On 11/19/24 15:26, Viesturs Lācis wrote:
otrd., 2024. g. 19. nov., plkst. 19:45 — lietotājs Peter Wallace
(<p...@mesanet.com>) rakstīja:
I would check the 7I36 output voltage when disconnected from the VFD. If that
seems ok, it may be a VFD EMI or grounding issue. This might require a line
filter on the VFD, a common mode choke on the VFD wired going to the motor, or
possibly a common mode coke on the analog signal/gnd wires going to the drive
I have attached a screenshot from VFD's manual that shows control
terminal wiring. Those are "active low" inputs.
I have connected terminal-11 to "minus" of 7i37 outputs and
terminal-14 and terminal-15 to "plus" of each of 7i37 outputs. And
terminal-4 and terminal-5 with separate cable to analog output pins.
Is there any chance it could be caused by config? I had it working
previously in this machine, but I did change to better pc as the
initial was quite old and the realtime performance on bookworm was
bad.
Ok, I will check with disconnected cable.
Is there a complete specification of the C signal protocol?
That link has the most detailed description I have found so far. I
could ask Alciera if there is more available...
I did a trace of C channel pattern with regards to A and B channel.
There are 6 different patterns that repeat 4 times on full revolution
of motor shaft. Those are 8 pole motors, 4 electrical revolutions, so
those 4 repetitions explain themselves. I have 24 csv files from
HalScope that show pattern of A, B and C signals. The thing is that
sometimes C channel can have 2x pulse frequency compared to A or B -
in some patterns there is separate pulse on both rising and falling
edge of A pulse. 6 different patterns sound exactly like 6 different
stages of Hall sensor signal. I do not necessarily need index pulse, I
have not used it so far (although homing on index might be a good
idea), rotor homing for bldc component is the main concern here.
otrd., 2024. g. 19. nov., plkst. 20:33 — lietotājs gene heskett
(<ghesk...@shentel.net>) rakstīja:
The first thing I'd look for is a ground loop.
Ok, thank you, I will try to check.
otrd., 2024. g. 19. nov., plkst. 21:19 — lietotājs andrew beck
(<andrewbeck0...@gmail.com>) rakstīja:
But I thought I would just try one of peters ency cards
It's a encoder splitter that takes in one channel and makes 2 out.
I did not know about existence of such a tool. The fact that I do not
see it available from any european reseller is just a minor
inconvenience. The actual issue is that I have 2 machines with total
of 7 drives (and there is one more of my previous retrofits that has
at least 1 servodrive not working so I would like to replace that as
well). Only 2 of them are used for actually driving the motor, 5 of
them are are replaced by 8i20 and now are used solely for chewing and
digesting the C channel pattern and giving out proper encoder signal
in a form of A + B + index. I use the index for bldc component only.
Fortunately for me the encoder output died on a drive that moves C
axis motor (at the moment it is optional and does not cause a lot of
inconvenience as I am now finishing toolchange and remap of M6
command). If it was _any_ other of those drives, that would be trouble
to say the least, so that is why I am asking about chances to read and
interpret C axis pattern in LinuxCNC.
Viesturs
there is another up and coming drive tech that may be of interest,
Viesters. Closed loop stepper/servo's. I am using several of them with
linuxcnc very succsesfuly.
pro's:
1. The motors have their own encoders outputting a/b in quadrature and
of course z for index.
2. this encoder signal only goes to the driver, linuxcnc never sees it
not even for feedback. So these motors do only what the stepgens in lcnc
tell them to do to a 1 microstep or less accuracy.
3. motors run cooler because the actual error determines the current
delivered to the motor on a microsecond by microsecond basis. That is a
power savings you can see in the electric bill unless the shops heat is
also electric.
4. faster, the rapids on the machines I run with these has typically
been doubled.
5. the dread of a layer shift (the motor losing home) has not occurred
on any machine axis using these motors.
Cons: your power supplies will need replaced, nema 17's run up to 90
volts, nema-23's need around 110V, which is used to improve response tines.
Cost: around $50 to $75 usd per axis by the time the higher wattage
power supply's are totaled up. Essentially zip difference if replacing
servo's. Some machines will need mist coolant retrofitted to make use of
the additional speeds. A mist so dry you can't see it will easily triple
tool life in alu, and double the machines thru-put.
Good luck finding your problem.
Cheers, Gene Heskett, CET.
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
