New GSSL Layout Proposal - Anyone Interested?

[ruckus328]

I want to stay away from the "thrust" circuitry as that is patented

But nothing will prevent you adding a "tilt" filter to the sidechain (shelved low cut and high boost together) and calling it "tilt" instead of "thrust".

Call it what you wish, it's still API's patent, is it not?  Maybe I'm wrong here with the legality of things though.  And I'll just be honest, as I'm not one to beat around the bush - I have no information on that particular circuit.  All information has been stripped from existance.  All schematic links are dead.  And as I didn't incorporate it into my GSSl build, I don't have any physical boards to look at.  The filter circuit I was planning on incorporating is very simple, small, and cheap to implement, should cost under 10USD for the associated parts.

 

[Kingston]

Call it what you wish, it's still API's patent, is it not?

No. It is impossible to patent a simple circuit like that. Whatever that specific API patent says is actually irrelevant to this topic.

We have a shelving (or wide Q bell) low cut filter, and a shelving (or wide Q bell, it's not awfully significant to the outcome) high boost. You are free to use those with any parameters, any setting (be it compressor, guitar distortion whatever), and in every way you can ever think of. No one will stop you.

 

[Igor]

Hi to all, after some thinking, I decided to help here as I can.
Well, one thing really hurted me-most things appeared here within short time after
it appeared at my 1RU thread. It made me sick evil for some hours,
well, things go by the way they go, and positive input always better than critics.

Anyway, I think, the man who deserves GIGATONES of respect with GSSL is Livingnote.
This is the best looking build I ever saw for this compressor, with PCB's drawn in Illustrator!!!
Really, put my hat off, Livingnote!

Well, some thoughts.

/everybody feel free to use

As there are at least 5 cases available for GSSL, all they have different
front panel holes and distances between xlr's. Many people use their own front designs.
One common thing is distances between ratio/attack/release switches on front panel.
This should be taken into account when making front panel PCB. It should be universal.
Same about pcb-mounted XLR's. I.e. there should be provision for simple wiring
for different cases, not affecting the audio.

Metering.
Most of people like analog appearance. But, as some like to see input/output,
it is good to have led bargraph meter at separate board.
Like:
http://www.groupdiy.com/index.php?topic=23977
At my old 1RU build, two LM3914's and some simple logic controls 2x 20 bicolor led's.
Feel free to use or modify the schematic. IIRC it is corrected version.
Gain reduction is red dot, output is green bar, or vise versa.
With some attention paid to ground layout, the whole comp is DEAD quiet.
It is not necessary to use bi-color led's.
Same can be done with 10-led DIP20 bargraphs, the dot can be "more highlighted" segment.
To make easy metering of audio, and save some space on PCB, I used THAT4301.
Detector, VCA, and some opams at same package-makes life easier.
Another goal is detector with volt/db output to use same v/db log scale for
gain reduction and output (LM3914 is linear-scale led bargraph).
In/out/GR switching can be done with simple 40XX logic or simple switch.

Paralleling the VCA's. One 2181C for colored compression, a bit misadjusted,
good for drum group. Four 2181A's can give very clean (0.005% THD or even better)
unit with signal floor down -100dbV unweighted.

Time constants. I very liked the sound when release was right after diode,
than attack, than to time constant capacitor.
But this makes absolutely different compressor.

Another thing, is  keep same threshold with GSSL or Oxford mode.
IMHO, mono should be -3db. This gives chance to both modes without changing threshold to
get same amount of compression in real life.

 

[ioaudio]

Hi to all, after some thinking, I decided to help here as I can.

good move.

-max

 

[tv]

Perhaps you guys could decide on "concepts"?

I mean theres three of you doing "SUPER-GSSL" layouts. IMO there is different "ground" to be covered, perhaps you each could cover one "ground" and sort-of "specialize" your designs for certain target uses? And enclosures.

 

[nickeveslage]

Or (and I can't imagine an efficient way to do this) work together and make one Super-Super-GSSL project....

At any rate...

I think the idea of everybody using a "standard" distance between the controls has been brought up many times and is a great one.  In a perfect world we could develop our own standards for a lot of things, not just a few (awesome) projects like the x51 racks.

Just imagine....costs would decrease because there could be group buys for just about everything.

But, like I said, that doesn't seem feasible. Plus, it's probably already been brought up.

The whole parallel VCA thing sounds interesting to me, and the idea of lowering the noise and everything sounds great, I just don't think that it would be noticeable enough, if at all (for me), to justify the extra dough for the extra VCAs.

I'll have to do some research into what the whole Thr@st thing is all about.

Like I said though, not fe

 

[Biasrocks]

Hi to all, after some thinking, I decided to help here as I can.

Nice move Igor.  8)

Always better to bring our resources together.

Mark

 

[Igor]

BTW, regarding the thr@st filters.
I saw elsewhere on net schem' for opto compressor with similar idea,
and it was some years before A*P*I came with it.
http://dt.prohosting.com/hacks/what1.gif
http://1176neve.tripod.com/id5.html

We can call it What filters!
Let's give a credit to What Compressor.

A bit different values of R/C's can be used...
it is all matter of taste, right?

 

[mitsos]

This thread has a much better vibe today. Guess I should have stayed away 2 days

I agree with Kingston about the "thrust"... The USPTO is full of crap giving a patent for something like that.  5170437 - hit google patents. Of course, I'm not saying copy their idea directly, and I prefer the word TILT.  I think this is the only version of the HPF that is somewhat appealing from an experimentation POV.  Maybe leave room for user experimentation with the curve (enough room for sockets for the filter components maybe)?

And thanks to Igor for seeing this for what it really is... DIY.

@Igor, not that it matters much, but I think that might be incorrect about the What comp... Seems the API patent goes back to 1990-92 (I don't know when they acquired it, who is Audio Teknology Inc?) The What is from 2000 or so?

 

[[silent:arts]]

@Igor, not that it matters much, but I think that might be incorrect about the What comp... Seems the API patent goes back to 1990-92 (I don't know when they acquired it, who is Audio Teknology Inc?) The What is from 2000 or so?

API don't (didn't?) have any patent on this.
ATI patented it.
once they have bought API, API used it too ...
I'm unsure if Tilt is the better name, isn't it a TM of Tonelux nowadays ... ;D
but call it what you want, I like it

Igor, nice to see you back in a better mood

 

[Biasrocks]

Definitely not a new idea.

It goes way back.


For example; a Quad 34 preamp had a tilt circuit.

http://www.dc-daylight.ltd.uk/Valve-Audio-Interest/Schematics/QUAD-34-Schematic-Circuit.pdf

And it's been discussed in detail here

http://www.groupdiy.com/index.php?topic=15732&highlight=tilt

And here, with contributions from Mr. W*lff himself.

http://www.groupdiy.com/index.php?topic=21126.0

Mark

 

[ruckus328]

Personally, I like "hump" instead.  They can be labeled "hump" and "big hump".  OK, all kidding aside, Tilt is a good name.  I like it.

Obviously will have to get rid of a few filter positions so I can make use of a 2x6 switch since it will be a dual ssc.  One thing I just want confirmation on if anyone could be so kind (I haven't implemented the hpf in my GSSL so I have no on-hands experience on it), is that each side chain goes through it's own hpf and then through a switch/relay to either be summed in GSSL mode or stay thier seperate courses for SSL mode, correct?  In other words, the L & R sidechains ALWAYS go through independant HPF's whether they are summed or not?  They are never summed and then go through the hpf in GSSl mode.

This is pretty much all that's left.  Just finished the power supply and almost done ripple calculations.  Then just some BOM issues to fix and on to layout.  All in all should be showing everyone the preliminary design in a few weeks.

 

[Igor]

Igor, nice to see you back in a better mood

Yep. I decided to help here without any rationality, just by reason I like this forum and most people here,
and I have gained some experience with this compressor. Why not share?

the L & R sidechains ALWAYS go through independant HPF's whether they are summed or not?  They are never summed and then go through the hpf in GSSl mode.

In GSSL mode, sum after sidechain filters. It is stereo music signal.

I used Lorlin 2x6 on my 500 Mixbuzz.

Practically, when I did listening tests with first 1RU build (it was 3 years ago), I got to next decisions.
After summing the signals in GSSL mode,(-6db) I put this down for about 3db than used +6db amp
which works as buffer as well. This way, the threshold stays almost at same place when passing from
4000 to GSSL link, at real audio signal. This gives a same chance to both modes sound good.

@Igor, not that it matters much, but I think that might be incorrect about the What comp... Seems the API patent goes back to 1990-92 (I don't know when they acquired it, who is Audio Teknology Inc?) The What is from 2000 or so?

Yes, checked again...you right, man.

BTW, the patent on thr@st done for implementing kinda tilt circuit at compressor's sidechain.
In other words, any tilt filter in sidechain can't be done without affecting the patent?
Or I got this absolutely wrong. Anyway, people can use their filters here...
Seems, the way to do it, the values of thr@st circuit should be avoided at schematic and pcb \and changed to TBD.

 

[Stagefright13]

It is a good feature to implement. Cause I never hardly used my GSSL 'till I added the super sidechain. Now I love the thing to death. And with the kind of music I record the "Tilt" REALLY kicks. Or "Hump" LOL!

It also is really good on a Jazz recording not just rock.

John

 

[Igor]

I personally found very useful 80-100 Hz HPF and double sidechain (4000 or Turbo-mode).
When we did some tests with my 1RU version some years ago, most people liked it.
When listening 500 MixBuzz, I sticked to 100hZ settings on most material as well.
We tried it at friend's studio in his Lunchbox, on different mixes from hard electronic to
soft jazz. Sometimes, we liked thr@st filters, but what 100hz and 2-3 db of compression
at 4:1 does with mix on electronic, omg, it is pure honey

 

[Stagefright13]

Good tip Igor I use the "Tilt" or 60 and 90 the most. Even on mellower music samples or drum busses. Altho I haven't tried the turbo board. I'll try the 90 more tomorrow on some stuff when I get time.

I am really down with LED peak hold metering on this one. Just my opinion. And as I said I am just an amateur in audio electronics. I leave it all up to you guys to make this stuff lol!

John

 

[Biasrocks]

I've just had thought for another very useful feature.

Dual Mono mode, ala Alan Smart.

This is very handy when your using the compressor for Kick, Snare compression duty.

It would require an addition set of controls on the front panel, but it
would definitely be a nice option to have and would help to separate your version
from the others rather nicely.

Mark

 

[Igor]

Except dual control, it means about 3-4 additional dpdt relays and dual threshold/makeup circuit.
I did it in 2007, here:
http://www.groupdiy.com/index.php?topic=24404.0
with DG40* series commutators.
Keeping both channel tracking absolutely same, within +/- 0.1 db at every ratio
and in every possible mode, was nightmare, but possible.
Today, there are very low Ron fet-based commutators which can work for up to +/-20V,
anyway, if sidechain and ratio commutation are high-impedance paths, the vca controls
paths are low impedance and additional Ron can add error.

Another thing. I suspect SSL did whole patcbay on FET switches;
they good only for high or medium impedance control paths.
For audio, fet switches sucks, if I can hear them on JBL6208 at my lab's workbench,
every normal-ear'ed person will hear this coloration with Focal/Adam/ATC/etc.

 

[Biasrocks]

I don't see the need to keep the control voltages matched up in dual mono mode since the whole reason for using this mode is to process two signals differently. It should be fairly easy to switch the control voltages between standard SSL ganged mode using one set of controls and dual mono with two sets of controls.

Mark

 

[Igor]

Exactly. Imagine or draw the circuit for

switch the control voltages between standard SSL ganged mode using one set of controls and dual mono with two sets of controls.

and you will understand why it is necessary to keep switches impedances very low (relays),
when switching from dual mono to single stereo. Paths going to control ports of VCA's have
low impedances, and 30R or 36R (for example) switch resistanse (Ron) DOES matters as well
as theirs mismatch.

It should be fairly easy

http://i251.photobucket.com/albums/gg291/diy33609/mixbuzz2007.png

I can agree with you, here is the schematic of MixBuzz in 1RU.
(c)2007, Kapelevich Igor

For education purpose only

It is relatively easy