deskthority

Was hoping for some help figuring out a dead column my RS84, I've done some preliminary work, but am out of ideas. -

-Continuity tests show that the individual switches are connected to each other and a pin on the controller. And lower pin shows connection to the controller if the switch is activated.
-Switches de-soldered and res-oldered.
-Firmware and layout reflashed.
-No response in Switchhitter.
Tools I have on hand: Decent MM, soldering iron, desoldering iron.

At this point I'm thinking that the problem might be with the controller itself, but am unsure.

I had the exact same problem (two columns not working but everything connected), and it came from badly soldered pins. Try to press the pins while the PCB is connected to see if that helps.

I did not get any reaction pressing the pin. I went ahead and re-soldered it a bit as well. Still no activity in that column.

You might check continuity between the pins of that column and ground/earth to see if you might have a short on that column.

Does your MM also measure frequency? Then you could watch the controller as it scans the matrix. You would expect a healthy controller to scan this column with the same frequency as the neightbouring columns.

Engicoder wrote: ↑You might check continuity between the pins of that column and ground/earth to see if you might have a short on that column.

There is continuity to ground on that column and no other. Investigating further.

OleVoip wrote: ↑Does your MM also measure frequency? Then you could watch the controller as it scans the matrix. You would expect a healthy controller to scan this column with the same frequency as the neightbouring columns.

It does measure Hz, but that's not something I've done before so I'm not sure what to probe.

bloo wrote: ↑There is continuity to ground on that column and no other. Investigating further.

Aha! Could be shorted to ground, as Engicoder suggested, or a broken controller output.

bloo wrote:It does measure Hz, but that's not something I've done before so I'm not sure what to probe.

WIth the board up and running, measure the Hz for this and the other column lines. Depending on your MM, it might work only against ground or against Vcc or not at all, because these are short pulses. I'd expect some hundred Hz or kHz.

I don't think that I am getting reliable frequency reading using my mm. Regardless of row or column I am getting 0Hz from the upper left pins on each switch and 60Hz from the lower right. I'm also getting 60 hz from just using one probe to the ground, so I think it's pretty dodgy. MM specs: http://www.kleintools.com/sites/all/pro ... nglish.pdf.

In any case, could I simply try to run bypass wires for work around the short (assuming that's the problem). I have some idea how matrices work, but I am unsure of the particular routes in this instance.

It seems that you do get a reliable frequency reading — of the mains frequency, not the scanning frequency.
It could be that the duty cycle is too short for the MM to accept it.

bloo wrote: ↑In any case, could I simply try to run bypass wires for work around the short (assuming that's the problem).

If you really have a short, this won't work since, conversely, the short will sort-of bypass your wires. You'll have to check very carefully with a magnifying glass the places where you have soldered all along the column trace. Is it possible to get the PCB layout document for this board? There you could identify the dodgy spots where ground traces pass near that column line. Where pins go through holes, the short can also be on the other side — especially if you desoldered and resoldered the switches. If you still can't find anything by visual inspection, I'm afraid there's nothing you can do but isolate the problem by removing the switch which is electrically nearest to the controller and cut the trace behind it so that just the pad of the first switch is connected to the controller. Later you can repair this easily by running a wire between this pad and the next switch.

Clean this pad from tin, check also the other side of the PCB. If it still has ground connection, I'm afraid you'll have to replace the controller.

I understand. Thank you.

I think that the board originated from china and I haven't gotten a response from the GB leader so I'm not sure about documentation or schematics.

At this point I'll give it one last go. I'm pretty sure that I can visually map the traces back to the controller. It is plate mounted so I'll have to de-solder everything and strip it down to get access to the other side of the pcb, but I'll need to have it back in pieces if I part it out in the end anyway. If it ends up being a controller problem I'll cut my losses; I dont have a setup or eyes good enough to solder something that detailed.

In any case, thanks for the help, people