Multiple profiles do not align

I have designed a Celtic knot, being a triquetta with a circle interwoven, the circle being in the centre of the triquetta, as per attached .pdf file drawing.

I made each of the 6 sections of the triquetta a separate profile, T1 to T6. These were engraved first on the MDF with a 1.5mm milling bit with a single pass depth of cut 1.0mm.

The 3 sections of the circle C1- C3 were cut immediately after T6 was finished.

The problem is, the circle is way off centre, having dropped down about 10mm, as shown in the cut picture attached.

The Celtic knot was designed with AutoCAD, then the tool paths and the G.code generated on Vectrix Aspire. The simulated cut is shown attached.
The G.code was imported into UGS to run the OEM controller.

Prior to this, I ran a test piece of a 40mm diameter circle , cut along the inside, and a 70mm squate surrounding the circle, cut from the outside with square corners. The cut result was a perfect circle of 40mm diameter and a 70mm square. You can see this on the picture showing the cut results.The machine feed rate , to get a decent circle was only 30mm /minute with a 1.5mm bit and a 1mm depth of cut. I had previously done a 40mm circle using 60mm/min feed rate, 1mm depth of cut, and ended up with a 37.5mm diameter circle

������������ ������������������������������������������������ ������������: Can anyone please advise why after the 6 profiles T1-T6 are engraved 1mm deep at 20mm/min feed rate, the circle C1-C3 profiles are way out of alignment, not even making a circle, as shown in the attached photo.
I am using a Chinese 3040 500W CNC table. The X & Y stepper motors are 2A rated, and the new Z stepper 4A rated. The table is quite sturdy

Hoping someone can point me in the right direction please

regards to all,
Jill

Sorry about the question mark logos - I cannot editi the post to delete them

Hi Jill - The only way that sort of thing can happen is if the motors are losing steps. Your feeds are very very slow. try 600-800mm/min. I usually run a 1.5mm bit at 2m/min and 1mm DOC. But that does not solve your mystery. Oh no pictures attached... Try again doing an air cut above the part to see what it does. If it is losing steps there maybe some mechanical issue with the machine to fix. Stiff bearings, loose frame etc. Peter

dropping down is interpreted by machinists as a movement in the -Z dirn. You probably mean fwd relative to the table? or down on the screen?

Hmmm as you have each of the programs as seperate is the origin of each the same? Do you return to 0,0,0 at end of each program so you know its starting at the right point? Why not program as one run vs so many?

Try these links for the missing images:
https://en.industryarena.com/forum/a...9&d=1677557057

https://en.industryarena.com/forum/a...1&d=1677557057

And the PDF:
https://en.industryarena.com/forum/a...3&d=1677557057

rhanks for images - comments still stand. Peter

Hi Peter, many thanks for this comprehensive and thoughtful reply.

If I speed the feed rate up to say 100mm/min, the whole table rocks and the motors chatter, and on a 40 mm diameter circle, it loses shape and size.

In the Celtic knot above, the circle tends to drop along the Y axis back towards the X axis Y=0.

The design of this Celtic knot is such that it is interwoven, The sections going under another section are deliberately kept 0.5mm from the top section. so it must be all done as 9 separate sections or profiles, with the ends of each profile 0.5mm from the nearest cut section.

The Aspire software is a bit weird. At the end of each cutting profile, it does not return to the Origin (0,0,0) but jumps to one end of the next profile to be cut, and not necessarily the nearest end of the next profile

hello i looking lantek expert please help me please

Hi Jill - Ok so you run slow to stop machine dancing that's OK. I understand the geometry of the triquetra I've done a few before. Does vectrix make one program for thw whole profile? or are the tris and circle different programs? if the triquetra is one program then the offset circle has to be caused by lost steps. If you repeat the program does the circle offset the same? I say this because if the machine is loosing steps my guess is it will be different each time.

Sounds like the machine needs checking out. Bearings belt tensions, bolted connections I'd do this first. Peter