My machine is making tapered squares

[bbergami]

I recently finished my second machine. I spent a little time trying to adjust all the settings. So I set it up to cut 3" test squares in mdf. On my Y axis...no problem...just make a minor adjustment in the steps per inch. On my x axis I was getting about a .012" taper in 3" from one side of the square to the other.
I've got dual ballscrews on "x" tied together with timing belts...run from a single servo. The screws are supposed to be within .004"/ft and the timing pullies are from StockDrive. So I'm assuming that's not the problem.
What else would cause this?
Thanks
Bert

[Al_The_Man]

Originally Posted by bbergami The screws are supposed to be within .004"/ft and the timing pullies are from StockDrive. So I'm assuming that's not the problem. Bert

Bert, I would not assume it untill you actually prove it and then eliminate it.
Use a dial gauge or better still at least a 6" digital micrometer, set up gauge block of some kind on the non-moving part of the table and use or make up a measuring block on the moving part. move the table to a position in the direction you intend to travel to take the measurement and set your mic on these points to zero.
Program a move in the X in the same direction a distance within the range of the mic. Do this a few times to confirm the reading. Do the same on the other side you will be able to confirm wether or not there is an error
If they are the same then the next step is to test for backlash in one side.
This can be done in a similar manner, only move the axis in the above direction and then back a similar or same amount and if a difference exists there is backlash.
Al.

[acondit]

Draw a picture of the "square" laying on the router showing the dimensions, maybe that will help me see what you're talking about?

You might try cutting a 3" square just about 0.100 deep inside a larger square maybe 6" to see if that identifies anything.

Alan

[bbergami]

I've used a dial indicator on the gantry to measure a 1" move. I went back out and checked it on both sides. While it is still out of adjustment by about .0035" they are within a half a thousandth of each other.
This time I loaded the dxf into Vcarve wizard instead of Deskcnc and generated the gcode for the 3" square. This time they were parallel and about 3.010". I'm still using DeskCnc to run the machine. Could the software be the culprit?

[JFettig]

How did you square the table up? it sounds like you made a gantry router? if so, how did you square everything up, it could be slightly cocked or one of the screws is turned a little bit pushing one side of the gantry to the side and when it goes across, it wont be square.

Jon

[bbergami]

Acondit, these are the dimensions in the x direction

[bbergami]

JFettig, I thought at first that it might be out of square. I set it to square by cutting a groove in x and y. Using feeler gauges to measure, then adjusting one side with a dial indicator while holding the other. But even if it was out of square the 2 opposing sides should be parallel, shouldn't they?

[dertsap]

could it be one screw has more backlash than the other ,or is the center of the backlash equal on both screws , what i mean is if you move one way then the other is the one screw contacting the opposite side of the nut before the other screw , this may cause binding , if you x and y axis are running perfectly perpendicular , i would think it s the drive system ,check the dia of the pulleys to make sure the are perfectly equal to each other , if the manufacturer has a +- .005 tolerance on the pulley dia , this could be your problem , if you have a pulley with a 1" dia =3.142 circumference add .005 you will have 3.157 circ. i don t know how muchof a factor this would play , but i would check it ,
is it possibly cutter deflection
even if it is out of square the two sides should be parallel , you will just end up wuth a parallelagram then a square part

[posix]

I would say 0.3048 millimeters is good enough for woodworking, or?

[Al_The_Man]

Originally Posted by bbergami Could the software be the culprit?

I can't see the software being the culprit if the G code indicates the same dimension on both sides of the rectangle and you are obviously using the same common motor to drive the x axis, but if you are moving in the opposite direction and one side is out, it may indicate a backlash problem on one side, or missed steps in one direction.
Al.

[bbergami]

I tried comparing the g code on each program. I really haven't gotten to the part about understanding g code yet...so it doesn't tell me anything.
I did check the backlash on each side. It's not but about 4 or 5 steps. The motors are servos so I don't think lost steps is a problem. The gantry runs smooth with no binding.

[High Seas]

Originally Posted by bbergami The gantry runs smooth with no binding.

Under servo power - or by hand?
Have you checked that no cables, leads, or other mechanicals are pulling it slightly out of whack? That bit me in the butt once - it was on the backside where I couldn't see it readily. When I moved to the center of the floor - WHACK right upside da head!
Good luck - Jim