1176 Build, Release knob affecting gain reduction without signal

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ottosteer

Member
Joined
May 16, 2021
Messages
11
Location
San Jose
Just finished my second hairball 1176 build and I’m having an odd behavior. The unit passes signal and works great when the release knob is set to the fastest setting. Without a signal when i set the release slower, the VU meter starts to indicate reduction. It goes from zero and all the way to 20db at the slowest release setting and everything else in between. The gain reduction is not just the meter but actual GR as confirmed with input vs output signal. Any thoughts? I’ve checked the obvious, like values and voltages. Also calibrated and that went as planned. I am very perplexed.


View attachment IMG_8034.MOV
 
Well hello mate, I came here looking an answer to the same question.

I was probing around the 1176 detection circuit and it seems to me that the Release pot can heavily affect the compression ratio.
It actually affects the control voltage to the FET's gate and it's starting point. From the circuit design it seems like this is normal behaviour since the release pot is directly tied to the Q' bias voltage which controls the starting point for the FET's control voltage.

Anyone can confirm this? I am curious on how much of a role it plays on the compression ratio. It looks to be affecting the threshold a lot.
 
Well hello mate, I came here looking an answer to the same question.

I was probing around the 1176 detection circuit and it seems to me that the Release pot can heavily affect the compression ratio.
It actually affects the control voltage to the FET's gate and it's starting point. From the circuit design it seems like this is normal behaviour since the release pot is directly tied to the Q' bias voltage which controls the starting point for the FET's control voltage.

Anyone can confirm this? I am curious on how much of a role it plays on the compression ratio. It looks to be affecting the threshold a lot.
Is yours a hairball build as well?
 
Without a signal when i set the release slower, the VU meter starts to indicate reduction.
Doesn't sound right. With no signal, there shouldn't be any voltage changing the Qbias as the release is changed.
You might have oscillations? You could try grounding the upstream attack pot / diodes junction to see if this stops the behavior.

Release pot shouldn't change the ratio or threshold. The ratio switches do change the threshold when changing the ratio. The original URIE manual explains this well.

I would check everything on the ratio board, make sure the resistances are correct, and that the -10 v is arriving correctly to R64.
 
Ive also had luck removing small leaks by replacing the 6.8uF tants in conjuction with the diodes to a different type, but Ive never seen more than 0.5dB movement or so on the full range of the release before replacement in those cases.

Gustav
 
perhaps leaky fd333 diodes?

have you searched our giant help thread for similar problems and solutions?

/Jakob E.

I didn't manage to get my hands on some FD333s so I went with a couple of 4148s as someone stated in the help thread, temporarily.
Do you think this might be the issue? I was actually wondering what the effect would be of a non low-leakage diode as the 4148s to the circuit. Someone suggested using a FET like the BF245 (?) instead of the FDs, maybe Ill try that when I get the time and see what's cooking.
 
If you are using new 4148s it probably isn't the problem, assuming they aren't backwards or something.
Most likely there is a wrong component value somewhere.
Have you gone over every resistor again and checked all the wiring?
 
Update:

I have just tried substituting the 4148s with BC546 BJTs and the result is amazing! I would love to know more about that, what's the issue with low-leakage diodes on this design?

Note: The BJTs make a good difference when the sidechain is triggered (signal goin to the peak detector) but if you disconnect, the release still messes with the FET's gate voltage. However, maybe I can do better with a different transistor.

OneMoreNote: By observation, my FET conducting point begins at about -3V gate voltage so that is where I have put my Q-bias point to be. Just above the limit. With no signal to the peak detector, the Release messes with this voltage quite a lot causing the FET to act ressistively (am I sayin that right? hope so).
 
Last edited:
An old thread but I had exactly this issue with a Purple Audio MC76 and found the comments useful. The salient word here is "leakage". Q bias adjustment voltage set for the FET gate is extremely high impedance. Any leakage will create a voltage drop across the 5Meg release pot and put the FET into conduction. Any attempt to measure this voltage will draw enough current to throw it off completely. The only way to monitor the effect is to measure the output signal. There are several possible leakage points. The first is the attack circuit. This can disabled by wiring out the attack pot. The second is the meter circuit which can be disabled by lifting the input resistor. The third is the stereo link. On the Purple unit this is a wire off the board and can be lifted. If none of these make a difference then it is the FET itself or the feedback from the output amp. On my unit it was the stereo link with a wiring fault.
 
I've been going through this exact same issue, it's a mnats rev g/f pcb...doing a rev g build.

Release knob affects gr and output level by like 20db, regardless of q bias setting. I've bypassed the sidechain (wired around the attack pot), lifted input resistor for meter circuit, and checked all my wiring atleast 5 times. I've checked all R values in the signal preamp, and tried different fees, all same results. I've also tried searching here and this was the closet topic I could find. Any ideas? I'm tearing my hair out on this one.
 
When you say "wired around the attack pot" you do mean that you have lifted the feed from the pot? Have you lifted the stereo link feed? If you have also lifted the meter feed then all that is left is the release pot going to the fet gate.
In desperation, just to prove a point, you could lift the resistor and capacitor feed to the fet and hard wire from the release pot.
Recheck the stereo link feed.... on the purple MC76 it is a bit of wire. My client never used this link so I just disconnected it!
 
When you say "wired around the attack pot" you do mean that you have lifted the feed from the pot? Have you lifted the stereo link feed? If you have also lifted the meter feed then all that is left is the release pot going to the fet gate.
In desperation, just to prove a point, you could lift the resistor and capacitor feed to the fet and hard wire from the release pot.
Recheck the stereo link feed.... on the purple MC76 it is a bit of wire. My client never used this link so I just disconnected it!
Yes, I lifted the feed from the attack pot. No stereo link option on the rev f/g board so can rule that out :)
Yes, its just the release pot going to the fet gate, I will try the release pot just straight to the fet.

Im sure I have made a mistake somewhere, just after staring at this for 7 hours yesterday, I cant seem to find it yet :ROFLMAO:
 
Well, release pot straight to the fet gate has the same behavior. I think I might need to write up a list of everything I have tried. Ive noticed that the fet gate appears to load down the voltage on the release pot. For instance, if I put my multimeter just on the output of the release pot, when its connected to the 2.2m resistor on the fet, It will vary from (rough values) like 100mv to 1.5v if I haveb the qbias set to 1.5v. If I remove the connection to the fet gate resistor, It will vary from like 1.2 to 1.5v. I think I havbe some more tracing to do.
This one is definitely driving me mad since I havent found any apparent errors yet.
 
Well, release pot straight to the fet gate has the same behavior. I think I might need to write up a list of everything I have tried. Ive noticed that the fet gate appears to load down the voltage on the release pot. For instance, if I put my multimeter just on the output of the release pot, when its connected to the 2.2m resistor on the fet, It will vary from (rough values) like 100mv to 1.5v if I haveb the qbias set to 1.5v. If I remove the connection to the fet gate resistor, It will vary from like 1.2 to 1.5v. I think I havbe some more tracing to do.
This one is definitely driving me mad since I havent found any apparent errors yet.
Okay I think I figured it out. C27 (Q1 FET gate to GND .22 film capacitor) was leaky! The leakage of the capacitor was inconsistent, when I built the unit it was find, now 1 year late the cap was leaky enough to essentially drag down the impedance and created a voltage divider with the release pot.


I'm going to go back towmorrow and confirm but after 14 hours of head scratching it's nice to learn something! I hope this helps someone in the future.
 
Okay I think I figured it out. C27 (Q1 FET gate to GND .22 film capacitor) was leaky! The leakage of the capacitor was inconsistent, when I built the unit it was find, now 1 year late the cap was leaky enough to essentially drag down the impedance and created a voltage divider with the release pot.


I'm going to go back towmorrow and confirm but after 14 hours of head scratching it's nice to learn something! I hope this helps someone in the future.
Hey iampoor1, can you confirm that this was really the solution to your problem?
Because I have the same prob here going like you through everything. I lifted c27 here but it did not solve it here.
So I am grabbing every hint I can get.
 
Hey iampoor1, can you confirm that this was really the solution to your problem?
Because I have the same prob here going like you through everything. I lifted c27 here but it did not solve it here.
So I am grabbing every hint I can get.
It was for me, yes. But I only got there after looking through everything 10 times and disconnecting every single part of the gr circuit I could think of. If your having a similar issue, I woundnt start here....as my issue was VERY odd indeed :D
 
Last edited:
Back
Top