. Adding my finger, assuming some high resistance, would have the effect of lowering the values of R73 and R74 would it not?

Ross

I would agree to that interpretation - youre resistance in parallel.

I think James confirmed that he measured resistors going in?

If not, it wouldnt be completely insane to hypothesise, that if one was packed with an error, maybe two were, which would cause you to have the same issue. I have had cases myself where my supplier sent a wrong value in a bag, and then did it again, where it was a stock, not just a packing error, and I let them know.

basics (resistors/caps) were packed by a third party, so unfortuantely, I cant refer to any resistor drawer to have come up with a wrong value first hand.



Gustav

 

[James Meder]

I think James confirmed that he measured resistors going in?

Yes, I measured all the resistors going in.

Perhaps we could try and deduce the missing connection, if there is one, using the file I've attached. For both James and me, touching the top of the first BC107 in the GR control amp causes the comp to work normally at the lower ratios (instead of needle-burying into heavy compression regardless of input level).

I might be interpreting this wrong but our issues might be backwards... mine seems to work correctly at ratios 8, 12, 20.
On ratios 2 and 4 I can't get it to reach 0vu while doing the calibration.
I am speculating a bit here as I do the calibration procedure to the best of my ability while having the ratio set at 2, so I'm not sure why it gets to 0vu correctly on the higher ratios.

Here's some measurements I'm getting around the area in question. Hopefully it's visible in the photo, but my resistor values are correct according to their bands.
R73, 1M - 4.4v - 334k in circuit (in red)
R74, 470k - 324k in circuit (in red)
R81, 47k - 14.5v - 44k in circuit (in red)
R141, 4.7k - 29.2v and 26.8v (in red)

R121, 47k - 6.4k in circuit
R93, 4.7k - 4.2k in circuit

T7, BC107 - 3.8v emitter, 5.3v base, 14.5v collector

 

[c1184701]

Yes, I measured all the resistors going in.



I might be interpreting this wrong but our issues might be backwards... mine seems to work correctly at ratios 8, 12, 20.
On ratios 2 and 4 I can't get it to reach 0vu while doing the calibration.
I am speculating a bit here as I do the calibration procedure to the best of my ability while having the ratio set at 2, so I'm not sure why it gets to 0vu correctly on the higher ratios.

Here's some measurements I'm getting around the area in question. Hopefully it's visible in the photo, but my resistor values are correct according to their bands.
R73, 1M - 4.4v - 334k in circuit (in red)
R74, 470k - 324k in circuit (in red)
R81, 47k - 14.5v - 44k in circuit (in red)
R141, 4.7k - 29.2v and 26.8v (in red)

R121, 47k - 6.4k in circuit
R93, 4.7k - 4.2k in circuit

T7, BC107 - 3.8v emitter, 5.3v base, 14.5v collector

I think I explained it poorly James. Mine also works correctly at 8:1, 12:1 and 20:1.

At 2:1 and 4:1, my unit goes into heavy compression with the needle buried to the left, with or without any input signal. Touching that particular BC107 restores the normal function of the compressor at 2:1 and 4:1.

 

[James Meder]

I think I explained it poorly James. Mine also works correctly at 8:1, 12:1 and 20:1.

At 2:1 and 4:1, my unit goes into heavy compression with the needle buried to the left, with or without any input signal. Touching that particular BC107 restores the normal function of the compressor at 2:1 and 4:1.

Gotcha, that does sound like we have the same issue then.

I just checked the finger solution again. My unit rests at 0vu throughout all the ratios when I'm holding the top of the resistor. This is while in G.R. mode fyi

 

[c1184701]

Yes, I measured all the resistors going in.

I also measure all resistors before assembly on all kits/projects I build.

I am speculating a bit here as I do the calibration procedure to the best of my ability while having the ratio set at 2, so I'm not sure why it gets to 0vu correctly on the higher ratios.

Gustav doesn't specify a ratio setting to use whilst calibrating the Q-bias, but out of habit I always set my 1176s to 20:1. This is how the MNATS/Hairball 1176 calibration document tells you to do it.

The ratio control is a double pole switch. One set of contacts act as a variable attenuator on the audio signal being set to the sidechain. When you calibrate the Q-Bias, you set the compression to bypass, which takes the sidechain path out-of-circuit. The second set of contacts act as a variable voltage divider on the 12V supply, and they seem to affect the control voltage appearing at the FET. These stay in circuit even when the compression is set to bypass. So it seems to me that there is a good reason to set the unit to 20:1 for the calibration, even though half the ratio switch is being ignored.

Here's some measurements I'm getting around the area in question

All my readings are similar to yours for these components.

 

[James Meder]

The ratio control is a double pole switch. One set of contacts act as a variable attenuator on the audio signal being set to the sidechain. When you calibrate the Q-Bias, you set the compression to bypass, which takes the sidechain path out-of-circuit. The second set of contacts act as a variable voltage divider on the 12V supply, and they seem to affect the control voltage appearing at the FET. These stay in circuit even when the compression is set to bypass. So it seems to me that there is a good reason to set the unit to 20:1 for the calibration, even though half the ratio switch is being ignored.

Interesting, I didn't realize this. For the hell of it I went through the calibration process with it set to 20:1 and it worked correctly... but then switching to the lower ratios is still has the same issue with far too much gain reduction.

I went through and checked all my resistor values on my ratio board and everything is correct. I've probably done this 4 times now :/

 

[JMan]

Are you feeding the same strength signal into the compressor when you switch the ratios? EDIT: and not changing anything else on the control panel, e.g. the input gain

EDIT again: my question was because of a normal behavior in 1176s where the threshold changes when the ratio changes and therefore appears to give less compression when switching to a higher ratio with the same input level. However, after looking at some of your previous posts, it seems that I've slightly misunderstood your issue, so forget I said anything!

 

[James Meder]

Are you feeding the same strength signal into the compressor when you switch the ratios? EDIT: and not changing anything else on the control panel, e.g. the input gain

EDIT again: my question was because of a normal behavior in 1176s where the threshold changes when the ratio changes and therefore appears to give less compression when switching to a higher ratio with the same input level. However, after looking at some of your previous posts, it seems that I've slightly misunderstood your issue, so forget I said anything!

No worries, thanks for checking in. We need all the help we can get!

And yes, we're working from the calibration procedure which is a constant 1k sine wave the whole time and the meter gets near buried when switching to the lower ratios.
This happens after completing the calibration procedure at 20:1.
I had tried to do it at 2:1 previously and I wasn't able to finish the calibration because I wasn't able to move the meter any higher than around -10 on the meter.

 

[karloff70]

Ok, so I had another pIay to see what I couId discern as I haven't found a tech yet. The right side does compress, but it seems to distort quite a bit quite, as in very difficuIt not to distort as opposed to nice, expected fet grit when Ieaning into it. It sort of sounds broken, chewing the signaI to pieces, but does compress. When switching to aII it does nice fet grit on top as expected.
The Ieft side does not compress at aII. I can't discern any differences of note inside.

I went through the caIibration before this anew, and it worked as in everything ended up as it shouId.

This is definiteIy beyond me now, need a tech. And some mood enhancers, as Iike someone eIse said, having a big, expensive, paper weight doesn't feeI amazing.

 

[karloff70]

In fact, if someone was to feeI to offer me a happy amount for it the way it is now, I wouId be very open to it. It is cIose, but I am pretty spent. It Iooks great, bar a coupIe of scratches on the front.

 

[Gustav]

Gotcha, that does sound like we have the same issue then.

I just checked the finger solution again. My unit rests at 0vu throughout all the ratios when I'm holding the top of the resistor. This is while in G.R. mode fyi

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

 

[c1184701]

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

I'm going to jump in here, since I believe James and I have the same issue. James can let us know if he agrees with the following.

After calibration, the needle rests at 0VU in GR mode at 20:1 ratio, with no signal input. Change to 12:1, 8:1, and the needle jumps and then returns to 0VU. All correct there.

Change to 4:1, and the needle jumps before moving over to the left, as though the unit has moved into heavy compression. The same happens at 2:1.

Touch the top of the BC107, and the needle returns to 0VU at 2:1 and 4:1. Touching it has no effect on the higher ratios that are working correctly.

 

[Gustav]

I'm going to jump in here, since I believe James and I have the same issue. James can let us know if he agrees with the following.

After calibration, the needle rests at 0VU in GR mode at 20:1 ratio, with no signal input. Change to 12:1, 8:1, and the needle jumps and then returns to 0VU. All correct there.

Change to 4:1, and the needle jumps before moving over to the left, as though the unit has moved into heavy compression. The same happens at 2:1.

Touch the top of the BC107, and the needle returns to 0VU at 2:1 and 4:1. Touching it has no effect on the higher ratios that are working correctly.

New information always leads to new avenues to finding solutions.

I thought this was a question of scale/actual ratio, while processing. The needle should definitely stay put when you are not feeding the unit.

I have sunday brain, and cant do a whole lot right now, but as a start, please measure DC the first test points (or just the output pin of the 1206 - you can find the pinout in the datasheet). (or even just remove them to see if this still happens)

Gustav

 

[c1184701]

New information always leads to new avenues to finding solutions.

I thought this was a question of scale/actual ratio, while processing. The needle should definitely stay put when you are not feeding the unit.

I have sunday brain, and cant do a whole lot right now, but as a start, please measure DC the first test points (or just the output pin of the 1206 - you can find the pinout in the datasheet). (or even just remove them to see if this still happens)

Gustav

Hi Gustav, I tested for DC at the 1206 pin 6, and TP2. Both measure 0V DC.

Some more info that I think is relevant, relating to the 6-pin header connector that leads to the ratio board: If I disconnect pin 1 on the 6-pin header, the problem remains. Pin 1 sends the sidechain signal to the ratio resistors. If I disconnect pin 4, which brings the output of the ratio resistors to the input of the GR amp, the problem goes away. Obviously I can't compress, because the sidechain signal isn't getting to the GR amp, but the needle is resting at 0VU GR at 2:1 and 4:1.

 

[Gustav]

Hi Gustav, I tested for DC at the 1206 pin 6, and TP2. Both measure 0V DC.

Some more info that I think is relevant, relating to the 6-pin header connector that leads to the ratio board: If I disconnect pin 1 on the 6-pin header, the problem remains. Pin 1 sends the sidechain signal to the ratio resistors. If I disconnect pin 4, which brings the output of the ratio resistors to the input of the GR amp, the problem goes away. Obviously I can't compress, because the sidechain signal isn't getting to the GR amp, but the needle is resting at 0VU GR at 2:1 and 4:1.

Ok, I have seen a handful of 1206 fail and just spit out DC, but they are not doing that.

Is the relay on the ratio board mounted on the back?
Double check att/rel pots mounting and values
double check resistor values on the ratio board.
If all the voltages around the transistors are good, I wouldnt do too much on that, but I might try replacing the 2N5457.

Gustav

 

[c1184701]

Ok, I have seen a handful of 1206 fail and just spit out DC, but they are not doing that.

Is the relay on the ratio board mounted on the back?
Double check att/rel pots mounting and values
double check resistor values on the ratio board.
If all the voltages around the transistors are good, I wouldnt do too much on that, but I might try replacing the 2N5457.

Gustav

Attack and release values/orientation are correct, as are all of the ratio resistors.

The relay is mounted on the back as per instructions - actually it has been removed for the time being, and replaced with a jumper so that the unit is hard-wired in compression mode. I did this to make it easier to measure resistances without relying on the 12v being applied to the relay.

It’s not the 2N5457, I’ve tried about 10 different ones there.

Ross

 

[c1184701]

Here are my latest findings, that have me totally perplexed:

With the ratio board disconnected, and set to 2:1 ratio, I read the following resistances between the pins of the 6 pin header:

Pin 1 & Pin 6 282k schemo: 56k+68k+56k+56k+27k+20k = 283k
Pin 1 & Pin 4 262k schemo: 56k+68k+56k+56k+27k = 263k
Pin 6 & Pin 4 20k schemo: 20k

Pin 5 & Pin 2 4.79k schemo: 4.81k
Pin 5 & Pin 3 4.79k schemo: 4.81k
Pin 2 & Pin 3 continuity schemo: continuity

So this all looks correct, but 2:1 ratio doesn’t work properly. As an experiment, I made a network of resistors wired directly in to a spare 6-pin header, with all of the relevant values for the 2:1 ratio:

Pin 1 & Pin 6 282k
Pin 1 & Pin 4 262k
Pin 6 & Pin 4 20k

Pin 5 & Pin 2 4.8k
Pin 5 & Pin 3 4.8k
Pin 2 & Pin 3 continuity

And believe it or not, with this construction plugged in the 2:1 ratio works correctly! What could be happening here? Both these assemblies are electrically identical as far as I can tell, yet they have different outcomes when put into the circuit.

 

[Gustav]

Thats god like level trouble shooting right here.

So basically, you replace the resistance you have over the switch in the board with equivalent resistance on a small resistor network, and it works as expected?

And you actually measured this on the switch, so its not a case of anyting on the switch shorting out?

So...is it me, or does this just leave the connector being mirrorered, an error in the sequence, or a wire not making contact?

Gustav

 

[c1184701]

Thats god like level trouble shooting right here.

So basically, you replace the resistance you have over the switch in the board with equivalent resistance on a small resistor network, and it works as expected?

That's right! Everything is pointing to the problem being around the ratio board/switch.

And you actually measured this on the switch, so its not a case of anyting on the switch shorting out?

No, I measured this at the pins of the 6-pin header, disconnected from the main pcb. This way every element of the ratio board assembly (the wiring, the resistors, the switch) are taken into account. The resistances between the pins are correct at every setting of the ratio switch. Both sides of the switch just divide a ladder of resistors, so it was straightforward to calculate what these resistances should be using the schematic.

So...is it me, or does this just leave the connector being mirrorered, an error in the sequence, or a wire not making contact?

The 6 pin header can only be inserted in one orientation. The sequence is correct, it is identical on my actual ratio board connector and the dummy 2:1 ratio connector I built. I would entertain a wire not making contact as a possibility, except that the comp works properly at 8:1, 12:1 and 20:1. If any of the wires in the 6-pin header are not making contact, it would totally funk up the operation of all the ratios in a similar way.

Is there any other electrical effect that could be at play here? I noticed that on the original 1176 schematic, the sidechain connections to and from the ratio switch are shielded. I believe they were shielded on my Hairball builds as well. Have you ever had any cases of possible interference into the sidechain in the MS76?

Maybe my wires to/from the ratio board are too long? They have certainly been trimmed to fit, but I didn't sweat over making them as short as possible.