Pentode pre/compressor (warning nube attempt at circuit drafting)

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Blue Jinn

Well-known member
Joined
May 28, 2009
Messages
406
Location
USA Northwest
I was thumbing through an old ARRL handbook, which had a 6AL5 based side chain for a speech compressor. It looked cool. the input was 'parallel' with the grid of the last triode, before the power stage, and the input to the suppressor grid of the first stage noting that it required a circuit with a pentode input. I'm not a designer by any means, and I googled the heck out of this. Did come across discussions using a 6AL5, and also  the "Li'l 4x4" which uses a similar concept, except not with the suppressor grid. I have no idea how to calculate the gain reduction using that method (I generally have no idea how to calculate a lot of things, which is why I hang out here trying to learn things) but I wanted to take a crack at it anyway to see if anything made sense. I took the pentode part numbers straight out of the RCA tables for a 6AU6 which seemed close enough to a 6BA6 (what I read suggested you use a remote cutoff pentode) the triode parts from Zugster's triode gain calculator, and the side chain straight out of the ARRL book. (Except I put in larger value pots in place of the fixed resistors for the attack/decay.)

I didn't find anything like this googling, so I am pretty sure I am barking up the wrong tree here, but thought I'd put it up for criticism anyway. Again, I haven't attempted to do any proper math here, and I have no clue how to calc the gain reduction. Also, I am not 100% certain there should be a pot on the input to the triode. The book called for a "single plate to push pull grids" interstage transformer, so I just took a swag at 1:2.

EDIT: I am also not sure I need all that gain on the pentode, so assuming I could strap the screen to plate instead, and other changes to part values. (And maybe do some math...)

EDIT 1: I did omit a blocking capacitor on the output by mistake.
 

Attachments

  • pentode_pre_comp.pdf
    230.7 KB · Views: 73
This is a 'single ended' compressor topology. If you look at all the classic tube compressors you will see they are all push pull types. There is a good reason for this. The control voltage derived by your 6AL5 does alter the gain of the pentode stage but at the same time alters the dc conditions of the circuit. What this means it that the control signal is effectively superimposed on the audio. For an ham radio speech compressor you can roll off the low frequency response so that the thump caused by the action of the control signal is not audible. If you want a response down to 20Hz then you cannot do this.

This thump occurs even in push pull compressors but it cancels out in the push pull output transformer. That's why all the pro vari mu compressors have a balance control.

Cheers

Ian
 
Thanks Ian.

The Dogzilla uses a similar pentode, but places the CV on the control grid, vice suppressor, and sends the suppressor direct to ground. The RCA handbook discusses this exactly, and recommends the suppressor to AC ground.  The circuit in this thread http://groupdiy.com/index.php?action=dlattach;topic=49415.0;attach=14564;image seems to use a somewhat similar sidechain with the control grid.

Like I said initially, the circuit fragment looked interesting. Would a push pull configuration make sense then, putting the cv on two suppressor grids, or beetter on the ct of an input transformer and take the suppressor to ground? (I'm assuming I'd also then need to transformer couple to the triode?) I'll sketch that out this PM.

Then there is the attached circuit I came across in another thread, which uses the suppressor grids in two 6BA6's in another configuration, which is supposed to cancel out the thump. If I'm reading all this correctly the ARRL side chain would seem to offer a bit more control over parameters. (I notice it also has a 6E5)
 

Attachments

  • vm comp.jpg
    vm comp.jpg
    151 KB · Views: 106
> a somewhat similar sidechain with the control grid.

Which is (typicaly) R-C coupled to the audio source. Now your control voltage is slowed-down by that signal input capacitor (unless the source is infinite-Z, which it probably isn't.) Not jst slow: two R-C filters in the gain-control loop (and enough sidechain gain) will oscillate. It can be done; may not be a fast-path to quality control.

Taking control to an additional grid isolates signal problems from control problems. But G2 G3 parameters are not touted big on datasheets.

> it also has a 6E5

Eye candy. You can put a 6E5 on ANYthing. Give me a 25-stack of 9V batts, I'll monitor my truck's spark or fuel injection on the magic eye.

> I have no idea how to calculate the gain reduction

First approximation: if voltages and Mus are not too wild, you "can" force the tube to cut-OFF, ZERO gain, "infinite" reduction.

But the distortion becomes gross long before that happens. Or if you keep the input voltage very-very small, you drown in universal hiss.

Computing the "useful" gain reduction depends on what you call "useful" (55dB S/N? 90dB? 10% THD? 0.1% THD?) and is very tedious.

> the attached circuit I came across in another thread, ...to cancel out the thump

That's a clever idea, wacky concept. It appears to give thump performance better than one tube but inferior to two tubes, signal performance more like one tube (nearly no 2nd harmonic cancellation, which severely limits signal level), for the cost of two tubes and no iron. Might be a dynamite dictation-machine AGC.

> a 6AU6 which seemed close enough to a 6BA6

AU6 has high gain and cuts-off suddenly. BA6 has higher gain and cuts-off very gradually. A sharp-cutoff tube gives a gain-range only about 2:1 before it just craps-out and distorts. Remote cutoff is kinda necessary. However the classic remotes are aimed for a 100:1 range of gain AND large 2nd harmonic distortion at low gain. Radios really do see 1000:1 range of signal and most modulation methods tolerate gross 2nd distortion (you can recover the whole wave from half the wave by flywheeling a tuned circuit). IMHO the 6BA6 is not the ideal audio squasher; that has not been designed yet (and never will be). I do think you should be trying BA6 instead of AU6.
 
there is one classic tube compressor that is not push-pull,

but it uses an EL panel and a LDR which might account for the lack of thumps,

 
Making it push pull doesn't *seem* to make it *that* much more complex, I've been studying PRRs varimu quite a bit to get an understanding of this.  If this would actually work, doing that would add about US$30, for Edcor transformers and an extra 6BA6. A couple of the examples I've run across, like the quirky one, use the suppressor grid. If I am understanding PRR's post, that would alleviate some problems.

I sketched it out  again (without calculating resistor values) and will post as soon as I can scan it. 
(OBTW folks, I'm *that guy* who starts one project in the middle of another, so I have a bunch of half completed projects on the bench...)

Oh and the 6E5 is just one of those shiny bright objects that always got me giggling as a kid! And that 1961 ARRL Handbook, I literally liked to look at as a kid because of all the cool pictures.
 
Scanning this and then fixing a mistake with Gimp was a bit of a hassle for some reason. Here is another draft. Resistor values are just placeholders. I am also not certain about the interstage transformer. This is still (for me at least) a "proof of concept"  and a learning tool.

If this is even viable, it doesn't look too complex to do point to point. from what I read, there are drawbacks to using pentodes in something like this, so I'm not sure of its application.



EDIT: Also, someone at diyaudio suggested several other tubes, and a few that have common cathode and grid 2 & 3, with separate suppressor and plates.

EDIT: Not sure how many times I can screw up this drawing.... Also, a different tube was recommended at DIYAudio
 

Attachments

  • 6AL6-pent-7.JPG
    6AL6-pent-7.JPG
    65.7 KB · Views: 37

Latest posts

Back
Top