Vari mu compressor with 6V6 tubes

[Heikki]

Here's something that some of you might find interesting. Compressor that use 6V6 tubes as vari-gain tubes. 6L6 tubes work here just as well and matching tubes doesn't seem to be an issue. Most 6V6 and 6L6 tubes bought as a matched pair for guitar amps seem to easily give thumpless operation. 

The principle is very simple just one variable gain stage and a sidechain amplifier, just like Fairchild 660.

Here's a simplified schematic. I'll post the full schematic after I've done last few tweaks.

Max output is 26 dBu to 600 ohm load with less than 1% THD, mostly 3rd harmonic. With over 10 dB of gain reduction the distortion figures aren't very impressive but that's already much more than most people need. Gain is 7dB to 600 ohm load and almost 14dB unloaded.

I stole the sidechain from here http://groupdiy.com/index.php?topic=11469.msg136132#msg136132 and made a couple minor changes.

Here's the compression curve with threshold at minimum.

Few pictures of the unit.

In the picture below you can see the toroidial 15k:600 output transformer that I ordered from Poland http://sklep.toroidy.pl/en_US/index

 

[DaveP]

I remember a thread some years back where PRR suggested doing this.

Well done!

DaveP

 

[Heikki]

I remember a thread some years back where PRR suggested doing this.

That's where I got the idea from.

Here's a more complete schematic

Edit: R10, FB1 and FB2 added to the schematic after PRR's comments.

 

[DaveP]

Thanks for that.

Dont you need another 220uF cap after the HT Bridge?  Or did you intend to make it choke fed?

DaveP

 

[Heikki]

There's only need for little over 100V for the screen grids and the anode voltage doesn't matter much. With choke fed power suply I got the voltages I needed from the power transformer.

 

[alexc]

Nice concept  and build - how did the noise floor shape up?

It can be a challenge sometimes with the proximity of an unshielded signal traffo to the psu.

Minimal signal path can be very classy and a match for those who like their GR in the lower dB range

 

[PRR]

I do not understand U3.

It appears to be biasing the cathodes to +24V.

However the direction of current flow is *wrong* for this type of regulator. I would expect U3 Out pin to pull-up until an internal parasitic diode clamped it to In pin voltage, maybe +35V.

A 24V Zener would clamp the cathodes well. A 12V Zener, a BJT transistor, and resistor trim would allow easy bias adjustment.

I suspect the meter could just as well be in the cathode return. That includes screen current but this should not cause much dynamic error (and there's no sure/exact relation between current and gain reduction). Me, I would prefer not to have "high" voltage on the back of a panel meter if there was another way. But as you say the plate voltage "must" be low and barely-fatal, and people should not break meters and stick their fingers in the hole.

If 6V6 work (I thought they were a little too linear), there are 12 Volt heat equivalents. That makes better use of your 24V supply, but also the 12V parts are in low demand, are not being made today, so the ones you find are lower price and all made 1950s-1970s when tube production was smooth and stable. 12V6 is a part; 12AQ5 is a 6V6 work-alike in mini bottle with 250V max plate rating. 12AQ5 is $4 each on thetubestore.com -- 12V6 is $8 each, while 6V6 runs $16-$34 all new-made.

 

[Heikki]

I forgot to draw the two lamps that are connected from U3 out to ground. The cathode current flows through the lamps, which I know isn't the smartest idea in the world. It'll work until one of the lamps die. U3 also powers the sidechain amp which biases the grids to -4V.

 

[Heikki]

I updated the schematic and put in a resistor parallel to the lamps in case a lamp burns out to keep the cathode voltage at 24V. For a stero unit I would only have the 24V regulator and connect all the heaters in series.

I like the idea of using 6V6 tubes here because if they don't match well I can always use them in other amps I have.

Nice concept  and build - how did the noise floor shape up?

The noise is -72 dBu to 600 ohm load, input and output volume turned to max, nothing connected to the input.

Edit: The noise is actually lower. My sound cards noise is greater than the noise of the compressor and the noise was measured with my sound card.

 

[midwayfair]

This is clever ... does PRR's original thread exist? I can't find it but I can find a couple threads referring to it, and I get mostly hits for PRR's 176.

Is there a way to control any of the time constants? I know in some old ones it they were just done with a high/low pass at the grid, but in a one-stage design like this I'm not sure what would be workable.

 

[Michael Tibes]

Looks very tempting  How does it sound?

Michael

 

[Heikki]

The release can be adjusted between 50ms and 2s with values shown for R13, RV2, C7. The measured attack time to get to about 10dB of gain reduction with values shown for R15, RV3, C7 can be adjusted from 1.5ms to 80ms. With R15 at zero ohms and C7 2u the fastest attack is little bit less than 0.5ms.

Looks very tempting  How does it sound?

Pretty transparent with sensible attack and release times and sensible amount of gain reduction.

 

[lassoharp]

With over 10 dB of gain reduction the distortion figures aren't very impressive but that's already much more than most people need.

How much distortion were you getting?    Also, what was max GR you could get with acceptable distortion?

In one of the old threads Rotheu claimed to get 25 db GR from 6L6 pair.

I like this idea, but it seems not worth the extra heater and HV cost if you can't get at least 20 db GR at presumably better distortion specs than the other 9 pin candidates (6ES8, 6BC8, 6BA6 etc)

 

[Heikki]

I'll have to do some more accurate distortion measurements and I'll post the results. The distortion isn't really any worse than most tube compressors have.

You can get more than 25 dB of gain reduction.

6W of heater and 10W of anode power isn't that much for a tube compressor. Most old tube compressors are way more power hungry.

 

[lassoharp]

I'll have to do some more accurate distortion measurements and I'll post the results. The distortion isn't really any worse than most tube compressors have.

You can get more than 25 dB of gain reduction.

6W of heater and 10W of anode power isn't that much for a tube compresson. Most old tube compressors are way more power hungry.

Great.  Look forward to seeing the results.  I wasn't sure if the distortion you had been referring to was something you saw as unacceptable.

25 db is plenty enough for momentary peak control.

Glad you took the lead to experiment with this. 

 

[Heikki]

I measured 1.5% THD at 15dB gain reduction  at 10 dBu output to 600 ohm load. The distortion is pretty much all 3rd harmonic.

I also measured a compressor similar to Fairchild 670 that uses 6 triode connected 6BA6 for each channel and has those expensive Sowter output transformers. It had 3% THD at 15dB gain reduction at 10 dBu output to 600 ohm load.

6BA6 can be found so cheap that they are almost free but you have to buy a lot of them to even find 6 that match well. In the end this was the cheapest tube compressor I've made and fastest to build.

 

[Heikki]

I did the distortion measurements at 1kHz and the distortion stays about the same even at low frequencies if the attack and release times are slow. With fast attack and release times the distortion at low frequencies can be high beacause of the imperfect filtering action of the RC circuit. Fluctuating grid bias voltage modulates the signal and causes distortion. Of course most people here probably are well aware of the low frequency distortion in tube compressors.

Here's what happens to 40Hz sine wave at 15dB gain reduction with attack and release at fastest setting.

And here's a schematic for stereo version I'm building at some point
http://www.kolumbus.fi/~kp5188/kompura/6v6tereokompura.svg

Like PRR suggested the meters in this one are in the cathode return. 6V6 screen current is 1/20 of the anode current so the meter will read just as correctly or incorrectly.

 

[DaveP]

I wonder if that distortion can be heard?  The ear re-creates fundamental bass notes from  harmonics even when the fundamental has been removed!  I suspect that some old saturating transformers give apparently great bass for this reason.

Best
DaveP

 

[Gene Pink]

Here's what happens to 40Hz sine wave at 15dB gain reduction with attack and release at fastest setting.

That's the waveform I would expect to see at those settings.

No compression until the slow 40hz sine hits the threshold, a bit of overshoot as the fast attack kicks in, only the peak gets compressed, and the fast release releases once the sine is below the threshold, and no gain reduction after that until the next peak in the other direction.

Gene

 

[Heikki]

I played some music from my phone through a Behringer mixer to the compressor and made a video.

https://youtu.be/-HZJnLgMD_w

Here are new input vs output curves with two different ratio settings. Also the input transformer has been wired for 6dB more gain.