Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: AiR_GuNNeR on July 04, 2011, 10:08:01 AM

Title: Is backlash compensation broken?
Post by: AiR_GuNNeR on July 04, 2011, 10:08:01 AM
I recently completed a Little Machine Shop X2 type mill CNC mill and have the latest Mach3 v3.043.022 installed.  I have for sake of discussion .005" backlash entered into the X-axis.  If I jog to the left, and then jog to the right, my first reverse direction will actually jog a little more the the left.  The next jog to the right will then kick in the compensation.

In the HSM forum, someone suggested v3.042.040 which I installed.  It at least moves the right direction on the first reversal, but the full backlash comp does not kick in until the second jog.

Am I wrong to expect the first reverse jog to command the stepper to rotate the full backlash amount plus the current jog amount?  I have backlash comp set to the default 60% max speed.
Title: Re: Is backlash compensation broken?
Post by: RICH on July 04, 2011, 04:34:32 PM
Backlash is implemented when a change in direction takes place.
You should initially remove any backlash by jogging a short distance in one direction before reversing the direction.
I don't remember anything wrong with BC in V3.043.022.
RICH
Title: Re: Is backlash compensation broken?
Post by: AiR_GuNNeR on July 04, 2011, 05:07:12 PM
Backlash is implemented when a change in direction takes place.
You should initially remove any backlash by jogging a short distance in one direction before reversing the direction.
I don't remember anything wrong with BC in V3.043.022.
RICH

That is what I thought should happen.  During the CNC build class, a couple of ArtSoft gents stopped by.  I showed one of them the odd behavior, and he was scratching his head on this behavior as well.  At the time, I didn't notice that the first jog in the opposite direction was actually advancing a table a bit in the wrong direction, as I was concentrating on the obvious sound that is made when the backlack comp. sets in.  I was very evident that the BL comp was happening on the second jog.
Title: Re: Is backlash compensation broken?
Post by: RICH on July 05, 2011, 07:22:16 PM
I am using version 3.042.033 and know for a fact that there is no problem in that version here.
Backlash settings that were used:
- 50% of max speed
- 0.005 Shuttle Acceleration
- Debounce interval 500

Can you post your xml file?
Do you have any break out boards, using the PP, etc?

You may want to try a slower percent, Shuttle Acceleration of say .010 of even .05 so you can watch the backlash get implemented at a snails pace.
Testing my memory as i don't use BC anymore...... ;)

RICH

 
Title: Re: Is backlash compensation broken?
Post by: AiR_GuNNeR on July 05, 2011, 09:07:48 PM
Hi Rich,
Thanks for the response.  I'll pull the xml file from the laptop.  
I don't have a shuttle, simply using the cursor keys on my keyboard to jog.  Does the shuttle setting affect that as well?
I'm not sure how to post it, so I put it on my server:
http://www.compufoil.com/HiTorque.xml
I am running the smooth stepper on my system.

I'll download and try out 3.042.033 as well.
Title: Re: Is backlash compensation broken?
Post by: RICH on July 06, 2011, 06:53:35 AM
Check that you have the latest SS plug in. Don't know how well BC works with the SS since it was implemented so maybe someone else can remark about it's BC functionality.
You may also want to visit  the WARP9 site since there are threads on BC  and the SS.

To post your xml fle give it a unique name like AIRGUN.XML and use then click the Additional Options which allows you to select the file from your pc to upload.

RICH
 
Title: Re: Is backlash compensation broken?
Post by: AiR_GuNNeR on July 06, 2011, 07:33:13 AM
Hi Have HiTorque.xml attached for you.
Title: Re: Is backlash compensation broken?
Post by: AiR_GuNNeR on July 06, 2011, 07:43:53 AM
Thanks for the hint on the Warp9 site.  This is from their downloads site of the latest development version.  It looks like the smooth stepper is the problem...so...the next question I have is can Mach3 be configured to use both the parallel port as well as the smooth stepper?  Parallel port for the motion control, and SS for the additional IP?

From Warp9 site:
"- Backlash compensation has been added. This version works great as long as you do not enter moves that are shorter than the amount of backlash you have. In particular, you will notice this if you use MDI or Step Jogs and you inch towards something and then back off. I am working on a version that will make that work properly. "
Title: Re: Is backlash compensation broken?
Post by: RICH on July 06, 2011, 09:56:29 PM
Quote
Parallel port for the motion control, and SS for the additional IP?
Don't know for sure but don't think so.

Quote
From Warp9 site:
Glad to see that Greg got something implemented for BC.
I sovled the BC problem by eliminating it from my machines and it was worth the effort.

RICH
Title: Re: Is backlash compensation broken?
Post by: microsprintbuilder on August 11, 2011, 12:07:35 AM
I can not get backlash comp to work either
no matter what I put in for a number nothing changes.
Title: Re: Is backlash compensation broken?
Post by: RICH on August 11, 2011, 06:01:51 AM
micro......
Quote
I can not get backlash comp to work either no matter what I put in for a number nothing changes.
We need some more info:
Version of Mach your using? Using PP or SS and any breakout boards?
DId you check the box in CONFIG>BACKLASH to implement it, save settings , and restart MACH?
Can you post your xml  file?

RICH

Title: Re: Is backlash compensation broken?
Post by: AiR_GuNNeR on August 11, 2011, 07:45:02 AM
It works on my setup to an extent.  The deal is you won't see the compensation if the distance of the reversal is less than the compensation you had dialed in.  For instance, say you have .005" dialed in.  If you jog left in .001" increments, then reverse directions, you would expect, and done properly, you should see the stepper twitch the .005" plus the .001" commanded.  As implemented, you won't see the backlash compensation kick in until the second or third reversed step.

In practice, if the reversal distance is greater than .005", then the BL would compensate properly.  If you are machining a round pocket, especially a larger one, the reverse step may well be less than that and your hole will not be smooth.

The parts I am making are not only functional, but also cosmetic.  They have six larger holes in them I used to bore with a boring head.  I had hoped to mill them to speed things up, but they didn't look good enough with the faulty BL compensation.  I ended up removing the SS, bought a C10 board for use with my old desktop with the parallel port until the driver gets fixed.
Title: Re: Is backlash compensation broken?
Post by: Hood on August 11, 2011, 07:49:50 AM
bought a C10 board for use with my old desktop with the parallel port until the driver gets fixed.

Are you using Backlash via the PP?
Hood
Title: Re: Is backlash compensation broken?
Post by: AiR_GuNNeR on August 11, 2011, 08:13:54 AM
Sheesh, I thought I was in the Warp9 (Smooth Stepper), forum...
The BL compensation is broken when using the smooth stepper.  The PP works fine however.
Title: Re: Is backlash compensation broken?
Post by: Hood on August 11, 2011, 08:26:49 AM
So you are getting perfect circles with no marks when using backlash in the PP?
 I would have thought with physical backlash that comp would never stop that happening unless you are taking a very fine finish cut, but have never messed much with the backlash in PP or SS. I do use it in the SS for rigid tapping and its great for taking the lash out of the spline drive on the spindle although I rarely do rigid tapping anyway.

Hood
Title: Re: Is backlash compensation broken?
Post by: microsprintbuilder on August 11, 2011, 09:03:37 AM
I have been using R3.043.022 after reading some post I am using the R3.043.039. I'm not sure what PP or SS is? I followed the procedure on setting backlash in the box and restaring the program. I am using a galil card and servos. How do I post my xml file?
Title: Re: Is backlash compensation broken?
Post by: Hood on August 11, 2011, 09:09:14 AM
Not sure if the Galil supports backlash, you would have to ask on that part of the forum.
Hood
Title: Re: Is backlash compensation broken?
Post by: AiR_GuNNeR on August 11, 2011, 09:43:30 AM
I have been using R3.043.022 after reading some post I am using the R3.043.039. I'm not sure what PP or SS is? I followed the procedure on setting backlash in the box and restaring the program. I am using a galil card and servos. How do I post my xml file?

PP is parallel port, SS is smooth stepper.  The SS allows you to use a laptop with Mach3 and runs through the USB port.  It gives you two parallel ports to work with and works real well for the most part.
Title: Re: Is backlash compensation broken?
Post by: Sage on August 11, 2011, 05:10:34 PM
Excuse me if I'm missing something here but what difference does using the PP or SS have in backlash compensation. My understanding is that Mach simply issues additional ordinary steps in a particular direction to complensate just before sending out the regular steps - NO?
 Why would it need to be handled any different through the SS. OR why would SS choose to handle it any different if it is handled already by mach when directly connected by the parallel port?

Sage


Title: Re: Is backlash compensation broken?
Post by: Hood on August 11, 2011, 05:19:14 PM
External controllers such as SS, DSPMC, Galil etc etc use a plugin to talk to Mach so things like backlash comp need to be handled by the external controller.
 I am not sure how many of the external controllers actually do backlash comp.
Hood
Title: Re: Is backlash compensation broken?
Post by: Sage on August 11, 2011, 07:56:43 PM
Got to wonder why they take on the task when it seems so simple and effective to let Mach take care of it directly. Especially considering (apparently in the case of SS) they screw it up and leave you high and dry with nothing that works.
 Of course, the best (and only reasonable) solution is to not have any backlash in the first place. I've found the compensation to be marginal at best anyway.

Sage

Title: Re: Is backlash compensation broken?
Post by: Hood on August 11, 2011, 08:06:45 PM
I dont think there is much choice or I am sure they would let Mach handle it, backlash comp is likely part of the Mach Driver so if you dont use the Mach Driver, ie you are using an external device, you have to do things like backlash in your device.

Backlash worked well for me with the SS in the testing I did, even slackened off the bearings to make more ;D Holes came out nice and round, just slight marks at the quadrants but I would imagine that would happen anyway as comp cant stop cutters pushing/pulling an axis around if its slack..

Hood
Title: Re: Is backlash compensation broken?
Post by: RICH on August 11, 2011, 08:07:17 PM
Quote
seems so simple

Sage,not sure that is the case with the SS , since if it was ,Greg would have solved the implementation of it a long time ago.
Yep, no practical backlash is the only way to go, but, BC served me well for a long time. In fact a friend used BC on his lathe which used 1"- 8 TPI
cheapo all thread rod  with about 1/16"  backlash and you would be amazed at what he made and the accuracy he was able to achive.

RICH