Calibrating a machine

What are the steps to calibrating a machine once it is finished?

1. Square up the X and Y to each other -- no problems
2. Square the Z up to the plane of X and Y -- no problems
3 Perform velocity tests to find max velocity -- Not bad at 5inches/sec
4. perform acceleration tests to find max acceleration -- again not bad 30inch/s^2
5. Calibrate the reduction to get highest accuracy?

This is where I am stuck in a loop. The timing belts are a given pitch. I have a set gear reduction. When I enter these numbers into my machine and let it move back and forth my dial indicator is showing that the machine is out 0.001 over a half inch and 0.003 over 3/4 of an inch. This is completely repeatable and if I move it back and forth over and over and at different locations of the machine I get the same results. What do you do to dial this in? In most of my parts this won't matter, but over the length of the table I could be out over an 1/8th of an inch. Of course I don't have a dial indicator that has a 48 inch swing in travel so I can only measure over an inch. I can send the table all the way down and back and measure, but how do you know it went where you told it to?

The longer the distance you can measure from, the more accurate of an idea you're going to have of how far off you are. Can you put a sheet of material on the table (preferrably something heavier so it can't shift around) and position the head at 0,0, use a framing square against the sheet and head and mark the position.......then command it to move...say 40 inches or so in the X direction and mark that position with the framing square and use a tape measure over that 40 inches to see where you're at? If you have someone who can hold the dumb end of the tape measure exactly at the 1" mark (instead of trying to use the hook) you should be able to get a pretty darn accurate measurement. NO, you won't be able to see thousanths, but over the length of 40", if one end of the tape is dead-nuts on 1" and the other is not dead on 41", at least you'll know which way you need to go and approx how far. Once you get both marks to be dead on with the tape measure, you'll most likely be accurate enough that....well, it just won't matter.

You're going to be able to measure the accuracy of the machine a lot more accurately over a 40" span with a tape measure than you can over a 1" span with a dial caliper. Using your example, if you're off by .004" for every inch, in 40 inches you're going to be off .160"......you'll definitely see that on the tape measure with no problem!

Last week I did just what vacextar said to do. I clamped a tape measure in place so that it could not move, only the tape. I then connected the tape to the router and sent it all the way to the end of travel. This proved to be very accurate. Later, I borrowed a 24 inch digital caliper set from work and retested the entire machine. I found that I was within .008 over a span of 24 inches. From there I went into the settings tab of mach and re-calibrated. I am amazed at how accurate these machines can be.

-Scott

I figured that I was over thinking this. I went home over dinner and used the tape measure trick. Basically I put the machine to 0 on X and Y. Marked both sides with a razor blade on part of the rail that isn't used for motion. Moved it to 40 inches and measured the error. I went back and adjusted the settings and tried again. After just a couple adjustments I am pretty much dead on. Each time I sent it back I double checked the 0 mark. Then moved it back to 40 and marked with a razor again. Both sides are now inside the mark on the tape. Actually they are both right at the edge of the mark.

I am going to put a dial caliper back at 0 after I finish the night shift. I am curious if it is returning to the same 0 or how close it is. Overall I was just over 9/16ths off at one end of the table with my initial settings. The great things is that I am using a single drive, with a driveshaft, and dual timing belts on the X. Both sides continually measured the same and I sent it back and forth a few times to check repeatability. I was sure that the shaft would cause some racking to happen as it twisted, but so far it isn't anything I can measure.

For completeness this is how I did it:

Step 1: set initial gear reduction settings in the software -- for me this is 10:44 for the main drive reduction and 0.360..... for the rack revolutions per inch

Step 2: Using a square, mark on the rail or a piece of material the 0 position

Step 3: Give a command to move the axis across the table. I simplified here and used a simple G1X40F120 command to make sure I wasn't dealing with any rapid move problems

Step 4: Mark the 40 inch position
Step5: measure
This gives you an error so if you measure 41 inches you can follow the rest.

Step 5: calculate adjustment: 41 inches/rack revolutions/inch = revolutions used. Then you can divide the error by the revolutions and use this number to adjust the rack ratio

So Distance Measured/ rack ratio = revolutions
distance measured - distance commanded = error
error/revolutions = rack ratio adjustment
Original rack ratio +/- rack ratio adjustment = new rack ratio.

The +/- dependso on if you are long or short. Short you add and long you will subtract. I did this a couple times and got it closer than I can measure.