Forssell Opto Compressor

[FredForssell]

Hey Guys...

My ISP has been down for a lot of today, but is currently working. I don't know what happened with them. Sorry for any inconvience this may have caused.

RE the Silonex... someone should make sure that these parts can handle large voltage swings with good linearity. I know that the VTL5C1 can handle 10-15 Vrms in this topology without problems, and I think you'll want to make sure that any sub parts can do the same.

I'll try to post the slope curves later on.

FF

 

[Guest]

was originally thinking about something api325-esque, but many components are too obscure (47k rev log pots!!)

http://www.google.com/search?hl=en&lr=&ie=UTF-8&safe=off&c2coff=1&q=50k+reverse+audio+taper&btnG=Search

50k reverse audio pots are common in some tube guitar amps. I suppose that is close enough.

Cheers,

elco

 

[sismofyt]

Standard pots are often +/-20%, so a 50kOhm instead of 47kOhm woulnd't make a difference :idea:

 

[sismofyt]

That one really sounds nice, but only works with a lowimpedance input. Checkout the 10dB line amp at ESP too :wink:

 

[sismofyt]

Depends on the Vactrol used, but I'd say a 100kOhm would do nicely. Fred?

 

[FredForssell]

Hey everyone,

The value of the series R working against the minimum on resistance of the LDR determines the maximum amount of gain reduction. I use VTL5C1 optos because a) they have 500 ohm on resistance, and b) because they have as fast an attack time as any opto I have seen (rate at about 1.5 ms).

The value of the series R that I use is 10k which gives a max gain reduction of 20 or about -26 dB. When the cell is completely turned off, I do not want the input impedance of the gain make-up/buffer amp to cause a signal drop across the series R, so I want the input impedance of the buffer amp to be at least 10x the value of the series R... in my example that means 100k. If you can live with some drop across the series R you can make the buffer input impedance smaller.

All of this brings up another point in choosing an opamp for your gain make-up/buffer amp. When the cell is off, the source impedance (seen from the input of the buffer looking back into the series R) is at least the value of the series R, or in this example at least 10k. So I'd recommend using a opamp with a JFET input stage, like our (gee, surprise) JFET-992/993 opamps, or any good low noise IC opamp with a JFET input stage.

Finally, here is a link to a PDF on my website showing an example of the slope curves for the circuit. I think this one used a 5k series R instead of the 10k series R, but you'll get the idea.

http://forsselltech.com/slope1.pdf

Remember, someone should check the lay-out of the PCB gerbers against the schematic. I haven't done that, nor have I made this circuit board (I don't do much in the way of through-hole designs any more), so I can't vouch for it.

 

[Bauman]

Another interesting option would be the hamptone jfet gain block...

 

[sismofyt]

Or a Gyraf SRPP :razz:

 

[tk@halmi]

For a simple solution one could use a 100k log pot between Pin 1 of J2 and ground with the wiper connected to the + input of an OPA604. Then fix the gain from 20 to 30dB or whatever you need on the opamp.

Tamas

 

[StephenGiles]

This is going to be a great compressor, great work being done by all here.

 

[FredForssell]

Hi everyone,

Just in case anyone is intereseted, here is a link to a PDF file showing the circuit that I use for the LED Gain Reduction metering.

http://forsselltech.com/OptoMeter.pdf

VR1 can be a fixed 1k, but typically I use a trim pot so that I have the option to tweak the meter response for a given opto cell. This happens less than 10% of the time. However, I hate having to grade and select LED/LDR cells and the trim pot in the meter circuit and on the main circuit (opto balance) make it very easy to use ungraded/unmatched optos.

Again, someone REALLY should double check the pcb layout for accuracy before making these boards.

Also on the buffer schematic that someone posted, I wouldn't use the blocking caps myself, but if you do, you do not need such a large input cap. It's working into 100k and can be much smaller and probably a good quality film cap could be used.

 

[bluebird]

2364 views of this thread !!

mc donalds is going to be calling us to advertise in this thread !!!

yea, so if you don't come through with this you will be hung by your toe nails at high noon. pictures will be posted at the brewery. :shock:

 

[mnats]

I've bashed out a cheap and dirty single sided PCB for Fred's metering circuit OptoMeter.pdf:

Checkplot PDF (12kB)

Bottom Layer PDF (9kB)

...in anticipation of Fred's double sided ground plane beautiful version.

Oh yeah, it uses T1 (3mm) LEDs rather than the 10 pack Fred's schematic shows.

 

[Eric H]

[quote author="kubi"][quote author="buttachunk"]builders; the version with large electrolytic caps will sound darker-- the version with only the .1U polyprop caps will sound cleaner.[/quote]

See both versions here.



kubi[/quote]
What you want is the small cap at the input, and the large cap at the output (unless you want a high bass-roll-off) all audio opamps have very low output impedance.

-Eric

 

[mnats]

[quote author="kubi"]
mnats, do you want me to add your latest files, too?[/quote]

Hmmm...maybe PMs aren't working? Please - go ahead and post my files to your Fred Forssell page.

 

[rafafredd]

So, there is a complete project now, with metering, all PCB´s, etc... And that was fast!

Who will be the firtst to put it together?

And for a all JFET signal path we could use a Forssell DLR1 balanced line receiver, what do you think? I was not able to get pricing info for the DLR1, but should not be very expensive.

Maybe Fredd will come and clear it out...

 

[mnats]

AFAIK there's no PCB for the makeup gain section...yet! I'm waiting for someone to breadboard it (not me - I'm the single sided PCB guy) before doing a layout.

I've discussed it with buttachunk and the idea is to make a board that will accept either packaged discrete and common IC opamps on the same board like a JLM Audio board.

I see this on Fred's site:

The DLR-1 is available factory direct from Forssell Technologies and is priced as follows?.
1-24 each $39.95
25-50 each $36.95
Over 50 $32.95

 

[soundguy]

could I suggest trying to layout a transformer balanced in/out version with a discrete audio path? I'd love to build this but dont really have a big need for another limiter that has chips in the audio path...

Is it just me, or would more peeps be interested in this if it was fully discrete?

dave

 

[CJ]

discrete!
who wants there signal going thru 10 zillion miniature transistors?
well, sometimes I do, but I digress.
where's that SSL board, cj has finally gone solid state!
where are those 5534's? 072's?
don't tell my friends!
:guinness:

 

[Svart]

for FET opamps are we talking about a tl082/92 or something like the opa604/opa134?