On Sat, 18 Jan 2020, Gene Heskett wrote:
Date: Sat, 18 Jan 2020 15:58:23 -0500
From: Gene Heskett <ghesk...@shentel.net>
Reply-To: "Enhanced Machine Controller (EMC)"
<emc-users@lists.sourceforge.net>
To: emc-users@lists.sourceforge.net
Subject: Re: [Emc-users] I'm back to a sseriel failure that can only be fixed
with a full powerdown of the whole system.
On Saturday 18 January 2020 15:07:35 John Dammeyer wrote:
Hi Gene,
I've run into similar problems with my system as I try and make it
intrinsically safe. The issues I've run into involve the various
levels of intelligence in the hardware. I don't know if your 7i76 can
solve your problem since I'm not familiar with it.
As I've said before I'm using the PMDX-126 Break Out Board which was
really more designed around MACH3 than LinuxCNC. Obvious because the
two ribbon cable connections and 4 mounting holes are designed to hold
the Smooth Stepper. (USB or Ethernet). Still haven't connected second
port on the MESA 7i92. But that's a different story.
One of the 'features' of this BoB is the Charge Pump input. MACH3 can
generate this on any parallel port pin with a checkbox and can even
keep it running during the ESTOP IN (Pin 10 normally) by checking a
box in the main config screen.
Under Linux CNC this is generated with stepper driver #4
net machine-is-enabled => hm2_7i92.0.stepgen.04.enable
The BoB, when configured to use a charge pump shuts off all outputs
when the charge pump vanishes. This includes the machine enable
signal (Pin 1). Under Linux this is: setp
hm2_7i92.0.gpio.000.is_output true
net machine-is-enabled => hm2_7i92.0.gpio.000.out
Since the BoB has a couple of relays, one of them (SPDT) normally
allocated to motor direction or coolant, uses the NO contacts to be in
series with the ESTOP switch on the machine and feeds into the ESTOP
input (Pin10) on th BoB. If the Charge Pump vanishes, the BoB itself
takes care of shutting off the outputs which is like mechanically
hitting the ESTOP.
Same with the limit switches. A FAULT (or NC ESTOP opening) input
into the BoB also causes all outputs to shut off. A small micro
removes that FAULT 4 seconds later but the system ENABLE (controlled
with the F2 power button) remains off after this. AC/DC power is
returned to the motors and the system.
The problem I ran into was the FAULT or ERROR output from the various
motor drivers. A missing power supply voltage caused a fault on the
STMBL AC Servo drive. Something wrong on the DC HP_UHU servo drives
causes a FAULT. In all cases that shuts OFF High Voltage power and
then ON again. To reset the STMBL I have to bring the ENABLE FALSE
and TRUE. That's Pin 1 and done with the F2 button on the user
display.
The HP_UHU drives were more problematic. They don't have an enable.
To clear a fault you either cycle 16V low voltage power or bring the
processor reset pin low. I did that with the NOT ENABLE signal and a
couple of opto coupler open collector transistors. Now when not
enabled the DC Servo drives are held in reset. And it's true, HV DC
and a shorted transitor driver pair could keep the motor running open
loop so still not the safest.
I first tried enabling the high voltage with the system enable. This
created a race condition for both drives. The STMBL was enabled but
High Voltage wasn't there yet so it created a fault which shut off the
HV. Circular problem.
The HP_UHU drive, when reset is released, blips the FAULT which shut
off high voltage again resulting in ENABLE going false which again
held the drive in RESET.
SO the issue is that just after a reset or HV DC power being restored
both the STMBL and the HP_UHU blip their fault line. That shuts off
the BOB which removes the HV. Then when the HOV is re-applied it
starts all over again. Vicious circle. Hence the 4 second delay
before power is restored but the enable is held off keeping all the
servo drives either reset or not enabled.
Once the ENABLE is set true, the FAULT from the drives is blocked for
a few seconds to allow the random start up faults to fix themselves.
All done with a PIC12F609 8 pin processor.
For real safety I need to add a latching relay circuit that requires a
manual energize button to latch it on. So anything that shuts off
the HV opens this relay. And a button must be used to relatch it.
The original mill hardware has that on the Spindle and Coolant
switches. An ESTOP requires both switches set back to OFF before the
released ESTOP restores power.
So for your system, look at a charge pump as a method of resetting
hardware. Then a relay wired in such a way that you have to manually
close the latching circuit to lock that relay closed. An ESTOP or
missing charge pump or even Limit switch can remove power. All your
switch has to do is momentarily close this relay and a second pair of
contacts will hold it closed. Since the primary contacts on this CAN
be part of the ESTOP NC circuit a machine FAULT like a limit switch or
motor driver automatically causes a removal of voltages. And, like
untwisting the ESTOP switch, you need to press a physical ON button to
relatch power.
Hope that makes sense.
John Dammeyer
For you it does, but without going to a separate parport, and using it to
control power to the interface box and resetting the 5i25 jumpers for
remote power from the boxes own supply, its a chicken v egg problem.
IOW I can't use a signal that has to pass thru the interface, to control
the power to the interface, And even then with logic power coming in
from the 5i25, will that leakage prevent the full drainage to zero that
is needed to fully recover the substrate scr someplace in the 7i76D.
Thats not what happened this time as the 12 volts remained good, whereas
on the 6040, the 12 volts was being crowbared by this not on the
schematic scr, but this time it was not.
Maybe Peter has conjured up a theory by now, but I've not.
Not really other than possibly some 5V issue ( a shorted cable 5V might cause
this kind of flaky behaviour due to PTC cycling and dropping the bus isolator
power, losing communication ) We have close to 10000 systems in the field using
the sserial code used on the 7I76/7I77/7I84 and we have never heard of such an
issue before. A SCR latchup would destroy the driver chips even with a .5A
supply so its not likely thats an issue
-----Original Message-----
From: Gene Heskett [mailto:ghesk...@shentel.net]
Sent: January-18-20 10:39 AM
To: Enhanced Machine Controller (EMC)
Subject: [Emc-users] I'm back to a sseriel failure that can only be
fixed with a full powerdown of the whole system.
Greetings all;
If then. While shaving the side of the BXA's foot, swarf built up on
the top of the y axis home switch, and when I went to brush it off,
some of it went below the switch and shorted the 12 volts to ground.
Lit the swarf strands up like a light bulb, which I don't quite
understand since the teeny little switcher supplying that 12 vols.
Short circuit protected, it shuts down at 2.1 amps. The switch still
has the 12 volts on it but I've removed it from its bolt down. I
have not used the halmeter to see if its traceable, but without
sserial, I doubt I can read it.
All the power leds on the power supplies are at normal brightness
and so are the leds on both the 7i76 and the SainSmart bob on the
5i25's p2.
The GO704 and its interface is wired like Peter recommends, but
apparently thats not good enough, without doing a full powerdown on
the computer and the separately powered interface box. So I think I
am going to have to concoct a method involving the actual unused
parport to act as a master switch for the whole thing so I can do a
powerdown reset on this interface combo by command from the
computer.
I can't believe I've blown the 7i76D, the available voltages just
aren't there, but Peter also cannot duplicate this fail. But I have
now encountered it on 2 separate builds.
Has anyone else actually invented this particular wheel?
Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page <http://geneslinuxbox.net:6309/gene>
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
Peter Wallace
Mesa Electronics
(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
