5i25 + 7i85S setup questions

Hello all, I've spent a few days googling and reading documentation but I've hit a bit of a wall and might need some help.

Firstly my questions:

1. How can I confirm a daughter card is functioning? I have been able to blink the LEDs on the 5i25, and the 7i85s is currently powered correctly by the 5i25 (CR1 is on and I've jumpered the 5i25 to supply power), but following the instructions on this page I do not see more pins available with the card plugged in or unplugged.


2. Which is the correct firmware to select in the PNCconf config page? PNCconf did not have a 5i25+7i85s option in it, eventually I worked out that I needed to download the 'support software' zip on the 5i25 page and paste it into the HM2 folder, I now have "5i25/configs/hostmot2" as a board name option and the firmware drop down now has many more options, including "5i25_7i85sx2", is this the correct option for my setup? I don't see a single 7i85s option, so I assume leaving all of connector 2's pins as 'unused stepgen' etc is fine?


3. The IO connector on the back of the 5i25 is I/O Connector 3, correct? I've gotten the same results when plugged into either connector


4. Can I unplug the daughter card while the computer is on, or will that damage things? It's probably paranoia but I've been powering down the PC to unplug the 7i85s or switch the connectors

Other info:

• I'm using a 5i25 + 7i85S plug and go kit on a bench with nothing else connected so far. I've been attempting to send some step pulses out. I've been using a real world oscilloscope to probe every pin on the 7i85 connectors and I can see 5v or GND where I expect it, but never any signals.
• I plan to use servos with step/direction drivers, and encoder feedback piped back to linuxcnc. I am currently trying to set up the system in a simple open-loop stepper style (ie linuxcnc just sends some pulses through the mesa cards, expects nothing back) as a test setup, adding complexity later.
• I'm using linuxcnc 2.7.8, installed as a clean install from the ISO from the website (ie the 'simple install method'). Since installation I have used the synaptic package manager to check and update everything it could

1. The 7I85S is a passive device so there's no way LinuxCNC can tell if its there or not
2. Your example files use stepgens from 04 up, this is not correct for P3 (P3 stepgens start at 00)
P3 is the DB25 connector on the back of the 5I25 (5I25man page 5)
3, You should verify the the 5I25 has the correct firmware installed with:

sudo mesaflash --device 5i25 --readhmid

4. Its best not to live plug/unplug the breakout card

Thanks PCW, I'll try that as soon as I'm back from work tonight.

Just to confirm the stepgens section, you're referring to any parts where it says 'hm2_5i25.0.stepgen.04.dirsetup' (for example) in the hal file? So anywhere I see an 04 it should be an 00, 05 should be 01, and so forth, correct?
Do i need to change it anywhere else, or just the hal?

Also, when using PNCconf I only get the option for starting at 4 for stepgens, is that a bug or do I have the wrong config selected in PNCconf?

I'll try those changes tonight and report back, thanks again.

Thanks again PCW, with that information I was able to see the issue and resolve it. I now have stepper signals out, and encoder signals in.

I have uncovered a new issue however: noise on the step pins and an erroneous square wave on the direction pins.

The problem follows these steps:

• Currently motors are on a bench, no sensors/switches enabled, no encoders. Servos responding to step/dir.
• When opening linux cnc motors are stationary but servoing
• Turn off the estop in software, motors unchanged
• 'Power on machine' in software, motors unchanged
• Home all axis, run the default 'linuxcnc' gcode, motors move and change direction when they should
• if I hit the blue 'stop' button, motors pick a direction and spin at a steady rate
• If i probe the step/dir pins on the 7i85s, I seem to be getting some noise on step pins, and a full square wave out of the direction pins
• If I hit the 'estop' button in software, the square wave stops and the noise is reduced.
• This is repeatable with all the wiring disconnected, ie a bare 7i85s on the bench, connected to the 5i25 and nothing else. which eliminates issues from my wiring, motors, motor drivers, etc. This suggests that it is a configuration issue, though I don't yet understand the linuxcnc system well enough to diagnose it.


I've had a look at the hal file, though I don't really know what I'm looking for. Thoughts?

The hal/ini values in your file set seem quite odd,

Why are you not using the default 1 KHz servo thread and P value of 1000?

What does you "blue stop button" do?

I'm really not sure, I just used pncconf to create a really basic stepper system to test stuff out. I used the default values for pncconf for almost everything except axis limits and steps per mm. Where do I change those values?

The blue stop button ('stop program execution - esc') i assume is supposed to stop the current gcode program (see attached photo). When I click it the step pin gets a bit noisy (peak to peak of 2V at its worst), and the direction pins just sends a square wave. When I probed it just now, it was 3.8khz, but it picks a different frequency each time.


I also just noticed that when, during the linux text example gcode, a motor is idle, it moves arbitrarily too. ie if X and Y are idle while the Z lifts, X and Y will continue to spin. As soon as they get a legitimate command in the gcode they will follow it smoothly, then go back to spinning when idle again.


To clarify what I was saying before:
Motors are normal at power up.
Motors move mostly normally during gcode execution, but spinning oddly when they are supposed to be idle in the program.
If it hit the 'stop program execution' button (blue stop button) they pick a direction and a speed and spin indefinitely
It also happens if I jog a motor (as soon as I stop jogging the motor I jogged starts spinning. Only that motor, others remain unaffected until they are jogged)
If i close the program, they stop
If I hit estop (in software, no actual buttons hooked up) they stop
Probing with an oscilloscope shows a 5v-0v square wave on direction pins, and 2v noise on step pins.
It occurs with all my wiring disconnected.


Thoughts?


PS I've reattached the hal/ini/pncconf file as it may have changed since I started the thread.

Those ini values are _not_ defaults
default servo thread period is 1000000 ns (1 ms) not 100000 ns (100 usec)
default P term is 1000

I've no idea how those values got changed. Odd.

I changed the values to what you said and the square wave on the dir pins changed to an even 45hz on all motors that had previously been active, and some random step signals were going out.

I bumped up my stepscale on each axis to make motor motion more obvious and the problem went away the next time I opened linuxcnc. I'm not sure if it's coincidence though. Thoughts?

I'll keep poking around the configs to see if I can isolate what I did to make the 45hz go away and report back when I do.

45 Hz on direction pins most likely means you had oscillation from bad tuning/servo thread
values (which you did have), Once this was fixed, the problem went away