MESA 7i76U inputs changing state without changes to config

I have a DIY RouterMill that I've been fighting with for many years. I recently ditched MACH3 and paraports for the MESA 6i25 <=> 7i76u combo.
After flashing the 6i25 with the 5i25/7i76u x 2 bitfile and updating LCNC to 2.94 and configuring with PNCConf, all of the steppers worked as expected and were tuned for optimum movements for my machine. Turning my attention to the INPUTS, based on what I understand from the docs, the INPUTS are sinking so I configured the machine with all INPUTS as follows:

VF(24v) -> switch(NC) -> 7i76.0.0.input_nn

After starting LCNC, I notice that the ESTOP was triggered but, the button was not closed. Checking the INPUT pins with halshow, I could see that the board was showing all INPUTS as TRUE, therefore I changes all of the INPUTS to input-nn-not and restarted LCNC. Now all of the INPUTS seemed to respond correctly. When I attempted a HOME_ALL, the Z moved to positive, so I reconfigured it to home negative with PNCConf. That worked and after more reading and adjusting of some parameters, I got the Z homing correctly backing off the trigger position by .050 in.  Then I directed my attention to X.  However, I had to adjust a few more parameters for the X axis then restarted LCNC. *)#$!! All of the inputs had switched and were now reporting as TRUE. Frustrated I changed all INPUTS back to input-00 and restarted LCNC. With this session I was able to get both the Z and X axes to home correctly, however, Y was now the problem child as it would not back off the trigger position. After several attempts I finally decided to swap DIR-/DIR+ at the microstepper so the Y config was exactly as the X config. (My Y axis is at the back of my machine so I had reversed the motor in the config for correct movement). After making the change and while running PNCConf I confirmed that the Y axis was now responding correctly then proceeded to resume verifying  the HOME ALL function, The attempt failed, so I tweaked the HOME parameters but, when I restarted LCNC all of the INPUTS had again switched their state (*)#$!!,*)#$!!). More bizarre was that now no matter how I set the PROBE input (input-nn or input-nn-not) the halshow remained the same.  After several additional restarts, the input pins again flipped the input state.

At this point I'm half inclined to suspect the 7i76 board as it has one INPUT, input-04, that is always stuck in the same status in halshow much as the PROBE input is now responding.

If anyone has any ideas that I can try, I would greatly appreciated the support.

Regards,
Stephen

Very odd, the only thing I can think of is that
there's something wrong with the supplied
field voltage (not filtered?) or the field voltage
sensing on the card is faulty.

What is your field voltage and what does
the 7I76 field voltage hal pin read?

You would need to have the sserial mode set to 2 to
be able to read the field voltage hal pin, so in the
hal file, change:

sserial_port_0=00xxxxxx

to:

sserial_port_0=20xxxxxx
 

I'm using a dedicated 24v switching power supply which is holding a constant 24.1 output. Setting the sserial_port_0=20xxxxxx and reading the fieldvoltage I'm seeing a constantly fluctuating value from 22.01 - 23.9 v

That does seem like a lot of ripple (nearly 2V)
that should not directly cause an issue but its
odd.

Probably the card should be returned for evaluation/repair/replacement

 One other clue might be to look at the analog input hal pins.
They should be very close to 0V with nothing connected.

The way the 7I76U digital inputs work is that they are read as
analog voltages and the threshold/hysteresis is set by the
field voltage. That's why I suspected a field voltage sensing issue.
 

I did check the analog inputs which are not connected. All are sitting at exactly 0v.

Thanks for responding Tommy.

As I mentioned the 24v power supply is holding steady at 24.1v. I've had my DVM connected for several hours without a single deviation, so I'm confident the PS is good. I've just sent the results of my findings to MESA tech support to get their input on the issues.

Well, tested the card and found no issues.

All inputs read low at 50% of field voltage and high at 65% of field voltage
applied to an input.

Heated to +70C and cooled to -40C and still no issues. Ran at 5V to 32V field
voltage and no issues.

Here is an image of all disconnected inputs in halshow (@ 24V)

 

 

I'm just tuning in to this discussion, so maybe I missed this, but: On my Mesa 7i96 the Mesa inputs aren't like most digital inputs. They're not self-referenced. They all share a "common" pin, which should be grounded if the inputs are active high, or pulled high if the inputs are active low. Is it possible your 7i76 inputs are the same? Have you done that? If you don't, the your input might not be floating, but its reference is, and could easily lead to erratic/intermittent behavior.

paul