Tool diameter accuracy

I was doing some cutting earlier in a really difficult material - 3mm MDF. I was cutting some small components which needed to slot into each other, and after the first attempt none of them fitted. Measuring, I found that I was getting dimensions consistently about 0.7mm out. That was averaged over a dozen or so different dimensions, internal and external. I regenerated the gcode using a cutter of nominal dimension 2.3mm instead of its theoretical 3mm, and all the pieces cut and fit perfectly. Question is, is it my crap machine bending with the strain of a 1.7mm DOC in MDF, or is it likely to really be cutter size? It's a single-flute carbide cutter from CNCRouterParts, but I can't check its diameter directly as it is single-flute. I don't have much experience with these kinds of cutters - do you need to do trial cuts to find out their effective cutting diameter before use? I'm happy to blame the machine, but curious about cutter accuracy as well.

Hi Neale,

Doesn't seem like a particularly heavy cut so it's a bit odd that one. I cut a fair bit of 3mm ply, liteply, and the occasional bit of MDF and use a single flute 3mm TCT router bit without problems. Here is my home design Spitfire which all locks together nicely . . .

Attachment 11747

Can you do a plunge into a semi-hard material (e.g. hardwood) to drill a hole and then check the diameter? This gets around the single flute problem for diameter.

Some of my bits are quite worn, especially the lowest part of the cutter, and they are less than the original diameter. Should throw them really but they are still OK for roughing or non-critical jobs. How old are your cutters?

That Spit looks like it going to be lovely - which Mk are you going for?

Sorry for the hijack....I have a thing for Spitfires (except maybe the Mk5 with that intake)

Haven't you by chance turned on/forgotten to turn off tool radius compensation in the controller program? 0.7 mm is too big difference in tool diameter. You can measure the diameter at the stem, where the flute ends. Accurate enough to give you an idea. This if the shank is of diameter greater than 3 mm.
Drilling a hole or milling a groove is also a good approach.

If the machine was flexing then in one direction the error would be positive and in the other direction it would be negative.

Also +1 re. Washout's comments.

No, I don't think that it is a tool compensation problem as the gcode preamble includes G40 which should turn it off. That amount of error is strange - too big for cutter diameter error, I would have thought, but too small for any significant axis calibration error (which I have checked in the past). The error is also pretty much the same, within the limits to which you can measure MDF anyway, along both axes, which I would have thought meant that it's probably not machine bend as I would expect that to vary with axis. However, this is my MDF router, so anything's possible. At least the error is consistent enough that I can tune it out by tweaking cutter diameter in vCarve. Sometimes I have to pretend to be an engineer and not a scientist (it works, dammit, so forget the theory!).
Thanks for the thoughts, guys, and I shall try putting a mike across the plain part of the cutter (3mm single flute on 6mm shank) as a check.
This is what I'm playing with.
Attachment 11750
It's a cable chain idea based on something I found a while back; it needs reasonably accurate cutting to make it all stay together (no glue or loose fixings) while still being able to articulate.

There's a few of those designs on YouTube if you search for 'diy cable chain'

Indeed - that's where I found the original. However, I've redrawn it to be a little bit bigger as the original was tight to take CY cables, etc. I was just taken by the idea of hinged links that didn't use any kind of conventional hinge, and it's an interesting CAD->CAM->machining exercise.

Wow, that spit looks great. G.

Quote:

---

Originally Posted by Neale

It's a cable chain idea based on something I found a while back; it needs reasonably accurate cutting to make it all stay together (no glue or loose fixings) while still being able to articulate.

---

Don't see the point in cutting these other than the exercise.!! . . . They will cost more in Elecy than they cost. . Lol

The cutting differance will be MDF frame swelling in the morning dew try cutting in afternoon when it's warmed up. .:hysterical:

Now seriously I would check for backlash or end float on the screws and because it's consistant the differance could be the amount of backlash. Try cutting something similair with a larger cutter and see if error is same amount I'd wager it will be.!
After this then check steps per again and make sure it's moving the exact amount it's told. That amount of differance can really only come from either cutter comp or wrong amount of movement, either thru backlash or steps.

Jazz, now you're taking the proverbial! The frame doesn't swell at all with heat or moisture. The bed goes up and down in the middle instead!

I'll try chasing backlash and leadscrew float. However, the leadscrew is something like 1.25mm pitch so I'm seeing more than a half-thread backlash if so. I have anti-backlash pairs of delrin nuts on each axis so I'd be surprised if backlash were that bad but I shall check. The nuts are running on commercial stainless studding so while it's a bit smoother than BZP studding, I know that the nuts are going to wear.

Steps per - I did check a while back with a dial gauge (ever tried getting a magnetic base to stick to MDF?) and it was pretty much spot on. Those parameters should be locked in to the software and hardware so I hope they don't drift!

I'll also try a different cutter or two cutting a standard size piece and see what happens.

Measured the shank of the cutter just above the flute. Converting from imperial to metric (only have an imperial mike) that came out at 3.08mm. Plunged a hole straight down and checked size with the shanks of twist drills and that came out between 3.0 and 3.1mm. Pretty conclusive, then.
I was going to measure backlash, leadscrew float, etc., but when I did a quick "shove it and see if it clunks" test, I realised that the whole thing bends so much that it wasn't worth the effort of digging out the dial gauge, etc. As a further demonstration of general bendiness, I changed from two passes of 1.8mm DOC in MDF to a single 3.5mm pass, and the dimensional error changed significantly to the point that the pieces didn't fit any more. The machine works fine for engraving nameplates, or anything involving light cuts where there's no need of great dimensional accuracy, but ask it to machine two components that will slot into each other, and you're on a loser. Those of you who can machine aluminium on a CNC router are allowed to laugh. I can't machine 3mm MDF...

Washout, Geoffrey,
Thanks for comments on the Spitfire - its a mk9 (nicest shape I think!) and has taken 3 years on/off (mostly off) to get to this point.


Neale,
Have you got any photos of the machine you could post? Might help gauge stiffness.

Are you sure you are not loosing steps - the deeper cut is higher load on the leadscrew and delrin nut. I confirmed my machine was missing steps a while ago by touching the stepper motor body during cutting - I could feel the kick every now and again. I think mine was an extreme case so instead you could also set up a DTI on the tool, gantry, or Y axis plate when it is all at the home position, cut out the shape and then see where the machine homes back to. DTI should be back at zero, otherwise there has been a step or so missed.

What is the control system spec, voltage, drivers, BOB etc?

I won't bother with a picture. It's too embarrassing... The machine is basically the JGRO design, and there are lots of references to it on the 'net.
I'm using (all from Zapp) a 68V power supply driving 3 M752 drivers, fed by a ZP5A-INT BOB. Motors are all SY60-something 3Nm steppers. Given that they are driving relatively fine-pitch leadscrews (1.25mm) at modest speeds (max rapid is about 900mm/min to avoid whip) with 400microsteps, I don't think that the motors, etc, are being overpowered by the cutting forces. 3mm cutter, 1.8mm DOC, in MDF? FWIW, I also use LinuxCNC, which seems to be a minority choice on this forum, with clock rate within the max suggested by PC latency tests. I do seem to get good repeatability; I can cut several sets of those cable chain components and pieces are cleanly cut (as clean as MDF cuts) with good registration of first and second passes, as far as I can see.
Don't build your CNC router from cold-rolled cow dung...
(Indeed, the Spit looks good so far. One of the projects that I shall get round to one day, and one of the drivers for building the router in the first place, was to cut components for a RC glider. That project's still way down the list, though!)

OK, if it's one of these then I guess flex is still a possibility if backlash is low:

Attachment 11775
JGRO machine

The electric spec seems very good, leadscrews are low/medium, and machine is . . . . well, MDF on gas pipe bearings. The next step is up to you - whether to upgrade the machine or stick to engraving / simple jobs.

If you do get around to the glider you'll love cnc cutting out the bits and slotting them togehter. No more fret saw! and everything fits perfectly.

I've had enough frustration and I'm working on the design of the Mk2 which will be welded steel, ballscrews, and profile rails. I'll take the electronics and motors across from the current machine as I reckon that they'll do nicely (and it's why I overspecced them in the first place). In the meantime, I'm doing some interesting projects with some local artists/craftspeople who are more interested in artistic interpretation than accuracy! I spent some time today with a friend who wants me to turn a bunch of photographs of her daughter's head into a full-size 3D version routed into a bundle of wooden sticks. It's the kind of project that no-one doing this kind of thing professionally would even look at :smile:

(and yes, my machine looks somewhat like that one. Oh, to have a machine that will take drawings and turn them into fitting components without three false starts while I calibrate out the bend!)