New GSSL Layout Proposal - Anyone Interested?
- Thread starter ruckus328
- Start date
Help Support GroupDIY Audio Forum:
ruckus328
Well-known member
Igor said:
Thx. Turns out I was blind, it is on the datasheet. 1.3mm comes out to 0.05" btw.
Guys, I have another question. I'm curious - does anyone know the reason that Gyraf chose to run the sidechain off 12V instead of justy running it directly off of the 15V rail? Has anyone done any experimenting with running it directly off 15V? I know the IC's (THAT/TL074's/etc) will run fine at 15V, I don't know if there's some catch I'm not seeing though. I mean, there has to be a reason he did this right?
ruckus328
Well-known member
But what if the +/-15V supply doesn't fluctuate?
I'm my own worse enemy lol. I was damn near done when I started to ponder getting rid of that 12V rail. I don't want to shoot myself in the foot here though. I searched the forum high and low and found one post that talked about running the sidechain off of 15V. I guess maybe that's telling me something?
I'm my own worse enemy lol. I was damn near done when I started to ponder getting rid of that 12V rail. I don't want to shoot myself in the foot here though. I searched the forum high and low and found one post that talked about running the sidechain off of 15V. I guess maybe that's telling me something?
Harpo
Well-known member
The GSSL sidechain VCA used to be implemented as DBX2150. This Chip has a recomended supply voltage of +/-12V or 24V across IC and absolute max.30V across IC. Tolerances in 7815/7915 voltage regulators might exceed this VCA abs.max. voltage rating. With THAT218x you could supply it with +/-15V, adjust the Ibias resistor value, connected to VCA pin5 for required current and recalculate the ratio setting network for the added DC-offset. Adjusting threshold and makeup for extended range shouldn't be a big deal.ruckus328 said:
ruckus328
Well-known member
Harpo said:
Thanks. That makes sense. I wasn't thinking about the 2150's. This leaves me in a dilemma. 15V regulation is currently off board. As well as a dedicated 12V rail for LED's, relays, and something else that will make you all happy
. With all of the other modifications and unknowns I have at this point, I'll just admit I'm moderately paranoid about changing the sidechain voltage. If I leave the sidechain voltage at 12V, it doesn't seem logical to bring that off board as well, since it will only result in additional wiring to the main board. And since nothing else besides the TL072/74's are running off of 12V now, sticking with the small 78L12/79L12's shouldn't be an issue, although I think I could allow enough room to even use regular TO-220 packages there. Wouldn't be room for heatsinks, but there's no reason why you would need them anyways. Hmm.....
livingnote
Well-known member
Well if you're going separate supply for "rough stuff" like relay and LED power anyway, what I just did was to
cascade some 12V regs right off the 15V audio rails - works really nice, I guess without the power on LED
dangling off the little regs you could just as well power the turbo circuit+sc right off the 78L and still have room
to spare.
cascade some 12V regs right off the 15V audio rails - works really nice, I guess without the power on LED
dangling off the little regs you could just as well power the turbo circuit+sc right off the 78L and still have room
to spare.
ruckus328
Well-known member
livingnote said:
LOL, About 2 minutes ago I just thought: "I wonder what Lukas did", since I assumed you also went seperate dedicated 12V supply for all those LED's lol. That's pretty much what I was just thinking, run the 78/79L's right off of the 15V rail on board. Although I generally don't like to daisy chain regulators, in this case there's so little current draw on the SC rails, it's not really an issue in this case. (I saw in some post somewhere that sombody stated the 12V regulators were running off the 15V supply, but I just double checked and that's not the case - In Gyraf's design, the 12V reg's are running off of the main +/- rails just as the 15V regulators are, which so I'm confused by his statement that they're there to provide predictable gains in case the 15V rail fluctuates?)
Anyways, since the regulator circuitry is so minimal, it doesn't make sense to roll the dice and change the SC circuity to run off of 15V and have to deal with all of the value experimenting and fustration that will ensue. Especially since I already have my hands full with the other modifications & enhancements I've done. Plus this will keep the PS simplified and smaller. And now, it can actually be a multi-use Power supply, and be used with existing versions of GSSL's if someone wants.
livingnote
Well-known member
Yeah plus you're right in that 3V drop the reg likes and it's rock solid on the incoming, so just give it some 22µ
or so caps and you have yourself Dr. Overkill re sidechain precision without even wanting to
or so caps and you have yourself Dr. Overkill re sidechain precision without even wanting to
you're loading down the main "audio-rails" regulators this way. not wise at all. this is not a "true seperate rail" for leds.
if you want to "decouple" the "leds" rails from the rest of the circuit, do it exactly like gyraf did it (BEFORE the regulators, i.e. directly from the graetz bridge). perhaps try to increase the in-line resistor to reduce the regulator heat dissipation. Personally, I would use lower voltage than 12V to power the leds.. but this is not my project.
Additionally, take care that ground current return paths from LEDS on the PCB don't mess with the "clean" audio ground..
>>> I'm confused by his statement that they're there to provide predictable gains in case the 15V rail fluctuates?
There is absolutely nothing confusing in this statement.
if you want to "decouple" the "leds" rails from the rest of the circuit, do it exactly like gyraf did it (BEFORE the regulators, i.e. directly from the graetz bridge). perhaps try to increase the in-line resistor to reduce the regulator heat dissipation. Personally, I would use lower voltage than 12V to power the leds.. but this is not my project.
Additionally, take care that ground current return paths from LEDS on the PCB don't mess with the "clean" audio ground..
>>> I'm confused by his statement that they're there to provide predictable gains in case the 15V rail fluctuates?
There is absolutely nothing confusing in this statement.
Attachments
ruckus328
Well-known member
tv said:
The dedicated 12V rail does run off of the graetz bridge, it is a true seperate rail. It's a second 12V rail. It is not associated with the sidechain 12V rail. It is a LM7812 1.5A TO-220 located on the power supply, and will be pulling somewhere in the ballpark of 200-250mA of current with all features on (It's also driving multiple relays as well as a 3rd item which is going to pull 100mA minimum). Returns for this rail are only tied to ground at a single point (the power supply ground input connector), so it is completely out of the audio return path.
So that being said, If I daisy chain the sidechain LM78L12/79L12's off of the 15V rail, do I still risk loading down the 15V rail? Because now the SC rails are pulling next to nothing. Lukas - I believe from what you said this is exactly what you're doing - and no issues? Also, I believe all of the guys using MNat's PSU with the GSSL are in effect also doing this. And at the end of the day essentially what I have is the same as MNat's PSU, only with increased bulk storage and additional fixed 12V rail. But if there is a risk, I can easily move the 12V regulators to the power supply, and run them directly off of the graetz bridge (which is what I would usually do anyways - as I said, I usually don't piggy back regulators), it just means 2 additional wires to run between the 2 boards.
livingnote
Well-known member
Ah...yeah. Redundant post of mine here.
ruckus328
Well-known member
Yea, that was the additional step I was contemplating - bringging the tie point between the logic and power surger stuff (LED's / Relays) back even further - having them tie directly at the XFMR input, through seperate bridges. That's as isolated as it gets.
livingnote
Well-known member
Might there be any merit decoupling the LED one from the other through a resistor at that point
or is that wasted care?
or is that wasted care?
I don't think you would need a full bridge for powering the LED rail, since it is a single rail, in this case two 1N4007's would do.
The C-R-C filtering is cheap and works bi-directionally (you can "look" thru it both ways, from the brige towards a regulator and from the regulator to the bridge, backwards) so it will decouple any surges/hash happening on the LED rail from other rails (adding a resistor could work as a "bridge" in single sided PCBs).
I would advise that you make a separate "ground" rail for the "dirty" return currents from the LED rail. If you must reference control voltages from sidechain etc, you can use "ground sensing" (with an opamp/diff-amp) to decouple the clean "audio" ground from this "dirty" LED ground, where the two would otherwise need a hard link.
The C-R-C filtering is cheap and works bi-directionally (you can "look" thru it both ways, from the brige towards a regulator and from the regulator to the bridge, backwards) so it will decouple any surges/hash happening on the LED rail from other rails (adding a resistor could work as a "bridge" in single sided PCBs).
I would advise that you make a separate "ground" rail for the "dirty" return currents from the LED rail. If you must reference control voltages from sidechain etc, you can use "ground sensing" (with an opamp/diff-amp) to decouple the clean "audio" ground from this "dirty" LED ground, where the two would otherwise need a hard link.
Harpo
Well-known member
Why dirty ground at all? Powering relais and LEDs from a floating maybe 24V regulator across bridge rectifier with regulator both sides input/com voltage dropping/heat sharing resistors in front should keep 0V clean and keep any hash in front of the audio rail regulators.tv said:
Harpo
Well-known member
Good call. Might get tricky with circuits referencing audio gnd, but as you already said, it's not our project...
>>> Might get tricky with circuits referencing audio gnd
Not at all! In a "GSSL" compressor, you have negative CV, so for the GR meters etc you would need to invert it - which is an ideal place for "ground sensing" the "clean" ground via the noninverting input of the opamp that "processes" the CV to be used in the LED meter...
Not at all! In a "GSSL" compressor, you have negative CV, so for the GR meters etc you would need to invert it - which is an ideal place for "ground sensing" the "clean" ground via the noninverting input of the opamp that "processes" the CV to be used in the LED meter...
Harpo
Well-known member
Sure, but I was thinking the floating supply. The usual LED metering suspects LM391x don't seem to have differential inputs and the chip internal diode at signal input buffer pin5, V- pin2 and RLO pin4 seem to connect to the same potential, usually gnd. This supply then would not be floating anymore and LED currents get dumped to audio gnd again.tv said:
