squarewave

Re: Bi-color leds and shift registers
« Reply #40 on: January 13, 2021, 02:23:20 PM »
Just to be clear, so relay current is returned on the cable shield?


boji

Re: Bi-color leds and shift registers
« Reply #41 on: January 13, 2021, 04:24:12 PM »
Quote
Just to be clear, so relay current is returned on the cable shield?

Relays are normally energised while unmuted/unsolo'd. They return to dcom via the relay daughter boards housed within each of the channel/group cards (no current on i/o pins). 

When cv pins on daughter boards are pulled low via i/o cable, the relays de-energise, and 'digital' voltage source returns over i/o cables to logic pcb. BC550 emitter delivers about 900uA to pour.

Edit:  Going to assume then max draw (all buttons in) pulls about 14mA total per bucket. Probably didn't need the transistors, as led's pull far more directly off their respective i/o... ::)

Buffers for nothing...when BC chips are free. :-[

« Last Edit: January 13, 2021, 05:30:32 PM by boji »

squarewave

Re: Bi-color leds and shift registers
« Reply #42 on: January 13, 2021, 07:05:30 PM »
They return to dcom via the relay daughter boards housed within each of the channel/group cards (no current on i/o pins).
So supply for the relays is delivered by the control cable but returns to the power supply by a completely different "dcom" path. That's not ideal because sudden changes in current will emit an electromagnetic field that could invoke currents in nearby high impedance and / or high gain circuitry. For example you could get spurious digital events like you engage / disengage a relay and a bunch of LEDs flip erroneously. If return followed signal, the electromagnetic fields would cancel and that wouldn't happen. The caps on the transistor bases will probably slow things down enough that it will be ok but you might try to not run those lines next to digital lines if possible.

boji

Re: Bi-color leds and shift registers
« Reply #43 on: January 13, 2021, 07:56:22 PM »
Quote
So supply for the relays is delivered by the control cable but returns to the power supply by a completely different "dcom" path.
Thanks again.

Just checked current draw of com pin of the relay sub-pcb.
Active pins are 12v, dcom, and cv.  When relays are energised, 23mA goes to dcom at console backplane, and source follows sink.
When cv is made active, relay sub-pcb dcom pin current drops to almost zero. So it is as you say, cv does becomes the only return path during relay de-energising. So both dcom and cv should physically follow source, even if it means extending source however far to cv's interface?

Quote
you might try to not run those lines next to digital lines if possible

Indeed, relay cv definitely gets their own shielded control cable (2 per bucket), distinct from button sensing & illum cable runs.
« Last Edit: January 13, 2021, 07:59:30 PM by boji »

squarewave

Re: Bi-color leds and shift registers
« Reply #44 on: January 13, 2021, 09:40:01 PM »
Actually I was wrong about the direction of current in my previous description. I see that your relay transistors are current sinks and not current sources. Of course the direction of current doesn't matter but because I'm not sure I understand the identifiers you're using to describe your circuit, I will just restate my concern:

If the current used to power relays is not supplied by conductors physically adjacent to the conductors sinking that current, then the "source follows sink" rule is not satisfied. And because all 8 of the conductors in what appear to be RJ45 jacks are occupied by current sinking conductors, the only conductor in that cable bundle that could supply the source current would have to be the cable shield. Meaning unless the cable shield is supplying the 12V used to power the relays (and it most certainly should not), then it must be coming from somewhere else and that is not ideal. But again, it's not a show stopper because the caps on the transistor bases will limit the intensity of the magnetic fields generated when a relay switches on.

boji

Re: Bi-color leds and shift registers
« Reply #45 on: January 13, 2021, 09:50:17 PM »
Quote
I'm not sure I understand the identifiers you're using to describe your circuit
Perhaps I should make a block diagram as I'm probably not using the correct descriptive terminology.  Thanks for your time, still good food for thought.


boji

Re: Bi-color leds and shift registers
« Reply #46 on: January 14, 2021, 05:52:09 AM »
Quote
the only conductor in that cable bundle that could supply the source current would have to be the cable shield.
I think this is where the confusion lies, as no positive voltage needs to get sent out of the M/C logic pcb to drive relays. Their 12v source comes from the channel pcb's. The logic pcb only serves to sink cv so to disengage relays, and both 12v and 5v sources eventually share dcom bus.

I do see your point about proper EMI practice.  :o  Unfortunately don't know of any other way to satisfy it except to possibly take the cv return going into logic pcb (racked under desk) back out of rack to relay sub-pcbs.

As is, the plan was to join the four logic pcb dcom's together within the rack, and then send it over one cable back to the dcom bus bar at the end of the desk.

Mostly a redraw, but hope it clarifies:




One of the relays at the front of channel / group pcb's:

« Last Edit: January 14, 2021, 06:48:25 AM by boji »

squarewave

Re: Bi-color leds and shift registers
« Reply #47 on: January 14, 2021, 11:53:26 AM »
Oh, that's no problem. I thought those transistors were sinking the current of the relay coil. But if you're just using them as an inverter (not sure why that's necessary since I would think you could just change the code but whatev) to send a control voltage then there's no significant currents over those cables. Carry on.

boji

Re: Bi-color leds and shift registers
« Reply #48 on: January 14, 2021, 03:18:49 PM »
Quote
you could just change the code
True. They even have a provision to invert a side's pins all in one go.

Quote
Carry on.
/Salute
PCB's ordered.
« Last Edit: January 14, 2021, 04:57:35 PM by boji »

boji

Re: Bi-color leds and shift registers
« Reply #49 on: February 19, 2021, 05:26:05 PM »
Is there ever a valid use for delay()?

For example

Code: [Select]
while (!Serial) delay(500); //wait 1/2 second for console voltages to stabilize

I'm figuring that the pausing of code is desirable since the code ought not proceed until serial is available, which ensures debugging notifications.
« Last Edit: February 19, 2021, 07:20:09 PM by boji »


squarewave

Re: Bi-color leds and shift registers
« Reply #50 on: February 19, 2021, 07:44:06 PM »
Is there ever a valid use for delay()?
I would never use delay(). Your code is no different from just:

  while (!Serial) {}

But I wouldn't do that either. And I assume !Serial is pseudo code. There's probably a way to detect if the serial port is ready by checking some register of some sort.


 

Related Topics

  Subject / Started by Replies Last post
1 Replies
2083 Views
Last post August 04, 2004, 07:52:49 PM
by volki
15 Replies
3977 Views
Last post August 23, 2004, 02:54:18 PM
by
8 Replies
11686 Views
Last post October 04, 2006, 06:16:19 PM
by Brad McGowan
1 Replies
1350 Views
Last post October 26, 2007, 03:28:34 PM
by CJ