My New 1U GSSL Design (AKA SB4000) RELEASED!!! - GOTO PAGE 9 FOR ORDERING INFO!

[sonolink]

Why not Mono/Stereo? Am I wrong or Mono is "2 independent ch" and stereo is "settings apply to both ch"? Something like "Link/Unlink"....
Sum/Average? Now I'm starting to get confused...

 

[ianmgull]

I'm going to build a couple of these myself and am also a bit confused by the various sidechain options. Would someone more knowledgeable mind giving a brief description of each of the available sidechain modes?

 

[SSLtech]

No drop when switching between Oxford & Aarhus modes (I compensated for the gain difference when switching)  
...
Thoughts anyone?

My thought is... Impossible.

You can't have a 'compensation' which is correct for all applications.  Whatever you choose is going to be wrong for something.

Put narrow mono stuff in and the difference is going to lean one way, put wide stereo in and the difference will lean another. -It's just a fact.

I'm sure what you've done is laudable, but be careful NOT to mislead people by claiming the impossible.

Keith

 

[ruckus328]

No drop when switching between Oxford & Aarhus modes (I compensated for the gain difference when switching)  
...
Thoughts anyone?

My thought is... Impossible.

You can't have a 'compensation' which is correct for all applications.  Whatever you choose is going to be wrong for something.

Put narrow mono stuff in and the difference is going to lean one way, put wide stereo in and the difference will lean another. -It's just a fact.

I'm sure what you've done is laudable, but be careful NOT to mislead people by claiming the impossible.

Keith

True, it's source dependand, no argument.  I didn't mean any claim as to achieving mistical feats or intend to mislead, if anyone thought that my apologies.  All I meant was there's not a 6db drop when switching between modes as would be the result without adusting the first gain stage on the SSC's when switching.  Yes, it will still lean one way or the other some depending on the signal you're feeding it though, that's just the nature of the beast.

 

[ruckus328]

I'm going to build a couple of these myself and am also a bit confused by the various sidechain options. Would someone more knowledgeable mind giving a brief description of each of the available sidechain modes?

Mono Sidechain (AKA GSSL Mode, AKA Aarhuus Mode) - L & R control signals are summed through a single sidechain.  Drawback is the low end is often overcompressed, so this mode is often used in combination with a high pass filter to compensate.  Most here seem to often like this on the drum buss.

Stereo Sidechain (AKA Original SSL Mode, AKA Oxford Mode, AKA Turbo) - How the original SSL Works - L & R control signals each go through their own sidechain

My design does both modes and is switchable between the 2.  It also has 60/90/120Hz high pass filters & 2 other special settings incorporated which can be used while in either mode.

Hope this helps some.

 

[ianmgull]

It absolutely does. One final question though:

I assume the two extra settings you mention are the two "tilt" modes you mentioned earlier in the thread. Care to share what that does? Is it simply a different filter after the control signal?

 

[Igor]

Keith, I reading this thread as full of adventures science fiction reading.
I did -3db sum for GSSL mode at Mixbuzz500,
to let more chance to both modes sound good when switching between them, see how it treated.

No drop when switching between Oxford & Aarhus modes
(I compensated for the gain difference when switching)

but be careful NOT to mislead people by claiming the impossible.

Just my $0.25,

Noise levels from 5Hz to 50KHz are at -130db on an RMAA scale.  Not too shabby.  Too be honest the power supply is overkill at this point, but NOBODY will have noise/power issues.  The grounding scheme is very simple, just a simple wire from the power supply to the star ground, pin 1 of the inputs/output XLR's to star ground, and IEC ground to star ground as well of course.  

That's interesting. I am not best in numbers. But, at first glance,
there's no a/d or d/a audio chips with 130db dynamic range.
OK, maybe, it is souncard from alliens.
Let's get INTO numbers. Maximum output of GSSL can be +22dbV.
I mean 0dbv=1V RMS. OK, in DBM ref 1mw/600Ohm this can look 2.2db more.
Than, if we measuring maximal headroom, OK, we feed signal till cliping,
this is our max ref, than shorting the input, this is noise floor.
What I really can't understand, 130-22=108dbv
I don't remember what is thermal noise of single 20k resistor, IIRC a bit more.
-108 dbV noise floor with one THAT chip and 20k resistor.
Without taking into account other stuff. IMPOSSIBLE!

I've got THD readings of .002%.  

This is PURE BULLSHIT!
Or, maybe, you measured the THD with bypass in. Happens
THAT Datasheet shows 0.01% for A grade.
Distortion does not goes down significantly in case of paralleled VCA's.
S/N maybe. I can assume that wit 8 paralleled A grade VCA's distortion will be arround
0.01% but .002%.
Yes, it was typo again

 

[snipsnip]

Igor, I think its fine to raise questions if you dispute the data. But have a little respect. You and your fanboys have a field day when someone challenges your designs (check remonators 2254 build thread for one of many examples).

So just act like a grown up will you.

 

[JoeMorris]

just because this lovely looking project may put a dent in the sales of your ridiculously overpriced designs doesnt mean you need to act like spoilt little girl, igor. you dont strike me as being in the wonderful general spirit of this board at all.

 

[Kingston]

I think the whole issue is due to the fact the -130dB noise floor given by ruckus328 is an RMAA number. It's dBFS, and thus not referenced to anything but the soundcard.

noise should be given in dBU. ruckus328 should calibrate that RMAA number to something in the real world. Measure 0dBFS 1khz sine output from sound card (ie. full blast) with a multimeter, you will find out how much above an equivalent dBU number it is that way. It'll be something like 12-25dB depending on your DA converter. Subtract that from the RMAA result and you now have noisefloor in dBU.

It'll still be impressive.

[edit]

corrected a stupid mistake. it's dBFS, not dBA.

 

[snipsnip]

I agree - We need proper figures... and need the imput of the experts to spot the errors where the less experienced cant.

But theres a way to do this.

This guys is GIVING the community something as a no / little profit operation.

There is no need to be such a douche.

 

[diyfanatic]

here we go again!  :-\
a bit childish isn't it?

I think I'll to take more than one board just for support!

 

[[silent:arts]]

... It's dBA, and thus not referenced to anything but the soundcard.

for sure you mean dBr (dB relative)
dBA is referenced to an A filtered dBspl, which has nothing to do with electronics

noise should be given in dBU

correct

however I vote for no SSL bus compressor war.
stupid.
let us keep this thread clean for ruckus328 design.

 

[Kingston]

... It's dBA, and thus not referenced to anything but the soundcard.

for sure you mean dBr (dB relative)
dBA is referenced to an A filtered dBspl, which has nothing to do with electronics

Oh crap what a brainfart. I meant dBFS of course.  :-[

 

[ruckus328]

Man, you can always tell when Igor's been out drinking.  Seriously dude, your attitude is beyond rediculous at this point, you're only making yourself look bad.  I feel like the ultimate broken record here.

Guys, yes, the RMAA noise scale is relative, I assumed everybody knows this which is why I clearly stated these were RMAA numbers.  

YES!, it was a typo.  And boy am I sorry.  It's .02%.  Not 3 decimal places.  Obviously you can't go lower than the capabilities of the VCA's.  Honest mistake, my sincere apologies to everyone and thank you for pointing it out, I'm surprised nobody else noticed that.  I would have gladly corrected myself as I'm doin now, but you didn't have flip your noodle.  Would you like me to start pointing out all of your typo's?

I can configure my Lecroy at work to measure THD but haven't gone through the hassle.  I'm sure 99% of the people here use RMAA for all of their readings and calibrations.  Power rails/ripple were all measured using my Tektronix, should I not trust that?

I'll post my RMAA readings for everyone to see, as I had already intended to do.  At no point have I tried to be deceptive, it does little good as I pride myself on honesty, and fudging numbers makes no sense, as after the first round of boards ship, things like that would become very obvious, which wouldn't help me very much with credibility for future projects, so why the hell would I do that?

Just to be clear, this is a hobby for me.  You claim it is for you Igor, but to that I say "BULLSHIT".  I don't need to do this.  I make a damn good living and am making nothing off of this.  On the contrary when all is said and done I'll have to sell a couple hundred boards just to break even for all the money out of pocket it cost to develop it.  Proto boards, machining costs, constant shipping charges, tooooons of components, may of which I ended up not needed, etc etc etc.  That's not even considering all of the time (hundreds of hours), which to that I'll be making $0.  In fact I've lost money because I've put off paying studio work because I don't have the time to do both.  The wife's been neglected, my house is a disaster, yard work hasn't been done.  Many of you have no idea the time and stress it takes to do something like this, and many of you do.  You guys here that have been supportive, PTown, Kingston, Volker, Mitsos, Lukas, and TOOOONS of others (too many to name everyone), THANK YOU!  You guys are the reason I've pushed forward with this.

 

[Kingston]

I'm sure 99% of the people here use RMAA for all of their readings and calibrations.

Yes, but unfortunately the numbers it gives mean little without a reference. RMAA works on dBFS only. It only knows what it sees between the digital 24bit (depending sound card) range. That varies between almost every AD/DA converter and manufacturer. Until you calibrate your specific AD/DA dBFS scale to a real world measurable dBu scale (RMS volts) the RMAA numbers are somewhat meaningless.

RME HDSP for example outputs 1khz sinewave at 0dBFS (maximum digital volume) to 12.6dBu (measuring RMS volts at the output) on its +4dBu XLR balanced setting.

 

[okgb]

I say Bravo to ANYONE who gets something done
could be worse , the chinese or people from other grps
ripping stuff off
Not everything survives and it's either growing or dying
and we want this grp to last some time eh ?
[ support , nurture and hopefully prosper for all - elect me , ha ! ]
so everyone back to your corner to think about what you did
and i'll try to finish building something

 

[Ptownkid]

Igor....man, give it a rest bud. You seem to have made it your personal mission to try and crap on Mike's efforts. Why don't you stick to doing what you do best, which is creating and improving YOUR OWN projects. Every time you pop up in this thread, you turn another handful of people away from your designs. No one wants to support someone who is arrogant (albeit talented), condescending, and flat out childish.

Mike, you are doing a great job and this has been a very productive, insightful, and well discussed project.

Keep it up and ignore the rest.

Cheers

 

[HammerheadAudio]

As soon as you have some price guesses on this baby please post it! Assuming I could afford it, I would love one of these to try out next to my original Gssl, and to use in the studio.

Kudos on the awesome work so far, the fact that you have remained very professional through the ranting of a few is amazing too!

 

[mikefatom]

I'm in for one or two sets. This project deserves support! I am amazed at how willing you still are to push this forward considering everything.

Great job Mike!