Vari mu compressor with 6V6 tubes

[ruffrecords]

Have you made any tests to see how the two channels track when the stereo link is engaged>

Cheers

Ian

 

[Heikki]

Just curious, what taper is that ratio pot?

Looking at the response graph quoted a few posts above, it seems most of the ratio control is in the first ~20 percent or so of the pot. Mid position is almost all in.

If linear, perhaps an audio taper would spread it out better?

It was 100k linear pot. Since then I've changed it to 50k and now the new max is the mid position in that graph.

Have you made any tests to see how the two channels track when the stereo link is engaged>

The channels track very well if the tubes in channel 1 are well matched with tubes in channel 2.

 

[rankot]

Is it necessary to use 6L6 tubes, or any other pentode (EF86, for example) will do the job?

Also, could some Russian tubes, for example 2P1P, be used for this? Which parameter to look when searching for substitutes?

 

[Heikki]

Is it necessary to use 6L6 tubes, or any other pentode (EF86, for example) will do the job?

Also, could some Russian tubes, for example 2P1P, be used for this? Which parameter to look when searching for substitutes?

You can use just about any tube for vari-gm compressor but for best results you usually need remote-cutoff tube.

Are the links to the schematic working for others?  I'd just like to study it.

I accidentally deleted the folder while back where I had all the schematics and pictures for this thread.

Here's the schematic

Compression curves

 

[abbey road d enfer]

Is it necessary to use 6L6 tubes, or any other pentode (EF86, for example) will do the job?

I would think a small-signal tube, such as the EF86, may have difficulties in handling signal with a large GR. That's one of the stumbling points in the Vari-mu concept. It has to deliver its highest output when its transconductance is at its lowest. I believe the larger geometry of power tubes is a positive factor in that respect.

 

[bluebird]

NIce design Heikki! When you say the screen voltage is critical, do you mean in relation to the cathode or grid voltage?

You have the screen at around 150V referenced to ground correct?

 

[Heikki]

When you say the screen voltage is critical, do you mean in relation to the cathode or grid voltage?

To cathode

You have the screen at around 150V referenced to ground correct?

The screen grid to cathode voltage is somewhere around 100V and voltage relative to ground is about 125V. 6L6 tubes have huge tolerances when it comes to screen current. Almost all new production 6L6 tubes have higher screen currents than the old  datasheets suggest. The grids probably aren't aligned as well as back in the old days. If very specific screen voltage is needed regulator or a stiff voltage divider should be used. In my design a zener and not so stiff voltage divider is good enough.

I think triode connected 6L6 pair would be almost a drop in replacement for for Fairchild 660. It would save money and time for clone builders. You could get 10-15dB of compression with about the same amount of distortion as Fairchild. With 200 - 240V on the plates the amount of control voltage needed would be in the same ballpark and only 2 tubes needed for the job.

 

[lassoharp]

Heikki,

What was your final choice for a transformer to load the 6L6's with?

 

[bluebird]

Thanks, Heikki.

So I'm assuming you are using the 6L6's in pentode mode for more gain then, but you can get the same compression performance out of them in triode mode, just less overall gain? is that the deal?

 

[Heikki]

In triode connection gain is low so using them as pentodes made more sense. We don't need much power so the screen voltage can be fairly low. With low screen voltage the control voltage needed is also sensible.

 

[bluebird]

I've played with this for a couple of days and can't get decent distortion specs under 10 db or so GR.  I've only been using the tubes in triode mode. I've tried matched JJ 6v6, Sovtek KT88 and some RCA 6L6's.  I'm basically using the old Poor Man's 660 setup and the cathodes are pulled down to -8v or so. I can get good sounding compression down to about 6 to 8 db GR but then the signal get really distorted. The tubes are matched so they don't thump but they obviously don't have the power to drive a load at 15db let alone 20db GR. My B+ is 200v. I've tried different output transformers with different ratios 8:1 down to 4:1 with different load resistors down to 100 ohms on the output. I've been able to get way better performance out of 8 matched 6BA6's.

So I'm not sure if 2 6L6's would be a drop in replacement for a Fairchild 660 circuit. I'm sure you could get 6db of nice reduction but beyond that it kind of falls apart. Also I think the grids of the 6L6's are higher impedance than that of 4 triodes in parallel, so a diode bridge side chain acts quicker on the 6L6's and will not sound like the original circuit.

There must be some element to your circuit that is allowing the deep reduction to happen. Is it the fact that you are raising your cathodes instead of dropping them down? Seems like that would be just a relative thing and not matter. Or perhaps it is the screen voltage that is the secret ingredient to keeping the tubes supplying more current close to cut off?

It would be so nice to just have to match two tubes instead of eight! And I have not tried your exact circuit so I can't comment on that but so far in a more traditional circuit I can't get the performance you are.

I'm going to try again with regulated 120V screen and 200V anode and see how it goes.

Thanks!

 

[lassoharp]

Heikki,

What was your final choice for a transformer to load the 6L6's with?

 

[Heikki]

Heikki,

What was your final choice for a transformer to load the 6L6's with?

I'm not sure if you're asking what is the load on OT secondary? If you are it is around 300 ohms.

 

[lassoharp]

Oh Sorry Heikki,  No I just meant what transformer in general did you end up using?  Something commercial or did you wind your own?

 

[Heikki]

Oh Sorry Heikki,  No I just meant what transformer in general did you end up using?  Something commercial or did you wind your own?

I've used Lundahl 1689 and  15k:600 custom transformers from Toroidy.

 

[Heikki]

I've played with this for a couple of days and can't get decent distortion specs under 10 db or so GR.  I've only been using the tubes in triode mode.

In my design I can get about 16dB compression before it starts sounding bad and 6V6 always distorts a lot more than 6L6. I think you should be able to get decent distortion figures in triode connection too. Try different loading on the tubes.

 

[bluebird]

In my design I can get about 16dB compression before it starts sounding bad and 6V6 always distorts a lot more than 6L6. I think you should be able to get decent distortion figures in triode connection too. Try different loading on the tubes.

I spent a lot of time with different loading. I can wind my own transformers and I've tried a couple different winds on this, and like I said everything from 1K to 100ohms resistive loading on the output.

I'm not saying this is not worth doing but I feel like multiple remote cutoff tubes will perform better than two 6L6's.  A while back I built a compressor with one 6ES8 and a 15K;600 output transformer with no make up amp and it preformed very similar to this set up. Basically a PRR varimu with a Traid A-65J and no make up. If I can remember correctly it had a similar sound being that the sidechain was working into a higher impedance.
IMHO to get really down in the -20db range with a single gain stage (660 topology) your gonna need remote cutoff tubes in numbers.

Again, this design is really great and sounds good in a reasonable range that most sane people would want to use a compressor in. BUT whats so cool about the Fairchild is its ability to go down to -20db and still hold together.

 

[Heikki]

Again, this design is really great and sounds good in a reasonable range that most sane people would want to use a compressor in. BUT whats so cool about the Fairchild is its ability to go down to -20db and still hold together.

Does Fairchild hold together at 20dB compression? Maybe if the output level is low? I doubt it but I've never even seen a real Fairchild so I don't know. Some measurement data from a real Fairchild http://kompura.ghr.fi/sources/schematics/Fairchild/670_RBT.pdf. At 12 dBm output, 10dB compression, with time constant 3 on, at 1kHz 2nd harmonic is 0.32% and 3rd is 0.48%. Under similar conditions with well matched tubes my compressor has about one third the distortion.

When matching tubes for compressor matching for low thump isn't enough. Even  if the plate current is equal at a certain control grid voltage there might be big difference in transconductance. Low thump pair of tubes might be unusable in a compressor because of high distortion.

 

[bluebird]

Does Fairchild hold together at 20dB compression?

Yes I work at a studio that has two 670's and one 660. I have seen on an Audio Precision less than 1% distortion throughout the whole 20-20K range with -20db GR. It was a long time ago so I may have to redo the test to make sure I'm 100% correct on the distortion figure. I have noticed that the JJ 6386 are not always matched as good as they are supposed to be, and when people complain about the Fairchild not "sounding quite right" is usually inconsistent tubes.  I really wish there was a good answer to not having to match a bunch of tubes or spending $100/tube but I have not found any. Your circuit seems to be the closest.

When matching tubes for compressor matching for low thump isn't enough. Even  if the plate current is equal at a certain control grid voltage there might be big difference in transconductance. Low thump pair of tubes might be unusable in a compressor because of high distortion.

This is true, I have a curve tracer and the JJ 6v6's cut off at around -25V grid, but some old Phillips 6v6's still manage a couple mA's down to -40 grid. The Phillips seem to sound better in deeper reduction but thump more because they are not as matched as the new JJ's.

Again, not trying to discredit this circuit at all, just want to put the limits into perspective from the results of my personal  experiments with it for other readers to consider.  I also have not tried your exact circuit so I maybe missing something important.

I will post some AP graphs of my 6BA6 compressor when I get home. Its not as good as the Fairchild but better than what I can get with the 6L6's

 

[bluebird]

Lol, My studio just got one of these to try...

https://www.tegeler-audio-manufaktur.de/Schwerkraftmaschine_der_Kompressor

Sounds like they made the "perfect" varimu compressor. Controlled by plugin motorized knobs - stereo and under $3k.

Time to throw out all my designs and just go buy one of those