Bridgeport Retrofit - keyboard input delays noted

Thanks again andypugh. You did my research for me once again....

I did manage to fumble my way through getting a backlash component installed and loaded into my custom hal file. Feeling rather giddy about that. You did have a few typos but I figured them out from the error messages when trying to install.

Now I just need to use it. I just love breaking new tools.

Mark

I assumed that. I appreciate you writing this piece for me. It took some looking to figure out this is written in C. May be obvious to many but not to me. I have very little programming experience and never even looked into C. I think i follow your logic though. Maybe...

I rest my case your honor. Anybody who can't even sort out the language has no business writing his own hal components. Yet.....

I am pleased to report that using the backlash component cleared the dither noticed on the scope. I rewound back to my original custom hal and inserted the dither compensation (backlash) function and have not been able to reproduce the original symptoms (keyboard delay). I also don't see any unexpected debug messages showing tasks which I didn't command. And finally, when halscope shows dither in the hm2_5i25........analogin0 pin it does not appear at the output of the backlash.0.out.

I have not backed out the change to axis to comment out those few lines when contention (2 different inputs used to control a single function) was being considered. I think I will do that and then go make some parts and conduct a thorough shake down cruise.

Thanks again for all of the help.

Mark

I though I might mention that I tested a couple 7I77s for analog input noise and the noise is less than 1/2 LSB.
so if you are getting 2 LSBs of noise (~280 mV) this is likely from an external source. You may need shielded wires
and careful wire routing (away from noise sources) to get minimal input noise. The analog inputs on the 7I77 are filtered but
have a fairly high bandwidth (20 KHz) so are susceptible to high frequency pulse noise from axis drives, VFDs etc

I don't have a real o'scope available so I can't really look at the electrical noise into the 7i77. I can tell you that the wires to the pot supplying 24V and from the wiper are less than 20" long total distance. I have the 7i77 and the computer mounted in a metal enclosure (old heidenhain pendant) and not physically near the servo drives. No vfd is used. I am still using the original KTK spindle drive which is in another metal cabinet on the other side of machine.

I am using the original Interact 24V regulated supply that resides in the axis cabinet which supplies the original 24V relay logic to provide the field voltage on the 7i77 also.

As part of this adventure, I installed some bypass caps (.1 and .01uF) directly onto the 7i77 input pins (used TB-6 pn 1 as signal gnd) hoping to eliminate noise (again no scope so t/s in the dark). I put in additional 1000uF electrolytic on the 24V supply thinking I had high ripple. When neither of those had any noticeable effect, I ran a 3 term 12V regulator to supply the pots. As a last resort I stuck that 9V battery onto the mesa input pins. None of those changes had any noticeable effect on the observations.

Not saying there is no noise. Just clarifying my physical configuration, methods and observations.

I might just route some shielded cable to and from the pots as you suggest.

Thanks for the interest and assistance.

Mark

TB6 pin 1 would not be a ground for the analog input pins
only TB2 pin 8 is a suitable ground

The 9v battery was between TB2-8 and analog input.

I can redo the bypass cap scenario with proper ground this time.

Mark

PCW wrote:

I though I might mention that I tested a couple 7I77s for analog input noise and the noise is less than 1/2 LSB.
so if you are getting 2 LSBs of noise (~280 mV) this is likely from an external source.

I just re-read this and have a question. In my halscope screenshot attached, does the ~150mV spikes represent 1/2 LSB? If so, I am chasing gremlins. I am not seeing ~280mV transitions with any regularity.

Mark