melted connector last time got loose inside?
which pins are constantly at 5V? and are they shorted to 5V power rail (resistance measurement with system off)?

rasz_pl wrote on 2024-03-02, 13:20:

melted connector last time got loose inside?

Connector looks fine.

rasz_pl wrote on 2024-03-02, 13:20:

which pins are constantly at 5V? and are they shorted to 5V power rail (resistance measurement with system off)?

Almost all on the top side, of the image I posted. Not sure that I understood what to measure? But, it can't be that all pins are powered with voltage.

There's a difference between a pin being directly connected/shorted to +5, or pulled up to +5 via a resistor. A pull up resistor means that if the pin isn't being actively driven by the chip then it'll measure at +5V, but the chip can drive the pin to Gnd. For example, on the PS/2 connector, one pin is directly connected to Gnd, one to +5V and 2 pins (clock and data) are pulled up to +5V. So the Clock, Data and power pins will all measure as +5V when no data is being transmitted, but only the power pin is directly connected to +5.

A way to check is to turn the board off and measure resistance between each pin and a known +5 point (easiest on the power connector). I'd guess that one or two will measure 0 ohms, but others something like 1k - 10k. At the same time it's probably worth checking your PS/2 connector and measuring the resistance to +5 of the clock, data and power pins, and resistance to Gnd of the ground pin.

snufkin wrote on 2024-03-02, 19:06:

There's a difference between a pin being directly connected/shorted to +5, or pulled up to +5 via a resistor. A pull up resistor means that if the pin isn't being actively driven by the chip then it'll measure at +5V, but the chip can drive the pin to Gnd. For example, on the PS/2 connector, one pin is directly connected to Gnd, one to +5V and 2 pins (clock and data) are pulled up to +5V. So the Clock, Data and power pins will all measure as +5V when no data is being transmitted, but only the power pin is directly connected to +5.
A way to check is to turn the board off and measure resistance between each pin and a known +5 point (easiest on the power connector). I'd guess that one or two will measure 0 ohms, but others something like 1k - 10k.
At the same time it's probably worth checking your PS/2 connector and measuring the resistance to +5 of the clock, data and power pins, and resistance to Gnd of the ground pin.

PS/2:

Ground vs Data Pin = infinite -> this is strange?
Ground vs Ground Pin = 0 Ohm
Ground vs Power Pin = 8.61 kOhm
Ground vs Clock Pin = 9.38 kOhm

Update 1
Resistance on controller:
All pins are in range from around 1.3 kOhm till around 9.5 kOhm except:
Pin 4, pin 10 & pin 11, pins from 23 till pin 33 and pins from 37 till pin 40 have infinite resistance
Pin 8 = 15 kOhm

Update 2:
I didn't measure data pin correctly on PS/2 connector first time. It is ok, it's not infinite resistance, but around 9.61 kOhm. I was measuring not connected pin. Stupid me.
This means that the PS/2 connector is fine, I guess, but I would like someone to confirm it.

Does all of these measures sound like faulty chip?

Update 1:
The keyboard works ok now. What a hell did happen?

Ok, narrowed down the problem. It is either connector again or keyboard cable. Because there is error again and moving the kb cable a bit fixes it. Will investigate further.

boby wrote on 2024-03-03, 23:20:

Ok, narrowed down the problem. It is either connector again or keyboard cable. Because there is error again and moving the kb cable a bit fixes it. Will investigate further.

"melted connector last time got loose inside" 😀 When soldering connectors its very easy to slightly melt them inside by keeping soldering iron for too long on legs. Its invisible from the outside, but inside connector you no longer have snug fit. It can still be keyboard cable tho.

rasz_pl wrote on 2024-03-04, 01:58:

boby wrote on 2024-03-03, 23:20:

Ok, narrowed down the problem. It is either connector again or keyboard cable. Because there is error again and moving the kb cable a bit fixes it. Will investigate further.

"melted connector last time got loose inside" 😀 When soldering connectors its very easy to slightly melt them inside by keeping soldering iron for too long on legs. Its invisible from the outside, but inside connector you no longer have snug fit. It can still be keyboard cable tho.

It is the connector. Never cross my mind that I could damage it with iron.
I put some WD-40 and for now it works. Even if I move cable around, it doesn't
brake.