Re: Amplifier fault w/steppers..?????

[Jon Elson <elson-at-pico-systems.com>]

P.K. Ferrick wrote:

>Hello all,
>
>I've been having a problem running EMC lately that I can't seem to
>crack, and I'm hoping that some kind soul on the list can help me out.
>First of all, I'm running BDI 2.10 on a 333 MHz machine with 128 Mb of
>RAM.
>
>EMC always starts up fine, but _sometimes_ when I click on "Machine On"
>(after pressing F1 to toggle off estop) I get the following error
>message:
>
>    Amplifier fault on axis 1
>
>This doesn't happen every time, but after it does the only thing that
>will make it stop is a full power-off reboot....not something you have
>to sit through too many times before you get tired of it!!!
>
>Examining the messages in the console window (from which I start emc
>using emc.run) I find the following:
>
>    minimilltaskintf.cc 962: Error on axis 0.
>    minimilltaskintf.cc 962: Error on axis 1.
>    minimilltaskintf.cc 962: Error on axis 2.
>
>The weird thing is that I'm using steppers and therefore steppermod.o
>and not the servo driver at all.  So where does an amplifier come in???
>It seems as though the state of the parallel port is related to this
>problem.
>  
>
Yes, definitely.  You need to figure out what digital inputs are hooked 
up to what, by the
software components and hardware devices specified in the xxx.ini file 
that is specified
in whatever xxx.run file that actually starts EMC.  It is possible that 
the ini file is
specifying a hardware address for the auxilliary inputs that is actually 
some other device
like a scanner, and not a parallel port at all.  minimillio.o is 
expecting aux
I/O on a second parallel port, whose I/O address is specified in parameter
PARPORT_IO_ADDRESS.  I don't know what it does if there is no device to
respond at that address.  I know that the real-time section of EMC 
doesn't know
or care whether a particular I/O device exists or not.  I thought that the
Linux side DOES care, and you'd get an error message if the address 
DIDN'T point
to a device, but I could be wrong on that.  But, if a device doesn't 
exist at that
address, then the status bits read from that port won't be reliably 1's 
or zeros,
they can fluctuate depending on what else was read previously.

As for the "amplifier fault", there are 4 software status bits for each 
axis.
One is an output, for axis enable.  Then, there is - limit and + limit 
switches,
and the axis fault bit.  Stepper drivers can have a fault condition, too.
I think with the stepper version of things, one input bit feeds the 
status condition
for all axes, so that's why you get the 3 error on axis n messages.

Jon