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

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
yea...that is what I did...you might need to refresh your browsers cache....or download the file

EDIT: OK NOW its fixed! haha
 
Ok...so more reading I find that this oscillator is known as a "blocking oscillator" as referred to in the manual. It outputs a square wave.

Food for thought :

http://mysite.du.edu/~etuttle/electron/elect37.htm


there are many ways to make an oscillator...many easier ways since when the PYE was originally designed. If it was designed today I imagine it might use a 555 chip as the oscillator or they even have ICs that's main purpose is to oscillate. I imagine we can't just pull the old oscillator out and put a new one in...or could we?
 
I think Kingston is just trying to reality check you. Enthusiasm makes it is easy to put too much on ones plate.
Good choice of project though. Don't think there are many diypwm comps. Keep it up.
 
Sorry for my frequent responses....want to update with what I've found...to get some ideas.

www.acsoundstudio.com/DIY/Pye/Information/schematic_NEW_OSC2.jpg
 
abechap024 said:
...there are many ways to make an oscillator...many easier ways since when the PYE was originally designed. If it was designed today I imagine it might use a 555 chip as the oscillator or they even have ICs that's main purpose is to oscillate. I imagine we can't just pull the old oscillator out and put a new one in...or could we?
You could always breadboard one up and test the differences between the two...  ;D

By the way, AC: which Pye compressor are you cloning?  I may be in for a pcb from ya!

 
With regards to the oscillator.  As far as I'm aware a 555 puts out a square wave or a triangle wave with a cap added. I'm no expert on discreet oscillators, but I think they normally put out a sine wave, which might rule out using a 555.

I think that many of these old style PWM compressors vary the mark/space ratio by changing the break point on a comparator, so that it triggers at different heights on a sine wave fed to the other input.  Thus varying the mark space ratio.

It would be useful to get a scope on a real Pye to see what the signals doing in different places.
 
Rob Flinn said:
With regards to the oscillator.  As far as I'm aware a 555 puts out a square wave or a triangle wave with a cap added. I'm no expert on discreet oscillators, but I think they normally put out a sine wave, which might rule out using a 555.

I think that many of these old style PWM compressors vary the mark/space ratio by changing the break point on a comparator, so that it triggers at different heights on a sine wave fed to the other input.  Thus varying the mark space ratio.

It would be useful to get a scope on a real Pye to see what the signals doing in different places.

Yes I also first assumed that the discrete oscillator would probably output a sine wave, but doing some reading in the manual, the oscillator type used in the pye was of the blocking oscillator variety. Blocking oscillators output a square wave/triangle depending on the transformer used. At least they should, but because of the use of the transformer are very prone to ringing and over shoot, I do think this would be our main huge stumbling block, requiring a oscilloscope to tune the oscillator for variations in the transformers. Really a huge time sink. Lets throw that old design out, as I'm sure the man who first designed the compressor would do if he had a stash of 555's available.

Granted if anyone happens to have a PYE laying around PLEASE open it and take note of the waveform type and amplitude!

Which is my next question....the whole side-chain looks to be run off one -12v rail, so at what amplitude should the oscillator output?

Also, Rob what your saying about the circuit changing the mark/space ratio....that makes sense that the oscillator would output a sine - -  -cause I cant seem to understand how it would do that with the oscillator outputting a square wave. I still have more reading and digging to do...good stuff though.
 
It maybe that the oscillator is set to output a triangular wave, because that will work in the same way with a comparator as the sine wave, all be it with a slightly different response.

I know someone with a PYE that might let me get a scope on it. I'll make some enquiries.
 
Rob Flinn said:
It maybe that the oscillator is set to output a triangular wave, because that will work in the same way with a comparator as the sine wave, all be it with a slightly different response.

I know someone with a PYE that might let me get a scope on it. I'll make some enquiries.

That would be great.  ;D
 
Hi Rob,
Actually I mis-spoke earlier. I can't seem to find where I read that a Blocking type oscillator can produce a triangle wave output. Seems that they only produce Square waves. Though there must be something we are missing here, It seems very important we get someone in this world to sit down in the same room as an oscilloscope and a pye unit. Lets hope you email works some magic :D
 
even if this doesn't end up 'pye' per se, i like the thought of venturing into a less ubiquitous form of compression.  i was actually just looking into shelling out for a 4060 manual when this thread got going, must be subliminal messages flying around...  haven't found a legible shot of the 4060 schematics (any revision), but i'm curious to compare them to these.

@ mr. bouncy- while we're on a pye kick, how did mr. purpose's eq end up comparing to yours?
 
Disclaimer : I'm far from a PWM expert, so I may be totally off base here !

The operation is explained in the Pyebook.pdf previously posted. The problem is that this book obviously refers to a different set of schematics than the ones posted along. (components numbers are different).

The oscillator (VT12 in the pdf, TR14 on the schemo) outputs a pulse signal rather than a square one. Actually it is a square but the cyclic ratio is not 50/50. Otherwise you would get a minimum attenuation of 50% of the signal at any settings.
Then TR13(VT11 in the pdf), switches out that pulse according to the sidechain and decay time setting.

Unfortunately no information is given about the width of the pulse. This would need to be measured on a real Pye. The Pyebook also cite 200 and 250KHz as the frequency of the oscillator. Given that this frequency must be filtered out by the LPF (L1, L2 on the schemo), I guess it's important to get it right, so this should be measured too.

Short of those measurements, you could experiment with a variable square wave generator with adjustable mark:space ratio once you got the rest of the circuit prototyped. Find the frequency that is filtered off nicely, and the ratio that gives the correct decay time.

Hope this helps.

Axel
 
I know it may be a long shot but just an idea: how about using uC and it's DAC output as an oscilator: you can generate any waveshape without changing a circuit itself so it would be easy to adjust even after pcb's are manufactured. I know that frequency like 250k may be difficoult to achieve, but those new chips can do a lot of fancy stuff.
 
Yes, that's definitely not a bad thought. But a simple 555 correctly configured can give a wide range of waveforms. IF we had a scope of the actual waveform we were trying to replicate would be great.

http://focus.ti.com/docs/toolsw/folders/print/tlc555calc.html  <--handy tool... the search continues
 
I have only had the opportunity to look at the schematics and the service manual. As Axel pointed out indeed the manual refers to different circuit but the principle is explained.

Michal_k, your idea is not a long shot, in fact it would be greatly more stable and smart. I'll have a stab at this and see what we can come up with but in any case it should not be a problem.

 
Sahib,
I have a really blurry schematic that the manual is referring to. The only thing that is different (as far as I can make out) is the output stage is a bit different, but everything else (including the side chain) is very much similar.
I made out enough to make these references:

TR16 = VT14
TR15 = VT13
TR14 = VT12
TR13 = VT11
TR12 = VT10

eidt:A little update on my end...I have the whole circuit in a simulator, I'm trying various voodoo and hudoo to coax it to preform and allow me to get some usable info. Slowly learning things though


 
Back
Top