All things G1176 - the new "repost" thread.

[a2d]

I removed   C27, R7 , and C3 and the problem goes away the meter will only sway .5db
could it be that the fets are not mached ???

I reinstalled C27 and C3 all still good, but as soon as i install R7 the problem starts

 

[a2d]

anyone?

 

[casrec]

what happens when you unlink them? I had this same problem with my boards and it was that the Fets Q1 and Q10 were not matched from board to board. This is extremely vital. Maybe this is not what you are experiencing, but when I would use a board by itself, it worked great, but as soon as I linked them together one meter would go up the other would go down and they would move with adjusting the release pot. The control voltage is connected to the attack and release pot, R48(which connects to Q10), C27, C3, R7(which connects to Q1). So by removing C3, C27 and R7 sounds like you have a mismatched fet problem. Buy some from hairball or google fet matching and there are some simple schematics that you can build to test.

On a the same note, does anyone know if there is a commercial transistor tester I can buy that test and evaluates not only NPN and PNP but also Jfet transistors? Thanks a bunch!

Brandon 

 

[Echo North]

On a the same note, does anyone know if there is a commercial transistor tester I can buy that test and evaluates not only NPN and PNP but also Jfet transistors? Thanks a bunch!

www.agilent.com/find/b1505a 

Wanna lend me $45,000? 

Mike

 

[casrec]

Holy bleepity bleep bleep!!!

Thanks Mike. I will write you a check right now. Better yet let me just go crack open my monopoly game and send you some cash .

So nothing any cheaper just to check Vgs and Vds values?

 

[Echo North]

Holy bleepity bleep bleep!!!

Thanks Mike. I will write you a check right now. Better yet let me just go crack open my monopoly game and send you some cash .

So nothing any cheaper just to check Vgs and Vds values?

Ya there are a few out there (that do Jfets too).  If you haven't already, do a search for "curve tracer".  All the usual manufactures make them but I have found them to be either too spendy or nothing more then a circuit and a pot.

Mike

 

[casrec]

Thanks mike,
Under a grand is definitely more in my price range. Found a few that seem pretty decent and flexible from $600-750. Do your stereo link PCB's allow the fets to have a wider tolerance. Or does it just make it more stable. Putting Q1 into the right conduction can make control voltage on pad 7 off in a big way from unit to unit. What exactly is this little guy doing?

Brandon 

 

[Echo North]

Do your stereo link PCB's allow the fets to have a wider tolerance. Or does it just make it more stable.

I would say both.

Check out Ed's analysis here:
http://www.groupdiy.com/index.php?topic=35656.msg437728#msg437728

The only way connecting the output of the side chains works is if the FETs are perfectly matched.  You can find some that are very close, but who knows how close they stay as they heat up/age?  Tossing a battery and a pot in there (1176SA) may line them up at a point along the curve...but only at one point.  If you've plotted FET curves you know how they might be the same at point A...but drastically different at point B. 

The purple circuit basically keeps the compressors working as separate mono units, but getting the same summed audio into both sidechains.  No combined release and attack settings and no need to have perfectly matched FET's because they are not tied together (no risk of DC potential). 

Since I've plotted about 2,000 FET curves, I have a good idea of what looks normal and what looks wrong.  Basically I did a lot of testing with the kits and found out the tolerances I need to maintain to have two units tracking within 1db up to 10db of GR.  I was surprised with how different to two FETs could be and still have very good stereo tracking.

Hope that makes sense...I mostly don't know what I'm talking about.  I just know people who do 

Mike

 

[a2d]

Mine is a single unit
Yes I was thinking that aswell in the end but I wasn't sure. I ruled that out since I purchased a kit from Hair Ball, maybe I damaged them when I had the in the wrong way round in the beginning???
I guess I'll have to give Mike a call and get 2 more

Thanks Brandon!!!

what happens when you unlink them? I had this same problem with my boards and it was that the Fets Q1 and Q10 were not matched from board to board. This is extremely vital. Maybe this is not what you are experiencing, but when I would use a board by itself, it worked great, but as soon as I linked them together one meter would go up the other would go down and they would move with adjusting the release pot. The control voltage is connected to the attack and release pot, R48(which connects to Q10), C27, C3, R7(which connects to Q1). So by removing C3, C27 and R7 sounds like you have a mismatched fet problem. Buy some from hairball or google fet matching and there are some simple schematics that you can build to test.

On a the same note, does anyone know if there is a commercial transistor tester I can buy that test and evaluates not only NPN and PNP but also Jfet transistors? Thanks a bunch!

Brandon  

 

[casrec]

Do your stereo link PCB's allow the fets to have a wider tolerance. Or does it just make it more stable.

I would say both.

Check out Ed's analysis here:
http://www.groupdiy.com/index.php?topic=35656.msg437728#msg437728

The only way connecting the output of the side chains works is if the FETs are perfectly matched.  You can find some that are very close, but who knows how close they stay as they heat up/age?  Tossing a battery and a pot in there (1176SA) may line them up at a point along the curve...but only at one point.  If you've plotted FET curves you know how they might be the same at point A...but drastically different at point B. 

The purple circuit basically keeps the compressors working as separate mono units, but getting the same summed audio into both sidechains.  No combined release and attack settings and no need to have perfectly matched FET's because they are not tied together (no risk of DC potential). 

Since I've plotted about 2,000 FET curves, I have a good idea of what looks normal and what looks wrong.  Basically I did a lot of testing with the kits and found out the tolerances I need to maintain to have two units tracking within 1db up to 10db of GR.  I was surprised with how different to two FETs could be and still have very good stereo tracking.

Hope that makes sense...I mostly don't know what I'm talking about.  I just know people who do 

Mike

That makes perfect since. I was considering adding a 1176SA circuit directly into my boards. I have gotten lucky a few times with matching a couple of pairs to to have no swing when linked. Most of the time I have a .5 to 1 db swing with with a .04 to .06 swing in the control voltage at pad 7. I guess this is tolerable and is what the SA is for. I will say that I do not like the combined attack and release and personally if you can get a good accurate calibration in each unit I think you can compress unlinked using controls, the meter and fine tuning with your ears. Your(purple) circuit, now explained is a much better way of going about it and since really learning the 1176 circuit the SA is more of a band-aid.

By the way, congrats on you store and your cases look great.

Brandon   

 

[Pine]

New PCB is finish everything working OK !!!!


I have now problem with calibration in STEREO    when switch on  stereo link once meter is -3db and twice is +2db.

 

[casrec]

New PCB is finish everything working OK !!!!


I have now problem with calibration in STEREO    when switch on  stereo link once meter is -3db and twice is +2db.

This is because the fets are not matched perfectly. You either need the SA circuit to compensate for this or preferably Hairballs stereo link boards.

 

[Ptownkid]

Mine is a single unit
Yes I was thinking that aswell in the end but I wasn't sure. I ruled that out since I purchased a kit from Hair Ball, maybe I damaged them when I had the in the wrong way round in the beginning???
I guess I'll have to give Mike a call and get 2 more

Thanks Brandon!!!

what happens when you unlink them? I had this same problem with my boards and it was that the Fets Q1 and Q10 were not matched from board to board. This is extremely vital. Maybe this is not what you are experiencing, but when I would use a board by itself, it worked great, but as soon as I linked them together one meter would go up the other would go down and they would move with adjusting the release pot. The control voltage is connected to the attack and release pot, R48(which connects to Q10), C27, C3, R7(which connects to Q1). So by removing C3, C27 and R7 sounds like you have a mismatched fet problem. Buy some from hairball or google fet matching and there are some simple schematics that you can build to test.

On a the same note, does anyone know if there is a commercial transistor tester I can buy that test and evaluates not only NPN and PNP but also Jfet transistors? Thanks a bunch!

Brandon  

You did not have the fets from hairball the wrong way around, you had the first pair that I gave you in backwards...

 

[Pine]

a2d  do you find your problem ?

 

[a2d]

nawww I haven't found the problem yet, I threw it in the trash,,, just kidding, I'm waiting for the frustration factor to settle down before I try again

 

[a2d]

Hey Bryan,,,

Now that you mention it you're  right I didn't have the HB ones in the wrong way,
It looks like a transistor matching issue
so then WTF????

 

[Echo North]

Hey Bryan,,,

Now that you mention it you're  right I didn't have the HB ones in the wrong way,
It looks like a transistor matching issue
so then WTF????

If it's a single unit I'm not sure it could be a matching issue.  Matching the meter and GR FETs just ensures the meter is reflecting what the actual GR is.

If anything it would be a damaged FET.

Send me an email with your address (info at Hairballaudio dot com) and I'll send you a new pair for free if you want to try that.

Mike

 

[beaversaber]

Hi,
Almost done building my mnats revision F with input and output transformers.  the unit passes audio fine with Q1 removed, (a little hot, but from what I've read this is normal?) but as soon as I put Q1 in, there is significant high frequency roll off (even with the compression bypassed), and won't compress when I try to engage it.  The meter doesn't show any GR and I can't hear anything being done.  However, with GR enabled, the GR meter follows the release knob and high frequencies get added back in as I turn the release knob clockwise

All my voltages look okay for Q12,13,14,15. And my resistor values are all correct (although i'll double check)  Ive checked for shorts/solder blobs and everything looks okay.  I'm going to keep troubleshooting, but has anyone encountered this problem before?

 

[a2d]

Hey Mike

I got thoes fets , I must say thank you very much,  I wish it fixed my probem, well it kind of did, the meter only sways 1 db when turning the release knob.
I'll get to the bottom of the issue.
Oh one more thing not realating to the 1176 does anyone know what value the trimmers should be on the PS board that came with the Rev J boards
Thank again Mike

Mike Smith

 

[casrec]

Hey Mike

I got thoes fets , I must say thank you very much,  I wish it fixed my probem, well it kind of did, the meter only sways 1 db when turning the release knob.
I'll get to the bottom of the issue.
Oh one more thing not realating to the 1176 does anyone know what value the trimmers should be on the PS board that came with the Rev J boards
Thank again Mike

Mike Smith

With both boards calibrated correctly and needle at 0 in GR mode how much does it swing when switched to link? If it's only 1-2 db this is normal and since all the controls are connected at this point adjusting the release will affect the meter. This is completely normal and to have proper stereo operation you need to make the SA circuit or preferably buy Mike's stereo link PCB's. If you don't have much swing when you link the two units and you feel that each channel is calibrated properly don't even use the link. Just use two mono channels and compress evenly using the meter and fine tune with your ears. There is nothing special going on with stereo linking. All it is doing is dividing the difference in the control voltage in half then adding and subtracting to make the same. Say one has 1.03 and the other has .98vdc at pad 7. There is .05 difference. Divide in half and you get.025 so each will have 1.005vdc. But since each is calibrated at 1.03 and .98 it will mess with the conduction of Q1. This is what the SA circuit is for. Mike's(Purple's) circuit is superior because it is summing the two outputs going into pad 15 and bypassing summing the control voltage for Q1 and does not sum the attack. They stay calibrated properly but maintain close compression. Hope this helps.

brandon