Advanced Search

My mill uses a Jetway JNF9C motherboard, that I am starting to have problems with. This board has a n old-school PCI slot, in which I have a 5i23 card. 

I don't seem to be able to find a PCI slot mini-ITX board at a sensible price, so I thought I would try a PCIe-PCI adaptor. 
Specifically this one:  www.amazon.co.uk/dp/B0CHM91KJB

It almost seems to work, in that the 5i23 appears to be detected. But then fails with "/DONE not cleared...." (I can probably find a way to paste the exact output, but it's not straightforward as the motherboard is not on the network) 

Is there any point persisting with this, or should I just give up and buy a 6i24 to replace the 5i23? 
If so, which 6i24 do I need to run the somewhat special firmware (15 or 25)? (firmware=hm2/5i23/harrison.bit) Which is compatible with 7i49 and 7i39 boards (though I could drop the 7i49s as they are currently not in use) 

One issue is that I don't know which way round the 5i23 should be in the PCI slot, it's keyed to go both ways round.... (But if it is being detected then I guess it's probably right?)
 

Hi,

At this point we can use the clothoid_3d.so library with the scurve motion planner.
What looked like a mission impossible to me, is succeeded in a functional 3d clothoid library.

I brewed an idea how to integrate this chunk of software into linuxcnc's motion planner.
Because one clothoid fit takes ~25 servo cycles we have to use an old men's trick.

1. We put the clothoid fitting algo in a seperate thread.

2. We need to have a modified segment buffer, wich request new segments from the interpreter when at ~10 segments of the buffer end.
If we don't do this, the machine will pause in the program for a short while, when recieving a new gcode load. We must avoid this.
We have to look how linuxcnc behaves when giving us the very last segment of the program.
I hope it will not stop the program directly but wait's until we are ready.

3. I think also to create some kind of segment ringbuffer. Wich is usefull for motion reverse. And usefull to avoid gcode dump wait periods.

For info:
Julian has done work on a back-lash solution using scurve motion. This is scurve on top of scurve, wich works!

Hallo, 
Problem 1: gelöst 

Problem 2: teilweise gelöst
Frage: meine CPU hat nur zwei Kerne, muß ich dann im grub-customizer   "quiet isolcpus=1"  schreiben ?

Problem 4: gelöst 

Problem 3:  nicht gelöst 
   hat Niemend eine Idee warum der Watchdog mit QTdragon nicht funktioniert ?
wie ich oben schon geschrieben habe, Konfiguration erfolgt mit Stepconfig.
schalte ich auf Gmocappy funktioniert es, 
schalte ich um auf QTdragon funktioniert es nicht, sonst keine Änderungen 


(deepl translation)Hello,Problem 1: solved  Problem 2: partially solvedQuestion: my CPU has only two cores, do I have to write “quiet isolcpus=1” in the grub-customizer ? Problem 4: solved  Problem 3: not solved does nobody have an idea why the watchdog does not work with QTdragon ?as I wrote above, configuration is done with Stepconfig.I switch to Gmocappy and it works,if I switch to QTdragon it does not work, otherwise no changes

langdons wrote:

I bought some DM860H and DM556 drivers off Ebay and Aliexpress, respectively; they worked, but have no flyback diodes.

As PCW mentioned above, they do have flyback diodes for sure, but over voltage will destroy them as soon as powered on, and so will wrong polarity, or worst case, powering them with AC power directly from the transformer, that can happen with a miss wired bridge rectifier.
Was it you that had a blown capacitor in another topic?

So, any help on ethercat finding my drives would be nice. When i try twincat, it sees the drive but then it triggers a communication error. so something is set wrong somewhere.

That did not take long! ?
This makes it usable with Mesa boards and even parallel ports as they can easily count encoder signals

Normally on a 4 layer card you do not need top or bottom fills
unless for heat sinking or shielding. You do need to be a bit careful 
about the 3.3V/1.2V plane cut/bypass capacitors/supply inductance,
and some under FPGA filling of the VCC plane can help here.

The SPI clock will be the fussiest signal so if you are routing it
a long distance, it may need parallel termination at the end of the trace
(series termination does not work with two endpoints)

thanks i will give this a try tonight.
i did not know ther was a "qtvcp copy" tool
i duplacated the directory and replaced the ui fie and renamed one ot two others.
that obiously did not work.
thanks for your help.

All modern step motor drives have flyback diodes, what they typically do not have is an overvoltage clamp.
This means you must never switch the DC power to the drives anywhere between the power supplies output
capacitor and the drive itself. This means no switches/relays/fuses/circuits breakers in the DC circuit. The
fuse in the  drive itself is just for fire protection, if it blows, the drive was already dead.

Hi all,

I worked on the sensorless homing this week-end and it seems to work, at least on a single axis. I uploaded this video where you can see it in action. The code is a mess so I didn't upload it on github yet, and will first try it on a gantry system before publishing it. I also tried to connect a ABN magnetic encoder (AS5047P) directly on the TMC5160; I can read the encoder position through the X_ENC register and use it as feedback for LinuxCNC, as expected.

I'm not sure what next step will be; probably clean up the code a little and publish a "main" branch. I'd also like to try a Pi Zero 2 version without a GUI, actually without X-Server at all if that's possible, to see if this could be an option for OpenPNP users as a cheap yet powerfull motion controller.

Hi,

i currently try to finalize the homing_comp for pokeys. it seem to basically work, but at the end where the sychronized axes should finish their homing. they are still shown as unhomed. (sometimes one of them got marked homed but never both of them)

 
 


Somehow it seems that i'm missing something, but i cannot figure out if i need to set a HAL flag or missed a step which is required to be done to finalize overall homing.

Can some one please point me in the direction where to look at? 

[TRAJ]
COORDINATES =  XXYZ

[JOINT_0] 
HOME_OFFSET = 0.000000
HOME_SEARCH_VEL = -1.000000
HOME_LATCH_VEL = -0.500000
HOME_FINAL_VEL = 0.000000
HOME_USE_INDEX = NO
HOME_IGNORE_LIMITS = YES
HOME_SEQUENCE = -3

[JOINT_1]
HOME_OFFSET = 4.100000
HOME_SEARCH_VEL = -1.000000
HOME_LATCH_VEL = -0.500000
HOME_FINAL_VEL = 0.000000
HOME_USE_INDEX = NO
HOME_IGNORE_LIMITS = YES
HOME_SEQUENCE = -3

[JOINT_2]
HOME_OFFSET = 0.0
HOME_SEARCH_VEL = -3
HOME_LATCH_VEL = 1
HOME_IGNORE_LIMITS = YES
HOME_SEQUENCE = 2

[JOINT_3]
HOME_OFFSET = 0.5
HOME_SEARCH_VEL = 3
HOME_LATCH_VEL = 1
HOME_IGNORE_LIMITS = YES
HOME_SEQUENCE = 1


if it is required my home_comp can be found here github.com/zarfld/LinuxCnc_PokeysLibComp...pokeys_homecomp.comp

it basically comminicates with pokeys.comp by a status and a command signal using following enumerations for each joint seperately

typedef enum
{
    PK_PEAxisState_axSTOPPED = 0, // Axis is stopped
    PK_PEAxisState_axREADY = 1,   // Axis ready
    PK_PEAxisState_axRUNNING = 2, // Axis is running

    PK_PEAxisState_axHOMING_RESETTING = 8,   // Stopping the axis to reset the position counters
    PK_PEAxisState_axHOMING_BACKING_OFF = 9, // Backing off switch
    PK_PEAxisState_axHOME = 10,              // Axis is homed
    PK_PEAxisState_axHOMINGSTART = 11,       // Homing procedure is starting on axis
    PK_PEAxisState_axHOMINGSEARCH = 12,      // Homing procedure first step - going to home
    PK_PEAxisState_axHOMINGBACK = 13,        // Homing procedure second step - slow homing

    PK_PEAxisState_axPROBED = 14,      // Probing completed for this axis
    PK_PEAxisState_axPROBESTART = 15,  // Probing procedure is starting on axis
    PK_PEAxisState_axPROBESEARCH = 16, // Probing procedure - probing

    PK_PEAxisState_axERROR = 20, // Axis error
    PK_PEAxisState_axLIMIT = 30  // Axis limit tripped
} pokeys_home_state_t;

typedef enum
{
    PK_PEAxisCommand_axIDLE = 0,           // Axis  in IDLE
    PK_PEAxisCommand_axHOMINGSTART = 1,    // Start Homing procedure
    PK_PEAxisCommand_axHOMINGCANCEL = 2,   // Cancel Homing procedure
    PK_PEAxisCommand_axHOMINGFINALIZE = 3, // Finish Homing procedure
} pokeys_home_command_t;

I have been taking a look at panelui, however I can not finde any simple an clear example of how to set up a simple demo to get going is there such a thing like an easy to follow example?

I also have hard time to find information  wether the panelui can work in conjuction with a ui like QtDragon or not?

Kind regards/Stefan

I would have brought the JTAG signals out for initial programming of the FPGA and not shared the serial flash signals.
May sound cumbersome but programming only needs to be done once via JTAG, after that mesa flash does the job fine.
I've thought along the lines of Andy that it would be an interesting project.
From my 7c81 clone experience I think there is the possibility of being able to bring out an extra 17 pin port, for a total of 68 io. Tho this would require firmware changes, new custom board name and and such. Also I made some changes to mesa flash and the driver to suit.
One thing I have noted with the RPi5 is that it's more forgiving in relation the SPI cable length and "messiness", IE it will tolerate longer DuPont cables that an RPi4 would. I thought I would just mention this.
Apart from not having the skill to design a board to suit, nor anyone to help during the process is that I like having the boards seperate, just in case my butter fingers cause one the voltage translation chips to blow a smoke valve and damage the PCB. Honestly this is likely to be a rare event.
On the whole good luck and keep us informed. If you have any questions regarding the Linuxcnc image I'm happy to help.

Thank you for your reply. I have read it. My question is whether this encoder is BiSS and uses a standard protocol that can be decoded by LCNC, because I understand that some of these types of encoders use their own protocol.