PDA

View Full Version : calibration



luke11cnc
24-01-2012, 10:06 AM
Hello all
I have a small issue with my y-axis and would assume the same with my z-axis as they are both on pulleys

the issue I have is calibration

this only came to light when I tried to cut my dragon out on Sunday night when I tried to cut a distance on my y-axis off 300 mm as that was the size of my model and when I measured it it was 286 mm. at first I thought it was the model software and the scale factor, so I set my spindle to cut at 286 mm and to my surprise my spindle came up short.

well today I have set it to cut/drill three hole in a triangle of 100 mm and it's cut it out at around 95 mm

how do I recalibrate my CNC Machine Mach3 is the software I'm using


James:exclaim::exclaim::exclaim:

Jon S
24-01-2012, 11:04 AM
goto the menu :CONFIG then :MOTOR TUNING under that.

In the left lower corner of the dialogue box is "steps per".

Your error is not consistent, so there is more to this, it looks like you need to move 5% more which needs 5% more "steps per" in the triangle example the error is closer to 10% presuming it is an equilateral triangle.

The same dialog enables you to set the acceleration, if that is wound up too high then it will cause you problems, so till you have steps per correct I suggest you wind the acceleration right down.

You need to remove backlash from the equation by moving in a direction before stopping and then continuing motion for the purpose of measurement. So dont just move from Y0 to Y100, first move to Y-10, then Y0 then Y100 and measure the Y0 Y100 progress.

Note the settings before you hack them ;-) Someone cleverer than I might be able to tell you exactly what to change it to but suck it and see has always worked for me.

hope that helps you.

Jon

JAZZCNC
24-01-2012, 06:55 PM
James ring me and I'll walk you thru sorting this it's not a problem just probably slightly over tuned or miss calculated steps.. . .Easy fix.!!

Jonathan
25-01-2012, 11:21 AM
It's a 2:1 ratio with 500mm pitch screw, so 10mm effective pitch and (I think) 1600step/rev. That means in mach3 it should say 160step/mm, or at least that's what I would have entered for you (or 80 if it's actually on 800step/rev). I can't see why it would have changed unless you accidently changed it when I suggested lowering the acceleration, but even so if the error is inconsistent something more interesting is happening.

Looks like it was a right angled/isosceles triangle as 95*300/100=285mm. If you're just using a ruler that easily accounts for the 1mm difference. I'm surprised as I checked it with digital calliper over 300mm and it was spot on.

I told Luke how to work these values out, see if he remembers :smile:

Feel free to call me too if Jazz or Luke haven't already sorted it.

luke11cnc
26-01-2012, 07:34 PM
I would like to thank Jazz for helping me calibrate my CNC by using Mach3 to work out the correct steps per mm

all calibrated now cheers Jazz

and I've cut my collet holder out now

James

JAZZCNC
26-01-2012, 08:10 PM
I would like to thank Jazz for helping me calibrate my CNC by using Mach3 to work out the correct steps per mm

all calibrated now cheers Jazz

and I've cut my collet holder out now

James

Pleased could help and glad your sorted and cutting true. And your further down the road than me if you've cut your collet holder.!! One of them I'll do it later when I've spare time job's, but I,ve only had them 3yrs so there's time yet. . :dance:

Jonathan
26-01-2012, 09:49 PM
I made collet holder by merely drilling a line of holes in a spare piece of wood with the right size (quite big) drill. Works perfectly well...the next one will be incorporated into the gantry and store them in a circular array. Prizes for guessing why!

What was wrong with the values in the end? Strange that they would change for no apparent reason.

JAZZCNC
26-01-2012, 10:27 PM
I made collet holder by merely drilling a line of holes in a spare piece of wood with the right size (quite big) drill. Works perfectly well.

Mine involves a piece of wood but it's even more techni-not than that.!!. . . . 13 nails with heads snipped.. :toot:

luke11cnc
26-01-2012, 10:39 PM
Prizes for guessing why!

what do we win Jonathan (tool changer for free) LOL

I haven't got a clue Jonathan maybe the belts have stretched a bit ?? but it was out about 6%

James

Jonathan
26-01-2012, 11:06 PM
what do we win Jonathan (tool changer for free) LOL

No idea...but chip won it. The ATC mechanism seems solid, just need to solve issues with the bearing arrangement and it'll be ready to go.
The general idea is to incorporate two tool racks hidden *inside* the gantry, one either side with 8 tools on each and have it so the Z-axis retracts inside the gantry, tool holder slides under spindle and rotates to the right tool ... tool is changed then the Z-axis drops down to carry on as normal. The only issue with this plan is it will add overhang to the Z-axis, reducing rigidity, and loose a bit of travel. I've got 400mm of travel to play with, so that's not a big deal. Using 760mm profile rails. I've had the profile rails for a long time but not been able to decide or afford to mount them properly.


I haven't got a clue Jonathan maybe the belts have stretched a bit ?? but it was out about 6%

Timing belts don't stretch by 6%! Even if they did it's the number of teeth that determine the ratio. What number is in step/mm for the Y-axis under config-motor tuning?


Mine involves a piece of wood but it's even more techni-not than that.!!. . . . 13 nails with heads snipped.. :toot:

So you've only got the metric collets.

luke11cnc
26-01-2012, 11:26 PM
step/mm is 85.46301769

James

Jonathan
26-01-2012, 11:51 PM
step/mm is 85.46301769

That's very interesting...I just don't see how with 30T pulley on motor and 15T pulley on the 10mm pitch screw it could be anything other than 80 step/mm. Unless the pulleys got mixed up somehow. Sounds silly but could you count the teeth on the Y-motor pulley. I reckon bearingstation accidently sent me 28T pulleys and neither of us noticed as that gets 85.71429 step/mm which is very close. If that's the case you should check Z and put the 85.714 in for Z.

Jazz did you arrive at that number by scaling and measuring, or some calculation?

JAZZCNC
27-01-2012, 12:19 AM
Jazz did you arrive at that number by scaling and measuring, or some calculation?

I could tell you but then I'd have to kill you.!!. . .:rofl:


I did nothing other than explain to James how to let Mach do all the hard work and use the Set steps Per unit feature on the settings screen. . . . Simplizzzzzz

JAZZCNC
27-01-2012, 12:23 AM
So you've only got the metric collets.

Nope they are long nails. . :joker:

Jonathan
04-02-2012, 04:25 PM
James did you check how many teeth are on the Y-axis motor pulley? The mach3 wizard has got you very close, but diagnosing the problem properly by counting the pulley teeth is better as that gets it spot on. Also this means the Z-axis is also out by 7% if you've not already changed it. If it's 28T on the Y and Z motors and 30T on the ballscrews put 85.71429 (i.e. 400/5*30/28=85.71429) in the motor tuning for both.

luke11cnc
11-02-2012, 11:01 AM
well I've been in my cold garage this morning Jonathan and the pulley is a 30t so there is some think going wrong then,? it is not stalling and the only thing I have changed is the set up as described by Jazz on motor tuning(witch is working)

any idea's Jonathan on this problem

James

JAZZCNC
11-02-2012, 12:41 PM
Yep I have.!!. . . . . . I think it's because the pulley diameters are not exactly to spec and therefore the ratio is different. Thats why I ALWAYS use the Mach wizzard because it works of the actual movement travelled and doesn't relay on calculated specs based on unknown specs of inacurate components like pulleys.!

So long as you measure the distance traveled accurately, preferably over the longist distance possibly then it works perfect.

Either that or the screws pitch are not exactly what they say they are.!!!. . . .(Very very unlikely)

luke11cnc
11-02-2012, 12:44 PM
I was thinking the same Jazz


James

Jonathan
11-02-2012, 12:54 PM
I ALWAYS use the Mach wizzard because it works of the actual movement travelled

No, you measure the distance travelled. But we can't ever measure it with zero tolerance, so it will always be slightly out. For most people that doesn't matter, but that doesn't make it all-right in my opinion.


I think it's because the pulley diameters are not exactly to spec and therefore the ratio is different.

Definitely not true. The only way the diameters could be different is if the number of teeth is different. If you look at the tolerances for the pulleys it's not possible for such a large error. Also having made several pulleys it's obvious that if you get the diameter out by the smallest amount the belt simply doesn't mesh with the pulley, so if the diameter was wrong the pulley wouldn't work.

James did you count both pulleys on Y and Z? It just seems too big a co-incidence that the number is so very close to making it 28T.

JAZZCNC
11-02-2012, 01:02 PM
Ok "Boy wonder" your the "Man" for calc's All I know is how to make it work correct.! :dance:

luke11cnc
11-02-2012, 01:03 PM
I will check both later as I'm going out but should be back around 2 ish

James