Advanced Search

hi cakeslob,

I checked the firmware version (stm32F446 octopus 1.1) At some point i was using minicom and cutecom to test the spi communication, I got [42] back at some point.

I just reinstalled the firmware using CubeProgrammer. What I dont get is why config.txt with the pin mappings should go in the sd reader of the octopus board. I did a secure format of my 8 and 16 Gb sd cards from realiable brands (Kingston and Sandisk) FAT32, but I never had the firmware being renamed to firmware.cur. For this reason I believe that having the sd card in there with the config file not being read is the issue.

So I'm assuming that CubeProgrammer did a good job and the MCU is flashed correctly. However when powering the board, it takes no motor current which I thought is weird. And the linuxcnc control still stuck in estop....

Gosh what else can I try?
Thanks

p.s. maye I didnt name all things I tried but I definitely went over the checklist you provided  

Does anyone want to buy my Mesa control cards? I have a single board pc with PCI/PCIe slot I can add in for cheap.

Located in France/Norway but I will be back in the States in April so I could ship them from there as well. 

They work perfectly, I am just moving away from LinuxCNC 

100 Euro/USD + shipping

​​​​​​​Luke 

 

I checked:
setp hm2_5i25.0.encoder.00.filter 0
It doesn't change anything.
In the evening I will try to take halscop photos of what they look like.

Hi Grotius,

I guess you calculated va1 directly as the vector between centerpoint P_c and Startpoint P_a0 and va2 by first calculating n via the cross product between va1 and the vector between P_c and end of the arc P_as?
In the colinear case this will not work as the circle is not clearly defined in this situation, so it is required to get the additional info from LCNC in which plane this arc is supposed to be so your solution is the way to go.

Sorry Timo, ich hatte mich evtl. etwas missverständlich ausgedrückt.
Es ist im Vorgang der Referenzfahrt unerheblich, wenn die X-Achse ein paar Zehntel verstellt ist!
Selbstverständlich ist es sehr wichtig, dass die Referenzposition exakt und reproduzierbar angefahren werden kann.

Wie schon geschrieben, Dein Tip mit dem Messtaster war Super und das werde ich unbedingt durchführen.
Interessiert mich wirklich brennend wie die Wiederholgenauigkeit bei meiner Maschine aussieht und, wenn ja, wie groß die Steuung ist.

Wenn Du Lust hast, Deine induktiven Sensoren mal zu testen, wäre das super. 
Dann könnten wir mal eine ergebnissoffene, praktische Gegenüberstellung machen. 
Ich hatte mich ja damals auch nur aufgrund von Annahmen für die mechanischen Schalter entschieden..

Gruß Jens 

@urvilsuthar
We can change feed rate in real time but acceleration changes need a queuebuster command in the gcode like M66 E0 L0.

In my opinion it does not make the mesa cards obsolete. yes digital servos don't need the step/dir or analog interfaces of standard mesa cards. But that is not the point. In my opinion mesa cards are not used in professional of the shelf solutions anyway. So it is the hobby and semi professional market mesa is addressing.
When your controller of an old machine breaks do you replace everything? or do you replace the broken parts with new stuff. Here having the option to go with ethercad would be great.
When you build your own machine you most likely will only use a digital spindle servo and not one on all axes to keep the cost down.
When you want to add stuff for automation the ability to add steppers is huge due to the low cost. And everyone has them laying around who works with automation. I once hacked together a system which pneumatically unlocked a clamp and advanced a square profile for the next part with two rubber wheels and two steppers. Cheap and fast.

Right now my main motivation is linear encoders.

Wenn es keinen Unterschied macht, ob die x-Richtung verstellt ist, dann ist doch auch egal, ob der Schalter schuld daran ist?

Der Unterschied zwischen Schaltpunkt und wann der Motor wirklich steht ist rel. gleichgültig solange a) das Zeitintervall immer gleich ist (das ist wohl der Knackpunkt) und b) der Anhalteweg immer gleich ist.

Ich habe an der Maschine induktive Sensoren. Das kann ich mal ausprobieren wie reproduzierbar das ist. Für mich egal, weil ich immer erst "home", dann mein Werkstück sowieso neu einmesse. 1 cm reicht mir im Prinzip, aber das ist besser.

Klatsch doch mal Deine INI hier rein.
Ich bin zwar nicht der Spezialist, aber es tummeln sich hier Gott sein Dank einige richtig gute Leute, die mir fast immer weiterhelfen konnten.

Gruß Jens

I found the parameter:     36  u32   RW     0x000007D0  hm2_5i25.0.pwmgen.pwm_frequency through halcmd
Now the motor partially spins. The PWM signal is set at 2 khz and scale of 5.0 and when I click the jog '+' button in axis, the motor had two distinct activation ( it spins for a quater turn, stops and then spins a couple of turns in the same direction) I measured the PWM voltage on mesa 7i76 board it matches the observations with a small 0.25 volts spike right after the click, it goes back to zero and then goes to 4.8 volts.
The same small voltage spike also exist in direction output too and this spike when I click '-' button makes the motor turn a quarter turn in reverse direction, stop and then turns in the same direction as the '+' jog button would make it spin.
I have uploaded the HAL config I used for this testing, it has fake position feedback to mitigate joint follow error which I tried widening the max error allowed but nothing worked.
Is there a way to suppress that initial voltage spike? Or am I doing something wrong in the HALfile?


 
 

unknown wrote:

The older versions are suitable for ethercat, the latest versions have a kernel that ethercat doesn't support.
 

That may have been resolved by the etherlabmaster guys by a recent commit on their  repo. We install from their repo so any changes flow through via their Buildbot without further action other than an apt update if an existing install or reinstalling the image.

I've found a spindle overload signal comes from the ancient VFD, 5v at similar level to the limit switches and encoders.
For the limit switches I am using encoder channels 3 and 4.
Analogue output 5 is for the spindle, and the spindle motor has no encoders.
Can I use one these unused encoder pins, say QA5, for this overload signal? Or is it out of bounds if the analogue side is used?
Thanks

You learn new features of Linuxcnc all the time! Well I did. Thanks

This is a bit left field, I wonder if you could use external offsets 
and modify ini parameter [AXIS_L]OFFSET_AV_RATIO)
linuxcnc.org/docs/stable/html/motion/external-offsets.html
To use external offsets, you need to allocate a percentage of velocity and acceleration to the external offsets so this by definition reduces what is available to the gcode

PCW wrote:

If by daughtercard you mean the 7I84U, any firmware with a sserial port will do.

(basically and 7I96S firmware that does not use the serial expansion for modbus etc)

Sorry for being daft, but I reflashed the board with the FW you uploaded here. So far so good, but just checking - does this FW have the serial port feature, so I can connect to the 7I84U without having do download anything else for any of the two boards?