What features are a must have on the GSSL?

[Stagefright13]

OOPS! http://www.expataudio.com/

 

[regularjohn]

If you have one built, and haven't noticed any hum, just try turning it on, and listening to the output with no signal present.  -Turn it up until you hear the hiss. -If you hear hum in with the hiss, that can be eliminated.

I think that builds which have more current consumption might also suffer more greatly; so if you have turbo, dual sidechains, relays, illuminated buttons, bargraph metering and lord knows what else, you might be causing more ripple on the lines, and the problem with the existing GSSL layout is that any modulation of GROUND reference tends to get picked up at the current-to-voltage converter stage immediately following the VCA.

So if you only hear 'inky-black silence' when you wind a ton of gain onto the output of a GSSL, you're in luck.

But BOM doesn't enter the equation as it's not possible to fix this with simple parts replacements. This is a problem with physical traces of the PSU area. The noise is polluted to the board from the VAC connectors next to the rectifier, and because the onboard filter cap + rectifier traces let some of this noise through. CRC board ensures this is cleaned before it ever enters the board.

Inky black silence is all I hear on all the gssl's I've built - and I've built a number of them.  With make up gain all the way, I've never heard even the slightest hint of hum, even with my monitors cranked so high that a mosquito fart would make them explode; and it's never measured anywhere near to where it would cause a problem (certainly not where Kingston said his are) - even when chaining many pieces of gear together.  

I disagree with you, Kingston, on the bom issue - and on your assertion that every gssl has this problem.  The changes I make to mine significantly stiffen the voltage rails.  I also take extreme care in the placement (and selection) of my power transformer, and in the selection of my components.  But like SSL tech said about having a bunch of add on boards and lighted switches involved, I typically only have the gssl, and one or no add ons - and if I use christmas lights they certainly aren't run off the 12V rails like the gssl is designed for.  Watching your grounding scheme is also very important - probably the most important factor.

But I can appreciate the fact that a very inexpensive pcb can absolutely eliminate the chance that you may have hum problems.  

 

[Kingston]

Please measure that thing before making any claims on the noise floor.

The fact you can't hear the hum is not helpful to an engineer, even "With make up gain all the way". It might be that listening environment is shot and you will never hear anything below even 100hz correctly. or maybe headphones don't go that far. That's why we take the physical airwaves completely out of the equation and measure VAC directly. Trust me, non-CRCRC GSSL noise floor is nowhere near flat like it should be. It might still be good enough for a particular studio set up where noise concerns are relaxed elsewhere as well. But this is DIY and we have the opportunity to optimise to perfection, so why not?

PS. I don't, or would not use the CRCRC. I don't like the GSSL stock regulator "application note" set up one bit. It's something I would only use for simple non-critical stuff like LEDs and relays. Also, a completely external solution is better for a PCB where ground traces are known to cause problems and distribute noise in unpredictable ways.

 

[Ptownkid]

I still don't really get the whole C&B thing as a built in feature...I mean we all should have tracks to spare in our DAW, so it's not like it's hard to do. Beyond that, i totally agree with Keith that it would be much more versatile as a stand alone.

 

[stereokillah]

to stay in CRCRC discusion,as SSLtech say there is a hum without it but it's not evrybody who found it, all the unit i built had

*around -86 dbfs ( -70dbu ) of noise without modif.
*around -90 dbfs ( -74dbu ) of noise with two or three cut on ground trace
* between -100 to -105dbfs ( -84 to - 89 dbu ) with CRCRC

ALL WITH FULL GAIN

if i've good remenber i built more than 20 unit with differents psu transformer even transformer out of rack and diferents option no signifiant difference

only crcrc give a fine signal, but some commercial unit have a signal noise like gssl without crcrc and whitout modif.

 

[jackies]

PS. I don't, or would not use the CRCRC. I don't like the GSSL stock regulator "application note" set up one bit.

So what are you using, Jung regulators?

 

[Kingston]

PS. I don't, or would not use the CRCRC. I don't like the GSSL stock regulator "application note" set up one bit.

So what are you using, Jung regulators?

I just meant I don't like the small cap->LM78XX->small cap set up. It's led and utility power, not pro audio in my book. the least one could do is a souped up LM317/LM337 text book app.

But in a related story, I actually did test Jung regulators on a GSSL. I built a batch of them for another project that wasn't ready, and GSSL happened to be the closest thing on the table and got to be the test bench. I haven't got tools good enough that could measure a difference between properly implemented LM317 and a Jung regulator. The difference is there, just not measurable by me (while LM78XX noise I can actually see).

 

[jackies]

I actually did test Jung regulators on a GSSL.

This is cool!
I always been curious about Jung regulators, even have a couple boards, which I never got around to build, and was wondering what would be the difference compared to usual LM*** regulators. What were you building them for, some DAC project or something?

 

[Kingston]

They are for the rkn80 AD/DA project, and also passed a quick test there as well. The difference is radically lower noise (would make even NASA happy) and low low output impedance. They are as close to pure DC source as is feasible for audio, not counting ridiculous car battery set ups advocated by some audiofools of course. Ground sense is also handy for the seriously anally retentive engineer as you can discount the effects of wiring on the regulators.

 

[SSLtech]

PS. I don't, or would not use the CRCRC. I don't like the GSSL stock regulator "application note" set up one bit. It's something I would only use for simple non-critical stuff like LEDs and relays. Also, a completely external solution is better for a PCB where ground traces are known to cause problems and distribute noise in unpredictable ways.

For what it's worth, I built two identical turbo GSSLs recently, Both with 78xx/79xx and CRCRC.

I tested them and measured them and calibrated them (they were both fitted with trim pots for ALL the usual suspects: -Ratio, channel gain, meter cal etc.)  to assure myself of correct and similar performance.

Then I took the 78/79 out of one and replaced them with LM317/337's, with the required componentry and a little additional decoupling to the regulation sense leg. -Then I re-measured/compared them both.

There wasn't any significant difference which I could measure, in the two which I built; one with 78/79 and one with LM317/337.

Now, I should point out that I'm not making a claim that there's never any difference... there are 78xx from different manufacturers, and maybe some are noisier than others... who knows?

But for the two which I built, there was no noise floor difference outside of a dB or so, which was within inter-unit variation

 

[ruckus328]

Keith, - did you have the crc board still fitted when you put in the LM317's?

 

[SSLtech]

CRC was pluggable. I tried it with and without. -I did find that the CRC still made an improvement, when the regulation was on-board. -It appears that the 120-cycle current 'surges' just need to be moved off-board with that layout, which is why the CRC works so well.

Actually, the second board set which I used for the 317/337 testing is the one which I just put into this customer build. -I returned it to 78xx/79xx regulators, simply because they're a neater fit without the associated Voltage-setting componentry.

Keith