|
Post by svart on Jun 30, 2020 21:43:01 GMT -6
So I've come back to this again.
I finally got a quality triple bench power supply to run the pieces.
So far nothing burns when I power the boards up!
EQ works, but one pot is hooked up backwards so the frequency range is bunched up at one end and backwards. Not sure how that happened..
Compressor doesn't pass signal through the vca, so perhaps the sidechain has a stuffing error, but I didn't really dig into it yet.
Haven't tried the interface PCB or the IO PCB yet. I ran out of solder so they're not quite finished yet.
|
|
|
Post by svart on Jul 3, 2020 20:48:08 GMT -6
Ok, compressor debugged. My original issue was that I had forgotten the convoluted signal flow through my switches.. It happens.. The switching for sending the signal to the insert can be switched to before or after the compressor. The sidechain can also be taken from before or after the board on a separate SC bus that can be assigned from anywhere in the strip.
I had simply forgotten that I needed to hook the Return to the Send for the signal to come out of the board's connector in BOTH modes, not just for one mode.
But I also found a PCB error in that I routed a trace too close to a through-hole and the annular ring bled over to the trace during etching at the factory. I'll need to edit the PCB file to get a little more clearance, although DRC didn't catch this error for some reason even though it caught other clearance errors. This was causing signal bleed even without power, but with power it was causing the output opamp to "fight" and draw a lot more current.
The interface PCB which contains the main interface to the bus cables as well as the muting and routing circuits seems to work. I haven't brought up the digital section yet as I need to figure out how to trigger the mute FETs without the data bus from the controller in the X2 mixer.
|
|
|
Post by matt@IAA on Jul 3, 2020 22:07:15 GMT -6
Ugh. Once I had a ground pour offset that didn’t get caught in DRC and effectively muted 2/3 of a circuit. I feel your pain.
Also, that level of flexibility in routing is awesome.
|
|
|
Post by svart on Jul 3, 2020 23:19:14 GMT -6
Ugh. Once I had a ground pour offset that didn’t get caught in DRC and effectively muted 2/3 of a circuit. I feel your pain. Also, that level of flexibility in routing is awesome. I've definitely found that I like certain channels with EQ/filters before compression, some after. I didn't opt to have a ton of routing buses in the channel strip, so the compressor is positioned after the EQ section for simplicity's sale, but the sidechain has it's own bus and can be sourced from anywhere in the channel strip. I'm not sure this is ultimately more useful, but should add some interesting abilities.
|
|
|
Post by matt@IAA on Jul 4, 2020 10:19:30 GMT -6
That’s really clever. I never thought about bussing a sidechain input.
|
|
|
Post by svart on Jul 4, 2020 11:08:22 GMT -6
Playing around with the silkscreen I made for this. I think the ink is not mixed well enough or the silk doesn't have high enough thread count because the middle lettering got smeared while the ends didn't which I believe means it's bleeding through too much. Anyway, there's changes I need to make to the artwork so a new screen needs to be made soon once I make sure nothing else needs to change.
|
|
|
Post by svart on Jul 5, 2020 22:05:21 GMT -6
|
|
|
Post by johneppstein on Jul 5, 2020 22:20:35 GMT -6
Very nice!
|
|
|
Post by svart on Jul 8, 2020 11:42:39 GMT -6
That’s really clever. I never thought about bussing a sidechain input. Here's the diagram of the routing for the sidechain and whether the compressor is pre or post insert:
|
|
|
Post by svart on Jul 10, 2020 11:21:41 GMT -6
Tested the compressor once again after finding another issue. A ground was left of of the netlist for some unknown reason. It shows if I highlight the whole netlist, but will not show in unconnected nets in DRC for some reason. I've changed it and imported it a few times and it just won't show up.
Anyway, I manually added it to the layout and wired it in on my test board. The gate release seems to work better, but I'm having a strange issue with the gate in general.
The Range control seems to be backwards to the silkscreen on the channel faceplate and the threshold seems to be be a lot more touchy than it should be which would first indicate that they might be backwards.. The potentiometer on the SSL schematic has CW markings on all the pots, and I've triple checked that my schematic and layout follow the SSL schematic in direction.
So either those two pots on the original schematic are wrong, or something else is happening that I don't understand.
|
|
|
Post by svart on Jul 10, 2020 13:57:42 GMT -6
As for the Range pot, it looks like at least one of the former attempts at cloning the 4k channel compressor ran into the same issue as it has the pot reversed in the final design.
|
|
|
Post by svart on Jul 14, 2020 8:17:05 GMT -6
Did some more testing over the weekend. The Range and Threshold pots for the gate work backwards to the original SSL design. I've checked my schematic against the SSL schematic and it's exactly the same. I checked my layout and it follows the schematic.
So at some point SSL must have changed this, or wired it differently from the schematic.
|
|