Advanced Search

Hi all,

Today we added another feature to the scurve planner : tangential knife

The planner outputs a hal float value in degrees.
This rotational value can be used by tangential knife machines to direct the knife in
the vector direction of the trajectory.

 

Those are step motor drives, 8A max
No idea how they are interfaced but
maybe worth a bit more searching with 
the drive PN:

214 072 01d

Remember to remove the short once you have updated the firmware.

Sorry for the problem

(luckily there are  no bad firmware files out there other than what was
installed on cards)

Hey! It worked
Thank you!

Now i am not sad anymore.
It must have been a bad connection. I solder shorted the 2 pins on the backside of the mesa board, then it worked.

I will be back with more stupid questions, probably.

Unfortunately this board has serious speed limitations, as it connects to the parallel port it is very slow and you have to adjust the Base Period Maximum Jitter according to the type of PC.
There is a latency test, which is important to check.
 

You need to synchronize the number of pulses from the hardware with the software. The photo below has the adjustment that increases the number of pulses.

 

The software has an initial screen to change the pulse speed. There must be a calculation for this, but I don't remember it at the moment. But you can try until you get the movement and then adjust it for the best performance.
 

Many thanks, I will give this a try and report back.

Thanks

Thanks for your response.
Do i need to change the parameters like step time for drivers in the first page pf stepconf? I think thats the main reason of motor not running. Since i can toggle the machine by clicking on PC, i think breakout board right now can send the signals.

That actually should work.

The faulty firmware has RPI GPIO 25 (40 pin header pin 22) 
as the SPI /CS pin.  The actual RPI SPI CS0 pin is RPI GPIO 8 (40 pin
header pin 24), so connecting 40 PIN header pins 22 to 24 should
allow access to the faulty firmware.

In fact  the only reason this was missed is that on a RPI4, GPIO25
happened to be low by default so /CS was asserted.

Is it possible you have the wrong pin numbers or a unrelated cable issue?

Next I tested a M.2 ethernet adapter, Intel I226 on a M.2 card and a cable to lead out to the RJ45 connector.
It's the same M.2 A+E key 2230 size card.
 
It turns out the cable runs over the screw also in this case. The cables are now much thinner and much easier to bend so it isn't particularly easy to make room for the securing screw but can be done. I think I broke the connector on the board but it continues to work.
 
 
 
 

The function was just as good with this M.2 network adapter. The adapter showed up in the pci listing automatically, showed up in the interfaces list also automatically and when I switched ethercat config to this new network adapter, ethercat worked just as well.

I decided to use this one, mostly because it was easier to make a nice installation. Here it is where the RJ45 connector is in a plastic housing, glued to the top of the T8.
 
The verdict: It works. Tricky with the cables on the M.2 board. Easier to make a nice installation of the RJ45 connector.

For both these ways of adding a network board to a mini-PC I must say that there are many ifs and buts and maybes that need to be figured out beforehand. The absolute recommendation is to buy the right system to start with.
I  tried to follow that advice and looked around for another N100-based motherboard and there are a few out there, even with Intel network adapter.
I test-fitted it in the cabinet, I used my old trusty J1900 board. And while there is room for it, there isn't much extra room. And the QtPlasma interface can use more CPU power me think.
In the end I thought this could be a fun experiment. And as i happens both products work.
 

Hi,
did you look through this thread where I got PathPilot running on a non-Tormach mill?

forum.linuxcnc.org/pathpilot/45846-pathpilot-7i92-g540-hal-file

cheers,

Peter. 

Hello friend
You can use Glade 3.8.6 software and edit the file responsible for the image in LinuxCNC

 

OK, thank you!

What is the shorting of pins supposed to do?
I have tried both suggestions, but the command still results in the same error.

I don´t have a rpi4.
I am a bit sad about this, i don´t want to buy a rpi4 just for flashing.
Stupidly i thought this would be a good Mesa card choice for me, i now wish i had bought something with ethernet. Then i wouldn't be stuck with rpi.

Hello friend
You can use an oscilloscope on the pulse output of this board, use the test axis feature in the image below

 

I recommend leaving the drives activated at this point, as remote activation may not be enabled in this test.
I also recommend not using the board indicated in your link, as it is not reliable, it does not have electrical isolation, and when the system loads it activates the outputs by itself, making the machine move before the releases and safety.
I recommend using the 7i96s electronic board, which is very safe and high quality.

I put in the MAC-address for the new adapter into /etc/ethercat.conf, and restarted ethercat service.
There were no initial datagram UNMATCHED messages, so I use the plasma table for some minutes.
Even after a day there were no datagram UNMATCHED messages, so I call that a win.
I even went back from isolcpus=1,2,3 to no isolpcus, it works just fine without that.

 
 

The verdict is: it works.
The stiff cable is a problem. And overall making a nice package of this will be a challenge.