New 4th axis

[spumco]

[ Play Quicktime file ] 20190303_155215.3gp [ 418.97 KiB | Viewed 14484 times ]

Just finished upgrading my 4th axis. Other than the trunion table plate, all machining was done using UCCNC and F360 for CAD/CAM. The trunion table was tool large for my machine travels, and I've got a friend with a 6030 Fadal...

Old one was a worm drive with unacceptable backlash - only good as a motorized positioner as I had to indicate & verify every move.

New one is based on a harmonic drive, so no backlash and very rigid. Drive is a hollow-shaft so I have a through hole for longer parts and a 3MT taper in the faceplate. Harmonic is driven by a Nema 34 stepper via an AT5 belt. 4th axis frame is steel, all surfaces in contact with table or harmonic were surface-ground and then bedded with epoxy during final assembly.

Tailstock spindle is from a 5C collet spindexer so I can hold stock or just about anything else in it (5C chuck?). Bearings are huge, so I may motorize it and have a home-brew RapidTurn type of thing in the future. The tailstock housing is on linear rails so I can move it up to the 4th axis for support without having to re-align anything.

Tailstock bearing pockets were bored in place using a boring head mounted in the 4th axis - poor man's line boring - to ensure the spindle was coaxial with the 4th.

So far, it's positioned within a couple tenths (0.0002") over 10 revolutions with the indicator at a 3.5" radius. I.E. G0 A3600 gets me 10 revolutions plus or minus two tenths at the outer edge. Runout at the MT3 taper is undetectable with a 0.00005" indicator.

Max speed with the trunion table is about 25RPM with pretty aggressive acceleration before the stepper looses steps.

Thanks to everyone on the forum and CNCDrive for the help - even if you didn't know you were helping!

-R

[Haik]

Very nice work! I constructed a simple 4th axis from a 8 inch turn table so can relate to your past worm-drive problems.
You've really expanded the capability of your machine. Thanks for posting this!!!

-Haik

[cncdrive]

Hi Spumco,

That's a nice 4th axis I have to say.
And I'm happy if we helped you in any ways.

Will you make any videos of machining?
And when you will get there it would be a good idea if you could share some informations about what CAM software and how you using it so others could learn.

[spumco]

I'll make some videos but they probably won't be terribly exciting. I'm still in the 'one hand on the feedhold' mode for quite a bit of the machining I do

I'm using F360 for both CAD and CAM, and I believe it's a bit limited for 4th axis 3D contouring type of work. It works well for 3+1 and certain continuous tool paths, such as engraving or pocketing on a cylinder. But because UCCNC doesn't treat the A-axis as a rotary axis I don't think the CAM + UCCNC will work very well for 3D rotary work with significant changes in diameter- at least for maintaining surface speed of the tools. I could be wrong, of course. Still learning.

The trick - so far - with F360 CAM is to accurately measure the 4th axis center of rotation and then use those dimensions for part offsets inside CAD. If the trunion table centerline (dowel pin holes) is xxx thousanths off the rotational centerline, I edit the trunion and fixtures in CAD to match that offset and zero my Z and Y work offsets to the 4th axis centerline. The CAM then figures out where the part is in space and adjusts locations. The result is that F360 CAM outputs toolpaths based on a trunion fixture surface swing which isn't coaxial with the 4th axis.

Because the trunion plate is bolted to the drive flanges I can nudge it close to alignment before tightening it down - but nothing is ever perfect. Once it's set the Y offset shouldn't move unless I remove the trunion, and the Z offset won't move at all.

It's the opposite of having a commercial controller with something like HAAS' tool center point control. I have to tell F360 where the part is in space (related to the 4th axis) rather than the controller keeping track of where things are (and will be).

Someone more clever than me might be able to write a macro which automatically adjusts the work offset, but I suspect it'd need to be a continuously-variable offset as the table swings - probably something better suited as a feature internal to UCCNC.

-Ralph

EDIT - oh dear. I just saw that the latest beta version has a rotary axis setting. Didn't see it as I'musing the 2017 screen - must go play with that now!

[spumco]

P.S. Anyone want to buy a slightly used worm drive 4th axis? It's cherry...

[ThreeDJ16]

Nice....got any vids yet? I sold off my old worm gear drive 4th axis as I never used it. But I want to build a trunnion table as one of our last group projects. We have so many projects going and the other guys in my group are selling the previous stuff we built, so time has been limited to move on to the newer projects. Maybe by next year this will be in the cards. Think my mill will finally be complete when I get the trunnion (after the linear rails, ATC, servo upgrades, possibly a new cabinet....LOL). Trying to setup my mini shop for retirement, which is still quite a few years off. But as slow as things have been moving on the last several projects and the limited time to do them, I might just slide in with everything done...

[spumco]

I don't have a youtube or facebook account so I have to convert to low-res stuff to upload here.

Here's a pic of how to use it for positioning an oddball fixture, as well as a clip that shows the trunion being used a a fixture plate. No sexy 4th axis moves...

[ Play Quicktime file ] 20190305_223621a.mp4 [ 1.91 MiB | Viewed 14340 times ]

EDIT - duh. I already posetd that fixture photo in the other thread. oops.

[Derek]

My 4th axis and trunnion table has been one of the better things I've done for the B-port. It is really nice to hit 3 sides of a part with weird angles.

Love those vises for sure.

Derek

[ThreeDJ16]

Crap, it states plug-in not supported. Seriously? Oh well. FYI, I held off on using Youtube for a long time, but now I just find it's so much easier. I can share it to Facebook or pretty much any forum. If you have a Google account then you already have a Youtube account basically.

Well, hopefully by years end I'll have my linear rails and a good start on the ATC. Somewhere in between I'll probably work on the trunnion table. Just want these linear rails first as they will increase my Z by almost double and my Y by 150%. So will make a significant difference in what I decide to make. Either way, I'll probably just get my friend making spindles to do a 5C, which he's starting soon and use a GT3 or GT5 timing belt and DMM servo. Should be a lot of fun learning how to do that in CAD.

So, is it possible to use a trunnion table and mill with the 4th axis rotating? Or is it just like pick an angle, it stops and hits the odd angles (which really maybe all I need). I assume now that A,B, C are all selectable as rotary axis, it will do real time milling while A is in motion? Sorry, I can't see any of your vids to see if that is true.

Gonna go try to find that damn plugin for quicktime...LOL...haven't seen quicktime vids in years. D'oh, didn't see the mp4 beside it....ok was able to see vids. Guess I still have the same question as before though.

-=J

[spumco]

So, is it possible to use a trunnion table and mill with the 4th axis rotating? Or is it just like pick an angle, it stops and hits the odd angles (which really maybe all I need)-=J

UCCNC can move the axis continuously during milling, although there are some factors to consider. Seriously not an expert here, and I haven't played with setting my A-axis as a rotary in UCCNC yet.

First thing to keep in mind is if your CAM supports continuous 4th axis (or 5th) moves. Second is to determine if your CAM & controller will adjust the axis feed rate depeding on the diameter of the part - if the diameter is changing over the part as it rotates (something other than a cylinder) then the rotational speed needs to change so the tool tip/flutes maintain a constan chip load. As most home-built (and some commercial) rotary axis drives are painfully slow, it's just not possible to keep the travel speed very high with 4th axis moves. This issue can be acceptable on large-diameter work where the desired feedrate is slow, but the problem can be bad on small stuff.

Think of it like CSS (constant surface speed) on a lathe.

Say you're cutting an aluminum part - imagine a cylinder with a single turn helical groove - and you want to cut at 100IPM and whatever RPM for your desired chipload. Assume the part is in a chuck, and the A-axis is along the X (typical). Also assume the part is 1" diameter, and Z-0 is at the center of the part. Groove is .125 deep, and the groove is 6" long from the start point - not the actual groove length, but along the part. Finally, assume the 4th axis is set up so the step/unit are some multiple of a degree so you can program in degrees.

You'd program the following:
G1 Z0.375 F100 (plunges in to stock at start of groove, OD is .5" from Z-0)
G1 X6 A360 F100 (cuts a single turn groove)

What happens (I think) with a slow 4th axis is that the controller (UCCNC) realizes that it can move Z and X at 100IPM, but it can't rotate A fast enough to keep up. So it calculates the time it will take to reach the end point based on the max speed of the 4th and slows down the X so that both X and A finish positions are reached at the same time. Someone else please check my math, but the result is nowhere near 100IPM. Given a diameter at the cut surface of 0.875" you have a circumference of 2.75". At a max of 10RPM (sort of typical for low-end worm gear drives) the result is a surface speed of 27.5" per minute, give or take.

So the controller puts the brakes on the X axis and the cutter creeps along. And the issue is compounded be the fact that the controller may not know that you're using a rotary axis, so it can't adjust the rotational feedrate up or down as the diameter changes (think of the helical groove along a cone).

And your CAM probably doesn't know what the upper rotational speed of the 4th axis is, so even thougth the CAM is changing the A-xis rotational speed to adjust for diameter variations at some point the 4th axis is bouncing off the rev limiter and the controller throttles the linear axis back. You get the desired chipload at the large diameter of the part, but by the time you get to the small end it's creeping (relative to the surface).

So the answer is yes - but. When you see videos of commercial machines with fast, swoopy 4th and 5th axis moves there's some serious computing going on in addition to fast rotational drives. I think HAAS claimed 1000 degrees a second for one of thier rotary positioners recently - like 170RPM.

You can see why I was thrilled to jump from about 8RPM with my worm drive to about 25RPM with the harmonic drive - and it should get up to about 100RPM when I put a servo on it (based on drive input speed limits).

-R