One-Tube Opto Comp

[OneRoomStudio]

This isn't groundbreaking by any means, but I've been thinking about a way to build a cheap-but-nice-sounding tube opto compressor, and it seems like something like this would accomplish most of what I want while being very affordable. The power supply would be a little annoying, but really only needs to produce B+ (~280V), +/- 15V, and filament (the 6.3V DC for the filament can also be used for the "+5V" in the meter-driving circuit). I based the side chain on Joe Malone's LA500a, which I know works well (I own a pair of them). This won't be able to drive anywhere near the level of an LA-2A, but it should do around 20dB of make-up gain (+14dB from the IT, +19dB from the tube, and -12dB from the OT), which is enough for me. For the cost of a pair of Edcor UMM transformers (under $25/ea), one 12AU7, a TL072, and a handful of passives, this seems like a nice project.

Thoughts on the circuit?

 

[rock soderstrom]

I am very curious about the answers, I had a similar idea with the same side chain. I wanted to implement it with an E88CC/6DJ8, so that it could cope with less B+ and the power supply (or better the transformer) could be simplified. A voltage multiplier that sits on the secondary winding of the bipolar power supply and a passive heating PSU on the other secondary winding.

However, I have never implemented this idea.

The version shown above is also interesting, in the end the PSU could also be done conventionally with two transformers.

I am looking forward to feedback.

 

[metalockpick]

If you use a high voltage, low current light bulb, you can often light it up with the tube, if the tube will drive sufficient current. High voltage, low current light bulbs react quicker to transients. You pair that with a good high impedance LDR element and use a precision voltage to current converter (I resistor or a pot) . I've always done it with tube power amps, but I think it could be done with nine pin tubes (and a transformer)..

 

[OneRoomStudio]

If you use a high voltage, low current light bulb, you can often light it up with the tube, if the tube will drive sufficient current. High voltage, low current light bulbs react quicker to transients. You pair that with a good high impedance LDR element and use a precision voltage to current converter (I resistor or a pot) . I've always done it with tube power amps, but I think it could be done with nine pin tubes (and a transformer)..

All very true. My goal with this project was to keep it as cheap as possible. An opamp and a vactrol is a lot cheaper than an additional tube and a transformer.

 

[ruffrecords]

This is very interesting because I am doing something similar. The main differences are first that I use an SRPP gain makeup stage using a 6922 tube. This provides about 28dB of gain and you can use a 2:1 OP transformer so you only lose 6dB of it. I am also building a stereo pair and wire the heaters in series so they work off 12V dc. I am also running the side chain off 12V dc which saves some power supply. My side-chain is also different. I use a voltage to current op amp to drive the LED and its input voltage is created by a conventional diode RC circuit driven by a side chain input amp much like yours.

A couple of queries about your design. First, it looks like you have an bout 46dB of gain in the side-chain amp which seems rather a lot to me. Second are you sure the TL072 can drive the diodes and LED through a resistor as low as 220R?

Cheers

Ian

 

[rock soderstrom]

I am also building a stereo pair and wire the heaters in series so they work off 12V dc.

You could add a simple microphone amplifier, also with an E88CC, which would allow you to heat these two tubes in series with 12.6V DC.

This could be a nice recording front end for an audio interface.

 

[OneRoomStudio]

A couple of queries about your design. First, it looks like you have an bout 46dB of gain in the side-chain amp which seems rather a lot to me. Second are you sure the TL072 can drive the diodes and LED through a resistor as low as 220R?

The gain setting in the side chain was what Joe used in his design to drive the VTL5C2. I figured if it worked for him....
As for the TL072, I think you're right to question that and a 5532 would probably be a better choice. Thanks!

 

[L´Andratté]

Not a biggie, but you need a resistor to GND after C5...

 

[Wavelength]

I second the 6922 idea it has higher gain and transconductance. I designed a 6922 tube guitar preamp one time with a 9V input and a 75V DCDC converter (12V NMT1272SC, 5V NMT0572SC). I used some custom inductors at 100H/5ma (DCR = 1K) that I had made and it worked nice.
Just a word of warning with the 1M gain control those things really cut the top end on anything. Go with a 500K they all sound better.
Another thing you can do is vary the gain by putting the OPT-R portion across the plate resistor that way you don't effect the input signal but basically the gain stage. You can cap couple so the OPT-R doesn't see any current, though I don't think that matters.
You can also do a reactor follower on the output which kicks ass. Basically use a hammond 60H or 150H from cathode to ground and then cap couple to the 600:600 output. I do that a bunch and it works really well.

 

[OneRoomStudio]

Thanks to everyone for the feedback. I like the idea of the 6922/E88CC, and many of the other suggestions. Here's a new schematic. This time, B+ is around 180V - much lower than before. I kept some things the same, as the goal is still to produce a good sounding tube opto for as little money as possible, and things like large inductors, additional tubes, etc. add to the cost. What do you all think?

 

[volker]

What do you all think?

Don't skimp on a resistor and bias the output stage properly. Rg can be your potentiometer.

 

[OneRoomStudio]

Don't skimp on a resistor and bias the output stage properly. Rg can be your potentiometer.

Thank you, I had read that before, but wasn't sure how to implement the gain pot. Would this work? Do I need a series resistor below the gain pot to prevent it from going to "0?" Also, do I need the grid stopper? If so, is 1K appropriate?

 

[rock soderstrom]

Also, do I need the grid stopper? If so, is 1K appropriate?

I always use 300R as grid stopper for E88CC in a cathode follower circuits, which has always worked well so far. But 1k will also work, I don't see any disadvantages at this point.

 

[rock soderstrom]

Okay, here's an attempt to build a power supply for the above circuit.

The idea is to work with only one readily available standard transformer (e.g. 200V/6.3V) for tube devices.

A Delon bridge feeds a LM317 circuit for a regulated 6.3VDC heater voltage by voltage doubling, the doubled voltage is also used as input for the 12VDC part (for VU driver and lighting for example). The 12VDC also serve as input for a yet to be found DC/DC converter to supply the solid state part with +/-15V.

Does anyone have experience with the Meanwell MDD01M-15 module? I do not , I am not familiar with such DC/DC converters. What would work here (preferably in TTH technology), what do I need to consider?

Does the sidechain really need +/-15V or would +/-6V from a simple virtual ground circuit be enough? I don't think so...

I gladly welcome any input, criticism and suggestions for improvement!

 

[scott2000]

Does anyone have experience with the Meanwell MDD01M-15 module

Does that say 34mA?

Here's something interesting from Heikki's alternate psu for the vari mu....Think series string heaters
post 185
https://groupdiy.com/threads/transformerless-vari-mu-compressor-build-thread.78043/page-10
Come back Heikki!

 

[rock soderstrom]

Does that say 34mA?

Yes, that's a bit low.

Here's something interesting from Heikki's alternate psu for the vari mu....Think series string heaters
post 185

Thanks for this link! Even without the meanwell SNT RS 25 24, this DCWN06A-15 is dang hot!

Come back Heikki!

++1 What's up? Has he left us?

 

[ruffrecords]

Here is my current side cahin

Cheers

Ian

 

[rock soderstrom]

Here is my current side cahin

Thanks! You generate the bipolar voltage for the op amp with a buffered virtual ground. The +/-6v seem to be sufficient for the sidechain. Very interesting. That would make things easier...(and way cheaper)

 

[MidnightArrakis]

All very true. My goal with this project was to keep it as cheap as possible. An opamp and a vactrol is a lot cheaper than an additional tube and a transformer.

[An opamp and a vactrol is a lot cheaper than an additional tube and a transformer] -- But.....not as much fun!!!

/

 

[OneRoomStudio]

Thanks! You generate the bipolar voltage for the op amp with a buffered virtual ground. The +/-6v seem to be sufficient for the sidechain. Very interesting. That would make things easier...(and way cheaper)

If only E88CC's could be powered by 12.6V, this would be even easier....