am I being acquainted with mr. lost step here or is something else the matter?
:confused:

p.s. I wanted shape on the left and instead I got, well, you guessed it... :(

oh, and btw this happened due to my material surface not being flat so the bit dug into it more and more towards the top (3mm or thereabouts at the top) instead of 0.3mm per pass.

90degree v-shaped bit

Did it actually finish in the same place it started?

yes, actually! or at least I think.

just to make things clearer, I wasn't drawing a rectangle so it couldn't really go back to where it started but I was doing surfacing, so it was a back-forth traversal along the x-axis and steping a small amount on along the y-axis on each pass. (like a square wave with X axis being amplitude and Y axis being frequency)

I thought it didn't but now looking at the workpiece you can see where it lifted Z to (what it thought was) safe height at the end of X and Y (top left corner) and traversed back to the proper 0,0 (bottom left corner).

The "safe Z" wasn't actually safe at all but was around -1.5mm so it left a pretty little trail going back home.

Now, looking at it some more, the new X would've been 0 but Y was shifted to -4 (or thereabouts) so the home position would've been slightly lower then the original home but X would've been the same.

Would a crummy photo make things easier to understand?

ok here they are. the first picture shows the "rectangle" on the block (it was meant to be an engraving of a rasterised photo so you can see some relief in the middle, it was a multi-pass job)

0,0 marks the original home, sneaking arrow is the direction of the router. end is where it was supposed to end the pass, go back to 0,0 and continue another pass.

the second photo shows where it actually ended and the red line shows the mark that the rapid back to home (0,0) left. so by this time Z was also screwed.

you can see the new home is slightly lower but approximately in line with the original home X. only Y seems to be off.

you'll also notice the original 0,0 is quite shallow at 0.3mm compared to the top which is 3mm. that's due to the block not being mounted horizontally but on an angle which I didn't spot untill it was too late.

So the 0,0 position has actually moved on the machine?
It's not simply the work piece moving due to insuficient clamping?

Paul

no no work piece stayed put.

I remember seeing quite a few posts from you about roller screws.
Did you go down this track? If so, can you lock the planatery threads as a test to see if the screws are causing the problems?
Paul

no didn't go down this route, I needed to concentrate on having a machine up and running first.

It looks as if you may be having a problem with the direction line and pulses. Instead of ;losing pulese you are always getting an extra pulse to the left. In other words the tool is always shifting in the same direction. This is a oribklem that others have had with timong of the Step signal along with the direction change pin. I am not expalaining this very well but it has been discussed quite a bit. To check you need to set up a reference poiont on one axies then write Gcode to move off the reference say 3 inches and then move back. Repeat this as 100 times. Or you can start with 10 and move up from there. See if you are losing reference in the same direction on every change in direction. Actually it is usually just on one change either when direction goes positive or when it goes negative. If it only happens under load then it is probablly lost steps. If it is timing them it will probablly show up even when cutting air.

Garry

PS
The curve on the sides of the piece makes some sense but the sloped line on th top would seem to have to be from the piece moving or bad Gcode as the maching has to be cutting a diagional on the last pass and I can't think of a driver issue that could cause that.

If your bit went back to the correct 0,0 (x,y) position for each pass, then you didn't lose steps in the middle of the pass. Lost steps accumulate with each pass, they don't recover without you re-homing.

What software did you generate your G-codes with?

Brian

Hmmm, kind of a stumper with out some more info. Have you tried simple squares and circles and what not? If it always comes back to the same place as it started, maybe just a lot of flex?

For free, you could generate some simple test shapes with SpiroCNC. (http://www.cnczone.com/forums/showthread.php?p=114958&highlight=spirograph#post114958)

Just download it and upzip it.

Steve

thanks spalm I'll try that today. but now I'm suspecting flex in combination with misalignment of the workpiece piece because yesterday I "surfaced" my work bed and it went ok but I was only taking off 0.3mm a pass. the amount of sawdust it created (not woodchips), now I'm contemplating starting up an MDF factory :D