New spindle location to reduce flex

[David Falkner]

There's no way a carbide bit could flex that much. They don't really flex at all, they snap. I saw similar looking results back when I used to rout 3/64 grooves (for ebony inlay) in solid wood with a dremel. It was the lack of rigidity in the tool itself, molded plastic bearing seats, etc. Bought a Dotco 1/4" die grinder and made a router base, problem eliminated.

That's kind of what I thought, Richard. The carbide bits don't feel like they would have much flex at all but I don't doubt they flex a tiny amount without breaking.

It's interesting what you said about using a Dremel. Before I built our CNC router I designed and built a pantograph to hold my Dremel so I could do inlay on headstocks and fingerboards. What I found is that my pantograph worked great but the Dremel had too much runout and flex to do anything except large inlay pieces and designs. It wouldn't handle intricate cuts because of the shaft runout.

David

[richard newman]

Dremels are great, but they are not precision tools. The Dotco cost me $200 back in the last millenium, when $200 was serious money. There were also small electric precision spindles, even more expensive. All these cheap chinese spindles are pretty amazing values.

[Mark Bolton]

There's no way a carbide bit could flex that much. They don't really flex at all, they snap.

That is in NO WAY SHAPE OR FORM true. I can run you a part today, tomorrow, a week from tomorrow, a month from tomorrow, that will show crystal clear tool flexure in a 3/8" solid carbide compression. Its apparent on a daily basis. And the tool will cut, and cut, and cut, and never break, and get swapped out when dull. Carbide tools deflect just like anything else. I can push a solid carbide compression hard enough that you can see the tool deflection naked eye plain as day when the parts come off the machine. Its the simple reason why you run a stepover on your final pass. The bander will show you the tool deflection way better than naked eye ever will.

No doubt David ditching that tramming plate alone will make a major improvement. Losing that flexure point and the additional moment leverage will make a huge difference.

[David Falkner]

Thanks, Mark. When I cut the Longworth chuck arc slots I set Stock To Leave at 0.007" for the side wall (Radial) and then a finishing pass to clean that up. I would think it should be taking 0.0035" from each side wall of the slot but when the bit enters for the final pass I only hear it cutting on one side of the slot and maybe ever so lightly touching the other side. It comes out the correct size but my guess (theory?) is that the initial full depth cut at 175 ipm has some flex in it and cut that 0.0035" in the first pass.

The first pass is a climb cut and the second is conventional cut so that may have something to do with it, as well. So either way there's some flex involved, whether it's the bit or the tramming plate but it's there. I guess when I get the second spindle mount in place and make the same cuts I'll see if I get the same results.

David

[richard newman]

That is in NO WAY SHAPE OR FORM true. I can run you a part today, tomorrow, a week from tomorrow, a month from tomorrow, that will show crystal clear tool flexure in a 3/8" solid carbide compression. Its apparent on a daily basis. And the tool will cut, and cut, and cut, and never break, and get swapped out when dull. Carbide tools deflect just like anything else. I can push a solid carbide compression hard enough that you can see the tool deflection naked eye plain as day when the parts come off the machine. Its the simple reason why you run a stepover on your final pass. The bander will show you the tool deflection way better than naked eye ever will.

Mark, I stand by the first part of my statement, there is no way a carbide bit could flex as far as in the photos, that is a lack if rigidity in the machine. I stand corrected on the second, there does have to be "some" flex, no matter how small.

But I really do wonder how much deflection you would get in a 3/8 compression bit if it was firmly clamped in a vise and indicated while force was applied up to the breakage point. The deflection you speak of can still be from flexing in the machine. Routers are wet noodles, compared to mills and VMC's. And I think material springback is as much a reason to run a final cleanup pass.

Do you have an old dull bit you could test to breakage to quantify this?

[Mark Bolton]

Mark, I stand by the first part of my statement, there is no way a carbide bit could flex as far as in the photos, that is a lack if rigidity in the machine. I stand corrected on the second, there does have to be "some" flex, no matter how small.

But I really do wonder how much deflection you would get in a 3/8 compression bit if it was firmly clamped in a vise and indicated while force was applied up to the breakage point. The deflection you speak of can still be from flexing in the machine. Routers are wet noodles, compared to mills and VMC's. And I think material springback is as much a reason to run a final cleanup pass.

Do you have an old dull bit you could test to breakage to quantify this?

Do you follow channels like Edge Precision that regularly run "spring passes" on relatively rigid tools and setups to cleanup a thou or more even when they creep up on a tolerance? I dont creep up on anything. Im trying to get parts off the machine single pass. If I can hold the part and the tool deflection is .003" I would simply offset and let the edge of the part be sloped. The place where it rears its ugly head is when I cant hold the part (very common) and have to two pass. The step/witness line is plain as day.

No doubt some amount of that deflection is in the machine and some amount is in the tool. In my trivial experience with CNC Im completely clear that tool deflection even in wood is dead real. I run the shortest tools possible. I run 7/8" tools on 3/4" material. You can slow the tool down and the deflection disappears. Its plain as day.

[Ron Bontz]

Interesting read. What CNC brand is this anyway?

[David Falkner]

Saturn welded and stress-relieved steel frame from Fine Line Automation; I designed, sourced, and built everything else. I had the most fun designing the electrical schematic and wiring the enclosure!

David

[richard newman]

I designed, sourced, and built everything else. I had the most fun designing the electrical schematic and wiring the enclosure!

That's the part that scares and intimidates me the most! Don't know anything about electronics, and even after spending hours online to decide what components to use, I can't figure out what kind of connectors to use with them. Structural design and precision metal working are so much more comfortable for me.

[David Falkner]

I posted the build thread here but the photos were hosted by Photobucket and they're no longer showing properly. I need to see if I can get the photos to Jim Becker and get him to update the thread. But here's the link and you can sort of make out the photos - https://sawmillcreek.org/showthread....st)-CNC-Router

David

[David Buchhauser]

In the referenced build link - Right click on photo, then left click on "view image" to view photos clearly without the blurred photo bucket marking.
David

saturn cnc.jpg

[David Falkner]

Oh, COOL!!! Thanks, David, I didn't realize you could do that.

David

Edit - which browser are you using, David? I just tried it in IE and View Image isn't an option. In Chrome I get Open Image In New Tab and that works. In Edge I don't see it as an option, either.

[Jim Becker]

With Chrome, right click and open in a new tab to get a clear image. That said, "a pox on Photobucket!!"

[richard newman]

Do you follow channels like Edge Precision that regularly run "spring passes" on relatively rigid tools and setups to cleanup a thou or more even when they creep up on a tolerance? I dont creep up on anything. Im trying to get parts off the machine single pass. If I can hold the part and the tool deflection is .003" I would simply offset and let the edge of the part be sloped. The place where it rears its ugly head is when I cant hold the part (very common) and have to two pass. The step/witness line is plain as day.

No doubt some amount of that deflection is in the machine and some amount is in the tool. In my trivial experience with CNC Im completely clear that tool deflection even in wood is dead real. I run the shortest tools possible. I run 7/8" tools on 3/4" material. You can slow the tool down and the deflection disappears. Its plain as day.

Not familiar with Edge Precision, and my cnc experience is limited to pearl inlay and fret slotting with sub 1/8" cutters, done very conservatively on a relatively stout machine. One thing I do know is that there are all sorts of places where things can flex between the work and the end of the cutter, and I believe all of them contribute more to the problem than actual tool flex.

Working to .003" on a router seems pretty ambitious, what kind of machine do you have Mark? What kind of spindle? And what are you cutting?

[Mark Bolton]

Not familiar with Edge Precision, and my cnc experience is limited to pearl inlay and fret slotting with sub 1/8" cutters, done very conservatively on a relatively stout machine. One thing I do know is that there are all sorts of places where things can flex between the work and the end of the cutter, and I believe all of them contribute more to the problem than actual tool flex.

Working to .003" on a router seems pretty ambitious, what kind of machine do you have Mark? What kind of spindle? And what are you cutting?

I dont disagree that any deflection is cumulative, Machine flexure, slop, backlash, and tool flexure. We have gone down this road here before and Im easily confident that Im in those tolerances confirmed in wood and non ferrous (brass and aluminum to 1/2" thick). I dont doubt that at a painfully slow pace I could hit .001 easily.

I have no delusion that pushing a 3/8" solid carbide tool at full depth (3/4") at 750-800 ipm (running at 150% or more) the result is a combination of machine flexure and tool deflection. But Ive spoken to enough tool manufacturers and done enough head scratching on our parts to know that tool deflection clearly there with solid carbide tools. The most rudimentary machining guidelines make it clear that all materials bend. Vortex would be a good source for conversation on the matter.

Machine here is 60x100, ShopSabre, ES919 10hp spindle. There is without a doubt flexure in the machine.