Excitech CNC Modifications Thread...

[Eclipze]

2mm is quite a lot, and I wouldn't expect "that much" as a result of flex. I only get noticeable flex when I'm slotting through acrylic with the 6 or 12mm bit at a good pace. But I put up with this as it's not critical to my work for roughing. Finishing passes are what I look to for tolerance, where the tool bit engagement is low and the load on the gantry is minimal.

For light work though, 2mm would be bearing freeplay. If it's not freeplay, then I'd look at the gcode or the DSP for this sort of problem. I'm wondering if it's the package you use to route with that is producing a lead in cut. Can you attach a simple gcode example and describe where the gouge occurs?

[BEERMAN]

I hear you.
I have looked at the Gcode and it seems fine. All simulations in artcam are fine.
I have sent a file to Big S and it has worked on his machine.
He has sent me a program that has worked on his machine but it does not on mine. I have tried my USB on another CNC and it ran a simple area clearance perfectly.
I have re booted the DSP
I have reinstalled ZHB981 version to the handpiece.
I then installed ZHB701 version to the handpiece.
i have reformatted the parameters many many times.
tried it in emergency mode.
The bearings seem tight and leadscrews etc have no noticable wear marks etc. There is that flex but i do have to force it to get it. And the machine was working fine. It also seems only at entry and then it cuts perfectly even at full depth of cut. The error can occur in even the softest of materials??
I agree it seems to be ramping into the cut but i have been unable to determine why.I design in Corel, Do the toolpaths in Artcam 8 then take to the cnc via USB stick.
Originally i thought it was USB or DSP???????????????????but i honestly am stumped, I will have got the DSP handpiece from my old(that i know is working) machine and test it on this machine.
i will run a new program when i get this thing back together and email you the gcode for a simple circle..
cheers
Beerman

[Eclipze]

No problem. Sounds like your on top of having a next-step to solving the problem. Given your getting perfect cuts otherwise, and it's only on entry, then it's easy to make the call and say that it's not bearings.

Given you have tested gcode on an independent machine, and it doesn't have the same problem, then it appears to be controller related.

Would be good to see a simple example of the gcode. I have very little experience with the DSP. I tried it, but then shelfed the whole CNC until I made an alternative Mach3 interface. The DSP was too basic and limited for what I wanted, and didn't want to spend the time bumping into it's limits.

If it is the DSP, the only thing I can think of at the moment is to include a dwell command at the right time (or at multiple times to be sure). It's against my own logic for how the DSP should process the gcode, but ignoring such logic, perhaps a G4 P3 might give it time to settle prior to the ramp in. Perhaps its an uncoordinated movement from the DSP getting ahead of itself. The delay may help it finish previous commands (such as M commands), before ramping in. Grasping at straws, but very easy to test and do an air cut.

As I mentioned, I don't have a DSP anymore and not much experience with it. If you're hitting a brick wall... any chance of a short video? Might be easier to identify by seeing it happen, or even just a picture of the resulting gouge might generate some ideas.

[BEERMAN]

Eclipze
Could you email you email address and i can send some photo's of the problem in an rar file. So you can visualize what is going on.

Beerman

[Eclipze]

The major limitation I'm finding with this machine is with the stiffness of the gantry. Something I'm hoping to upgrade on and wondering if anyone has found a solution.

Where the limitation kicks in with the gantry moving to the rear. Machining forward is no problem, however machining towards the rear the cutting bit will start chattering. It's a build up in amplitude with distance, and if I apply downward pressure on the spindle, it will dampening the chatter over time (i.e. the oscillation tapers back off until it's far less noticeable within 3..6 seconds). It ultimately effects the quality of the machined edge.

If I use a sharp 6mm bit, cutting acylic with a depth of cut of 3mm, feedrate 1200mm/min... the chatting in that direction is easily noticed. But only in that direction of cut.

It's believed that the problem is due to flex in the ganty. It's vertically quite rigid, however horizontally it can twist and flex. Some measurements with a dial gauge has shown this deflection. For example, gauge mounted on a separate bar clamped to the gantry struts and measure the deflection at the middle. With a bit of muscle applied to the gantry, flex could be detected. At the spindle tip, the deflection was approximately twice this measurement.

Planning to take the bottom thin angle plate of the gantry and replace it with a really thick steel angle profile, which will be bolted to the gantry plate at multiple locations. Hopefully it will be fit without having to modify the cowling. But if it doesn't fix the problem, then most likely the back of the cowling will have a bit section routed out to make room for square steel profile. The cable tracking will also have to moved.

Hopefully if this problem is fixed/reduced, surface finish on both side walls and on the face might see an improvement.

Has anyone looked into this issue?