[jlowe1224]

To start off, here is a link to the machine I recently built http://www.cnczone.com/forums/diy-cn...build-log.html I should mention also that the only other wood I have cut was mdf which cut just fine, nothing like this. Today I have been trying to cut out parts on veneer plywood, which I know can be tricky to cut, but I have been having terrible problems. Ill first discuss my setup then show problems. I am using a BRAND NEW white side 1/8 inch upcut bit and mach3, while generating toolpaths with cambam. To get this out of the way, I doubt cambam or mach3 is the problem.

So I starting cutting a mdf plywood type of wood that looked like mdf between two sheets of very thin wood, at 16,500 rpm at 30 ipm with that feed rate or anything calculated. The surface splintered so much that it looked like utter crap, but for the most part I could sand it all down and other than the missing spliters the edge quality was fine, not nearly as good as I have seen from others. I thought it might just be this weird plywood.








Tried 3/16 inch straight flute, better results (random design)




Thinking it was because of this weird plywood, I cut again on high quality 3/8 inch birch plywood, and it was the biggest mess I have ever seen. I tried with both 3/16 inch straight flute and the 1/8 upcut and same results that is in the picture.. It was barley even cutting, rather just getting all stringy and looking like ****.



At this point you are prabably cringing at that quality. Yet I am hesitant to say that this is not the worst.

I tried to problem solve by messing around with feed rate and spindle speed, and I found that I can get no splinters but still quite a bit of fuzz and stringy chips when I go 80ipm at 21,000 rpm. Good, but still not anywhere great. However these were simple squares. picture not uploaded


I went ahead and went back to the mdf plywood material and cut at the rates I found to work the best, and this is where my hopes were crushed. There was no splinting and surface quality was much better but still fuzzing up when cutting against grain (compare fuzziness to first 3 pics) but the actual accuracy and precision of the cuts was unbelievable.





The small slots...yeah those are suppose to be rectangles (with rounded edges of course) but I dont even know what the hell that shape is. All I know is that going from 50ipm to 80 ipm made the cuts go from crap, to well, worse. Its hard to imagine this is backlash, because I am using 2 start percision acme rods with anti-backlash nuts on all axis, so the entire machine has be structurally weak and move alot, although it seems pretty rigid. Whatever is causing it is preventing it from even completing the cut, not finishing the corners off. Also on the gear shown below, as well as every other peice, it leaves a nasty line where the separate layers are, regardless of the speed. I was hoping I could machine at least 100ipm but I cant even do 80ipm without rectangles looking retarded.



Any advice? I know a compression bit would probably solve the splinting and fuzzing, but it wont fix the problem with the inaccuracy going at 80 ipm and I dont want to spend $70 on a small bit that will probably break soon. I want to try to use the upcut bit I have, but I can get a white side 1/8 straight flute bit if it will improve it alot more. But more importantly I want to address the problem with the cutouts not being rectangular, and why a medium speed of 80ipm is affecting my machine so terribly.

[jfong]

I always use a down cut end mill when cutting veneer ply. You don't as much tear out. The quality of the plywood also determines how good it cuts. Sometimes the face veneer is so thin that it chips out badly no matter what you do. I've never used mdf core ply so have no idea how it cuts.

[pippin88]

Use a down cut bit!

[Eddard_Stark]

As others mentioned, your splintering is being caused by using an upcut bit. Upcut evacuates the chips upwards ripping the veneer on the ply. Use a downcut and that will be eliminated. (or a compression bit on a full doc) As for your rectangles not coming out....it looks like you have a massive amount of slop somewhere in your machine. I glanced at your build log and noticed you are using gas pipe for linear rails, check there for slop. Also do a backlash measurement. I measure mine with a 1" dial indicator; move say .75", of course adjust steps per inch until you actually move what you command. Then move back to zero. Any backlash will result in the machine moving less than commanded. Alternatively you can chuck a rod in your router and move a dial caliper.

[jlowe1224]

The reason I ruled out backlash and the slop you are talking about is because when I run the machine slowly, the cuts come out perfectly accurate, but when I increase the speed something happens and it goes out of whack. I did however realize that I used a helical coupler connecting the y lead screw to the stepper, so when the router moves right, it could possibly be stretching the coupler a bit, but I feel like it wouldnt affect those rectangle slots like it did. Ill try the dial indicator if I can get one.

[ger21]

The faster you go, the more the machine flexes. This is due to both cutting forces, and inertia.
A Solsylva isn't a very rigid machine to start with. Most people start making a lot of modifications to stiffen them up as soon as they get them running.

As for the slots. The one that is symmetrical definitely looks like backlash. The one next to it looks like a lot of machine flex? Mach3's CV mode may also be contributing a little, but I don't see these issues in your other pics.

Was the gear cut in two passes? Does it look like that all the way around, or just on one side? It looks like either the spindle is out of tram, or the second pass was made slightly larger. Are both passes cut in the same direction?

With the machine off, and no bit in the collet, grab the collet and start pushing and pulling in different directions. I'd guess that you'll see quite a bit of movement. Any movement at all is bad, so try to find out what is allowing movement, and take steps to eliminate it.

And yes, you need some downcut bits.

[Jon.N.CNC]

An upcut bit is always going to burr like that, really you want to use a downcut spiral with that thickness ply. Compression bits are great but you have to plunge over the depth of the upcut part which I know the onsrud bits the smallest upcut on their compression spirals is 5mm so no good for 6mm ply.

Too me looks as though your missing steps as well, either that or you have a coupling that is slipping, worth checking for first.

[jlowe1224]

[QUOTE=Too me looks as though your missing steps as well, either that or you have a coupling that is slipping, worth checking for first.[/QUOTE]

Thanks for the advice. I will purchase a downcut bit and hopefully it helps. Also I think I can rule out skipping steps (not sure if this test can indicate it) But when I slowed it down and cut a part out, I didnt set it to go deep enough, I ran it again slightly deeper and still didnt cut through (my bad job on zeroing z axis) then third time It cut through and each time I ran the code it went in the exact same spot and didnt cut away any extra material from the sides. In theory, this should mean it didnt skip steps right? Because then my home would be slightly off each time? However I did find I was using a helical coupling that can flex out 1/4 inch

[jlowe1224]

The faster you go, the more the machine flexes. This is due to both cutting forces, and inertia.
A Solsylva isn't a very rigid machine to start with. Most people start making a lot of modifications to stiffen them up as soon as they get them running.

As for the slots. The one that is symmetrical definitely looks like backlash. The one next to it looks like a lot of machine flex? Mach3's CV mode may also be contributing a little, but I don't see these issues in your other pics.

Was the gear cut in two passes? Does it look like that all the way around, or just on one side? It looks like either the spindle is out of tram, or the second pass was made slightly larger. Are both passes cut in the same direction?

With the machine off, and no bit in the collet, grab the collet and start pushing and pulling in different directions. I'd guess that you'll see quite a bit of movement. Any movement at all is bad, so try to find out what is allowing movement, and take steps to eliminate it.

And yes, you need some downcut bits.

I found significant backlash in y-axis due to helical coupler with no support so it could literally expand. Also the gear was cut in two passes, which the visble line is the spot between the two passes, and it goes all the way around.

[ger21]

You had the coupler supporting the screw? The screw should be supported with bearings at each end, with the coupler mounted outside of the bearing.

[jlowe1224]

You had the coupler supporting the screw? The screw should be supported with bearings at each end, with the coupler mounted outside of the bearing.

I phrased that poorly. No so I have two bearings on each end of the lead screw supporting it, and a collar clamp on the end opposite to the stepper. However On the right side I connected the lead screw to the stepper with a helical coupling, so you can tug on the lead screw and it will pull out stretching the coupling a bit, but the collar clamp prevents it from getting tugged to the right. I will fix this by getting a solid coupling.I hope it makes as big as a difference as i think it will. In the picture you cant see the bearing but it is in the wood, and there is about 1/4 inch space between the bearing and coupling so for the meantime I added another bearing in there so that the couple will directly push against the bearing so that it cannot get stretched under cutting forces. However there will still be some backlash from the twisting of the coupling I think.

[ger21]

I use helical couplers and have no issues at all. The bearings should be preventing your coupling from stretching. If you switch to a solid coupler, your stepper bearings may fail, because they'll be carrying the thrust loads of your leadscrew. The screw itself should not be able to move at all with no motor connected.

[Maglin]

Your lead screw mounting is the problem it sounds. You should have a sandwich of bearings on one side so that no matter how hard you pull on the screw it has zero in and out play.


Sent from my iPhone using Tapatalk

[Jon.N.CNC]

Thanks for the advice. I will purchase a downcut bit and hopefully it helps. Also I think I can rule out skipping steps (not sure if this test can indicate it) But when I slowed it down and cut a part out, I didnt set it to go deep enough, I ran it again slightly deeper and still didnt cut through (my bad job on zeroing z axis) then third time It cut through and each time I ran the code it went in the exact same spot and didnt cut away any extra material from the sides. In theory, this should mean it didnt skip steps right? Because then my home would be slightly off each time? However I did find I was using a helical coupling that can flex out 1/4 inch

Hmm it is a test to say that you do not loose steps when not under cutting load, once you have fixed your coupling problem, try before a job starts either making a mark where zero is setting up 2 dti's compressed about 0.1mm into the bit with bit at zero (one at X- one at y-, obviously not putting them in the way) then run a job and see how accurately it returns to zero.