Transformerless Vari Mu Compressor build thread

[Bo Deadly]

I recently got an email from somebody wanting my KiCAD design files for this project. I responded in a manner that I thought was appropriate for somebody wanting something for free on which I have spent a lot of time and money on. I really like the part where he threatens to shoot me with a gun.

You don't have to be a foresic document analyst to see that looks a whole lot like the user "midnight arrakas" or whatever the name is talking about defense contracts in all CAPS with triple exclamation points.

 

[Heikki]

You don't have to be a foresic document analyst to see that looks a whole lot like the user "midnight arrakas" or whatever the name is talking about defense contracts in all CAPS with triple exclamation points.

Looks like his marketing tactics have changed. His old account used to post only on the The Daily Grind section of the forum.

 

[rankot]

There was a lot more interest in this project than I expected, so here is a help thread.

Schematics, BOM and instructions can be found here. In case I don’t pay server hosting fees, backup is here.

Mouser cart which should have almost all of the components needed. Missing are tubes, tube sockets and the 36V wall wart. Double check everything before ordering! Does NOT have all the components needed for the rotary switch sidechain version!
https://www.mouser.fi/ProjectManager/ProjectDetail.aspx?AccessID=af87d40971
Here is a Mouser cart for the rotary switch version. Best double check everything is there! Lorlin CK1054 rotary switches can be probably gotten cheaper somewhere else.
https://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=1b27f61fbf

Some hints and tips

C9 on the power supply board can be omitted

C29, C30, C31, C32 on the main board can be replaced with jumpers.

Heikki, I've checked your page and there's no BOM there. Mouser cart is fine, but I see there are 10 3u3/250V film caps, while I can't find them in the schematic - there are 2u2 only, and four of them. Also, I wouldn't recommend shorting C29-32, because they protect expensive output circuitry in case someone plug output of the compressor into phantom powered input of a preamp!

Also, one question regarding tubes - is it possible to use Russian 6K3? They seem to be direct replacement for 6SK7.

 

[scott2000]

Heikki, I've checked your page and there's no BOM there

I see it here towards the bottom?
https://ghr.fi/main.html

but I see there are 10 3u3/250V film caps, while I can't find them in the schematic - there are 2u2 only, and four of them

Hmm. Well the (4) 2.2u caps can be replaced with 3.3u or 4.7u to get frequency response flat down to 20HZ. Maybe this is what they are, seeing as how the (2) 2.2u in the Mouser bom are for the sidechain. The quantity of (10)3.3u is maybe to make sure you get 4 that match closely. But idk. I can't even see the Mouser bom on his site. Just have it from earlier correspondence.

is it possible to use Russian 6K3? They seem to be direct replacement for 6SK7.

That's what he is using in his. I guess he refers to them as Soviet 6sk7 in an earlier post.

 

[Heikki]

BOM for each of the boards are on the website Variable Mutual Contuctance Compressor

Sidechain
GR meter
Power Supply
Main board

Scroll down and you should see them. First post of this thread has the mouser carts. The mouser cart has 3.3u/250V instead of 2.2u/250V. 3.3u should be used if flat frequency response down to 20 Hz is needed. With 2.2u the frequency response might be -0.5 dB at 20 Hz.

The russian 6K3 is equivalent to 6SK7.

rankot is right to caution against leavin C29, C30, C31 and C32 out and replacing them with jumpers. The worry there is that phantom powered mic preamp with large caps at the input can deliver a large current surge and destroy the output devices. Perhaps the main worry is that internal output protection diode (which I assume is there) in INA2134 could be destroyed easily. The 100 ohm resistors at the output do offer some protection by limiting current from the possible surge. If you don’t want to live dangerously like me listen to rankot and keep the caps there.

 

[paolomunarini]

Hi everyone, and happy new year!
I started my build. I'm going for the stepped (mastering) version.

Three quick issues:

• Mouser BOM is missing a 25V/100uF electro. I replaced C2 on the psu board with a 35V/560uF that I already had (should be plenty).
• Mouser BOM is missing 2x 1 Ohm 2W resistor (R1 on the sidechain board). I did a few calculations and a 0.6W should be enough here (BOM in fact includes 0.6W resistors instead).
• Again, Mouser BOM misses 4 of the 8 phantom power protection caps (C9-C10-C11-C12-C29-C30-C31-C32).

Regarding the second issue, @Heikki you know your design better, so, is 2W really needed here?

Best

 

[Heikki]

Sidechain R1 can be 0.6W. I think I originally used 2W resistor there because I have a bunch of 1 ohm 2W resistors. I’ll need to add the missing components to the mouser cart.

Scott is in the process of modding his compressor to the mastering version. I gave him some advice how to measure the threshold to make sure the threshold hi-low switch works as I intended. Now that I think about it, if hammond 107V is used instead of 107N the 5.6k resistor value might need to be different.

Edit: I think if 107N is used the resistor on the hi-lo threshold switch should be 4.7k and if 107V is used it should be 5.6k.

Here is the email I sent to Scott how to measure if the threshold is working correctly.

I’ll try to describe how to measure the threshold.

Turn gain to max, ratio to max, release and attack to fast. Make sure the threshold hi-lo range switch is NOT switched on. Turn threshold to min position/rotary position 1. Use your DAW to feed 1 kHz signal to the compressor. At first make the signal so low that there is no compression. Then increase signal until you start seeing the 1 dB compression LED lighting up. Write down the compressor output level where the compression starts happening.

Now turn the threshold switch to position 2. Now the 1 dB compression LED should turn off. Increase the signal from you DAW until you start seeing the 1 dB LED lighting up again. Write down the compressor output level where compression starts at position 2. Do the same to all the steps up to position 11. (Step 12 is always off)

After you have measured position 11 turn the switch on so it connects the 5.6k resistor to the circuit and turn the rotary back to position 1. Now there should be no compression going on. Increase the signal until you start seeing the 1 dB led lighting up and write down the compressor output level. Do the same again to all the steps now that the 5.6k is in the circuit.

Does this make sense?

The amount needed to increase the signal should stay fairly constant in dB between the steps and where the error might happen is when going from position 11 to position 1 with the switch flipped on.

On the front panel it might best to mark the position 12 for threshold OFF instead of MAX.

Edit 2:

The way the threshold and mastering modification is supposed to work is.

Threshold steps

Step 1 compression starts approximately at -10 dBu output
Step 2 compression starts approximately at -9 dBu output
Step 3 compression starts approximately at -8 dBu output
Step 4 compression starts approximately at -7 dBu output
Step 5 compression starts approximately at -6 dBu output
Step 6 compression starts approximately at -5 dBu output
Step 7 compression starts approximately at -4 dBu output
Step 8 compression starts approximately at -3 dBu output
Step 9 compression starts approximately at -2 dBu output
Step 10 compression starts approximately at -1 dBu output
Step 11 compression starts approximately at 0 dBu output
Step 12 compression starts never

Then you flip the threshold range switch

Step 1 compression starts approximately at +1 dBu output
Step 2 compression starts approximately at +2 dBu output
Step 3 compression starts approximately at +3 dBu output
Step 4 compression starts approximately at +4 dBu output
Step 5 compression starts approximately at +5 dBu output
Step 6 compression starts approximately at +6 dBu output
Step 7 compression starts approximately at +7 dBu output
Step 8 compression starts approximately at +8 dBu output
Step 9 compression starts approximately at +9 dBu output
Step 10 compression starts approximately at +10 dBu output
Step 11 compression starts approximately at +11 dBu output
Step 12 compression starts never

These are when Hammond 107V is used on the sidechain boards. With 107N the output levels where compression starts will shift few dB up.

Edit: 3

Scott has confirmed the mastering modification works as planned.

 

[scott2000]

Scott has confirmed the mastering modification works as planned.

Yeah. Thanks to Heikki being patient with my goof up. Had miswired the threshold switches to the wrong resistor on the sidechain board.

Room is limited on the front panel so I opted to mount the threshold switches and output trimmers/pots under the meter boards....and the highpass switches above..
Awesome..

 

[Heikki]

Here is a front panel by TomKelly. It is for the rotary sidechain boards and the GR meters are on top of each other.

https://ghr.fi/proaudio/varimut/3UTK.fpd

 

[Unit-8]

Hi Heikki,

I saw that you updated the sidechain schematic, but I'm not sure where the 2 connection points of the HI - LO threshold switch are on it, while they are visible on the latest revision of the PCB.
Are they J11 and J12 on updated schematic?
Then, on the previous revision of the PCB the points where to connect the switch are between R9 and R26 with the addition of a 5.6K or 4.7K resistor depending on the type of audio transformer used.
Using the series of resistors of the threshold control to get 1 dB per step, do you use the same values? Does the same apply to the new PCB revision?
Also, there is the 107N transformer pinout on the schematic that doesn't match the Hammond datasheet.
Nice project anyway.

Ta

 

[Heikki]

Yes the connection points are J11 and J12 or in other words from D2 and D3 anodes to V+. These points can be found at one end of R26 and R9.

In the updated schematic marked on the resistors are values for 1 dB steps and on the left side of the schematic are written values for 3 dB steps.

https://ghr.fi/proaudio/varimut/sidechainrotary.pdf
Thanks for pointing out the pin numbering on the transformer. I had not noticed it and I need to correct it at some point. Doesn’t really matter that much in practice because there’s only two ways to put the transformer on the PCB and if pin 1 goes to the square pad on the board then it is the right way.

 

[Unit-8]

Thank you for clarifying some points.
If I use the Hammond 107N does the resistor going on the HI - LO threshold switch have the same value of 4.7K for both the 1dB and 3dB steps, or does it change?

 

[Heikki]

With 3 dB threshold steps there is no need for the hi - lo switch and the 4.7k resistor.

 

[Heikki]

Here are some suggested modifications for mastering use.

There might be some instability issues with the sidechain high-pass filter. I'll investigate it further during the weekend.

 

[steves.mastering]

There might be some instability issues with the sidechain high-pass filter. I'll investigate it further during the weekend.

Just submitted a Mouser order .. exciting so I’ve got plenty of soldering ahead of me until I get to adding the hpf.
Will look forward to seeing your findings

 

[Heikki]

I got to testing the HPF today. It was not stable and made the compressor motorboat under certain conditions.

I updated the pdf with a simple and stable solution for the HPF. Adding 750 ohm resistors in series with the sidechain transformer makes a high-pass filter with -3 dB around 120 Hz. Then to disable the high-pass a switch bypasses the resistors. The disadvantage of this simple circuit is that when the HPF is on it knocks down the overall signal level going to the sidechain a couple of dB. I also measured and made sure the high frequency response of the transformer stays the same when the driving source impedance is increased, so nothing to worry there.
https://ghr.fi/proaudio/varimut/layout/steppedmastering.pdf

 

[Unit-8]

Hi Heikki, could you explain how you got the 120Hz frequency of the HP filter?

 

[Heikki]

The source impedance we are driving the sidechain transformer is normally 660 ohms (330 + 330). When the HPF is engaged we increase the impedance with the 750 ohm resistors making the impedance 2160 ohms. It is just a simple RL circuit where L is the transformer inductance.

In practice it's not just a simple RL circuit and because it is a transformer and we have to consider the reflected load. Improvement to this HPF circuit could be placing 0.47u caps parallel to the 750 ohm resistors so we wouldn't lose overall signal level going to the sidechain when HPF is engaged. I'm going to try it out when I get home after work.

edit: I'm using Hammond 107V transformer. 600 ohm side of 107V and 107N have slightly different inductance but there shouldn't be a huge difference if 107V or 107N is used. I'm going to confirm the results later with 107N too.

 

[Heikki]

Did some testing and here is the updated version how the sidechain HPF should be done.
https://ghr.fi/proaudio/varimut/layout/stepped.pdf

 

[manulaudic]

Hi Heikki, could the HPF be added to the original board? not the rotary switchversion?
Manu