Display and keyboard delay with 10.04

More
11 Apr 2014 01:40 #45813 by jcarpenter
I have a new install of 10.04 replacing 8.04. The Axis display has a 1/2 second delay from where the lathe is actually running at. Axis also jumps at each update instead of steady following the toolpath. Keyboard entry has a 1/2 second delay. Makes for interesting zeroing operation! I am using an Intel 1.5 Ghz processor with 1 Gb Ram. I have been running 8.04 on this Denford ORAC lathe for years. After the computer fell off a shelf (5 feet onto concrete) I had to replace the harddrive and repair the power supply. As I couldn't find my iso of 8.04 I used 10.04. Now everything is crap. Is it the new software? Do I have an undiscovered hardware failure? How do I diagnose this? I am a Linux noob.

Thanks,

John Carpenter

Please Log in or Create an account to join the conversation.

More
11 Apr 2014 03:28 #45818 by jcarpenter
Well, I went back and downloaded and reinstalled 8.04. And got most of my delay removed. Probably what is left I never noticed before. Why is 10.04 such a CPU power hog? Or is somthing going on that I could fix with a settings change?

Thanks,

John C

Please Log in or Create an account to join the conversation.

More
11 Apr 2014 06:30 #45827 by BigJohnT
If you have an older PC often 8.04 is better than upgrading to 10.04.

JT
The following user(s) said Thank You: jcarpenter

Please Log in or Create an account to join the conversation.

More
11 Apr 2014 13:41 #45845 by ArcEye

As I couldn't find my iso of 8.04 I used 10.04. Now everything is crap. Is it the new software?


It is the old story of, higher numbers are not better.
Microsoft have brainwashed people to believe the opposite so that they continue to buy their newest offerings, when in truth they haven't made anything that works halfways decently since XP.

8.04 is a Ubuntu distro number, nothing to do with Linuxcnc.
The important thing about the Live CD based on this distro, is that it contains a realtime kernel built for uni-processor machines.

10.04 based Live CD contains a realtime kernel for SMP machines (multi core or multi processor)

Choose what matches the computer, not the one with the highest number

regards
The following user(s) said Thank You: jcarpenter

Please Log in or Create an account to join the conversation.

More
11 Apr 2014 18:01 #45846 by jcarpenter
Ok, I will stick to the LinuxCnc from 8.04. Seems to operate fine. Now I just have to figure out why the lathe displacements are only 75% of the software displacements.

Thanks,

John C

Please Log in or Create an account to join the conversation.

More
11 Apr 2014 18:26 #45847 by ArcEye

Now I just have to figure out why the lathe displacements are only 75% of the software displacements.


If you mean that you lathe only moves 75% of the commanded distance, that will be simple.

Open the .ini file, each [AXIS_x] axis section has an entry SCALE =

That number is arrived at from the figures you entered in stepconf and equates to the number of steps required to move an axis by 1 machine unit

If your numbers are for example SCALE = 300, then they probably need to be 400

Hopefully it is an exact fraction of the current value, most likely caused by wrong gearing ratio being entered or similar

regards

Please Log in or Create an account to join the conversation.

More
13 Apr 2014 20:41 #45907 by jcarpenter
Arceye,

Both my axes are off by the same amount. My math checks out as far as I can see.Here is the info:

x axis - Gecko 540 at 10 microsteps
Stepper at 200 steps per rev
Motor pulley 15 teeth
Cross lead screw pulley 15 teeth
Cross lead screw pitch has 2.5mm per rev
That works out to be 10.16 rev per inch. So 200*10*1*10.16 = 20320 which is exactly what my scale value is from the .ini file.

z axis - Gecko 540 at 10 microsteps
Stepper at 200 steps per rev.
Motor pulley at 15 teeth
Leadscrew pulley at 30 teeth
Lead screw has a pitch of 5mm per rev
That works out to be 5.08 rev per inch. So 200*10*2*5.08 = 20320 which is exactly what my scale value is from the .ini file.

When I move the x and/or y axis .125 inch on the software the carriages move .100 inch.

I have an early Gecko G540 - is it possible that it is doing 8 microsteps? If not I guess the next step would be to tear the lathe apart and measure leadscrew pitches and count sprocket teeth. Or just use ratio the scale until it is accurate and not worry about what in the math is wrong.

John Carpenter

Please Log in or Create an account to join the conversation.

More
13 Apr 2014 21:03 #45908 by ArcEye
Hi

The Denford Orac should have 5mm and 2.5mm pitch leadscrews.

However you are in the US, I don't know how you came by it and whether Denford ever produced imperial versions for export

The G540 does come with 10x microstepping, from which it gradually switches to direct 1x drive with increasing speed, but this does not affect the calculation, even at high speed it still takes 10 pulses to trigger one full step.

The difference is compellingly even, it moves 80% of the commanded distance, if your figures are correct.

I would be tempted to use that ratio to increase the SCALE figure and check the result

regards

Please Log in or Create an account to join the conversation.

More
13 Apr 2014 21:27 #45910 by jcarpenter
I went back and looked at the original retrofit photos. The cross lead screw definitely has a larger sprocket than the cross lead stepper motor. This means the literature that I have declaring 1 to 1 and 2 to 1 leadscrew ratios is probably in error. I wish the original harddrive with the stepconfig files could have been recovered. But I am not going to tear the lathe apart to get the correct ratios. I'll just make the correction to the teeth number based on the error and go on. Thanks for the help.

John Carpenter

Please Log in or Create an account to join the conversation.

More
13 Apr 2014 21:50 #45911 by jcarpenter
Went back and found the calculation sheets for the retrofit 4 or 5 years ago. Astonished the paperwork was still around. Both x and z steppers motors have a 12 tooth sprocket, the cross lead screw has a 15 tooth sprocket and the leadscrew has a 30 tooth. Mystery explained. Good to go!

John C

Please Log in or Create an account to join the conversation.

Time to create page: 0.249 seconds
Powered by Kunena Forum