Vactrols with a trick?

Speaking of the company without TOKO inductors. :wink:

I was reading about their opto compressors.

htttp://www.manleylabs.com/containerpages/langdualoplim99.html

(remove one t in htttp)

Quote:

" A similar principle was found in the older Teletronix®** & Urei LA-2A®**, LA-3, and LA-4 units which used a slower electro-luminescent panel-light on a conventional LDR encased in a light-tight enclosure. Ours uses modern VACTROL technology (but with a trick!) for exact sample-to-sample tracking and no high frequency loss while digging deep into limiting-- problems commonly associated with older LA-series limiters and other opto-based units. "

What exactly is their trick? Anybody reverse engineered their Vactrol control circuit?

JH.

What exactly is their trick?

Click to expand...

Good marketing?

and no high frequency loss while digging deep into limiting

Click to expand...

Probably just "stacked" time constants.
Or maybe even simpler. Two LDR´s in paralell. One fast, one slow.

Vactrols are useful devices, but frustrating parts to design for manufacture because of their large sample-sample variations of transfer characterisitics over current range and ambient temperature. The technique that I have used in many of my Vactrol designs is to control a d.c. path through the Vactrol that is part of a separate low-frequency (sub-audio) feedback loop. By biassing the Vactrol to a specified d.c. resistance, sample-sample matching at a.c. is improved by an order of magnitude or more.

Oops sorry double post

I think one time in an email she told me they use two , together
can't remember if she said back to back [ series or paralall ]

so it would [ or could ] work outside the range or have a greater range
sorry for being vague i just don't rememeber

regards Greg

By biassing the Vactrol to a specified d.c. resistance, sample-sample matching at a.c. is improved by an order of magnitude or more.

Click to expand...

Thats very interesting, i like to learn more about that.
Care to post a schematics?

LAWO did something similar using optocouplers that included 1 LED and 2 seperate (probably matched) LDRs; with one LDR controlling gain, the other being part of a feedback loop adjusting LED current. :shock:

Kit wrote:

By biassing the Vactrol to a specified d.c. resistance, sample-sample matching at a.c. is improved by an order of magnitude or more.

Click to expand...

Thats very interesting, i like to learn more about that.

Click to expand...

The principle is to have a very-low-frequency servo to set the standing current through the Vactrol. The audio is coupled in and out via capacitors, and the LF servo and the compression side-chain both contribute to the LED current.

Picture a string of resistors hanging between a positive reference voltage and a negative reference voltage. From the top, they read: fixed R, Vactrol R, fixed R, variable R, fixed R. The audio is put in via a capacitor at the top of the Vactrol R, and tapped off at the bottom of the Vactrol via another capacitor. The ratios of the variable Vactrol resistance and the fixed lower R perform the attenuation needed for compression. The junction of the lower Vactrol R and the variable R feeds the summing junction of a slow integrator. The current through the fixed and variable Rs hanging from the negative rail set the Vactrol current. The integrator output drives the anode of the Vactrol LED (via a current limit resistor). The cathode of the LED sits on the output of the side-chain op-amp, which must pull negative for compression. The audio needs drive and buffer amps, of course.

I could get about 40dB of gain change using this technique, with just one set-up adjustment needed to cater reasonably well for the wide range of Ohms per Amp exhibited amongst Vactrols.

I originally had a non-linear function in the side-chain to correct for the x/(1+x) effect of the attenuation curve, but it was a pain to do and I found it wasn't audible.

Later designs had two Vactrols with their LEDs in series. The second Vactrol resistor was used in a feedback circuit in the side chain to linearise the compression law. I still needed the LF integrator for the audio, though. More recently, I extended the design and built both two-channel (stereo) units with a single side-chain and eight-channel units with multiple side-chains.

There is no special trick, they use two vactrols with the photoresistors connected in parallel and the leds in antiparallel. They drive the LEDs with a simple audio amplifier, like in the LA2A, no rectification, no time constants. This way one LED lights on the positive alternance of the signal, and the other LED on the negative, probably this is the trick called "sample-to-sample tracking".
The time constants are the ones "built" into the vactrols. They use some slow VTL5c2 for this (500ms release)

chrissugar

hiya folks interesting stuff... heres couple of questions......
how good do these modern optos sound then ??
and compared to what ?
Chris, could a couple of LEDs be connected to my homemade LA2A in antiparallel instead of the EL panel ?
is 2 cells better than one ?
the T4B had two cells in parallel i been wondering about that for a while never really found a good explanation for that.

[quote author="chrissugar"]There is no special trick, they use two vactrols with the photoresistors connected in parallel and the leds in antiparallel. They drive the LEDs with a simple audio amplifier, like in the LA2A, no rectification, no time constants. This way one LED lights on the positive alternance of the signal, and the other LED on the negative, probably this is the trick called "sample-to-sample tracking".
The time constants are the ones "built" into the vactrols. They use some slow VTL5c2 for this (500ms release)

chrissugar[/quote]

Thanks Chris - that's the information I was looking for.

LEDs anti-parallel. So they don't need a rectifier, and the lower impedance of the two paralleled LDRs will slightly improove noise.

You sure about two 5c2's? No combination of two different types?
(Just to be sure.)
And the drive circuit: Is it a (current limited) voltage driver, a current driver, or just a voltage amp with series resistor? (Which value?)

JH.

No different vactrols, two VTL5C2, absolutelly sure. :grin:
Driver (sidechain) is a small amp based on LM386.

chrissugar

[quote author="gary o"]
how good do these modern optos sound then ??
[/quote]
different

[quote author="gary o"]
Chris, could a couple of LEDs be connected to my homemade LA2A in antiparallel instead of the EL panel ?
[/quote]
Probably with some small mods, but It would not be an LA2A anymore. The T4B is a significant part in the LA2A sound because it has a "memory", and it change it's behaviour depending on how hard you hit it. That effect you can't obtain with vactrols.

[quote author="gary o"]
the T4B had two cells in parallel i been wondering about that for a while never really found a good explanation for that.[/quote]
They are not parallel, one is for gain reduction (audio) and the other for metering.

chrissugar

Ah, thanks for the explanation, Boswell. :thumb:

About the LDR´s in the T4B, there were indeed two paralleled LDR´s in the signalpath:

http://www.waltzingbear.com/Schematics/Urei/LA_3A.htm

I cant be sure, but i think the reason is to get an "auto" release.
One LDR would hav a fast rise and fall time, and the other one would be slow. This way most of the "pumping" artifacts will be avoided.

I don't think so.

Never seen a T4 with more than two LDR's.

Here's a link to a pic of a (working) T4 in a LA3A:

http://img.photobucket.com/albums/v172/gyraf/T4b_05.jpg

I think that schemo is somekind of a hoax to prevent people from realizing the simplicity, like the "T+C"-part in the 1176 sidechain.

Jakob E.

Also look at this page: http://www.vacuumbrain.com/docs/t5.html
Some reversing done by our own CJ. Very good info. :thumb:

chrissugar

Never seen a T4 with more than two LDR's.

Click to expand...

Yeah, most of them seem to be built that way. But CJ seems to have found a 3-piece version.

From CJ´s site:

"Notice that the module with the light shields has three LDR cells. One of the big ones is probably for metering, the other two probably have different attack and release characteristics, which affects the compressors characteristics."

You just have to find something you like and stick with it.
You can get an infinite variation in compression characteristics just by changing cells.
So there is unfortunately no way to accurately control the T4 specifacations.
Every batch of cells is gonna be a little different.

I have a T4 RS model that will not compress much until you really hit it hard.
But this deep compression does not suffer the high end that the other units have.
I do not know why it works like this.
If someone wants to play around with it and figure it out, drop me a PM.

That darn T4 site is a nightmare to load.
I have to shrink that SOB.
Here is the earlier three cell model.

Thanks guys, Just wondered how good these optos sound, I like the sound of my LA2A, and in the hype about the newer optos it kinda puts the the LA down a bit, I done a few experiments with my LA using two cells in parallel ...didnt seem to make a lot of difference to my humble ears...I do like to experiment tho just to find different sounds.

Thanks .....Gary O.