[PCW_MESA]

I'll bet that is the problem. I didn't realize that needed to be done so I'll try it. Need to study up on Linuxcnc and learn about it. Just experimenting around.

To answer Brian's question, the speed after using lincurve is excellent, within about 1 to 5 rpm at calibration points and since the original speed correlation (before applying lincurve mod) was basically 2 sloped lines with the exception of a weird blip at 4500 rpm,

Thanks everyone for your help. Great solution for threading accuracy. Is there an easy eay for me to view the contents of the lincurve.ko? Where do you recommend I start to learn about Linuxcnc to understand what's going on with these types of mods?
Pat

lincurve.ko is a loadable realtime component so is compiled code and not editable

The source is here: https://github.com/LinuxCNC/linuxcnc.../lincurve.comp

If you want a better understanding of linuxcnc configuration I would start with the hal and integrator manuals here:

Index of /docs/2.7/pdf

[GTOGuy]

I have great calibrated cure now in the fwd direction and ad only one slow speed in the rev direction What do I need to do to get a reverse capabily?

Attached is my modified hal file so far

[brianbonedoc]

Try adding (-) speed values. Not sure if that will work but thats the only thing I can think of that can give the one slow speed in reverse. Actually you just gave me that epiphany - because I am having that issue on the lathe when backing out of rigid tapping it only goes slowly out!

for example:


setp lincurve.0.x-val-00 -242
setp lincurve.0.y-val-00 -300

Let me know if that works!

[GTOGuy]

Brian,

Are you suggesting that I create a new calibration set of negative x and y values for the entire speed range? If so, am I still limited to 16 total calibration points so I should reduce the number of fwd points to limit the total number of points to 16 (limitations of lincurve)
Pat

[GTOGuy]

I attempted to add several negative value entries and the machine only went to full speed (5,000 rpm) in fwd direction and still around 245 in reverse

[PCW_MESA]

I attempted to add several negative value entries and the machine only went to full speed (5,000 rpm) in fwd direction and still around 245 in reverse

Thats pretty much expected, The way the PathPilot hal file and hardware are constructed, it looks like the spindle speeds are
always positive at the low levels and the direction is changed by a GPIO bit

[GTOGuy]

Do you know how I can get the spindle to reverse through the full range, preferably without losing the new found spindle speed calibration using the lincurve implementation.
Thanks,
Pat

[brianbonedoc]

Pat - try replacing

net spindle-speed-raw motion.spindle-speed-out lincurve.0.in

with this:

net spindle-speed-raw motion.spindle-speed-out-abs lincurve.0.in

I would remove the negative values from your lincurve table as PP doesn't use them for direction - according to the master guru Peter Wallace.


EDIT: OK, this amazingly worked! I just did it on my lathe and now my rigid tapping goes in and out at 600rpm - before this small change it would only back out at the minimum in my lincurve module (about 200rpm).

[GTOGuy]

Thank Brian. I just read the code for the lincurve function and see how it works . This overall approach implemented to recalibrate the speed curve merely takes the native "motion.spindle-speed-out" vale into the function and interpolates it on the series of slopes/straight lines created between the x-y point values (up to 16) input by the user. Then the lincurve function spits out a corrected "motion.spindle-speed-out" value that replaces the original value and then the program proceeds on its original path to process the information. Probably you already understood it but I wanted to understand how it worked before troubleshooting my reversing problem.

Given this approach, it should not be affecting the decision for direction (fwd or reverse) that the machine is directing. I got rid of the negative values when I figured out it did not work and as Peter mentioned, the program sets an IO bit for direction rather than using negative values. I still can't figure out why my direction is not working. I will reboot and see what happens. Scratching my head at this point.

Pat

[PCW_MESA]

Note that master guru's crystal ball may be clouded by lazyness reading hal files and
inability to divine functionality of undocumented components

[GTOGuy]

Just saw your reply and will try it with the abs addition

[GTOGuy]

Just tried it and it solved the problem. Thanks. The lincurve function must be altering the format of the "motion.spindle-speed-out" number that is input in some manner if it needs to be converted to an absolute vale since I am not inputing any negative values.

Thanks again Peter and Brian for the help. This is a great improvement for accurate threading

[brianbonedoc]

Great glad it worked out! Actually I should be thanking you as without u making me think about it - my lathe would be backing out at a snails pace.

EDIT: For other users here's the final .hal file with the spindle calibration:

[GTOGuy]

for those using the hal file for their mills, don't forget the strip off the "mod11-" from the beginning file name and the ".txt" from the end of the file name. The 11 calibration points were from my mill vfd so yours will be different. I will say that when used, this puts my actual rpms within about 3 to 10 rpm of the commanded rpm. Love it.

[GTOGuy]

After updating to 1.9.8, I went in to change the .hal file with the calibration mods and now I cannot get the pathpilot to enter the desktop. Did I go crazy and can't accomplish what I was able to last year or is there a different way to get to the desktop now? I am using the left SHIFT and ALT keys when the big TORMACH splash screen appears.

[brianbonedoc]

Try ctrl-shift x (or maybe alt?) from PP and at the prompt type gnome-panel

This will get U too the desktop

[GTOGuy]

Thank you, I will try it

[GTOGuy]

I could not get the desktop to load for anything. Then I remembered I had switched to a wireless mouse since the last PathPilot update when I used to be able to get to the desktop using the shift and crtl keys. So I changed back to the old wired USB mouse and was able too get to the desktop during the boot up as advertised. I suppose it is the way the wireless hardware is interacting with the system??. I could not get your suggested key combination to work for me unfortunately, but thank you for the suggestion.

[brianbonedoc]

No problem. It may ctrl shift z or x or c I can't recall. But it definitely gets u out of PP and to a command prompt to which gnome-panel loads the desktop.


Sent from my iPad using Tapatalk

[GTOGuy]

still no luck with the key combinations to gtet out of PP to a command prompt. I'll search the forums to see if anyone has had any luck with other key combinations