Pinball Switch Matrix 

The current isn’t flowing backward, it’s just finding alternate paths to ground. If we apply current to one row at a time, and we ground one column at a time, and if we see current getting to ground on that column, then we know that the switch at the intersection was closed. But that only works if the current is only allowed to flow through that one switch. If current can flow through other switches, then it can flow AROUND the switch at the intersection and find a path to ground through any other closed switches, and it looks like the switch at the intersection was closed even if it wasn’t closed. To prevent that, we want to make sure that current can only flow from row to column, not from column to row. Diodes ensure that. Check out 19:20 for an example of current finding an alternative path to ground and how diodes block that path.

@@kyle5574 watch at time 24:30 you mention when the spinners switch is CLOSED the DC current can flow from Column to Row its colored in RED, which is not true because DC current can only flow from +Vcc to ground, from from ground to +Vcc. Why are you saying that the DC current is Bi-Directional that it flows from Row to column in GREEN and can also flow from column to Row in RED?

The current is still going from source to ground, even when it's going from column to row across a switch. Look at the picture a few seconds later. Look at switch 21 (upper right in the picture). The rest of the switches are fine, but the current is flowing "backwards" across switch 21, from column to row. Since we don't have diodes in this picture, the current is allowed to flow in either direction across the switches. In this case it's finding an alternate path to ground, around the open switch, and through 3 other closed switches. If the switches had diodes, the current would only flow across the switch in the correct direction and there would be no path to ground in this case.

@@kyle5574 Yes true the current is going from +Vcc source to ground even when the current is going from Column to Row across a Switch, ( this isn't considered backwards current flow, its considered an alterative path to ground by using column to row instead of using row to column). When DC current is flowing from Column to Row across a switch it causing "Ghosting errors" that is what other techs call it. The diodes don't prevent Anti Ghosting errors so if multiple switches are CLOSED the CPU has to register ALL the closed switches but the diodes don't fix or prevent Anti Ghosting errors. I don't think you have made a video lesson about Anti Ghosting errors for pinball switch matrix circuits.

my question is whether i need a diode between the pos and neg trigger to the relay timer - jz801 timer relay.. it has reverse prevention built in which may do the job ?

The JZ801 might already have a flyback/snubber diode across the relay coil. If you can't find schematics for the board, maybe measure resistance across the relay coil terminals on the board and see if the resistance varies depending on the polarity of your leads. If it doesn't have a diode, then, yeah, I would add one. I don't think adding an independent ground on a switch is going to work the way you want though. Maybe think about adding your circuit in-line with the switch, so the wiring goes: row--diode--switch--circuit--column That way when the switch closes, the current has to go through your circuit to get to ground on the column. Keep in mind the switch matrix is scanned many times per second, so the relay is going to try to rapidly open and close whenever the switch is held closed. The timer might not handle that, or the duration of the scan might not be long enough to hold the relay closed. If you have trouble with the relay, I'd look into an optoisolator. They might react faster than relays.