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

[Rob Flinn]

OK, so I installed the 1M8 correctly, and there are things that are better about the way I have one of the channels working, but it's still not great.  The channel is definitely showing compression, but not a great deal.  The meter trimmer will however now scale the meter. 


The residual signal on the base of the chopper transistor is still present.  See attached pic.    This is with no input signal.    Tweaking the 500K trimmer will raise and lower the level of this residual signal.  Thus the meter won't zero. .......

 

[Rob Flinn]

this next pic shows that the signal is much cleaner than previously.

Yellow trace 1 is taken from TR4.  Blue from  C10.    You can see there is still RF on the blue trace.

 

[Rob Flinn]

I was just looking at some of the other PYE diagrams I have & noticed one with a slightly different arrangement for the meter feed compared to the pdf diagram that someone  reverse drew from the board  You will notice that there is an extra 4k7 resistor in series with C21, & both are in parallel with R41 the 15k.I don't think this is an error on the part of  you guys who reverse engineered the diagram from the board, because I have another original version that has the 4k7 missing.

 

[Rob Flinn]

I'm also questioning that we should see nothing on the base of TR11 with no compression happening.  I have some pics of scope traces here that I took from a real PYE that a friend owns, & it seems that you do have very narrow  downward spikes, See attached pic for little or possibly no compression.  I'm not exactly sure whether this is compressing, because I made these pics several years ago for Abe, when he was experimenting .  It's possible that I might be able to  scope it again if I can find some time, so that I'm absolutely sure what is going on with TR11 base with no signal present.

See next post for quite heavy compression

 

[Rob Flinn]

See previous post for details.

 

[mdainsd]

I got my information from several pye manuals, I have not had the luxury to probe around a real one. Also, the one you  got to mess with, how long since it had been calibrated? Maybe it turns out that there is always some compression going on, sure doesnt seem ideal to me, but it may be what it is.

Good luck on your trip!

 

[Rob Flinn]

Also, the one you  got to mess with, how long since it had been calibrated? Maybe it turns out that there is always some compression going on, sure doesnt seem ideal to me, but it may be what it is.

I've no idea when the original unit was last calibrated, probably not for ages.  However the units sound great, so I don't think that there was an issue with them.  The thing is if you did want to calibrate it, you would need some original documentation to tell you how to do it ........

Also have a constant  side chain signal pulse which might last microseconds, is not really a problem because it's hardly chopping anything out of the signal.  Therefore very little gain reduction is happening.  So with that in mind I suspect that the fact there is a constant very short pulse on TR11 base may be a red herring when considering meter issues.  It may be that there is just a d.c offset in the meter circuit.

One thing I will try when I get a chance is to add the 100k that is missing from ground to bases of TR15 & TR18.  That looks to me like it needs to be there.

 

[mdainsd]

Got things further along over the week end. Ive fully loaded the boards with the exception of the oscillator and the switching transistor. I want to get the signal amp sorted out by itself then work on the side chain, then add the osc and transistor.

No smoke, and signal from input to output with disclaimers (more on those later). Very very clean.
I started my evaluation using a 100Hrz signal. Have a power meter across the input and a second for probing. Plus the scope.
The side chain actually responds as far as developing a negative dc signal. The decay is close in time but I didnt even bother trying to zero that in. The ratio dc offsets work, although I have four different original PYE schematics and they all have different R values for these resistors. I'll dial those in after everything else is right.  Setting to 1:1 removes the compression DC signal as it should. Limit I didnt spend much time with so not sure about it.

I lowered the input to 10 Hrz and with the decay in minimum you can see the dc control voltage responding to the sine wave.

Back to the signal amp. As soon as I started going up in frequency the signal started to drop. By 500Hrz it was down 3 dB...rubbish. I took C5 out and the frequency response between the input and the first transistor was within the manual spec, actually a bit better. Ive got stepped pots (linear)  in the threshold and gain positions, but I dont like the way they work. Ive found enough original information that those should be 40dB/2dB per step, step attenuators. I will correct that in time.

Now im up to the LP filter that removes the artifacts of the 250Khz switching. More problems. It too is rolling the signal off rapidly as frequency rises (we are talking about 1Khz). I gave it a rest there. My thought is since the three capacitors in the LP filter and C5 were all the same value, that perhaps I got a bad or mislabeled batch of 1nF caps.

 

[Rob Flinn]

I managed to nip out to the garage briefly.

I just checked the F resp of one of my channels with the s/c transformer pulled & the oscillator board pulled.  I get a slight lift in level around 20KHz but no noticeable drop until around 27KHz.  It's flat to lower than my speakers would reproduce sound too.

The minute I plug the s/c transformer & oscillator back in the RF kicks in.

That's all that I had time for.  Hope it helps.

 

[mdainsd]

More progress!

Replaced the defective caps in the LP filter, Now that passes signal front to rear. Down about 1dB at 20Khz, fine for now and probably for good.

Installed the oscillator board, and? Nothing! OK another board problem, this time on the oscillator (IC) board. Its the trimmer center pin issue again but in reverse. Now if you use a pot there with the leads inline you need to jumper the middle pad to the center auxiliary pad. I cannot get my oscillator up to 250Khz, End of pot only gives about 200Khz. But again, it is good enough for now.

Installed the switching transistor.

Checked signals. Weird. Where I expect to see these PWM pulses, I get a DC level with the pulses. The more input level the more - the control signal to the switching transistor. I dont get the waveforms shown for the original  BUT, It compresses beautifully.  Im wondering if the switching transistor is operating more like a regular transistor and as the base voltage changes it is syphoning more signal to ground rather than chopping it, Limiting works too, but the limit level (output) is quite a bit higher than compressed output.

1:1 works, no compression as it should be.  2:1.,3:1 and 5:1 "look" reasonable on the scope, but I'll have to graph each of their response curves to be sure.

Biggest problem right now is absolutely no meter movement under any condition, Zero all the time 

Rob, as far as oscillator noise on the output signal, mine has a tiny amount at low levels. I also see it is very sensitive to the scope and sig gen leads.

SR1200, my unit does not display any issue with cell phones. I actually laid the phone on the circuit board and carried on a conversation on speaker while monitoring output, both below compression threshold and under heavy compression. I see nothing on the scope.

 

[mdainsd]

A picture of the patient...

 

[mdainsd]

Got the no meter movement sorted out, I missed one solder joint, damn.

I have an idea on zeroing the meter, but Im waiting on that until I verify if this thing is really working the way it is supposed to.

The lowest compression threshold mine will start compression at (currently) is -10dBm. manual says lowest should be -24dBm. Not sure why yet.

EDIT: Its going to take a while to get this thing just right. Yes, it compresses right now. Im sure it sounds sweet, but I dont even want to hear it until I get it looking good on the test equipment.

Im starting to get a feel for the controls and trimmers. Here is my initial ideas of what they do.

R23 - I believe this trimmer is to adjust the unity gain that this thing would normally have. I say normally because the original has the threshold control reverse gained to the output make up. Ours dont, we have separated controls. My plan is to use my power meters on the threshold and make up attenuators and set them at mid range (20dB), then adjust R23 for unity gain. If I cant get it there then I will revisit the output transformer and select a ratio that gets the unity gain within adjustment range.

R52 - When I first studied the schematic I  thought this might be to calibrate the decay times. I dont think so any more. I set my input level at -24dBm with the input attenuator wide open. Then while monitoring the control voltage I adjusted R52 until the volatage was just coming off of the no compression level. Now it handles threshold as low as -24dBm, just like te manual. Set like this the G/R meter now sits at about 1dBm with no compression. When all else is done I will zero the meter with a slight adjustment to the R45 R46 divider. I cant check the upper threshold (+14dBm) because my precision signal generator doesnt have the legs for it, LOL

R42 - Meter full scale adjustment. I so far just upped the input level to where the control signal travelled as far as it would and then adjusted the meter for 20dBm.

The above three trimmers all interact with one another.....it will take a little doing to figure out the procedure for a full calibration.

R44 - Still a mystery. Havent play with it enough yet

 

[sr1200]

phone calls dont seem to set it off too much, but when texts come in, its like joe zawinul is in the house lol.

 

[Rob Flinn]

Glad you seem to have made some progress. I have managed the odd half hour on this in the last couple of days, but have made no real progress.  I had one channel working, but still the RF on output, and not really in calibration at all.  The meter is working but won't sit at zero.  Main concern at the moment is ridding myself of the dreaded RF. 

Mdainsd.
Did you use the inductors for R68 and R1 ? 

Are you susing an off board PSU ?

I would be interested to see your layout of boards & transformers.  All mine are off board.

When I adjust R23 it looks like it has something to do with the asymnetry of the output, since it affects the signal in that way. Also it is affecting the biasing of TR5 which would have that effect.

 

[mdainsd]

Rob,

I used the resistors in those two locations (R68 & R1).

I have an off board power supply in there. But its for hotel functions i.e. Relay board, meter illumination and anything else I may want to hang on there.

How much shielded twisted pair did you use in your build? All the AC wiring in mine (mains and low voltage) is wired with shielded twisted pair, with the shield tied to chassis ground at one place. All the audio that runs anywhere is also shielded twisted pair (only exception the s/c xfmr wiring which is just twisted pairs) shield tied to analog ground. Be careful with test leads, They add a lot of noise to the output if you dont watch lengths and grounds.

I'll post a picture of the layout hopefully this weekend. Ive got the second board out of it right now so it doesnt look finished.

I do see how R23 affects the symmetry of the first stage output. I wonder if I can just put the distortion analyzer across the first stage  and just adjust for lowest distortion?

Keep at it we are getting close!

Like I said Im going to build new attenuators for my two controls, but that can come whenever. Need to find that cool calculator that was posted around here for determining R values.

 

[Rob Flinn]

I took some resistance readings from the stepped switches of the real Pye compressors I got to play with.  I have them in spreadsheet format. If you send me your email I can mail you the spreadsheet.  I'm not entirely sure whether it will make much sense, & it was 2 or 3 years ago so I'm not sure I can explain them very well.    Might be possible to do them again.

 

[mdainsd]

Up and RUNNING! 2 channels doing their thing. It has spent a lot of time on the test bench. I have yet to hear audio through it but thats next.

Going to go instal it into the drum buss, I really like what I have heard these do for drums.

I still want to rework the step attenuators for this thing. The switches arrived and are in the machine shop to modify....always something

 

[bieckmusic]

great job mdainsd!!

mind me asking, about how much the build cost, and how difficult??

looks gr8

 

[sr1200]

Mine hasn't come off my Drum Bus since it came off the bench.  It works really nice if used lightly on acoustic guitars as well.

 

[mdainsd]

great job mdainsd!!

mind me asking, about how much the build cost, and how difficult??

looks gr8

Thank you!

Im into it just about 800 Dollars. It could be done a little cheaper but Im anal about things like NOS 60's knobs (even though there are reproductions) and the pull handles. And I added relay bypass and a hotel P/S. Adding all Mil-Spec silver wire teflon coated added to it too.

Its not that the build is hard, it is labor intensive, there is a ton of harnessing to be done. The boards themselves are high quality. The errors on them are minimal, the worst being having to jumper some adjacent pads around the trimmers, hardly a problem.

You will need a scope and a signal generator at a minimum.

A friend of mine has been watching the progress and he is dying to buy it from me, but I dont want to build it again, so he wont be obtaining it, LOL, I want to get on to the next big project!