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

[mdainsd]

Yes Rob. I think we are talking the same language about it now, 

My input and output are stepped attenuators as opposed to pots.

Does it help you?

 

[Rob Flinn]

Yes, thanks that clarifies things.

I now have the oscillation problem return after it was behaving so well for a day ....  WHat a PITA.  about 44KHz saw tooth

 

[desol]

Ok. After removing and replacing tr8, my suspicions were correct! Indeed it was the component causing the noise issue on the right channel. I found it by chance when I touched a leg on the transistor with a little screwdriver and it went pop!...and the noise disappeared. ...In addition, with just the XLR output cables plugged in and both channel amps cranked...i noticed i still had a  teeny tiny little buzz in each channel. After some poking around, I sourced it to the bypass relay boards and hence the 12vdc power supply which needed to be grounded. After replacing the transistor and grounding the 12v supply, both of my channels are now dead quiet....no cell phone interference(i never experienced any of that), no buzz or hum.

Indeed I did have the thresholds wired backwards... Now they function as they should, increasing compression as they're dialed further into the counter-clockwise position.

I tried to follow Mdainsd instructions on setting r52(release). I set the threshold for -10db, maximum reduction...and input a -10 db signal(.24vdc sine) into both channels. I already had the release trims previously turned all the way down(meters at zero). The base of tr11 showed the normal pulses(spikes) on both channels which increased when I applied the -10db signal. I increased the release trimmers until the pulses just barely started to move upwards and then set them both identically and left them there. Hopefully I did this correctly!

Interestingly, I remembered that I had the Eddie Kramer Pye plugin on my computer, which is supposed to be a digital true replica of a Pye unit. So, I booted the workstation up and ran a drum loop through it to check the controls. Unsurprisingly, it was operationally identical to the DIY unit...including the volume bump/offset that happens when one switches to limit. Only real difference is, this hardware unit sounds incredible....and the plugin, well....sounds like a plugin.

So...both these channels now are working properly(as far as I can tell!). Other than waiting on the knobs, running a few more tests, etc. I'm feeling like they're close to being buttoned up!

 

[mdainsd]

Great to hear! Another one is alive.

Yep you aligned yours the way I did mine. If you compared it to the plug and got the same results, then I'd say that way of aligning these is a good one.

Mark

 

[Rob Flinn]

Wish I'd known about this when I was populating the boards

http://buildbutler.org/build/pi-3141-rev3-chip-version#_22K

On another note I'm still getting quite a lot of 44KHz on the outputeven after the tranasofrmer.  I tamed it a bit by soldering a 10k acorss pin 2 & 3 on the XLR.    Still not really sire where it's coming from.  Other than that both channels seem to be working.  I wish we knew for sure what all these trimmers do. 

 

[desol]

I don't understand! You left the 22k's out? I just checked my unit and the sine's do look normal at the xlr's with compressor switched in and different ratio's. 1k sine at .77v.

Edit: Actually, I just tested them again and I think I just had the output knobs off a bit. Checked with outputs aligned, and they seem to switch the same on the scope. However, I may change makeup pot to stepped switch in the future.

 

[Rob Flinn]

I don't understand! You left the 22k's out? I just checked my unit and the sine's do look normal at the xlr's with compressor switched in and different ratio's. 1k sine at .77v.

I did notice something about my unit tho, and something I kind of suspected. My ratios are different on the scope between the channels. They both work good...but the right channel produces more compression per click. I'm thinking something isn't matched in the sidechain or isn't working quite up to snuff on the left channel. I'll have to study the schematic more to hone in on a likely component.

No I didn't leave the 22k's out.  It's just an app to help you find the components on the board easily.  You click on a compinent in the list on the left & it shows you where it is on the board.

I made a double attenuator, which isn't working quite right, probably because I rather rushed it to check it out.  Both my channels work after a fashion, but I don't think they're quite right.    I'm going to rebuild the attenuator in the next couple of days & see what I get. 

I believe R52 is actually a decay adjust.  I have a copy of a pye manual with a circuit diagram where R52 (which is R48 on this diagram is lablelled decay.  See attachment.

So if that is the case, which seems likely, what do the other presets actually do ?

R23 on Abe's board is not an adjust on test on the real diagram I have posted.  So who knows what that does.
R42 on Abe's board looks like it does meter scaling perhaps.  But that part of the circuit is different to the original, probably to cope with the meters you guys are using.  I need to probably restore to original because I am using some Pye type meters.  But this coukd be causing the meters to sit away from zero.
R44 & R48 I'm not sure about, I gues the thing to do is scope the area & see what effect they have on the sidechain signal.

 

[desol]

I noticed that neat piece of software that everything was dialed into.

Yes, I'm pretty sure it's in the build guide as well. R52 is release(decay). The way I have mine set is that it's basically turned all the way down(counter-clockwise till it clicks)...which lowers the meters to zero. Then calibrated up from there, using the scope. It still leaves the meters at zero. I didn't change any other trimmers from the screen print on the boards. I tried to dial R42 but it did nothing for me...so I left it at 5k1.

Would be nice to have an actual calibration procedure. Not exactly sure which Pye unit Abe's is based off of.

 

[Rob Flinn]

One other thing is that some of these PYE's have a threshold range from -24dB to +16dB.

 

[mdainsd]

Yep, they are all over the place. There are variants, One with a more substaintial output amplifier. Another with a gate and/or gain section in the front end.

The step attenuators are 40 dB 2 dB per step. That matches the -24 to 16dB version (or face  plate) The case I have is marked -10dB to +10dB, so i guess next time I have it open do a full measurement and see which is closer. I did use -24dB as my base input signal level. It probably can't output +16 on the more modest output section.

 

[Rob Flinn]

Mdainsd

Can you clarify which preset you used to set the threshold since R52 is definitely a release trim?  You said in a previous post that you made some notes when you calibrated,  this would be very useful to the rest of us if you would care to share.

 

[desol]

The threshold is fixed isn't it? At -10db? If the input exceeds the -10db, it gets compressed.

 

[Rob Flinn]

The threshold is fixed isn't it? At -10db? If the input exceeds the -10db, it gets compressed.

OK, I didn't twig that was what the case was.  I'd still like to know what the other AOT (trimmers) do.  One looks to adjust the meter, but does it set for zero or scale the meter so it reads the correct GR.  If it scales it & the meter is sitting about zero with no signal there is an issue somewhere else. For example another AOT resistor is causing gain reduction with no signal.   

 

[mdainsd]

Rob, I am not understanding  "If it scales it & the meter is sitting about zero with no signal there is an issue somewhere else. For example another AOT resistor is causing gain reduction with no signal. 

When there is no signal the meter should be at zero as there is no compression going on. These are not like Fairchilds and others that the meter sits way up at 0dB on a VU meter when there is no compression.  Edit: At least that is how mine is working, go to "AC Sound PI-3141 Stereo Compressor Pie Pye" on youtube, it shows the prototype with the meter in operation

Maybe Im misunderstanding what you are saying?

 

[Rob Flinn]

I fully understand the meter doesn't sit up like on a UREI .  I have 2 origianl PYE meters here for my build so I need to dump  R65 & R67.  I'm just trying to get my head round the operation of the meter circuit

The preset R42 is either for scaling the meter, i.e making it so that if you have 10dB gain reduction happening it shows 10 dB on the meter, or to zero the meter so there is no voltage across it with no GR.

Looking at the diagram I posted a couple of posts above
The meter moves when there is pwm action, but is the meter movement derived from the current change through VT11 & R44/R74 or from the other branch through R41, or both.  It looks to me like R41 is trimmed so that the node R44/R74 is at the same voltage as bottom of R41 for no signal to zero the meter.    I am still getting my meter off zero with no signal. Since I can't zero it with R41, something must be out of trim so that some PWM action is happening with no signal, which is what is getting the meter to sit up,
This is why I would like to know what the other trimmers do.

Sorry if I'm harping on about this, but I'm close to getting this sucker working, & understanding this seems to be key.

 

[desol]

Rob is yours sitting just off zero like in Abe's video? If you dialed the release all the way down and left R42 at 5.1k it should be like Abe's. All your other trimmers are at suggested values right? I used a little of the meter adjust screw to fine adjust even closer to zero after the calibration.

Glad you're close to getting it running. The more signal you feed into it, the better it sounds.

 

[Rob Flinn]

Rob is yours sitting just off zero like in Abe's video? If you dialed the release all the way down and left R42 at 5.1k it should be like Abe's. All your other trimmers are at suggested values right? I used a little of the meter adjust screw to fine adjust even closer to zero after the calibration.

I don't have the same meters Abe is using.  The thing I'm not getting is why one uses the release trimmer to zero the meter ???

 

[desol]

R52 states 1k....so that's where I had the trimmer at first. That put my meters close to center on first power up. I turned R52 down cause I read Mark's post relating to how he had calibrated the unit + I tried r42 to move the meters...which didn't seem to do much. The position of the release trimmer didn't seem to affect the sound of the unit. The only thing that seemed to change was the overall volume and meter position. Pretty sure I listened to it with the trimmer at 1k and turned down and there wasn't a major difference. So I just left R52 adjusted to minimum position(ccw) and carried on with the rest of the calibration suggestion from there, which when done still left them close to zero.

Not sure why he didn't print the trimmer functions on the board. But you can more or less kinda figure out by where they're located(I think). If we had an actual procedure that would be cool. The units probably should be -24, +16 on the meter scale like you mentioned before. It doesn't bother me much if the scale is a little off, cause I love the sound of the unit.

 

[Rob Flinn]

R52 states it's a 1K trimmer on the original diagram I posted.  That doesn't mean it's necessarily set to 1K, although it could be.  When I scope the base of the chopper transistor adjusting R52 changes the level of the Pulse, with it set at zero I'm getting about 300mV pulses.  I find it difficult to believe that they would bother putting an AOT resistor in there if in actual fact it needed to be shorted for the unit to operate correctly.

R44 & also R42 change the width of the pulses.    R52 has more influence on the meter level than R42

Can you point me to "rest of the calibration procedure ? "

 

[desol]

I wish I could. I've been looking for a calibration procedure myself. Strange it wasn't in the Pye manual, etc.
My r52 trimmer isn't shorted. After calibrating it back up on the scope with the threshold cranked, it was probably about 2 or 3 turns in from the bottom click where the control spikes began to move upwards.

The other trimmers were left on suggested settings.