Pye Compressor

[CJ]

Awesome! Thanks so much!!! 

could you try and measure the core dimensions?

just measure over the brackets,cans,or frames, close enuff,

can you see what color the lams are?

do you have a signal generator and an ammeter and voltmeter?

inject 20 Hz into the primaries at  a level close to what the transformer might see, probably 1 to 5 volts,

write down voltage and current for each xfmr,

here is the test set up

 

[DE]

Hi CJ, yes I can do this tomorrow morning (UK time!).
Will try to take photos with dimensions as well.
It looks like the compressor is now working to full spec but as I said, I'm still concerned about the output transistor temperature which seems too high for long term reliability.
DE

 

[CJ]

check your rail voltages,  reducing Vcc will drop heat,

 

[DE]

Hi CJ, here are the results.
Feeding 1v RMS sine wave at 20 Hz into the output transformer primary (pins 1 and 3) gives 2.56mA (at 1kHz it gives 0.47 mA).
(All other windings including the Zobel network disconnected.)

Feeding 1v @ 20Hz into the side chain transformer primary (pins 1 and 3) gives 0.160mA.
(Secondary disconnected.)

Feeding 1v @20Hz into the input transformer primary (pins 1 and 3) gives 0.007mA. Yes really, I double checked it! (Secondary disconnected of course).

I'm not sure how accurate the current reading of my multimeter is at 20Hz but the volts were measured on a scope with RMS voltage reading so are good.
Laminations of the op transformer are shown in the attached drawing, as are the other transformer sizes. I failed to get decent photos today but will post some as soon as possible and will include a picture of the oscillator coil which others might be interested in.
OP transformer laminations are silver/grey under some very thick yellow varnish. Overall size is around 52x33mm with a slightly unusual clamping arrangement.
As to the heat problem, I'm seeking advice elsewhere but am currently running at 15.5 volts instead of 16 and have fitted heatsinks which seem to help. I think I might increase the emitter resistors and see what that does, who needs to run at +24 dB these days?

 

[CJ]

Great Work!  Awesome!  This should be enough info to wind these things,

Here is the primary  inductance-1 V-rms 20Hz

Output: 3.11 Henries
Side Chain 49.76 H
Input 1137 H.

workingon the lams...

did you post the side chain and input DCR? can not find it..

 

[CJ]

here is an estimate of what is going down here,



output>

 

[DE]

Great work!
Here are the DCRs of the other two transformers that I thought I put in an earlier post but must have forgotten.
Input Trans Pri=821ohms Sec=864ohms
In the manual, the L is shown a as 135H at 50Hz, my meter (Peak Atlas LCR) measures L at 15kHz so is completely wrong.

Side chain Trans Pri=67ohms, Sec=200 ohms.
Manual shows 10.4H at 50 Hz (Pri I'm guessing).
Hoping to redo photos later today to give you a better idea of construction.
DE

 

[CJ]

Thank again! Inductance rises as frequency goes down so thos figures for 50 Hz make sense,
that input transformer is probably done on an E-E core as the inductance is pretty high.

 

[CJ]

here is a guess at the input based on the can dimensions, inductance and DCR,

10K:10K 

some crazies may want to wind this, but they will have to get the EE lams somewhere,

or they could buy a 10K:10K from Jensen and save a lot of hassle,

variance in DCR and computer assisted MLT analysis leads us to the pri-sec-pri structure

that middle section will run above the average of the inner and outer sections due to insulation and screens, thus the DCR offset,

on to the side chain and we are done...

DE, , can you do one more test?

inject a signal into the input transformer at 1 K Hz at as close as you can get to 24 db,  8 dB is fine, and measure the AC volts going into the side chain transformer, pins 1 and 3 > we want to see the ratio between the input level and the level going into the side chain transformer primary,

guessing it will be about 1/4 of the input V.

full props going out 

 

[DE]

Here's the test CJ.
2v RMS (so around +8dBu) at the input of T1 gives 840mV at T4 pin 1 with the attenuator in its mid position, ie threshold = 0dB (and compression set to 1:1).
For interest, there is unity gain between input and output in this condition as I would expect.
Heatsinks on the output transistors seem to cure the overheating so all is looking good for a fully functioning unit.
Photos will follow - I promise!
DE

 

[CJ]

Great work DE!

got the side chain figured, same core only 50 Ni instead of 80,

had a misprint on the input, we edited it to 80 Ni and 5 KG.

tried using a 25 EI lam for the side chain but the numbers wouldn't crunch, not enuff perm and cross section area to get the needed wire on there,

we bumped the max level down from 12 volts to 10 = 22 dB instead of 24 for the side chain as with the atten set at max and 24 dB going in, we would have about 10 volts on the primary, not likely to happen, but better to error on the low distortion side,,

here is the side chain:

 

[CJ]

ok we wound this prototype OPT for the Pye,

our turns estimate was not too far off to give us the same measured inductance,

we said 236 which gave us a bit more inductance than the original due to core perm estimates being a bit off,

turn out that 220 turns of total primary wire gives us the right inductance,

also, we had a hard time with the #24 pri wire being too big so we wound #25 and the DCR was still lower than the original ohms that were measure, so we can go with #26 instead which will give us less stress in trying to fit all the turns on the bobbin, and we can add extra insulation to extend the freq response beyond the 20K Hz that we have right now,

feedback DCR turned out like the original so we will stick with #36

that f/b wind has a lot higher ohms than the rest of the winds so that leads us to believe that it went on last, and that they therefore ran out of winding room and had to shrink the wire size down to #36 to get the turns on,

here is the revised print, we will send this out to Bruno2000 to see how it works,

 

[bruno2000]

ok we wound this prototype OPT for the Pye,

our turns estimate was not too far off to give us the same measured inductance,

we said 236 which gave us a bit more inductance than the original due to core perm estimates being a bit off,

turn out that 220 turns of total primary wire gives us the right inductance,

also, we had a hard time with the #24 pri wire being too big so we wound #25 and the DCR was still lower than the original ohms that were measure, so we can go with #26 instead which will give us less stress in trying to fit all the turns on the bobbin, and we can add extra insulation to extend the freq response beyond the 20K Hz that we have right now,

feedback DCR turned out like the original so we will stick with #36

that f/b wind has a lot higher ohms than the rest of the winds so that leads us to believe that it went on last, and that they therefore ran out of winding room and had to shrink the wire size down to #36 to get the turns on,

here is the revised print, we will send this out to Bruno2000 to see how it works,

Can't wait!  PYE comp and test gear all warmed up.......
Best,
Bruno2000