Jon,

I have tried both 1.9 and 1.7 on the bios epp setting.

I tried the epp_dir=1 setting as you suggested.

The cable is the same one I have been using since I converted the machine back in 2015 or so.

The load fail red led does come on and stay on when I boot up linuxcnc and get the error messages. also the green estop ok led is lit as well. Prior to starting linuxcnc the light is off.

I decided to try and update to 2.9 linuxcnc and see if that helps but I am running into issues with that also. I need to get python3.7 installed but it is not playing nice. I will keep working on it but I suspect it is a hardware issue.

Do we need to try another eprom?
 

This is catching my interest. I think I will try it. I have at my fingertips a Pi 3 Model B+ but it is not working, not sure why. Maybe I dropped it a while back. I also have a Pi 3 Model B V1.2, I have it running a fresh install of Debian 12 Bookworm.

Now it is installing LXDE.
Next up is realtime kernel, I'm looking at linux-image-6.1.0-41-rt-arm64
And then LinuxCNC....
Fingers crossed...

Today i recognized a problem with radiobutton, hal_io . The button is set to checked in designer, but the hal_pin is not set to true at startup. A single radiobutton is even unchecked at startup.  To click an already checked radiobutton doesn't help. I need to check one of the unchecked radiobutton (to activate) set the hal_pin to true.

I think the language barrier made the text above a bit confusing, so adding some screenshot


 



 





 

On a 7I92T you can change the  GPIO  resistors from pullups (default) to pulldowns
to solve this issue (W2 down)

fixed the spindle with a R1000 resistor between pin1 of the db25 plug and ground. all good now!

So I built a CNC plasma cutter out of garbage.... and since I built it out of garbage, I used a very old computer to run it.  It's a 32bit computer, so I'm stuck running Linuxcnc 2.8.  I got a base configuration working and running fairly well, but because it's made out of garbage, it really needs torch height control.  I got a THCAD and got everything wired up to my Mesa 7i96s, and ran the PlasmaC configurator.  The buttons on the bottom of the screen work, I can test the ohmic probe, and the probe cycle seems to work, but I can't get any of the GladeVCP components to load.  The tabs are there, but they're blank, and if I try to configure it with a Run panel instead of the run tab, it doesn't show up.  The other issue I have is I can't load a G-code file, it seems to get stuck "Filtering."  On top of that, I'm now having latency issues I wasn't having before.  I've been staring at and trying to troubleshoot my configurations for way too long and I can't find anything obviously wrong, so I suspect this 20+ year old Dell just isn't up to the task.  Part of the "art" of this machine was that it's made of junk, so if I can keep using this computer, I'd like to... but I'd also like the machine to work properly, so I figured I'd at least upload my configs for someone more knowledgeable to look at before I buy something more modern.

I'm afraid that the 56U is not supported. Or at least I did not add explicit support for this.

I do not know if the 56 series supports the PulseEngine V2 which I'm using to control the motion. You could try though.

In the source code there is a check , if the device is a 57CNC, which only logs a warning.
And there is another check if the "external pulse generator" is used.

Excerpt from the documentation

Pulse engine v2 supports two different pulse generator modules:
- Internal: similar to basic Pulse engine, limited to 25 kHz pulse frequency at 3 channels, uses built-in circuitry and pins
- External: new in v2, limited to 125 kHz pulse frequency at 8 channels, requires external circuitry to deserialize the data to pulses

yes, that actually helped! Machine is running with a stock cable now and all commands work as expected. Thank you very much!

Except when I shutdown, the spindle starts. That is very uncomfortable. Still trying to understand why this is happening here. Any hints?

Oh and in uccnc the step scale was 160 for xy, 320 for z, here it is 80 for xy and 160 for z. Well, that's not a problem I guess. Maybe loosing some accuracy?

I am only at the start stages, building the machine. Actually not dealing at the moment with CSS. I was wondering if the beta version ready for use. Eventually I installed 3.1.3. The Machine is DMC2 MINI from here: shariffdmc.com/product/dmc2-mini-cnc/ It needed many mechanical modifications that took some time to overcome.

Hi jjdege, by changing the pos-scale value, all the movements don't correspond to reality.
I solved most of the problems by changing the SERVO_PERIOD from 2000000 to 1000000.
I've also a problem in display section: MIN_LINEAR_VELOCITY = 0 MAX_LINEAR_VELOCITY = 50.00

Many thanks to all of you

bebe tart

Hi,
try changing the value in the .hal file

LearningLinuxCNC wrote:

John,

Yes it is the UPWM board. It is driving your pwm servo drives. The CMOS battery did die at one point but the parallel port is set up in epp mode. I am not sure why it is doing this. If you have any troubleshooting tips I would be glad to hear them. Could the eeprom be bad?
 

OK, so the scan of the parallel port is getting some random stuff, since it thinks it sees a bunch of PPMC boards there.
The FPGA on the UPC runs a checksum on the data loaded from EPROM, so if the red (load fail) LED flashes and goes out when power is applied, then it is accepting the firmware as valid.  If that LED stays on, then the firmware did NOT load.  First, check that the BIOS settings are for the parallel port to be at address 378, as that is what LinuxCNC is using.  If this is a PCI or PCIe parallel port, then the address will not be 0x378.  There are several versions of EPP mode, if the BIOS allows to select between 1.7 and 1.9 versions, try both.  If your parport cable is not rated for IEEE1284 operation, that can be the problem.
Oh, one other thing is there are some ppmc driver options that may need to be changed for the newer LinuxCNC version.
In your config file set, in ppmc_load.hal, on the line that loads the ppmc driver,
loadrt hal_ppmc timestamp="0x00"
try adding this to the end of that line:
< space> epp_dir=1
That might fix it.  This is documented in a comment in that file.

I'm not too sure what else to suggest.
Jon

subscribed .. curious about the outcome and further experiences.

venturing into embroidery myself lately, got a 2010ish 12needle chinese machine (in reasonably good working condition) standing round but neither time nor space to work with it. in the meantime learned a bit about design and software for it - a whole ecosystem by itself obviously - curious how and if you'd try to translate this to the (linux)cnc-world..

cheers,
markus.