Homing weirdness

Hi all-

An unusual homing problem has cropped up, and having spent a day or so staring at it I think it's time to seek some insight from the forum. Hopefully somebody can shoot me a clue to get back on track.

Here's the behavior (in AXIS, btw):

1. Pressing Home All starts the homing sequence, and all three axes successfully seek, find and back off their home switches.
2. The little round "homed" icon appears next to X, Y and Z in the preview screen.
3. Halmeter reports that: axis.0.homed = TRUE (same for Y and Z)

which would certainly lead one to believe that the machine is homed, wouldn't you say?

However...

a) several (GladeVCP) HAL buttons that should become active upon homing do not
b) MDI sequences initiated by GladeVCP buttons fail with "Can't issue MDI command when not homed"
c) programs will not run, failing with the same error

So apparently the machine is somehow NOT homed, despite evidence to the contrary! This has been working fine for literally a decade, so the obvious question is: what's changed? The answer is: I've added a homebrew MPG pendant, according to the standard instructions- which works exactly as intended, and which is connected (via a 7i76) along these lines:

net mpg-a encoder.0.phase-A <= hm2_5i25.0.7i76.0.0.input-16
net mpg-b encoder.0.phase-B <= hm2_5i25.0.7i76.0.0.input-17
net scale1 mux4.0.sel0 <= hm2_5i25.0.7i76.0.0.input-21
net scale2 mux4.0.sel1 <= hm2_5i25.0.7i76.0.0.input-20

I've tried completely skipping the custom_postgui.hal, thinking that the MPG code must somehow have stepped on the homing logic, but that had no effect on the problem.

Basically I'm at a loss to think of how homing could have stopped working, given that it seems to be doing all the right stuff.
Anyone have and ideas?

thanks,
Patrick

It would help if you posted the hal and ini files and the linuxcnc version.

You should be able to look at the homing state via the "axis.N.homed" bits

Not related to your homing issue but:

net mpg-a encoder.0.phase-A <= hm2_5i25.0.7i76.0.0.input-16
net mpg-b encoder.0.phase-B <= hm2_5i25.0.7i76.0.0.input-17
net scale1 mux4.0.sel0 <= hm2_5i25.0.7i76.0.0.input-21
net scale2 mux4.0.sel1 <= hm2_5i25.0.7i76.0.0.input-20

Is not a very good way to implement a MPG with a 7I76, since it will have a limited count rate since its only sampled at the servo thread rate, and it will not work with 5V MPGs. A better way is to use the 7I76s built in MPG counters.

Thanks, Clive for the quick response. I just found a typo in my Glade file (the one I thought for sure that I had only opened up to examine...!!!] that was gumming up the works. Everything is OK now.

PCW, thanks for the MPG advice. I will probably migrate it over to the dedicated inputs at some point, but so far reading the encoders at the servo rate is working just fine. And my homebrew MPG setup isn't operating at 5V, so no problem there.

Pat

pferrick wrote:

PCW, thanks for the MPG advice. I will probably migrate it over to the dedicated inputs at some point, but so far reading the encoders at the servo rate is working just fine. And my homebrew MPG setup isn't operating at 5V, so no problem there.

Pat

Pat, I did a pretty detailed post on how to set up an MPG on my 7i76e here
forum.linuxcnc.org/show-your-stuff/32029...build?start=20#85279
Its really quite simple