Mesa 7i77 Outputs stuck on?

Running Mesa 6i25 + 7i77.

I have all jumpers configured correctly.
I'm supply 5v power via power supply, and sharing field power and VIN for 24V power.

I'm running 5i25 + 7i77 firmware and I have the 6i25 jumpers configured for 33mhz to be 5i25 compatible.

Problem:
When I run LinuxCNC and I activate a digital output on the 7i77, the output seems to STAY ON, even when deactivated.
Each time I power cycle the system it behaves correctly once, and then repeats the behavior.
If I monitor the OUTPUT in LinuxCNC via HalMeter I can see it turn on and off, but measuring with a multi-meter on the actual board shows 22v, 18v, etc coming out of the board on a specific output.
I DOUBLE checked the 24v relay I'm connected to and it says it should take 50-100ma to switch the relay.
As the 7i77 supports up to 350ma for outputs this shouldn't be a problem.

I have a backup 7i77, and I tested with it, and they both behave the same exact way.
Nothing is shorted,
nothing is over-voltaged,
24v power supply is working correctly, with no switches, breakers, filters in-between.
the inputs work just fine.
I was VERY careful to ensure there is no reverse voltage or other issues anywhere else.
What do I have mis-configured to cause this?

Just a mute point but Re:

I DOUBLE checked the 24v relay I'm connected to and it says it should take 50-100ma to switch the relay.

Have you or is there fitted a flyback diode across the relay to stop back EMF as that could damaged the board.

I think the manual states any inductive load above 40mA requires it.

Note that you cannot check the 7I77 outputs with a 10M Ohm input digital multimeter
since the outputs have a bias source of a few uA for detecting loss of load.

Basically this means, your 7I77s are fine, you just need to have a real load to measure the outputs.

Clive wrote:

Just a mute point but Re:

I DOUBLE checked the 24v relay I'm connected to and it says it should take 50-100ma to switch the relay.

Have you or is there fitted a flyback diode across the relay to stop back EMF as that could damaged the board.

I think the manual states any inductive load above 40mA requires it.

Thanks Clive,
No I don't have diodes installed on the outputs. And it makes sense to me now that you would need them.

The original controls (before my retrofit) must of had some sort of protection built-into the board itself.

Can I get away with wiring a diode in series with the output from the card? I realize the coils potential has to go somewhere but this is also how the machine was running for years before I came along. I'm assuming a series diode would at least protect the 7i77?

For future reference are there any Digital I/O you're aware of that have this protection built-in?

Thanks for the help.

No, the overload and overtemp bits are monitored (and cause faults) but not the load presence bits

Note that the 7I77 outputs are fine for driving relays up to 60 mA or so without
a clamp diode.

The driver chips are designed for this service (relay/inductive load driving) and
have built in clamping (at about 40V).

The disadvantage of this on chip clamping is that the stored energy in the coil
is dissipated in the driver chip itself so you can overheat the chips if you switch them
very frequently say 100 Hz (as tommylight mentioned). This will be detected and the chip
will enter thermal shutdown so will not cause damage.

Diode clamps have the advantage that most of the stored energy in the coil is dissipated
in the coil itself, not the driver chip. This is the reason that diodes are required for higher
current inductive loads, the driver chip cannot dissipate the stored energy of larger loads.