DIY 12AU7 Compressor Troubleshooting

Hi and all the best for the DIYers

A while ago I decided to bring an old project of Kent Stevenson and PRR of a simple affordable DIY 12AU7 compressor unit back to life.
I gave the project a little update, using the components available in my country, updated the PCB a little bit, and did my own take on the power supply, which I designed from the basic components available in my local electronic parts store, no need for winding my own transformer. Which I can do but not in my current life situation

I am an electrotechnics and communications engineer, two years fresh off school. Audio has been a great passion of mine, since I also make and record music.
But please don't be too harsh on me, I am no expert in audio systems or measurements, and I know, I sometimes ask a dumb question...
So here goes...

First the pictures:






Second the description:
The compressor works, as far as I can tell. I can measure the output signal, the input is differential, but I feed it with a single ended signal generator.

What bugs me is HOW TO MEASURE A COMPRESSOR???
I can measure bandwidth of the unit, limited by the cheap transformers, but I don't see no (or very little) change when turning the potentiometers of ATTACK and RELEASE.

Another thing that bugs me is the VU meter (old VU from a vintage cassette recorder). I can't get no response from it, only when I turn the 100 Ohm trimmer the needle rises.

I measured the Ohmic resistance using DMM of the VU meter, which is 27,9 Ohm when connected in circuit and 634,3 Ohm unconnected.

Third is the info on the transformers, and all the materials I was able to get on this project, so that future generations can find it.
Transformers I used:
http://www.tme.eu/en/details/tr-1_1/pcb-transformers/

And last thing I want is to share the .brd and .sch files from Eagle 5, in exchange for the community help 

brd

brd

sch

sch

sch

pdf

> HOW TO MEASURE A COMPRESSOR???

If you can measure bandwidth, you can measure gain.

Is the gain the same for 0.1V input as for 5V input?

(With those too-too-teeny transformers, you better try this above 1KHz. They may not pass 1V at 300Hz, and less in bass.)

If the gain appears to be the same, verify that you have audio at the sidechain opamp outputs, similar to the output audio.

If you have more than 1V audio out of the opamps, you should get a negative DC voltage out of the diodes (junction of Attack and Release pots), which increases as audio level increases.

Then a sensitive DC meter across the common cathode resistor should show about 100-200 milliVolts with no signal, decliing to maybe 10mV with strong signal.

How does it sound?

I got back to this project and I am still having trouble with this compressor.
I will do further testing tomorrow to answer all the questions.

Output signal is significantly higher at 10kHz ( around 200mV) thank at 1kHz (barely 20mV)

The VU meter does not work... I dont know why...

Can someone check my schematic, maybe draw in the connection of the potentiometers??? Because I am not quite sure if I have it right...

big thanks

Why would you use a 12AU7 for gain reduction element???

radardoug said:

Why would you use a 12AU7 for gain reduction element???

Click to expand...

Perhaps...

Because it demonstrates that vari-mu stages are really vari-gm stages.
Because the gm of all triodes varies widely with current.
Because the 12AU7 is cheap and readily available

Cheers

Ian

Ok,

To answer PRRs questions/comments:
Negative Gain for Left channel: 1,7
for Right channel: 3,7-4,53

The Right channel has more gain reduction, I think this is because of different gain of the tubes. Should they be matched? If so how to match them? Is there a way to implement tube matching to the circuit?

The 22 Ohm Cathode resistor:
With or without signal the voltage (measured only with DMM at 10kHz, so it might not be accurate)

Left channel:
19mV
Right channel:
10,5mV

All the Potentiometers (Attack, Release, Output) are at max.

The output pot (1k/Log) does not affect the output amplitude. Why???

The VU meter does not move, but it can be set from -7 to -22 dB.

I measured with 10kHz signal, the input of the transformer was connected to signal generator, one of the input leads of the transformer was grounded.

Thanks for help

> 22 Ohm Cathode resistor:
Left channel: 19mV
Right channel: 10,5mV

Something is way wrong. The no-signal idle current should be several mA per tube, not sub-mA per pair.

Measure DC voltages, see what isn't right. It may be on my old-old notes. IIRC, plates over +30V, grids near Zero at no-signal. And of course the tube heaters should light-up dim orange.

Measured,

it does look bad... I double checked the components, I couldnt find a problem.

Values in brackets are for the other channel.

There is no dc path I can see for the grids to 0V. Whatever is connected to the three Xs might be the culprit.

Cheers

Ian

This was the thing that was bugging me, how to hook up the potentiometers for attack and release, I found some notes on the websites dedicated to this project, and I hooked it up this way. Best to my knowledge.
DELETED PHOTO

Put a jumper across C4 and measure tube P G K again.

> I can measure bandwidth of the unit, ..... no... change when turning the potentiometers of ATTACK and RELEASE.

If you are running steady constant test-waves, of course Att and Rel will not have much effect.... they are for dynamic start/fade sounds.

Of course if you have mis-wired the pots, they also may have no effect. This simple circuit has been re-re-drawn so often that _I_ can't see what is what any more. (Without sitting and thinking more than my sore back can stand.)

.

PRR thank you for replies, I appreciate it very much!

The pots are wired like this:



I understand the ATT and REL not having an effect on a steady test wave, but what about the output pot? It should regulate output volume, no?

What about the voltage on the sidechain opamp inputs, it should be 1.5 V, my supply gives me 1.8 V. Could that be a problem? I could probably tweak that by using a RED LED.

I rewired the pots, now to my best knowledge they are wired right. Changed the LED to RED now i get about 1.6V.
I am using only mono potentiometers for ATT and REL.

Put jumper across C4, ballanced and measured the tube:
a2=45,1 V
g2=0 V
k2=129 mV

f-f=11,92 V

a1=45,1 V
g1=0 V
k1= 132 mV

Voltage at the 22 cathode resistor (parallel with 100 Ohm Trimer) = 53,5 mV and sloooowly rising. =>Ik=2,9 mA

How should the THRESHOLD pot (1k/lin) react to a steady sine signal?
Hooked up according to:
https://web.archive.org/web/20110219124725/http://www.diyfactory.com/projects/diy12au7comp/diy12au7comp.htm

If you are making a stereo version with pots instead of rotary switches you need to use the resistors shown but only for one side. You then connect the FROM ATTACK of channel 2 to a point after the 1k circled in green (off of that long jumper is a good spot).

Click to expand...

Measured:

Channel 1:


Channel 2: