PYE Compressor/Limiter Thread *boards shipping* BOM up!

[sr1200]

This thing has gain for DAYS, and yes, turning up the gain is going to  distort quite a bit, at modest levels, I usually dont get above 2 on the font panel markings.  The threshold im still not sure IS a threshold, it almost seems to act like  half of an 1176 input knob (increasing this without touching the output results in a hotter level out).  As for the meters, mine sit at around the 4 (4.2) mark on the meters when idle (which was my only "problem" that remains with the cirucuit).  RF is a byotch with this unit.  Don't have ANYTHING that might even think of generating any kind of radio freq. cause this thing will pick it up and use it (iPhones are the worst, i literally have to put my phone on the other side of the room or this thing will start talking)  From what I've read though, this behavior was observed on an original PYE comp as well, so im not that worried about it, when i use it i just airplane mode or turn the phone off, the time im disconnected from the world is well worth the tonal golden squish that this thing delivers.

 

[mdainsd]

Did you hook up the missing shield on the input transformer?

 

[Rob Flinn]

The threshold im still not sure IS a threshold, it almost seems to act like  half of an 1176 input knob (increasing this without touching the output results in a hotter level out). 

As I said previously R44 seems to have something to do with threshold or at least how much the signal is chopped at R17.

Sr1200 how did you set your trimmers ?  Did Abe give you any info on this ?

 

[sr1200]

Not sure what you mean by set my trimmers.  Most of the trimmers were supposed to be "pre trimmed" the only ones that would have needed adjustment after the fact were the voltage regulator trimmer, the chopper (which i used a DMM to calibrate) and the meter adjust which did nothing for me.

 

[mdainsd]

A lot of work, but no cigar. I really buckled down on this project over the weekend. The devil is in the details...

Finished the harnessing, wired the xlrs, fine detailed the mechanical things like mountings, handles etc.

Got the hotel power supply up, bypass boards working, meter illumination dialed in.

Then got the boards installed and hooked up. On board power supplies checked, adjusted and investigated for noise (none).

Its been doing a P/S burn in. I'll give it 72 hours then start through the signal amp....so far not a thing gets warm so am optimistic about reliable operation.

 

[Rob Flinn]

Not sure what you mean by set my trimmers.  Most of the trimmers were supposed to be "pre trimmed" the only ones that would have needed adjustment after the fact were the voltage regulator trimmer, the chopper (which i used a DMM to calibrate) and the meter adjust which did nothing for me.

So apart from the Vreg trimmer & the chopper frequency, you just left all the other trimmers as they came from the supplier ?    If so, how do you know they are set correctly ?

Did you install a trimmer in position R48 or just put the 1M8 fixed resistor as Abe suggested ?    As far as I can see R48 is the resistor that scales the meter.  I'm not sure that R42 & R44 are to do with calibrating the meter.    Maybe someone a bit cleverer than me could elaborate ?

 

[sr1200]

The BOM states what the trimmers were supposed to be set at before installation.
I did use the fixed resistor, and those two resistors i believe are at least partially to blame for the meter calibration not working.

 

[Rob Flinn]

The BOM states what the trimmers were supposed to be set at before installation.
I did use the fixed resistor, and those two resistors i believe are at least partially to blame for the meter calibration not working.

It actually says "Value selected on test " on the original diagram for the presets.  Which normally means that they test the unit and then select a value for the resistor.  This suggests that the resistance value on the diagram is just a nominal value, & that it would change depending on the operation of a particular unit.  It probably means that the nominal value works perfectly for most units, but on the other hand it may not.      I think, however I will try your method  of presetting them to the nominal value & see what happens.  It's probably likely to get me more in the ball park.

With your meters do they just sit away from zero ?    My meters do move depending on how much I turn up the i/p level, but they never completely go to zero, even with no input signal.  This suggests either that there is a residual signal hitting the meter or that somehow there is compression happening even with no input.  Possibly a trimmer thing .........   

I'm also wondering whether I need the extra resistor R65 because I'm not using a Hairball meter as i have some original PYE meters.  So the F.S.D may be different to the Hairball ones.

 

[sr1200]

I have a feeling you're on to something with the meters, and no mine dont sit at 0 either, from what i understood, the original meter on the pye sat in the middle not at 0, either way, my adjustment pot does almost nothing (full up or down the swing on the meter is like around 1db of movement according to the stock hairball meter)  Im wondering if i should adjust the set screw on the front of the meter to compensate, though i was told thats not generally a good idea to play with that...

 

[Rob Flinn]

The adjuster on the front of the meter is used to mechanically zero the meter when the unit is not turned on.

I tried shorting the added resistor (R65 I think from memory).  But it seems to reduce the meter action that I have to maybe 1dB, as opposed to quite a large amount of swing with it in place.  It's possible that one of that one of the transistors is out of bias & is putting a small residual voltage across the meter with no signal present. 

 

[mdainsd]

Rob, if you are using original PYE meters (and are certain they are the correct value) then you should remove R65, not short it. The other  resistors  (R46 & R45) are the same values that show on original PYE schematics. The added 1K series resistor (R67) is in the original design but it was not on the circuit board. It is now for convenience.

Disconnect the S/C transformer so that there is no signal for it to act upon.

1) Is the circuit generating pulses at the base of TR11? It should not be. If it is I think it is time to look at the bias on the transistors in the output of the S/C as you mentioned.

2) More to follow...

 

[Rob Flinn]

Rob, if you are using original PYE meters (and are certain they are the correct value) then you should remove R65, not short it. The other  resistors  (R46 & R45) are the same values that show on original PYE schematics. The added 1K series resistor (R67) is in the original design but it was not on the circuit board. It is now for convenience.

Right, hadn't noticed that.  I just disconnected R65 & it actually made the meter sit higher (further away from zero) than with it connected.  I then put a lower value resistor in place & this makes it sit closer to zero than the specced value of R65

Disconnect the S/C transformer so that there is no signal for it to act upon.

1) Is the circuit generating pulses at the base of TR11? It should not be. If it is I think it is time to look at the bias on the transistors in the output of the S/C as you mentioned. This leads me to believe that something is out of bias, & is creating a sidechain signal when there is no signal present.

Please see attached scope trace for the base of TR11 with the s/c transformer pulled. 

If you look at the post after this I will attach a scope trace that shows what happens when I turn up the input signal (a sine wave at 1KHz) with no s/c trafo attached at the base of TR11.  The more I turn it up the bigger the second trace gets.   

 

[Rob Flinn]

Attached pic relates to my previous post.  Only 1 attachment per post ...

The trace shows an AF signal with the pulse from the attached pic in the previous post superimposed on top of it.  I'm not quite sure why I should be seeing any sign of the audio signal here with the s/c transformer detached.  You will also notice that with the i/p turned up this much the sine wave is no longer symetrical, perhaps suggesting a bias issue, although the input signal was quite big.

 

[mdainsd]

Removing R65 SHOULD increase any deflection the meter had before removing the resistor. That resistor is a parallel shunt across the meter to scale for the 1mA Hairball meter. So I see no problem there.

So we have the meter indicating a small amount of G/R, AND we have pulses at the chopper with no S/C hooked up. I suspect the meter circuit is actually representing what it "thinks" is going on, so Im not sure the problem is there. I dont have my PYE manuals here. I will retrieve them tonite, but I remember in the foldout that explains the chopping actions that with a signal UNDER the threshold no chopping occurs.

If mine were running I think I would start by looking at that 1.8m resistor installed instead of a trimmer at R48. Possible lift one end of it and also one end of R50. If this is done TR13 shouldnt be conducting (no base current). Then check for pulses at the base of the chopper, if they are still there we have a problem in that network around TR12.

Have you messed with R44?

Edit: Your second trace is at the base of the chopper? If you have audio there and the S/C is disconnected I tend to wonder about the actual chopper transistor?

 

[Rob Flinn]

mdainsd your analysis makes sense

I agree that the meter is displaying what it thinks is there.  It appears that if the s/c transformer is disconnected that there are only 2 ways for the audio signal to get into the s/c.  Via the chopper transistor or through the power rails.  Out of the 2 through the chopper seems most likely.    That would also account for a signal on the meter perhaps.

With regards to R44 can I refer you to my post no. 454.

Unfortunately I start a contract for a few weeks tomorrow, so I may not get much time on this for a while. 

 

[mdainsd]

How big is the test signal you are using as an input? I would think starting around .77Vrms?

 

[sr1200]

I could be way off base here, but with the IC version, shouldnt the "chopper" be constantly running?

 

[Rob Flinn]

It's 4am here and I just woke up and had a light bulb moment.  I think I have the 1M8 installed incorrectly.  This would definitely account for some of the issues that I'm personally encountering.    Unfortunately I can't do anything about unless I can snatch half an hour this evening, if I don't get back from London too late.    Looks like it's going to be a long day. ...

 

[mdainsd]

I could be way off base here, but with the IC version, shouldnt the "chopper" be constantly running?

No. The oscillator located on the daughter board should always be running (250kHz). But the base of the chopping transistor gets no pulse if the signal is lower than the threshold. From the manual: "switch inoperative (100% open)"

This where we should be with Rob's. He has disconnected the side chain its the same as no signal in or less than threshold so we need to get to no pulses on the switching transistor.

Hopefully he just installed that 1.8m resistor wrong and the circuit was off, so it was passing pulses. Fingers crossed!

Sr, have you ever scoped the base of the switcher in yours? I wonder because you have said that yours shows a little gain reduction even with no signal. Sound familiar?

 

[sr1200]

Thats quite possibly the issue with the meter, ill have to take it out over the weekendand look at it.  Both of my channels are functioning as expected compression wise and if i set my threshold above the level thats coming in i dont get compression, just the sound of the compressor not doing anything.