Official Forssell Limiter Build thread

[strangeandbouncy]

Hi Guys,


  Peter was not only kind enough to lend me one of his LA-Z pre-amps for appraisal, but also one of his Forsell Optical comps as well! So sorry to wax lyrical yet again, but I just can't help myself! I can only call it as i find it.

 Now you must all know what it is and what it does by now. I am happy to report that it does EXACTLY what it says on the tin! If you still don't own a 500 series rack of some kind, this could well be the final temptation to tip you over the edge. It seems that pretty soon you will be able to build your own 500 series rack. Fred's fabulous compressor sounds absolutely amazing with Peter's DOA, and IMHO benefits from the output iron. I am sure that Fred would disagree about the iron, but hey, each to his/her own. I did try swapping out the DOA's for 2520's. Certainly it made it sound a tad more API-ish, but I really really like the sound with Peter's DOA. Just as with the LA-Z Pre and the LA-Z EQ, the sound is full, fat, yet sweet and SOOO clear. The definition top, mid and bottom is nothing short of extraordinary. I have been recording with the pre and the Forsell comp for a couple of months now, and the whole experience has been so unfatiguing. There is nothing meek or mild about these, they are very bright and clear, just smooth and sweet. Not like anything else, really! So much dynamic control with so little artifact. Smooth and transparent . . .

 I could also use one of these comps on each channel in the mix quite easily! Drum separates, drumbus, bass, guitars, and especially vocals, just everything I throw at this sounds better in the mix, either instead of my usual suspects or, even better, WITH my usual suspects. SO  flexible, you can make it pump like crazy, but with a kind of transparency I haven't yet found before, or gently take out the peaks.

   Fred Forsell has designed an absolute classic here if you ask me, and I don't understand quite why they aren't in every darn studio in large numbers. Peter's module is a fabulous version thereof, . . . . . - and you can have 'em too, you lucky so-and-so's!


  If anyone would like to ask me about actually using this compressor both for recording or in the mix, I'd be only too happy to oblige.


  Kindest regards,



      ANdyP

 

[Svart]

I still need to post my hacking of the forssell opto comp.  I used the FETblokes but I redid the whole input and output stages around them. 

 

[khstudio]

Does Peters unit use FET based DOA's for the "Make-up" amp?

It's been a while but I remember Fred mentioning the benefit of using a FET because of the high impedance of the LDR when it's full on or off.... I forget which?

I've been meaning to try the FETblokes in mine too... would love to see your setup for that Svart.

 

[peter purpose]

Does Peters unit use FET based DOA's for the "Make-up" amp?

Peter's does not, but that is not to say that you can't.
I supply two op amp boards with the kit, but use what you want.

peter

 

[deva]

Hello everyone,

I have just built a dual unit, and it works quite good, except for one thing.
I get an audible pop (especially when using make-up gain) when I move either of the attack controls from the fastest setting to every other position and vice versa, with or without signal applied to the input jacks, in both compress and bypass mode.

I took a few readings and found out this:
with both attack controls on fastest setting, my supply rails read +19/-19 V
with one (doesn't matter which) attack pot in a different position, my supply rails read +19/-15
if both attack pots are in a position different from the fastest setting, the negative regulator shuts down completely and stays that way until I set fast attack on both pots or I cycle the power switch (but I need to set fast attack on at least one pot anyway).

One of the two units was tested a while ago with a +12/-12 lab supply and had no problems. I'm quite sure of this because the voltage drop is noticeable on the analog vu meters, they fall out of zero when there is no signal.

The only differences from the original schematics are that I'm using NE5532 dual opamps and LM317 / 337 for regulation.

If anyone has ideas about what to check it would be extremely useful.
Thanks!

 

[briomusic]

I don't really have a logical explanation for what's happening, but seeing that your machine was happy at +/-12V my first step would be to try reducing the operating voltage, which is easily done with the LM317/337. (You only need to change one resistor, maybe you already have a trimmer installed in it's place?)

What is the voltage you are feeding into the voltage regs?

 

[deva]

What is the voltage you are feeding into the voltage regs?

There are around +/-30V going into the regs. (Yes, wrong transformer choice)

 

[briomusic]

Even though it is not necessarily the root of the problem, it sounds like a good idea to start by sorting your psu out. you want to be able to have a regulated dual supply of 15-18V. Regulators like at least 3 V 'reserve' if I am not mistaken, so make that about 21V rectified. 21V/1.414=14.85V so a transformer with a secondary voltage of 15-18 Volt seems appropriate. The more the voltage reg has to burn the better your heatsinking has to be. (also dependent on current).

Next: tweaking your voltage reg. There are two resistors around the LM317, one between Vout and ADJ which we shall call R1. The other one, from Vout to Ground, we'll call R2. Take the value of R1 you are using and enter it into the following calculator:
http://www.elektronika.opatnet.cz/storage/www/lm317/
Now try different values for R2 to achieve a comfortable voltage for the FSC (15-18 is my guess). you can even replace R2 with a trimmer (or resistor+trimmer in series) to be able to change the voltage with the circuit up and running. you can switch the website to do the same calculation for the LM337.

Be aware that an unloaded PSU will have a higher voltage output than a loaded one, so either connect your FSC or a dummy load (3k3/2W resistors should do) while testing.

 

[deva]

Even though it is not necessarily the root of the problem, it sounds like a good idea to start by sorting your psu out. you want to be able to have a regulated dual supply of 15-18V. Regulators like at least 3 V 'reserve' if I am not mistaken, so make that about 21V rectified. 21V/1.414=14.85V so a transformer with a secondary voltage of 15-18 Volt seems appropriate. The more the voltage reg has to burn the better your heatsinking has to be. (also dependent on current).

Next: tweaking your voltage reg. There are two resistors around the LM317, one between Vout and ADJ which we shall call R1. The other one, from Vout to Ground, we'll call R2. Take the value of R1 you are using and enter it into the following calculator:
http://www.elektronika.opatnet.cz/storage/www/lm317/
Now try different values for R2 to achieve a comfortable voltage for the FSC (15-18 is my guess). you can even replace R2 with a trimmer (or resistor+trimmer in series) to be able to change the voltage with the circuit up and running. you can switch the website to do the same calculation for the LM337.

Be aware that an unloaded PSU will have a higher voltage output than a loaded one, so either connect your FSC or a dummy load (3k3/2W resistors should do) while testing.

My 5532 datasheet states +/-22V as maximum rating, so I should be safe running them at +/-19V.

I am using a dual 18+18V transformer - I said I chose a wrong transformer because it is specified for 18V at full load (~400mA). Since I'm drawing much less than that, the trafo is putting out ~22VAC each secondary.

Anyway I did a few tests. First, both units work without problems with the +/-12V lab supply.
If I connect only one unit to my psu and move the attack pot, the negative regulator shuts down.
I tried lowering the output voltage of the regulators down to +/-15 and +/-12V, with no effect. LM337 keeps shutting down.
I also tried using a variac to bring down the incoming voltage from 30VDC to lower values - I adjusted it to be more or less 4V higher than the regulated output, this too with no effect.
I measured the current draw on both rails when using the lab supply. My multimeter says that a single FSC draws around 35mA each rail with no signal, and that value doesn't really change much when moving pots or when signal is present. I don't know, but I think it's a reasonable value.

Well, I'm thinking about a faulty 337 or something like that. During my tests with the variac the reg was cold to the touch, and it has to supply only ~70mA, so it shouldn't go neither in termal shutdown nor in overcurrent protection.
Tomorrow I'll try replacing that and I'll see what happens.

 

[SIXTYNINER]

bump

 

[jensenmann]

Hi Gents
I´m trying to find out how to calibrate the FSC500 but it gives me serious headache. The output DOA was oscillating when gain was low (pot lower than 9:00). To resolve this I soldered a 3k3 resistor across the gainpot and replaced R61 with a trimmer to have the possibility for adjusting unity gain with a centered gain-pot.
Oscillation gone, so far so good. But how is this opto matching to be done and how should R28V be adjusted for maximum slope. Maybe it´s a language problem, I don´t get what is meant with slope.
Interestingly the GR-meter shows GR, but the signalpath doesn´t perform compression. Lowering threshold increases output volume but no compression. Turning the ratio pot affects output volume, too. Same thing happens with both modules I built. Is this a calibration problem or did I built the same fault in both channels?

 

[MikeClev]

I remember being puzzled by that 'select for maximum slope' resistor, I didn't know what order of magnitude I should be looking for, 1ohm, 1k, 100k? Well one day I found a post where someone mentions that 100 ohms worked for them. I think 'Select for max slope' means that it determines the maximum compression ratio the unit can give you. It sets the current passing through the Vactrol LED and thus the amount of 'gain reduction per volt' (An oversimplification, but you get the idea). 

I would try 100 ohms and see where it gets you. I built one on vero board with TL072s that works well with a 100 ohm resistor.

If lowering threshold is raising volume, is there a chance it is wired in reverse? Or if it is PCB mounted... is it meant to be a rev log pot?

Does the GR meter bounce in time to the rhythm of the music, or is it steady? If its steady, the compression may be working but just with a very slow release time (I had the opposite problem, too short release, when I mistakenly put a 49ohm resistor where the 49k release time resistor should be).

PS I love this circuit and plan to build a bunch of them production line style as soon as get my other projects out of the way!

 

[jensenmann]

Thanks, Mike
After actually having listened to it (I only watched sinetones on the scope before) I found out that everything works more or less. I simply misinterpreted the meters (GR moves downward for rising GR - I thought the other way around) and threshold labeling.... note to self: don´t trust your measurments
Calibration then was easy with turning the knobs in the right direction.
Further oscillations disappeared whith proper grounding of the scope and distortions finally disappeard when I disconnected the scope while listening to audio. Man, I feel like a noob.
One never stops learning. 

 

[sonicwarrior]

Today I finally started stuffed my tommytones upside down boards (Rev 1.c) and compiled the important postings from this thread regarding the full DOA version into a PDF file (attached).

Now my head is burning but I noted the following changes to the parts list compiled by Gegory W. Stein:

Omit: C32, C33 [tommytones #41]
R28 => 100r [tommytones #73]
Make up gain pot: R6 => 10K Log [DrFrankencopter #215 + khstudio #315]
R22 => 10K, probably something between 604r and 10K [cstella #237]
R24/R25 => set higher [clintrubber #272]  10K???
Threshold => 100K Log [khstudio #308], eventually wired wrong [khstudio #315]
R11 => 3.3K [khstudio #315/#318]
R6 => 10K [khstudio #315/#318] pot

The # numbers are the numbers of the postings from which I got the information. Don't know from where I got the information that R24/R25 should be 10K instead of 1K.

Something else is not clear to me:
What do I do with the 'J' part? Is that just a link?

Did anyone who built the discrete (DOA) version so far drill the holes for the pin receptables to make 'em bigger or are there pin receptables that fit in there?

 

[sonicwarrior]

And my combined schematic pic for the DOA version.

A shame this project seems to gain no interest.

 

[detonator]

A shame this project seems to gain no interest.

Agreed, but I'm interested.

 

[jensenmann]

After having used these for a year now I have to admit that I like them. I tried to sell my pair some time ago but I´m glad that I found no buyers. In the meantime they are back to use and turn out to be something very usefull. They don´t win shootouts on single sources but work very good for leveling out dense material in the mix.

 

[sonicwarrior]

@jensenmann:
Which variation did you build? Did you use modifications or the straightforward values of the 'documentation'.

 

[jensenmann]

It´s the Peter Purpose 500 version without modifications IIRC

 

[sonicwarrior]

Here is a picture of how I fit the DOA female pins (Mouser 575-034420) into my Forssell PCB by using the male pins (Mouser 575-310320) as an extension:

I used a DOA for the alignment while soldering the pins to the PCB.