Feeler: EZ Tube Lunch Box

G-Sun said:

Now, the questions remains:
- If I'm in, will you develop the EZComp further?
and
- What directions would you/we go for?

Personally the combination of fast with slow would maybe the most optimal.
Could there be a two stage comp?
Some (in the territory of) 2-4:1 fast and 2-20 slow (if knee), with two controls for threshold, and one for makeup?

Click to expand...

I have been meaning to do this for some time so a spur is no bad thing.

One thing you need to realise about simple opto compressors is that the compressor characteristics are determined almost entirely by the opto itself. So the opto defines the attack and decay (which are program dependent) and to a very large extent the ratio. As in the LA2A, the best you can do is alter the threshold and gain make plus a switch to alter the ratio a bit. If you want to go much further than that you are into the realms of the longer development.

If you are OK with that then I am prepared to take it further.

Cheers

Ian

ruffrecords said:

I have been meaning to do this for some time so a spur is no bad thing.

One thing you need to realise about simple opto compressors is that the compressor characteristics are determined almost entirely by the opto itself. So the opto defines the attack and decay (which are program dependent) and to a very large extent the ratio. As in the LA2A, the best you can do is alter the threshold and gain make plus a switch to alter the ratio a bit. If you want to go much further than that you are into the realms of the longer development.

If you are OK with that then I am prepared to take it further.

Click to expand...

Thanks!

That would put the final result in DAOC/ JLM LA500 -land I believe.
Not a bad thing.

However, as I understand the usability of the DAOC it's a little unflexibel.
Then, of course you're right about the longer development for further features.
If I believed the vision for the comp is good enough, I'm ok with doing rack, psu, pre and eq first,
then doing the comp when it's ready.

For more flexibility, yet simple: Is using 2/4/8 vactrols with different characteristics for a two stage compressor  a good idea?
Another idea is a FET-stage then a opto-stage, but I'm unsure if a fetdesign works within the current concept.

Another way to go is along the lines of pm660/670, but that's a different style comp, right?

What do you think regarding one or two module width?
and mono vs. stereo?

When thinking stereo, it's tempting to aim for something usabel on the masterbus.
For me that would mean slow opto with HPF I guess
BoB Katz on the subject: https://www.gearslutz.com/board/1670718-post2.html
(kind of disagree with my statment above )

I would like to elaborate on these ideas further,
to see what could be achieved within a reasonable time, labour and cost-frame.
I'll put my dual PQ1549 eq on hold

Or,
is it posibel to have 3 vactrols with different characteristics, the a switch fast/med/slow?

http://www.banzaimusic.com/Optocouplers/

G-Sun said:

Or,
is it posibel to have 3 vactrols with different characteristics, the a switch fast/med/slow?

http://www.banzaimusic.com/Optocouplers/

Click to expand...

Yes and no. Electrically it is possible to switch between three different Vactrols. Trouble is, not only does their attack/decay change but also their resistance with current and their slope. So you would get three completely different characteristics and accurate gain reduction metering would be a nightmare .

Haing said that, I was toying with the idea of having a simple gain reduction meter that monitor the current in the opto diode - more current equals more compression. The scale would be arbitrary - I have seem 1mA gain reduction meters on some old tube compressors so maybe something like  this would be acceptable? We could use a VU meter and set it so that 0VU meant 20dB of gain reduction for example.

Changing from comp to lim on each one also presents another switching problem. However, on the opto drive side, they all get fed the same amount of current because the LEDS are the same. The comp/lim slope depends on where on the current versus resistance graph you operate the opto. So another idea I had was that instead of changing the signal pot divider (on each one if there were three) to go from comp to lim, we could simply divert some LED current into a resistor in parallel with the LED to switch from lim to comp. The one resistor and switch would would for all three optos.

I'll give it some thought but let me know what you think of the gain reduction metering idea.

Cheers

Ian

Nice

Ok. Vu metering is nothing I know to much about. As long as we get an accutate picture of what's going on, I guess we'll be ok. If it was an IC option we could just gain it up and attach a trim-pot, right? But in this design there's maybe not such an easy option.
The meter should preferable show in gain as well (in/gr-switch).

Vactrols: I like your ideas.
What would happen if we put 2 vactrols in parallel?
(Fender style pick-up switch).

Getting some ratio and time optios would be highly beneficial.

G-Sun said:

Nice

Ok. Vu metering is nothing I know to much about. As long as we get an accurate picture of what's going on, I guess we'll be ok. If it was an IC option we could just gain it up and attach a trim-pot, right? But in this design there's maybe not such an easy option.

Click to expand...

Pretty much the same is doable with the tube version. Full gain reduction needs about 20V  rms applied to the gain reduction circuit. We can measure this voltage using a standard VU meter with a series pot that can be adjusted so 20V rms reads 0VU. Whether this will provide an accurate picture of what is going on is difficult to say but I expect this is the same for classic opto compressor too.

The meter should preferable show in gain as well (in/gr-switch).

Click to expand...

Yes, that should be possible. I will have to show the level at the unbalanced output since the output transformer is external but we just calibrate it for the internal level.

Vactrols: I like your ideas.
What would happen if we put 2 vactrols in parallel?
(Fender style pick-up switch).

Click to expand...

We could possibly do this instead of switching between a couple or three - so we would have a, b or a+b in true Fender fashion. The main thing that would happen would be the threshold would change and the timing would be some complex combination of the two - could be interesting.

Getting some ratio and time options would be highly beneficial.

Click to expand...

Timing options are no problem - just change optos. Unfortunately they all produce more or less the same ratios. They tend to have soft knees and the ratio is fairly constant at about 3:1.

Cheers

Ian

ruffrecords said:

Pretty much the same is doable with the tube version. Full gain reduction needs about 20V  rms applied to the gain reduction circuit. We can measure this voltage using a standard VU meter with a series pot that can be adjusted so 20V rms reads 0VU. Whether this will provide an accurate picture of what is going on is difficult to say but I expect this is the same for classic opto compressor too.

Click to expand...

Ok, you'll have to see what's best here I guess.

Yes, that should be possible. I will have to show the level at the unbalanced output since the output transformer is external but we just calibrate it for the internal level.

Click to expand...

cool

We could possibly do this instead of switching between a couple or three - so we would have a, b or a+b in true Fender fashion. The main thing that would happen would be the threshold would change and the timing would be some complex combination of the two - could be interesting.

Timing options are no problem - just change optos. Unfortunately they all produce more or less the same ratios. They tend to have soft knees and the ratio is fairly constant at about 3:1.

Click to expand...

Ratio: What would happen if we put those vactrols in series?
You mentioned comp/lim resistor switch. How do you plan to achieve limiting with 3:1 vactrols? (isn't the resistor gonna lower the ratio?

2 vactrols: a; b; a+b parallel, a+b series
3 vactrols..
Seriously: It all depend on how complicated/ costly/meaningful it is. ( Is this how the Fairchild 660 does it's 6 time constants -3tubes? )

How do you think of the input gain/threshold?
One tube-amp-stage before the vactrols?

G-Sun said:

How do you think of the input gain/threshold?
One tube-amp-stage before the vactrols?

Click to expand...

Input gain is neither necessary nor desirable. To keep distortion to a minimum it is important not to put too much signal on the Vactrol resistive element. So we put the gain make up control pot right after the input transformer then feed it to the Vactrol and the Vactrol feeds the gain make up amp. The input transformer also feeds the threshold control which goes straight to the side chain amplifier.

Regarding your other comments about options, including FETs, Fairchild features, double width modules, stereo versions etc; to keep this simple single  compressor development within reasonable bounds it will be based on the twin line amp card plus  an add-on PCB containing the compressor specific bits - optos, pots and switches. One half of the TLA will be used for gain make up and the other as the opto driving side chain. We can include wrinkles within that scope, like switching optos and trying different ways to alter the ratio. Anything other than that is really in the realms of the much longer full blown development. So for now think LA2A using a Vactrol with a few tweaks.

Edit: It will probably take at least 6 months to develop as it is given everything else I have to do so.

Cheers

Ian

Yes, I understand development needs to be taken in steps.
So, first step is maybe a mono one-vactrol design?
To see if everything works fine?
Then maybe a dual vactrol with switching and possible ratios?

Myself, I feel that just one vactrol, one time-constant, one ratio-slope,
is maybe not what I'm looking for (Yes, I were positive to the la2a, but that's a proven winner).
But if we agree that a more elaborate version is the goal within let's say 12months,
then that's perfectly fine with me.

And I'll maybe need a clearer vision regarding stereo-link and hpf?
I'm fine with a good mono EZ-strip -maybe a lot more than fine
Yet, if the TLA is to be stereo, then I might need to plan for that (6modules space and external psu)

It's not like you have to promise anything. Who knows what can happen in the process.
But a clarified vision can be very helpful.

Thanks a lot!

OK, we will take sensible steps. First thing to do is pick a Vactrol and get the basic  topology, including the metering, working. Then we can experiment to see what ratios can be had and try alternative Vactrols. Once we understand what does what and works and what doesn't we can proceed to the next step.

BTW, the HPF is very simple to implement. We can include that if you like in the first round.

Cheers

Ian

OK, my build will be something like this:

6 module space EZ Tube Lunsh Box
- Preamp
- Pulteq 3 band
- Comp
- External psu

Preamp as separate module with transformer in/out.
Eq + comp together, transformer in pre eq, transformer out post comp.
Compressor planned for stereo-link.
Optional pre to eq link switch.

This means fitting 8 x xlr, 8 x transformers and pwr in on the back I guess.

Stage 1
1 preamp, rack and psu (ext)

Stage 2
1 Pulteq 3 band, 1 TLA Comp

Stage 3 (optional)
Second channel of preamp, eq and comp

If there will be no comp stereo-link, then I need to know and plan for 4 module box with internal psu, pre, eq, comp and one spare slot.

Note on workflow:
I work much alone i studio. And when tracking I have a lot to focus on (what to play/sing, how..). Thus I'll set preamp-level and track a rehearsal-take. Then I'll loop the track back for eq and comp-adjustments. Then patch it in for the real deal. 
My signal will typical go like this EZpreamp > Harrison Ford HiPass > EZ pulteq and EZ TLA > AD
(at least, this is what I think today )

ruffrecords said:

BTW, the HPF is very simple to implement. We can include that if you like in the first round.

Click to expand...

I believe HPF is mostly needed for programmeterial and stereo.
So, it's maybe not so important for first development.

Houston we have a problem.

I am getting a strange 1MHz oscillation with cards used in the 3U lunch box with the new built-in PSU. A bit reminiscent of the problem we had with the 4 into 2 mixer and that Holger also reported, but the normal fix of 100nF at the HT connector of each card does not work. I thought it might be the heater regulator oscillation due to long leads from the rectifier but I went back to the original scheme with the on board rectifier and it actually got worse. At first I thought it was just a problem with the new 'classic' card but I built a regular Eurochannel card and that does it too.

Tomorrow I will try each card out in the 6U test rack with an external supply that I know does work then I'll try an external supply n the 3U rack.

Everything else on hold until I sort this out.

Cheers

Ian

OK, we will take sensible steps. First thing to do is pick a Vactrol and get the basic  topology, including the metering, working. Then we can experiment to see what ratios can be had and try alternative Vactrols. Once we understand what does what and works and what doesn't we can proceed to the next step.

Click to expand...

Ok, i need to catch up here with all the compressor related posts over the last days!
So the plan is now bulding an opto like the AOC again?
Or are we talking about a base design which then get all the actually non existing features like Attack, Release, Ratio and so on?

I thought, anything else would be only a "copy" of actual really good working units and not so exiting.

Cheers, Sven

Sven,

We are basically talking a Vactrol based opto comp/lim with some tweaks. To cut development effort, it will be based on the twin line amp board plus a small  add-on PCB for the compressor functions. If I can, I want to include the following:

1. Full wave rectification of control signal so a single opto can respond to both halves of the signal waveform.
2. VU based monitoring of LED current to indicate compression as this avoids the need for a dual resistor opto and makes it straight forward to read output level as well.
3. Ability to alter ratio by diverting LED current or by using the LA2A method
4. HPF in the side chain
5. Ability to switch between two different opto or the two in parallel

Hopefully this will make a useful compressor  in its own right but also answer basic questions that will help a later full blown comp development with attack/decay etc etc.

Cheers

Ian

Looking at some documnetation I see the PMT Preamp is designed for stepped pot.
That means I could drop the in/out-transformers, saving 2 transformer per channel,
and still have gain-structure being intact for "offline" tweaking, utilising the line in.

BTW: I belive the JLM LA500 uses 2 vactrols, has 3,5,10:1 ratio and a slow attack mod (fast and slow attack, fixed release)

G-Sun said:

BTW: I belive the JLM LA500 uses 2 vactrols, has 3,5,10:1 ratio and a slow attack mod (fast and slow attack, fixed release)

Click to expand...

I am sure it does. There are well documented op amp based side chain circuits for driving Vactrols and changing ratios, none of which I plan to use!!

Cheers

Ian

BTW: I belive the JLM LA500 uses 2 vactrols, has 3,5,10:1 ratio and a slow attack mod (fast and slow attack, fixed release)

Click to expand...

Correct. I built two of them in the past. An amazing 500 module! Joe is a really genius in building such small and extremly good sounding devices. Not to mention his extreme component density of the pcbs he is developing.

Cheers, Sven