Cycle start/resume pause and external buttons problem (Mesa 7i76e card)

Here are three attachments for the hallscope captured 4 signals (start button input pin..input-00, idle-on, run-on and run-sig)  when the machine starts itself: 

So at a quick glance it looks like that this is unrelated to the 7I76E signals and just
an issue with the logic.

(you might check at a lower time per division to make sure you are not missing a
~1 servo cycle glitch)

Here is an attachment for the photo of the probed 8 signal for :
 run-on
 hm2_7i76e.0.gpio.028.in
 hm2_7i76e.0.gpio.031.in
 hm2_7i76e.0.gpio.043.in
 hm2_7i76e.0.gpio.044.in
 hm2_7i76e.0.gpio.045.in
 hm2_7i76e.0.gpio.046.in
 hm2_7i76e.0.gpio.048.in

It does look like the GPIO is detecting some noise.
I think that you need better electrical installation, but you can probably mask this with a debounce function in HAL.

It isn't 100% clear, but is something happening that should happen if a button is pressed, but without a button being pressed?

As PCW says, wiring direct to FPGA pins is rarely a good idea.

Let me explain further, my MESA 767ie expansion connectors are connected with CNC4PC's C25 breakout board, that is the external button lines go through 74HC14N IC first before entering to Mesa card

Having further done the followings but all in vain (that is the cycle start button still turns on itself):
> setp deounce.0.delay15~150 and letting those gpio's input to and go out of them
> adding 47k ohm pull up resistor on those pins
> adding two 0.1uF capacitors on each those pins

Before using the Mesa card and LinuxCNC these couple of years, I used to use Mach3 Ethernet SmoothStepper card for a long time, it haven't this noise interference problem. For the Mesa card, is it really this problem this time?
As I have done all the hardwares: include building the keyboard pcb, key button switches and panel housing building. Those external buttons are of momentary switch type. I might need to change all of them. Change to normal closed type, that is they are normally at 0V voltage level, so the randomly positive voltage noise don't affect those input lines. Will this method work? Not yet test and i don't know? Please advise? There is no momentary switch of normal close type, I need to change them to toggle switches (2 positions spring back type) or limit switch (as it has NO and NC position for selection to use) but need to add separate piece of plastic button cap on top of it...Anyway, trouble...

Thanks for the reading, my English is not good

Missed to remind, CNC4PC's C25 breakout board socket pin outs do not match with 7i76e expansion connector pin out, i need to swap most of them when using it