[Ken_McD]

I’m new here, to CNC (I am truly a newb to CNC with solid technical skills and no experience) and certainly glad that I found this forum. I’m looking for some help / input / advice on an issue. I’ve recently bought a business that produces aluminum products out of flat sheet. The core of the production process is a Thermwood C50 router with two 15 HP heads, built in 1988 and the hour meters report about 10,000 hours. The controller is a 9100 series Cartesian 5 beast that is based on a 286 processor with about 640k memory. We use this router to cut different blanks or profiles from fairly thin (0.040 to 0.090) Aluminum sheet – where we then form the items with press or box brakes.

My problem is this – all of the existing programs were written by hand, that is the previous owner wrote the G code. The programs were written in an incremental format. When we bought the business, the first order of business was to buy a CAD/CAM package and start pulling the old designs into a package that would allow things to be easily manipulated. The CAD/CAM (BobCAD) package writes G code in absolute positioning and we find that the Z and W axis starts to lose its resolution and slowly start to eat into the table about 1/3 to ½ way cutting down a 12 foot sheet of aluminum.

We did find that there was a very simple backlash compensation move that was in the Z and W axis that we incorporated into the post processor which seems to help somewhat – but after running a couple of sheets it start to drop again. The funny thing is that on this old controller – the position doesn’t drop; it reports the correct depth even while it’s obviously off. Additionally there is some slight “jerking” (for lack of a better word) sometimes when the gantry changes direction in the x axis and y axis. (hard to tell which when the machine is making parts) Slowing the machine down minimizes this issue. There are ball screws on the Y axis (across the gantry) and Z and W Axis (routing and drilling heads) X axis is rack.

I fear that the controller could just be toast – my posted G code programs to cut parts is much longer than the hand written subroutines of the previous owner and I’m wondering if I’m just asking too much of the DOS based system with limited memory. OR I’ve got serious ball screw issues – though I’d think the old programs would be affected by this also. The old programs still seem to produce decent parts – though they are way inefficient and time consuming (many require the repositioning of material).

I’ve talked to an aftermarket controller manufacturer who states that my issues could be cleared up with the a new controller that has systems built into the software that would eliminate my backlash and inconsistent movement issues. Can a controller do that? Any help or input would be greatly appreciated. Thanks!

[cabnet636]

any time i have had incremental z drop (and i have had this) it was motors or drives. i would not replace the controller as my first test. call thermwood they are still in business and suport thier machines.

[Ken_McD]

Thanks James. We did replace the motor on the vertical axis and went thru a slew of amp board replacements. No change. Again - the incremental programs run fine - the longer absolute programs are the ones causing me problems. The guys at Thermwood have been great. Unfortunatly - My controller is so old that it cannot be upgraded. According to thermwood if I were try and upgrade thru them it would be cheaper to buy a new machine. I can't afford to do either.

[cabnet636]

formwood has a machine owners forum, have you tried there i am sure other owners may have had this problem as there are a lot of these machines in the market.

[ger21]

If the Incremental code works fine, then most likely it's an issue with the Absolute code.

[Ken_McD]

Thanks Gerry - I would have thought the same thing - except when I run the absolute code I end up with the same reported postion on the controller as the incremental - it says it in the normal cutting postion above the table. Except - its working it's way into the table. Any where up to 3/8ths to 1/2 inch off. Really wierd. After talking with Thermwood, it was explained to me that the controller doesn't 'exactly" check the positioning from the encoders - at least what is reported on the screen.

[ger21]

Then it must be some kind of bug in there control??

[maq_vega]

Cant promise anything but are you able to attach your incremental and absolute code (if possible for same job as not to confuse) for us to see I have been running Thermwood 3 axis and 5 axis machines and have operated for 14 years and could take a look at it for you. If the incremental programs work I would think it to be code/controller related. If you wish you could also email me: maq_vega at hotmail.com