Advanced Search

I tried with new consumables because I was thinking about swirling due to a worn nozzle but ithe issue remains the same.
The air pressure/flow is automatically adjusted by the powermax and it’s air supply respects the manufacturer’s recommendations (pressure, flow, water/oil contamination etc…)

Regarding the cutting direction, I let Pronest decide for me but I will try on next cuts.
Do you think compensating the torch inclination could work?

Yes, if you decouple the linear axis (and comment out the net statement that connects
the PID output to the 7I77 analog output pin, you can simply setp the output voltage
with halcmd.
 

Saturation takes a lot of overload, and you find out after 30 seconds to 1 minute that is it overloaded to much as it starts smoking, a lot.
And yes, toroidal transformers are metal core, wound in a doughnut shape, while normal transformers have E and I sections alternatively pushed inside. Microwave transformers usually have all the E on one side and all the I on the other side welded together, but that is not the best way for efficiency, they are on for several minutes at a time so not important.

Could you share your complete setup for doing that type of work on the lathe?

Can you make a new basic config, with just PWM's and encoders and test again.
I recall a few years back i made a config that absolutely did not put anything on analog outs, wasted an hour looking through hal, could not find it, made a new config and it worked, never had the same issue before or after that.
Another thing to check is the Mesa supply voltage, the 5V side, if 7i97T uses 5V, many of 5V power supplies are really bad, but that would also trigger lost connection, not always, though, so worth checking.

OK if you have a significant PWM channel 0 value number, PWM channel 0 is enabled
and there is no watchdog fault, that does point to a hardware issue. Its just very odd
because of the way the offset PWM works, any kind of hardware issue tends to result
is output stuck at some multiple volt output level, not 0V.

I changed the encoder filter to 0 on each axis, it had no effect. I wonder if I can somehow drive the servo without a closed loop and see if it works.

i believe all floating-point values in linuxcnc are doubles:

typedef double real_t __attribute__((aligned(8)));
#define hal_float_t volatile real_t

should be enough to hold 32-bit integers without precision loss

Hello guys,
bad news for me.
Today i installed the VFD to the wall a bit far away from the control box. After i was done, i wanted to check if everything works and check if i still have some noise problems from the VFD. 
Due to the fact that i am again totally confused from 5-6hours of trying to fix it, i am not anymore sure if i got the unexpected real time delay on first start up (but later i got it all the time), but i have found out that when i jog by keyboard or mouseclick on the jog buttons quickly, lets say couple of short and quick jogs, then i get a joint following error 0 and the machine is then off. When i put it on and do the same quick and short jogs then i get following error 1. Then again the same and following error 2. (i would not take the sequence serious, i later also got following error 1 first, at least i believe, but i am not 100% sure)
I GET NO FOLLOWING ERROR WHEN:
- Jogging normalny, no quick clicks
- running a program

Is it possible that USB is the problem?
I havent changed anything since i posted the screenshot of the very-good latency in latency-histogram.

More info:
It happens also that when i want to start Linuxcnc i get a error (see picture), but its possible that it starts after trying second or third time.
Also sometimes Linuxcnc got stuck (freeze) or the whole PC, but of course i started to play around with the Bios settings and specific bios settings could create more problem.

I dont have PS/2 keyboard / mouse available to test. Changing two USB settings in BIOS did not help. (legacy usb something and one other option)

So, if you will have no idea, then my only option might be the 7i76eu, BUT i am also thinking to go back to Wheezy and Linuxcnc 2.8, IF its possible to do file sharing with e.g. Samba, but i believe that all tutorials are before windows did something to block the file sharing. Rodw wrote something about it, here in this thread.
PS: I need nothing but working Linuxcnc, mousepad/geany and file sharing or a browser that works so i can send files by mail or open Microsoft OneDrive.


 

thx there will be a DB37 breakout board in the box and 10 overrule switches on front side

It all took a little longer, but I have a presentable development status for version 6.0.

In Probe Basic, the tool measurement routine can now be fully configured via its own user tab.
The page also contains a subpage with a little help on the settings.
The settings are no longer communicated via parameter file (.var), but via HAL pins. The parameter file now functions as a backup. Either because you want to test other values ​​or as a security backup.

Using the installation script, the files and folders can be quickly edited and moved to their location. Of course, a backup of all files is created. In the test phase, however, it might not be a bad idea to back up the machine folder completely.

LinuxCNC min 2.9.3, Probe Basic min 0.6.0-37.dev and QTPYVCP min 5.0.2-5.dev are required.

Nothing has actually changed in the routine's processes and no new functions have been added yet. Nevertheless, the routine for the connection to the GUI has been revised.
What has changed, is the way in which the offset is created for tools with a large diameter. An offset is no longer created across the board from a certain diameter, but a value is entered in percent in the "FNT ANGLE" column in the tool table. This means that each tool can be offset individually.

You can find everything here: github.com/TooTall18T/tool_length_probe/tree/develop

The routine itself can also be used without Probe Basic (user tab). Information about this can be found in the manual.

You just have to respect the max current of the transformer ! Tommy say metal core transformers can withstand a lot of overload, he's right, but I would think a toroidal one is not this type, and you could have magnetic saturation.

If you don't know your load current, just put a fuse to protect the output and you will see.

Yes, that how i tested it.

Thank you Aciera for sharing tips, will definitely convert it to motion pins.

I have gone thrugh the motion section pins in  linuxcnc.org/docs/html/man/man9/motion.9.html , but i am unable to find the feed override inputs. can you please share more details on this...