Pi pin out connected to pin in on other PI

[BoyLens]

I have one Pi which needs to respond to another Pi's change of a certain pin.

So let's say I have pin 7 from one connected to pin 7 of the other. This is al working; toggling the pin allows the other one to see it change.

But after a while, it looks like the second Pi (getting the input) is being destroyed. We changed it 2 times now. Is this setup maybe not allowed?

The first Pi has quite some more devices on it, via GPIO expansion boards (https://www.abelectronics.co.uk/p/54/IO ... ame=0&c=18). First time we destroyed the Pi was because I connected it to one of those boards, but my hardware guy told me that should not be done because the output is 5v instead of 3.3v. But now it seems it is damaged again (3rd time).

Any ideas?

[joan]

Have you connected the grounds?

A photo may be useful.

[BoyLens]

Sorry for the late reply! I thought I would get a notification when someone responded.

No, the second Pi is not grounded, could that be the issue?

[ElEscalador]

Both Pi's must share a ground (AKA - negative power supply connection) to complete a circuit in order for data to flow.

[BoyLens]

so that could damage the hardware, if it is not grounded? We are experiencing hardware issues and still trying to find the cause.

[joan]

I doubt a lack of a ground between two Pis would cause hardware issues.

A lack of a ground between two Pis would make communicating via the GPIO unreliable.

Connect pin 6 to pin 6.

[drgeoff]

I doubt a lack of a ground between two Pis would cause hardware issues.

A lack of a ground between two Pis would make communicating via the GPIO unreliable.

Connect pin 6 to pin 6.

I can envisage situations where a lack of a ground connection between two RPi which are independently powered by non-grounded PSUs could damage the GPIO input.

Totally agree that lack of a ground connection between them would make communication between them unreliable.

[procount]

Make sure the signal wire, but especially the grounds, between the 2 RPI are connected before powering them on.
If they are powered (for example) by 2 separate DC adaptors, there is a risk that both grounds are at different potentials (voltages) relative to each other.
Connecting them together when powered on with different potentials means a large surge of current will flow as the potentials try to equalise.
This voltage/current can be high enough to damage GPIO pins, or even blow a hole in a chip package (letting the magic smoke out) as I found to my cost in my early days!
It may not happen every time, but when it does, you tend to remember not to do it again.

The length of the wires between the 2 RPI may have an effect too, as long wires can pick up additional noise and present extra load and capacitance to the pins.
If they are excessively long, consider twisting the signal and ground wires around each other to reduce any interference pick up.

Or consider using the serial ports to connect the RPI together. Those circuits tend to be designed to be more robust, especially if you put appropriate line drivers between the two devices.
Using a serial connection can make your system more stable because you will need to receive a specific message instead of just detecting a voltage level, therefore it would be less prone to false triggering, although obviously it is a little more effort to program. Just some of my rambling thoughts that might help.....

[BoyLens]

Thanks guys! I'm going to look into this and consult the hardware guy (and reference to this thread)

We did see issues that the GPIO output was sending high/low for no reason, but when we connected the monitor to the sending Pi, the issue was gone. So it was grounded via the monitor. The second Pi is connected to a TV.