I've just made my first test cuts with my OX build (OX Tale). The job is a scaled-down version of the program I plan to use to level my spoilboard. For now it is surfacing an area of 170mm x 170mm taking a 1mm deep cut from the MDF. Cutter details and speeds are listed below. As you can see from the photos, most of the cut area is nice and smooth, but at the right hand side I am getting hockey stick shaped marks. This is where the spindle stops moving to the right, before moving away from the camera and then going left. Can anyone suggest what I could adjust / where I should investigate to try to avoid these marks? Many thanks Ray Details Kress 1050FME-1 spindle running at approx 12,500 rpm (setting 3 on the dial) 25mm bottom trim 2 flute cutter 3000 mm per minute feed rate
The error looks like it is in the Y axis. It's as if there is something lose on the right side gantry area which is causing a change in the cutters face to the board. What size cutter are you using? Out of curiosity, can you try surfacing an area in the Y axis instead of the X axis? Start on the right hand side this time. Can we have some more pictures showing more of your CNC, including the gantry area. A video would be even better. Gray
Thanks, Gray. I'll post more photos and a video tomorrow. I should have been clearer about the direction. In my photos above, the Y axis is left to right, and X is from the camera to away from the camera. But I take your point about trying the surfacing in the other axis - I'll do that and post the results. It's a standard OX but larger than usual (1000 x 750) and with double belts on X and Y.
Won't be able to cut further test pieces or post a video until tomorrow. Here are photos of the gantry. I've also written some test gcode to surface the test area following a square spiral path (both CW and CCW) as opposed to the "up and down" of my original test.
More test results and photos. Plus a video. In all the photos, I am standing on the negative X side of the machine, with the X axis running across the bottom of the picture. I repeated the original "zig-zag along the Y axis" test at 3000mm/min X,Y feed. The spindle moves left - away - right - away - etc. Same results. Next, Gray's suggestion of rotating 90 degrees. The spindle moves away - left - towards - left - etc. This time, the heavy marks were where the router is doing a right turn at the end of a pass. There were also some smaller but noticeable marks for the full length of each pass (hard to see in the photo). I changed to a "square spiral from outside to centre" pattern, first CW. This was much better. I think my Z feed rate is too high, because there was a problem with the initial plunge. But the rest of the pocket showed no marks. Same pattern but CCW. This had marks along the first couple of passes on each side but disappeared as the paths got shorter as the spindle moved in towards the centre. Same CCW spiral but at 1500mm/min. Much better, probably acceptable if I can sort out the initial plunge, but the CW spiral at 3000mm/min was slightly better. Thinking that slower speed would solve the problem, I tried the original "zig-zag along the Y axis" test at 1500mm/min. No better than the 3000mm/min version Conclusions so far There is a solution to my immediate need to surface the actual spoilboard: use a spiral CW pattern. I want to experiment with a lower Z feed rate to see if that fixes the initial plunge problem. But I still would like to get rid of the tool marks on the linear patterns. So suggestions welcome on what to try next. Video
I have just recently been watching Youtube Videos in relation to Tramming your CNC. It looks very similar to what people describe when the tramming of the cutter is out and requires adjustment by shimming. Have you looked at checking the Tram of the cutter, you would require some dial Gauges to do the check properly. I am no expert in this area so I could be barking up the wrong tree completely but do some searches on Tramming your CNC on youtube and it may be helpful. Cheers Glenn.
I don't know why it happens at one end of the movement, (as per video, the right,) and not the other, but you can see the curve is caused by the cutter tilting and digging deeper as it moves up, (as per the video) onto the next cross movement. When it returns on the next movement it removes a lot of the dig, but leaves a slight dip in the wood for us to see. Now we need to find out why it only does it on one end of the travel. Is it because your table is out so much that it is deeper at that point than the other? Is the cut deeper at the end where the dig occurs? Did you say you tried reducing the depth of each cut? Have you got a smaller cutter to try out? Gray
Hello Gray. I get the same pattern and depth of mark (as far as I can judge by eye and touch) at various Y positions on the board. So it doesn't seem to be due to the board not being level. Partly by accident I have made test cuts at various depths from 1mm to 3mm. That hasn't made a difference. Good question about the depth of cut at the marked and unmarked ends. I'll measure that tomorrow, and will make some test cuts with a 5mm cutter (the only other size I have for this spindle at present). The other test I'll run is the mirror image of the original. The original moves (from start) +Y, +X, -Y, +X, +Y, etc. I'll try to make it start from the top of the photo and move +Y, -X, -Y, -X, +Y, etc. Other things I have tried today with no effect: Found that one V wheel on each of the bottom axles of the X carriage were not tight against the underside of the X rails - I could turn them by hand and the X carriage wouldn't move. Adjusted offset bushes and tightened everything up so all four wheels underneath are gripping the rail. Changed grbl setting $1 to 255 so that stepper power is not turned off after moves.
is your bit/spindle/Z axis assembly is solid when you wobble it by hand? Does it make the same problem when you jog it manually?
Hi Dany, I have tried wobbling the spindle mount and it feels solid. Not tried the spindle or bit. I will do that, and also try your idea of cutting using the manual controls. Thanks.
Have done some tests and measurements this morning ... Depth of cut the same at both the marked and unmarked ends. No wobble in the bit / chuck that I can detect by wiggling it. Cutting manually using the controls in bCNC gave the same marks but less deep. Also tried a cut where the initial plunge is halfway along the first side - no difference. Ran the original pattern at a higher spindle speed (16,000 vs 12,500) - no difference. Reseated the spindle in the spindle mount and checked it was tight - no difference. Made a mirror image cut (starting away from the camera and moving towards it). It gave the same marks at the same end (negative Y end). The depth of marks seems the same as with the original path.There were also some very light "ripples" along the Y direction - not too bad but worse than the other tests. I'll make some cuts with a smaller diameter cutter later today. Apart from that, I think I need to tram the machine before I make any more dust
Made six test cuts with a 5mm spiral cutter. Depths of 1mm and 2mm. No tool marks in any of them. But ... In five of the six cuts, the cutter moved outside of the intended boundary of the pocket on the initial plunge (bottom right of the pocket in the second photo). In only one of the cuts (one of the 2mm deep ones) was the boundary correct. Don't know if this is related to my original problem.
What's the details on the 25mm cutter? Is it an insert style cutter? If it is, it may be possible that an insert is lose and performs fine when loaded (cutter is moving), but moves slightly when the cut stops. Just guessing here.
Hi evilc66. This is the cutter: Bottom Trim The blades are fixed rather than replaceable inserts. When I hold it up against a straight edge, the two blades form a very shallow V - with the cutter the right way up, the centre is slightly higher than the outer ends of the blades.
Well, that shoots that idea out of the water What about end play in the spindle? I know that the Kress spindles are supposed to be pretty reliable, but there's always the chance for a dud. It just seems odd that the 25mm cutter has problems where the 5mm doesn't.
This is mechanical flex. Because the bit is rotating it puts a side load on the spindle. It is going where it is not supposed to because something is bending or stretching. what to do? Tighten the belts. make sure all pulley grub screws are tight shorten the spindle, if you don't need a 4 inch gap for an actual part, then build it up with spare bits of MDF to minimize how far down the Z has to travel. (as explained in the OX build video!). shorten the tool. short tools are stiffer. stiff is good. make sure the tool is 'end cutting'. those are not end cutting, the cutting edges do not go all the way to the center! after all that slow down the vertical feed rate. In SketchUcam (and most other CAM packages) you can specify a Plunge Rate and a Feed Rate. Since router bits are not drill bits, ie they can drill but are not good at it, it is best to plunge slowly. Use Ramping instead of plunging, this prevents plunging, ramping down into the work instead. Router bits are good at this. make sure the feedrate suites the wood and the bit. This should be calculated from the RPM and type of wood. (I just googled and linked the first one with a nice picture The smaller the bit, the fewer flutes it should have. 1/8" and below, use single flute, above that 2 flute is good (for woods and plactics). Do not use 4 flute on wood or plastic.
Thanks for the ideas, David. Some of them I've checked already but not all. For example, I hadn't realised that FreeCAD can do ramp / helix entry, so I am going to try the pocket again with it on. I also have a DTI on the way, so can check if I get flex when pressure is applied to the bit - and what is flexing. EDIT Can't feel any end play in the spindle, evilc66.
Some progress. I checked the Y pulleys again and found loose lock nuts - very embarrassing With all lock nuts tightened and checked, I still get the original problem tool marks, but the plunge problem with the smaller cutter has gone. I also tried out the same pocket with helix entry in the middle and, not surprisingly, this gave a tidier edge to the pocket all round compared with the zig zag path. Next step, tram the spindle and see if that helps with the main problem.
