Machsupport Forum
Mach Discussion => Mach4 General Discussion => Topic started by: Dan13 on March 07, 2015, 04:04:51 AM
-
In the announcements it says that version 2251 has added screw error mapping ( see here ) (http://www.machsupport.com/forum/index.php/topic,27039.msg205259.html#msg205259).
But I can't see this... where do I find it? Any info and instructions available?
Dan
-
It is new and being tested. When it runs out good, we will document it in the setup guide.
Steve
-
Thanks for the information Steve. Waiting forward to playing with it.
While at it, just would like to make sure we are talking about the same thing. I have a 4.5m long cylindrical grinder. It is real hard to get the 4m long linear guides absolutely straight by mechanical means. I have done so to about 0.05mm. So when grinding a 4m long shaft the fluctuations in diameter are up to 0.1mm. I would like to be able to map the Z axis so that for each Z value, the X moves to correct the straightness error. Is this the feature you're working on now? Asked this on the Yahoo group and had Brian answer this was it, but now I am afraid we didn't understood each other.
Dan
-
I don't think you two understood each other. What you are talking about is kinematics. Screw mapping is the mapping the error in a screw or rack in only one axis of travel.
Steve
-
:( So no kinematics? Is it possible? I know I can do it with a PLC - the pulse train from Mach would go through a PLC which would have the mapping data and add or subtract a few steps to account for the error. But having this in Mach would greatly simplify things.
Dan
-
Eventually we will have a kinematics module. But for now we only have screw mapping and that will not do what you want. :(
Steve
-
Thanks Steve.
Dan
-
Hi Steve,
Will screw mapping apply to rack and pinion systems as well?
Also, I want to make sure I understand how it will work. I am looking at routers now and a typical accuracy I see is: .02mm/300mm which over an 2500MM x axis equates to .0065" . With screw or rack mapping, if I can get someone in here with a laser or some other super accurate measurement system and I can map the screw or rack accurately, then the error I see should be limited to the backlash in the system plus the mapping error. Is that correct?
Thanks,
Bill
-
Screw error mapping will work for the rack. But it as yet untested. So I would not say it is ready for prime time.
Steve
-
if I can get someone in here with a laser or some other super accurate measurement system and I can map the screw or rack accurately
Look into folks that do optical allignment. ( 0.00078" per foot accuracy, well, make sure you understand the spec )
Screw mapping takes time and if opticaly done must be done using calibrated scales and equipment, else, maping is is only as accurate as the uncalibrated system.
Get ready to spend some money to have it done.
RICH
-
Any update on this?
-
Have any of the new mach 4 updates added screw mapping yet?
-
I wondered if maybe screw mapping was coming soon.
Without it Mach4 is just a toy to me since I can't make anything precise. I'd really like to switch from my current cnc software but I need this. I'm including a chart to show what I mean...
Thanks.
Mark
-
Unfortunately, it is difficult, not tested, and not documented. :( So it is not ready. And I don't know when we will get to it. We have to prioritize. :(
-
:( So no kinematics? Is it possible? I know I can do it with a PLC - the pulse train from Mach would go through a PLC which would have the mapping data and add or subtract a few steps to account for the error. But having this in Mach would greatly simplify things.
Dan
hey Dan. you say you are correcting the other axs using a plc script?
i also would like to do the same thing but... is plc fast enough to compensate while gcode is running?
thanks.
-
Never eventually did it. Was just a thought. I think it would depend on what you've got for a controller (defines the signal width) and what is the max step pulse frequency. Modern PLCs can have scan times in the range of 0.1mS. My feeling is it could be not enough if you have high frequency step signal. But you could do simple math calculation to determine.
Another option could be an Arduino type microcontroller, which has much faster processors. The fastest I have seen from Arduino is 84Mhz. I think it should easily accomplish the task if you program it correctly.
Dan