Lapic error causing real time to stop.

That depends on what does and what doesn't work

Was this a previously working config that stopped working?

if this is a new configuration, have you verified that the pinout matches your I/O devices?

if your motors dont turn have you checked that your step/dir timings have sufficient margin?
(setting these to the drives minimums is never a good idea)

Its still a new setup.

I have the parallel port version working and latency that is nice and low.

Transfer of the pinouts to mesa but I still missing something and I'm not sure what.

See attached files.

Oz

First I would verify the the 5i25 --> breakout board pinout is correct
(step/dir on the right pins, any enable signals in the correct state etc)

Then if you dont get movement, i would widen all the step/dir timings
(say to 5000,5000,5000,5000 for steplen,stepspace,dir-setup,dir-hold)

Thanks,

The first point I'd like to verify is the charge pump as my controller won't work without it.

How would you approach that.

Regards,
Orrin

Ahh this is comng back to me,

what pin is the chargepump on and what frequency does it require?
Also what FPGA configuration are you using?

According to the manufacturer 1kHz square wave on pin 16 does the job.

Here are the files.

I cant write your HAL file for you , but one easy way to drive a 1 KHz chargepump is to

1. Set the servo thread rate to 2 KHz (this is needed to generate a 1 KHz waveform in software)

2. Load the siggen component (man siggen)

3. Set the GPIO pin that is routed to the chargepump to output mode (man hostmot2, GPIO section)
The GPIO pin number can be determined by the dmesg pinout print
that list both the GPIO number and the physical DB 25 pin number

4. Connect the siggens clock output pin to the GPIO output with a net command in your HAL file

I'll give it a go.

Thanks,

the chargepump is on pin 16 , but the problem apears to be Lapic is not stable , not the bit file
is their any indication led on the controller for the watchdog ?.

without the circuit of the controller it may be difficult to locate what the watchdog does , however their usually is a overide on the board somewhere to allow to switch the chargepump off , or disable it .

i would have expected an on board led for chargpump operation , do you have a circuit diagram ?

if you have not altered the origional files sent , the chargepump part in hal is working at 1Khz

until the lapic timing is sorted that will shut off the realtime , which is the error you are seeing in dmesg .

Until you give us some detailed information there is nothing further we can do.

There is no error in dmesg that is stopping realtime, the mesa card is recognised and initialised.
The LAPIC issue is a red herring.
There are no realtime errors, linuxcnc is running.

The whole problem presumably ( because you haven't told us anything directly ) is that your stepper controllers ( whatever they are ) require a chargepump / watchdog signal to enable them and you haven't provided it, therefore 'nothing works' (eg no axes move? We still guessing even on that )

Either you disable that requirement or provide one as previously described

If your parport setup worked, I would assume that the software charge_pump signal provided by linuxcnc was perfectly adequate for the job, without going to special means to provide another one?

regards