[forseral]

Hello All,

I've built a dual x-axis machine and I'm running emc2. I'm using the keiling kit with the gecko 540 and 387oz-in motors.

I've mapped the output of the x-axis to the a-axis in emc's config. I'm finding when I advance the axis, one side advances a smidge (~.5") further than the other. As a quick test I switched the output from the g540 on the motors but same axis continued to advance further.

I realize this is not a lot to go on but I figured it was worth asking.

Anyone have any thoughts?

[Al_The_Man]

If you have eliminated the electrical/electronic possibilities then it comes down to mechanical, No?
.5" is a bit more than a 'Smidge' in my book!
Al.

[forseral]

.5" is a bit more than a 'Smidge' in my book!
Al.

ha, ya, i suppose im a little delusional

[CarveOne]

Here are the EMC2 (8.04) Configuration Wizard settings that I use. Make sure that the Steps Per settings are exactly alike for both X axis motors. I don't remember if they have individual settings or assume that for the second X motor.

CarveOne

[doorknob]

Over what distance of movement do you see the discrepancy?

Are you using ballscrews or acme screws to drive the axis? Are you absolutely sure that the pitch of each screw is identical to the other? Can you test that manually?

[forseral]

Over what distance of movement do you see the discrepancy?

Are you using ballscrews or acme screws to drive the axis? Are you absolutely sure that the pitch of each screw is identical to the other? Can you test that manually?

I'm use 1/2-8 (8start) acme screws from mcmaster-carr. I ordered both at the same time and have validated they're the same.

I see the discrepancy in under 12 inches of movement

[forseral]

Here are the EMC2 (8.04) Configuration Wizard settings that I use. Make sure that the Steps Per settings are exactly alike for both X axis motors. I don't remember if they have individual settings or assume that for the second X motor.

CarveOne

are you using your Y axis for your secondary X motor? I'm using the A-axis and i'm curious the gecko treats it "secondary" so-to-speak.

[CarveOne]

Both X motors are moving the gantry. The Y motor moves the Z assembly along the gantry.

EMC2 is using the XYZ gantry machine configuration. I never could get the XYZA configuration to synchronize the two X axis motors (X and A). I also found it difficult to find out how to get it working online. None of the descriptions I found gave anything but vague information. When you configure the wizard it will initially show four motor options. After you define X, Y, and Z motors and go to the next page of the wizard you won't see but three motor entries if you go back to that page. If you define four motors the first time they will still be there when you go back later. Weirdness. I haven't seen what later releases of EMC2 have on these pages so I don't know if there have been changes to make this slaving setup more obvious.

CarveOne

[forseral]

Both X motors are moving the gantry. The Y motor moves the Z assembly along the gantry.

EMC2 is using the XYZ gantry machine configuration. I never could get the XYZA configuration to synchronize the two X axis motors (X and A). I also found it difficult to find out how to get it working online. None of the descriptions I found gave anything but vague information. When you configure the wizard it will initially show four motor options. After you define X, Y, and Z motors and go to the next page of the wizard you won't see but three motor entries if you go back to that page. If you define four motors the first time they will still be there when you go back later. Weirdness. I haven't seen what later releases of EMC2 have on these pages so I don't know if there have been changes to make this slaving setup more obvious.

CarveOne

Odd, I dont quite understand how yours is working. Can you post your .hal file?

I configured mine manually by editing my .hal file after a basic config. Notice the bottom section highlighted in green. I appended the output pins for my A axis to the xstep and xdir commands.

net xstep => parport.0.pin-02-out
setp parport.0.pin-02-out-reset 1
setp parport.0.pin-03-out-invert 1
net xdir => parport.0.pin-03-out

net ystep => parport.0.pin-04-out
setp parport.0.pin-04-out-reset 1
setp parport.0.pin-05-out-invert 1
net ydir => parport.0.pin-05-out
net zstep => parport.0.pin-06-out
setp parport.0.pin-06-out-reset 1
net zdir => parport.0.pin-07-out


net xstep => parport.0.pin-08-out
setp parport.0.pin-08-out-reset 1
setp parport.0.pin-09-out-invert 1
net xdir => parport.0.pin-09-out

[CarveOne]

When I first run the wizard, there are the four entries. Slaving only works when the first two entries are used for the two X axis motors. They must be named X and X. X and A did not work. I found a reference to it somewhere that finally gave me the clue. I assume that both X motor entries get the same signals. It wasn't at all intuitive as to how to do this. I know nothing about messing around in the hal files.

I'll fire up the machine today and copy the .hal file for posting here.

CarveOne

[CarveOne]

Here is the .hal file for my big machine that uses dual r&p drives on the X axis to move the gantry. It has been renamed to .txt file extension for uploading.

CarveOne

[forseral]

Here is the .hal file for my big machine that uses dual r&p drives on the X axis to move the gantry. It has been renamed to .txt file extension for uploading.

CarveOne

Thanks a lot. as I suspected, your configuration uses the X&Y axis. what controller are you using, a gecko? reading your configuration made me realize I need to revisit my hold timings and such. I guess could be losing steps on motor?

i did some additional testing tonight and found 2 strange things:

first: I did some additional testing today and noticed a few strange things: When I dropped my IPM to 1ipm, the gantry stayed aligned. jogging at high speeds caused it to become out of square.

second (off topic) When i ran the gantry from 0 to 20" back and forth without pause, the gantry would not make it to the 0 and 20" position until the last

[CarveOne]

Both of my machines use Gecko drives. The big machine has four 495 oz Keling motors and G203V drives running on a 72vdc PSU. The smaller machine uses G540 and 380 oz motors with 48vdc PSU. The EMC2 wizard has been configured to run both of the newer steel machines, if needed, as well as the original wood machine with three 425 oz motors and 3 axis Xylotex drive and 24vdc PSU.

As Al_The_Man implied, make sure that your mechanical parts are moving easily when the lead nut is disconnected. (If that is what you are using) Mechanical misalignment must be corrected before anything else or the motors may not be able to overcome it.

CarveOne

[JerryBurks]

Sorry if that is a dumb question....but why do the x-motors need to be connected to 2 separate slaved pulse outputs? I was planning for my new project to wire a single x-axis control output in parallel to both x-motor drive inputs to guarantee the same pulses (I have separate motor drives).

That is, if nobody here tells me that is not a good idea...

As for losing steps, on my current CNC I found going overboard with the microstepping (like 1/50) made the machine lose steps randomly even at moderate speed. At 1/8 it works just fine.

[doorknob]

Sorry if that is a dumb question....but why do the x-motors need to be connected to 2 separate slaved pulse outputs? I was planning for my new project to wire a single x-axis control output in parallel to both x-motor drive inputs to guarantee the same pulses (I have separate motor drives).

That is, if nobody here tells me that is not a good idea...

As for losing steps, on my current CNC I found going overboard with the microstepping (like 1/50) made the machine lose steps randomly even at moderate speed. At 1/8 it works just fine.

Wiring the motors in that way will make it difficult for you to reestablish synchronization in the event that the motors ever become unsynchronized (for example, due to missed steps on one motor).

If you use a separate axis and driver for each motor, and have separate home switches, then Mach3 has the capability of resynchronizing the motors.

[JerryBurks]

Wiring the motors in that way will make it difficult for you to reestablish synchronization in the event that the motors ever become unsynchronized (for example, due to missed steps on one motor).

If you use a separate axis and driver for each motor, and have separate home switches, then Mach3 has the capability of resynchronizing the motors.

Ah, thanks, learned something new. Not sure if my Planet-CNC controller can do that but I was anyway planning to force-synchronize the 2 x-motors mechanically with a timing belt. Besides the accuracy issue I am worried that a serious x-offset (like a completely stalled motor) may bend the gantry.

And sorry for the partial thread highjack....

[forseral]

.

As Al_The_Man implied, make sure that your mechanical parts are moving easily when the lead nut is disconnected. (If that is what you are using) Mechanical misalignment must be corrected before anything else or the motors may not be able to overcome it.

CarveOne

I validated this and they move at the slightest push (im using the glacern sbr rails). More testing today showed that the lead screw with the problem (6ft Acme 1/2 8 start from mcmaster-carr) has a very slight warp in it. the warping is on the opposite end.

Do you think this could be the culprit?

[CarveOne]

Yes, it can be, especially if the worst of it is at one end of travel near the mount. The rod is not as flexible there and won't bend as easily because the mount restrains it. Move your gantry to that end, loosen the mounting screws to let it self center, then re-tighten the screws. It may help.

Do this to both ends. Then you may need to re-adjust the lead nut mounts to get them centered with respect to the end mounts as well. All you can do other than straighten the screws is to replace them with straighter screws. It is a little tougher to get them aligned when the screws are this long. Any serpentine spiral bend in the screws will be more noticeable by eye. This comes from the stresses created by grinding or roll forming of the threads during manufacture. A simple bend usually happens during handling.

CarveOne

[forseral]

Yes, it can be, especially if the worst of it is at one end of travel near the mount. The rod is not as flexible there and won't bend as easily because the mount restrains it. Move your gantry to that end, loosen the mounting screws to let it self center, then re-tighten the screws. It may help.

Do this to both ends. Then you may need to re-adjust the lead nut mounts to get them centered with respect to the end mounts as well. All you can do other than straighten the screws is to replace them with straighter screws. It is a little tougher to get them aligned when the screws are this long. Any serpentine spiral bend in the screws will be more noticeable by eye. This comes from the stresses created by grinding or roll forming of the threads during manufacture. A simple bend usually happens during handling.

CarveOne

thank you so much. i ordered a new screw and replaced it yesterday. the problem is solved. the misalignment dropped from .5" to 1/16". I'm guessing this 16th is something else!

[forseral]

And sorry for the partial thread highjack....

no worries!