Advanced Search

For a bit I made poor progress tuning the R axis. No wonder, it has a brake in the servo and I overlooked that. Yesterday I had a bad day with that little brake. I was manually driving that axis closed loop by setting j6.pid.command and j6-unlock via halcmd in a terminal. I parked it with the brake locked, jiggled pid.command around till pid.output went to 0.0000 then looked at my brake hal logic....

About 2 hours later I was submerged in hal logic. 30 feet away a guy was welding. Somewhere in the back of my mind I'm blaming him for welding through paint. 5 minutes later the boss comes worrying. "I smell hot wires!" He stuck his nose in the big cabinet... nothing there. He walks around the back of the press. Phew! I hit e-stop, cutting power to the servos, but that nice old Baldor DC servomotor is now junk. Good thing I earlier forbade the boss from promptly scrapping the PSA and PSB axis. It had an identical motor.
It appears the brake had at some point allowed the axis to drift down about 7 encoder counts. At that point the pid slowly wound up but the motor was unable to fight the brake back to 0.000 position and gradually overheated and then smoked.

I now have hal logic automatically operating the brake for me. But it is interfering with the homing. I just now saw into it: my servo enable after brake release delay must be slightly too long. Homing starts then after my 0.1s delay the pid has some catch-up to do and overshoots just enough to fault the homing sequence.

I'm attaching my complete config. Most of the working part of it at this point is joint 6 (almost working). More to come...

In the ini file for a test I just did I see the following
JOINT_0
SCALE = 1000
ENCODER_SCALE = 10000

In the main.hal file
setp hm2_[MESA](BOARD).0.pwmgen.00.scale [JOINT_0]SCALE
setp hm2_[MESA](BOARD).0.encoder.00.scale [JOINT_0](ENCODER_SCALE)

JT

pncconf does not know about the field output options of the 7I76EU
(it only knows about the 7I76E)

The 7I76EU  has individually settable output modes for each output.

The default setting is sourcing mode on all outputs.

Sourcing mode means the 7I76EU outputs supply positive field voltage to a ground referred load
This is identical to the original 7I76E.

To set up sinking or push-pull outputs you must set the source or sink
parameters. These parameters are 16 bit numbers where each bit determines
the output mode of a single output. Bit 0 in the source and sink parameters
control the operation of output 0 and so on.

The default setting is source = 0xFFFF and sink = 0x0000 meaning all outputs are set to
sourcing mode.

To set to sinking mode, you set the desired sink bit to 1 and the corresponding source bit to 0
 To set push-pull mode, you set both the sink and source bits for the corresponding output.

That is if you wanted the first 8 outputs (0..7)  to be sourcing and the last 8 outputs (8..15) to be sinking
you would set the source parameter to 0x00FF and the sink parameter to 0xFF00

This is described on page 17 of the 7I76EU manual

Note the these parameters cannot be set dynamically, they must be set with setp commands in the hal file.

 

Die Eingänge laufen bei mir wie gewünscht, leider bekomme ich die Ausgänge nicht in Betrieb. Ich möchte gegen Gnd schalten, passieren tut nichts

Nachtrag: Verdrahte ich Output 0 auf TP6 als Schalter gegen + 24V (d.h. die Last liegt auf der anderen Seite an Masse) dann funktioniert es. Ich dachte die 7i76EU hat Universal Outputs, d.h. kann beide Modis?

Fastfred wrote:

I did try reversing the input, Does my custom.hal file look correct? I get a joint 0 amplifier fault instantly its the only joint i am trying to figure out at the moment i haven't tried to add the other joints yet. Is there other ways of doing this that might be easier to add in my case ?

Thanks

First thing to do is make sure the Clearpath's HLFB function is programmed the way you want/need.  Teknics permits a multitude of conditions which result in <some> sort of signal from HLFB, but I suggest you copy @djdelorie's setup and program the HLFB pin to be HIGH when 'enabled and following.'

Next, test the HLFB programming and your wiring to make sure LCNC is getting the HLFB signal properly.  With any amp-fault stuff commented out of HAL, start the machine.  The Clearpath should be enabled, and you should see (in halshow) the input pin which is connected to the HLFB output go HIGH (conducting).

You can also take this opportunity to estimate how long it takes the Clearpath to switch the HLFB signal from LOW to HIGH when it is enabled.  You will want this delay estimate later when configuring the amp-fault time-delay in HAL.

Once you've satisfied yourself that you have everything outside LCNC sorted, you've got a decision to make regarding the other servos' HLFB signal.

- If you want LCNC to 'know' which specific axis/motor trips, you will be using up an input pin for each motor's HLFB output signal.  You will also need a HAL configuration the same as (or similar to) @djdelorie.

- If you don't care which motors trips and just want the machine to stop when any motor trips, then you can wire the HLFB signal from all motors in a series and have a single 'amp-fault' input to LCNC.  This will reduce the number of inputs used to one, and will make the HAL configuration somewhat simpler at the expense of slightly more fussy physical wiring.

Do the tests (and HLFB programming), and think about the 'decision' I mentioned and come back here... we'll help you sort out HAL so everything works like you'd prefer.

Thanks for that fix.
But this works only for multitouch screens. For resistive screens you can only rotate the view.
Furthermore this also affects the setting for the mouse - you have always rotate on the left button and zoom on the wheel. Dragging does not work in any setting.

I am looking for two things.  I am retrofitting a Gerber 408 and I want to keep the main board in place and just replace the processor board.  The main board has all of the connectors to the servos,encoders, limit switches and also very simple interface buffering, level shifting 7400 series integrated circuits and the like. I am trying to keep this board in place and only change the board that sits on top by designing a custom printed circuit board that interface with the existing main board thereby simplifying the retrofit. 

The two things that I am looking for are:
1) any overall schematics for the entire router or any of the circuit boards.
2) I could use a main board in working or non-working condition so that I can reverse engineer the traces.  I would do partial disassembly so the board would be rendered un-usable. If any one has a board that they are willing to give up, please let me know.  I can pay for all shipping, etc.
 

Hallo zusammen ich habe mir überlegt welche mesa karten ich nutzen möchte , und zwar war der gedanke als hauptkarte die 7i97t zu nutzen und für das bedienfeld die 7i73 , für weiter i/os dann die 7i84 ( maho relaisplatine) . Aber ich habe mir die frage gestellt da ich das noch nicht ganz verstanden habe , wie diese karten dann alle miteinander verbunden werden also hardwareseitig ? Da die 7i97t dann nur diesen idc stecker hat für weitere i/ o karten also wie wird das gemacht oder ist das überhaupt so möglich , ich stehe irgendwie auf dem schlauch . Gruß tom  über anregungen oder vorschläge das anders zu machen würde ich mich freuen . 

Muzzer wrote:

@andypugh - could you suggest why your macros aren't showing up in my gmoccapy GUI? I've got no errors showing

If you start LinuxCNC from the command line are there any clues? 

My first guess would be that you don't have rsvg installed, but as that is a dependency of LinuxCNC it should be there automatically. 

yes, I saw it on YT that you were working, I just don't know which linuxcnc it was, mine is now 2.9.0 and when I export the sim configuration for racktoolchange, it puts it in the configuration, but Linux at the start shows some errors about gladevcp or something similar related to gtk and it won't run the simulation, maybe I should try to insert my ini and hal and leave the routine in the folder, or take the routine from that configuration and insert it into mine, I don't know.
 

Billy wrote:

Is the rPi GPIO interface considered fully baked, or is the Mesa 7i96S the smart bet? I would totally love to avoid the $150+ for the Mesa, but only if that does not then cost me an additional $300 in labor time.

 

It depends on what you need. I don't think that the GPIO interface on the Pi is quite as good as a parallel port, and the parallel port is only supported out of habit and as a way to get new users started cheaply. 

You will get more speed out of steppers with a Mesa card, and this is not solely down to step rate. It comes down to the resolution of the increments in step rate as the system accelerates.  If the Pi can do 20kHz flat-out then the slower rates are quantised at 10Khz, 6,66kHz, 5kHz... The motor can't physically follow such large speed steps. The Mesa card can do 20kHz, but also 19.99999 kHz too. And 20.00001

But, you could start with the GPIO and see if it's acceptable for the application. 
 

tommylight wrote:

In short:
RPI can do at most 20000 steps per second, if lucky and with a bit of fiddling (this is from what i read here, i stopped using RPI since version 3B)

The 3B wasn't very useful for LinuxCNC. The Pi4 worked a lot better (and can be used with Ethernet Mesa hardware). The Pi5 is better again. 

The reason I suggested the Pi5 in this context is that if anything goes wrong with the Pi you can just slot in an identical Pi with the _same_ SD card and be back to exactly where you were. 
If using PCs it can be difficult to find an exact replacement. Yes a _similar_ PC might work, and might run properly with the old SSD. But it might not. 

I fixed the missing 7i76eu from the stepper boards variable which caused a partial build of the main.hal file and leaving out the enable and the stepper setp lines.

JT

andypugh wrote:

What's the GUI? It looks specialised.

@Andy

Franco's been teasing us with a Siemens-like panel and GUI for a couple years.  I believe his company does retrofits:

teknix-com-ar.translate.goog/en/control-...en-US&_x_tr_pto=wapp

...and he appears to have done some major-re-writes to LCNC (both code and GUI)
forum.linuxcnc.org/show-your-stuff/48621...e-interpreter#267742
forum.linuxcnc.org/show-your-stuff/52484...control-panel#299250


Looks like he's to the point where his company has user-manuals available:
manual-teknix-cnc.readthedocs.io/es/main...achineInterface.html

But, sadly, he hasn't shared anything more than 'teasers.'

Can you attach your ini file?

Did you skip Drive 0 and Drive 1?

JT