SSL Talkback Compressor Anyone?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ruckus328

Well-known member
Joined
Apr 14, 2010
Messages
906
Location
Philadelphia, USA
Just wanted to see if there's still an interest for this?  I have a current want for this for an upcoming mixing project so in process of doing up a design.  Plus my schedule is lightening up and I'm jones'in for a new project anyways and you all know how I loves me some compressors.  I know somebody else was working on it but 2-1/2 years with a no show probably means it won't get done.

Anyways, my plans:

-Original circuit of course.

-Options for Jenson & Cinemag 4:1, possibly original Sowter available, waiting to hear back from all of them.  Switchable for mic or line input.

-I can add a mix/blend function if you guys want (since I know you'll ask :) ), just a cut and paste from my version of Keith's CnB schematic so it's fairly easy.  Debating between sniffing the dry input right after the input transformer, so In/Out knobs on the compressor will have no affect on the original signal, which normally would be the right way to do it for line levels since you can adjust the input level with previous piece of gear, however if you just plug a mic into this it could cause a problem as dry signal would probably be too low and not adjustable.  Sniffing the dry signal after the input knob would mean that by adjusting the dry signal level until desired volume would be simultaneously driving the compression level, which might not be an issue since it will be offset by the mix function itself, but it's almost like the best way would be an additional gain knob for the dry signal.

-48V Phantom?  If so probably would be a good idea to disable it when line mode switched in to avoid killing any gear.  Thoughts?

-For me this thing is useless without some kind of metering action.  Would like to see In/Out and Gain Reduction, undecided on using a good ol VU meter or going with some LED metering.  In/Out is easy (I think, actually not sure now that I think about it as might get different readings with the input tranny switching)  Not quite sure where best place to sniff the circuit for Gain Reduction would be, thinking somewhere in the vacinity of around the FET biasing.

Also, since I'm sure everyone will want to do this in pairs seems logical to me to make it a dual board.  Dimensions shouldn't be that large as it's a pretty simple circuit, so if you just want a single channel you'll just populate one channel on the board.
 
There is another thread floating around that covers just about everything you want to know. Do a quick search and see what you come up with. Might save you a lot a time.
 
Thanks, yes I know of the other thread you're talking about but 90% of it is OP trying to learn fundamentals of the circuit and what the CnB is (no offense intended of course).  In any case, that thread's going on 3 years old and OP hasn't been on prodigy in 4 months, so going to move forward with my own version.

I've already drawn up the majority of my schematic, the original circuit is pretty simple and CnB implementation is pretty straightforward as well, so this thread isn't too much of a quest for knowlege but rather a feeler of interest and what (if any) additional features/options people would like to see.
 
ruckus328 said:
Not quite sure where best place to sniff the circuit for Gain Reduction would be

Measure GR across C14

ruckus328 said:
-48V Phantom?  If so probably would be a good idea to disable it when line mode switched in to avoid killing any gear.  Thoughts?
You could incorporate this into the mic/line switch... or go Neve old school and have a separate input for your line in. A 1:4 tranny on the input is probably gonna be too loud when you jam a line level into it.

check out the Ridge Farm Boiler for some inspiration since it's a super pimped out LMC.
http://www.ridgefarmstudio.com/rfinew/bprod.htm
 
ruckus328 said:
-48V Phantom?  If so probably would be a good idea to disable it when line mode switched in to avoid killing any gear.  Thoughts?

You could incorporate this into the mic/line switch... or go Neve old school and have a separate input for your line in. A 1:4 tranny on the input is probably gonna be too loud when you jam a line level into it.

That sounds like a good solution to me too.

I would definitely like to see one of these.
 
Never... -NEVER have phantom available on any.. ANY line input socket.

Having the mic/line switch disable the phantom may SOUND like a good idea, until someone accidentally plugs in a piece of line gear with it in the wrong position... or moves the switch by mistake.... or any other scenario.

Simply having the ABILITY to switch phantom onto any line-level connector is INVITING disaster.
 
Keith, I whole heartedly agree, but don't alot of modern interfaces allow this to happen anyways?  For example my profire2626 sure does, I can engage phantom power on any of my xlr inputs, actually it's even Sh*ttier than that as there's only 2 phantom buttons, one controls channels 1-4 and 5-8, so I have to be reeeeeally careful before I plug in if I'm using an XLR cable as I might forget I have something plugged in on the back of the interface on one of those channels that I forgot about.  For the talkback, only safe way would be to have a completely seperate connector for line input (which is fine), but would really like to avoid having 2 seperate input xfrms as that'll drive the cost significantly.  And then I can see a problem with doing this as well - chances are most people will usually opt for the line connector being an XLR connector instead of TRS anyways (I know I always do), so they'd end up with 2 xlr conectors (one for mic and one for line), so there's still the potential for the someone to plug the line gear into the mic xlr accidentally anyways.

This is why I'm ensuring phantom protection on every design I do from now on.  The chances of gear getting hit in the real world are very probable.

Only thing I can think is have an xlr for mic, specify a TRS for line, have them both go through the same iron and phantom disabed when the line is engaged so phantom power can never be present on the TRS.  If someone wants to use an xlr for the line connector then they're on thier own and good luck.  Of course can see a problem here too though - 48V might be present for a microsecond depending on how fast and what order those relay contacts close.

Only real way to protect against any scenereo or the utter brain dead would be to have no line signal capability, mic input only, but then what's the fun in that?  ;D
 
I would like to see this in 51X formfactor.
I don't need a MicIn, but if wanted and regarding to Keiths concern you could move this to the front, as an independent connector.
I would like a (switchable?) stereo link. ;D ;D ;D
 
[silent:arts] said:
I would like to see this in 51X formfactor.

Hehe.  Yea, I've been telling myself to get on that bandwagon for a while now.  I guess this is good as an excuse as any.  I'll do it up for both formats (1U and 500/51X).  I already have mechanicals/PCB footprints made up for 51X format so this will be an easy first offical project for me in that format.  Of course, I need to order a rack for final test (you the guy to talk to for that?), but in the meantime I can move forward with the design.

[silent:arts] said:
I don't need a MicIn, but if wanted and regarding to Keiths concern you could move this to the front, as an independent connector.

I think in either format the Mic input should stay, as that's what this thing was originally for, but yes, seems like good idea for Mic to be in front in the 51X format. Now that I think about it, maybe in both formats this would be a good idea.

[silent:arts] said:
I would like a (switchable?) stereo link. ;D ;D ;D
How to do this has been bouncing around in my head for 2 days now.
 
Well, I'll probably end up having a slightly different board for 1U as it will need its own supply (although going to try and start standardizing my Power Supplies across designs where possible), different front panel scheme and different power interface/xlr connection scheme but I'm going to focus on 51X/500 format first and adjust for 1U after, which in reality is probably a days worth of work for me.

So it's official, this is now for 500/51X format, 1U will be done in parallel to allow for access to everyone.  Adjusting to run at +/-16V shouldn't really involve much if not anything.  Both systems will be schematically identical.

I don't want to overcomplicate this, but after more thought here's the features I think should be there:

1) Going to go with a simple analog VU meter that monitors output level and G/R.  Don't see a reason for Input Level as it's almost pointless and won't mean much IMHO.  Take for example a mic plugged in - If I sniff the input signal pre gain knob it will hardly register.  If I sniff it post gain knob it will be off the charts.  I need to figure out how to translate the G/R to the VU meter.  I guess sniff the voltage across C14, through a buffer feeding some small DC voltage to the VU Meter for 0 G/R reading when no gain reduction present.

2) Front panel Mic connector, rear panel Line Connector.  Need to figure out a way top ensure Phantom Power is never on the line connector, without having to use a seperate input transformer.

3) Keith's infamous Mix/Blend circuit.

4) If it doesn't start to get overly complicated, I'd like to explore some variable attack and/or release times.  I'll admit I'm not certain if I'm correct, but I think a matter of adjusting R33 for different R/C values?  I dunno, this is similar to the buss compressor but different enough to throw me a curveball.

5) Stereo link (If possible)..... surely anything's possible
 
ruckus328 said:
4) If it doesn't start to get overly complicated, I'd like to explore some variable attack and/or release times.  I'll admit I'm not certain if I'm correct, but I think a matter of adjusting R33 for different R/C values?  I dunno, this is similar to the buss compressor but different enough to throw me a curveball.

That's the first place I looked, but R33 seemed to act more like a threshold control no matter what I did with it.
I also thought that R35 was the ratio, but toying with that acted like a threshold control as well. I'd be interested to see what the answer is for attack and release.
 
Do you mean you have an original one to toy with?  I'm in process of spicing it, hopefully done tomorrow, maybe might get some answers, maybe not, but worth the try, at the very least it usually helps me get a clearer picture of what's going on with things.

So I've been thinking about CnB implementation.  Sniffing the dry signal after the input gain pot initially seems like the logical thing to do as if you're running a mic the level would probably be too low to sniff it pre-gain, however after some though I'm not sure it's going to be that simple and wet/dry levels could get really out of whack.  Picture this:

If you sniff it pre gain:
Line levels - you should be fine, increase your input for desired G/R, adjust output for appropriate makeup, wet and dry levels are the same, blend away.
Mic - Output is probably going to be much hotter than the dry Mic signal.  The more dry you blend in the lower your output level gets.

If you sniff it post gain (for sake of example let's say we're feeding it a +4dbu signal and we're right at the threshold):
Line Level - As you increase the input gain (let's say 10db) the dry signal will increase to +14dbu, wet signal will ensue 10db of gain reduction, so we're at -6dbu now.  You'll have to add 20db of makeup gain on the output to get your wet and dry equal levels, otherwise volume will increase as you blend towards the dry.  Your output is now 10db hotter than the input, and basically you're being forced to do this if you want your blend knob to be relatively even.  And who knows what you might be feeding this thing into, and +14dbu could be potentially clipping some A/D converters.  Personally I don't really like this or having no real control of my output level.

Only way I can really see to remedy this is to have 2 input gain knobs - one for the dry signal, and the other for the wet to feed the compression.

Thoughts anyone?
 
Here's my 50 cents:
I'd mostly use the line input but IF I'd want to record a microphone straight trough it, it would probably use something SM57-like.
No need for fancy condenser mics to record 'dirty attitude tracks'.
If you really want phantom on the same input as the line input I'd suggest you use a relay to switch between line and mic.

For the same reason I wouldn't be bothered with different attack and release times. This really seems like a compressor you want as an effect. Smacked real hard.
No need for tweeking.
But then again, you can implement the attack and release pots and everyone can decide whether they want them or not. Same goes for the phantom.
Like I said, this is just my vision...
 
Looks like a good project...
Speculation...isn't R33 more of a ratio control? and I'm curious on hearing how that att/release work...very strange.
 
Hank Dussen said:
Here's my 50 cents:
I'd mostly use the line input but IF I'd want to record a microphone straight trough it, it would probably use something SM57-like.

Well, the thing is I don't think you're gonna have enough gain to run a dynamic mic like the sm7/57 through this.  I'm simulating some things now but doesn't look too promising at the moment.

I've been thinking about some thing since last night - reality is, in a dry state, provided some things are thought and and executed accordingly, this thing will make a pretty nice mic pre as well as a comp, so you're getting 2 things in 1 really.
 
I have a couple of thoughts; I have an original one here, and I've racked them in the past... I contemplated much of the same stuff as you have:

GR indication is going to be tricky, the FET is nonlinear. Making it work will mean a bit of matching, and even then it won't track very well... translating that GR into a VU meter (which one inherently sort of 'expects' to be at least vaguely 'accurate') is more complicated. The circuit is feed-BACK topology to address the nonlinearity by self-correction. Any changes in settings or adjustments of the threshold or FET bias just make the attenuation onset and curve even more difficult to track.

GR won't simply be metered across C14... That tells you the rectified voltage coming out of the detector, but will show a rising voltage even BELOW the threshold when no GR is being effeced (because the voltage is below the pinch-off voltage for the FET).

Change C14 if you want to alter attack/release time constants. There's no elegant way to add separately adjustable attack and release controls without changing the nature of the circuit so radically that it's really NOT the same circuit with the same behaviour any more... Attack speed is limited by the cap charging through the 1K resistor... release is controlled by the DISCHARGE rate, which is also back through the same resistor...

Ratio: forget it. this is a simple feedback infinite ratio limiter. You'd have to mess with the circuit A LOT to be able to control ratio, and by that time, it's behaving differently.

If you added ratio control AND attack/release controls, not only would it be different to operate, it would SOUND different even if you tried to make it operate 'the same'.

GR indication is possible, but a little involved in terms of set-up, and would require matching FETs.
 
Thanks, well even if G/R has to get thrown out the door why would output level be difficult?  Isn't it just a matter of sniffing and maybe buffering the audio level at the output?

One other thing, been thinking about this all morning, if using this with a mic wouldn't the output gain need to have some series throw to get the output volume to usable levels?  1:4 input transformer gets you a little but other than that there's nothing boosting the signal, input gain does nothing more than increase the compression.  What are they doing in the console?  I'm not sure what this feeds into.

As far as using it with line levels, would just adding maybe a pad on the input be a better method of adjustment as opposed to reversing the transformer?

Ratio I knew wasn't possible, but based on what you said I think I'll leave well enough alone any tinkering with the attack/release.

The more I start to dig the more complicated this becomes to use for both mic and line levels, but I really like the idea that this thing could have so many uses.  I'm kind of looking at it from the standpoint that it can be a nice and dirty compressor, or a nice pre with adjustable amount of grit.
 

Latest posts

Back
Top