3d Mold divot issue - SOLVED - Loose screw holding pinion gear

[salewis]

Looks like it is time to chase some ghosts.....

My build partners and I have racked, raked, or is it whacked our brains in order to get a plan to go chase this ghosts.

The divot is 0.030 +- 0.006"

We are about 90% complete but this is a killer issue for us.

[Alan_c]

Does it appear in the same place everytime you run the file - possible code problem, if not check for a loose pinion or rack on your Z axis (if you have belt drive check all the pulleys on each shaft), possible loose spindle mount or even Z motor mount.

[smreish]

Are you using a vacuum holddown?
Foam has a tendancy to "climb" when shaping like this.
It has a stability issue and when you release the tension on the fibers on the top surface, parts will concave or convex on the release.

"short of like if you weld, the material shrinks and changes the shape of the base metal"

Have you tried cutting the mold out of something more solid?

[Gerald D]

After a divot, did the machine recover to the correct position by itself, or did you have to do a manual reset? For the time being, I assume that it recovered by itself, and that points to good electrics, good programming, bad mechanicals.

10643

My guess is that the red line is the deepest part of the mold, the point where the z-axis has to reverse direction. That would be another indicator of a mechanical problem related to the z-axis.

First suspect is the grub/set screw in the pinion. Then the rack mounting to the slide, the router mounting to the slide. Maybe even the hold-down roller at the back of the gantry is not doing a good job (stuck against y-car)

[salewis]

Let's see if I can get all the answers...

The problem doesn't appear in the same place/same time. This points away from gcode. Haven't checked the rack/pinion on the z, but this is a good/easy check.

The stock is RenBoard 440 which is more like a big slab of bondo (polyester body putty). While it is "technically" a foam (syntactic to be precise), it is very stable dimensionally. But the answer is no there is no vacuum in use on this part. BTW, when cutting this part out of pink eps (extruded polystyrene) the problem seems worse. So this line of reasoning is still plausible.

The "deepest" part is just to the right of gerald's line.

and now for the "piece de resitance" (don't you just love a southern boy who can't even speak proper english tries french :D )

No manual reset. The machine recovered itself, this is very evident in the "scratch" in the last picture above.

Can we rule out: z-axis motor/driver, wiring, interference as I don't want to look where the problem is not ;)

stan

[Travish]

what about checking the rails for build up or mis-alignment? A mis-alignment of the rails would only cause the gantry to rise not gouge or or drop as shown in your photo. It appears to be mechanical though...

What about putting an indicater in the spindle and run it along a pc of flat stock on the top of the surface. place the stock in the same area as you had your mold sitting and see if you notice anything dips or drops from the machine.

Usually when I see marks like this, it is due to not leaving enough stock from the roughing pass for the finisher or the rougher path has some issues with programming. The rest of the mold looks beautiful though.

[Kobus_Joubert]

I will put my money on the workpiece not secured 100% to the table. I had this before.

Did you LEAN on the table or pushed on the piece during the cut ?

[jwt]

To my eye the Y-axis is wonky.

Look at the vertical line, its as though the piece moved to the right causing the cut to be deeper.

[Alan_c]

Well spotted John, I think you're onto something there.

[salewis]

Update.. We have removed and sent the spider to a local machine shop to be surfaced (on the 6 posts that support the eccentric bushings). We were planning to do this when we move/paint the machine. This will eliminate issues there. So until next week, more thinking and thinking...

Thinking of ways to test various scenarios.... I'm going to create a simple gcode to move the bit back and forth with small steps (simulating the same steps as in the mold), first using 45 degrees but without the complex z, then 0 degrees (mainly x) and then 90 degrees (mainly y) to try to get a particular axis nailed down.

Yes, the y did look wonky in that spot. I'll look at the other spots, but it could the result of something like a sticking gas shock that really pushes in the x direction but appears y because of the 45 degree finish cut (again back to some slop in the z mechanicals).

More thinking and tinking..... We are 90+% home on this so I am trying to check this in perspective. The last 10% is more difficult and may be more rewarding in the long run.

Stan

[salewis]

Still thinking as the machine shop didn't call me today (will be next week now).

John's pointing out of the "wonky y" has been causing me to do some thinking... So I will ask a question...

Could a build-up of crap on the z roller (running 4 rollers at the top) cause enough of a shift of the bit to cause this issue? If so can I tighten the rollers (will be going back to 6) or rig up a wiper/scraper to keep it cleaner?

Cut a fair amount of styrofoam with pink bits everywhere especially the z rollers. Ted and I do not believe this occurred when the machine was "new."

[smreish]

Stan,
When I was cutting HUGE blocks of foam on #5, I did notice this problem.
Though the skew of the tool bit was not an issue with the parts I was creating - so I didn't pay attention to it.

I didn't notice it on the z axis, but it was evident on the x axis and caused the gantry to rise up on one side only that was full-of-crud. Thus, showing about a 1/64" change in depth of cut.

SO, it is entirely possible you need to add cow catchers to your machine.

I added them to scape the rails. a little 2" x 4" piece of delrin profiled to match the rail and "scrub" it. Fixed the problem instantly.

The static from the machine really grabs the foam particles,...I never noticed this issue with wood.

Sean

[salewis]

Sean, got pictures of the cow-catchers especially mounting?

[smreish]

http://www.mechmate.com/forums/showthread.php?p=12096&postcount=4

http://www.mechmate.com/Forum/messages/2/185.html


I don't have the actual pictures, but this should get the point across.

I just used a piece of 3/16 plastic (UHMW or Delrin - can't remember) and bolted with 2 - 1/4-20 UNC socket head cap screws to the Gantry end mounts. Total of 4 units installed on the machine in each direction.

The one that mounted (under) the cable chain entry point required an extra tab of metal added, but was easy. Depending on your cable chain mount, you might not have a problem.

My machine #5, was pre-mamba parts and the end caps had a slightly different profile that did not accommodate the cable chain.

[crashingsucks]

ok fellows, I think I found the problem. The z pinion gear on the transmission has just the right about of play, combine that with a sticky z strut..... and as if by magic, z-boo-boo's!!!!

[salewis]

Ted, aka Crashingsucks, indeed found the problem...

We're VERY HAPPY now...

Moral of story.... check the grub/set screws...

Thanks to everyone who helped.. It was a team effort.

[crashingsucks]

Well I feel a little stupid, if I would have read stan and my own thread I would have seen that Gerald D pointed the grub screw out on post #4....


One should learn from the master!

Thanks all, Ted :)

[Gerald D]

Alan was the first to suggest that you had a loose pinion ;)

Funny thing is that the grubscrew issue occurs close to the point of the motor changing direction, but never at that exact point.

[smreish]

...and now I don't feel so bad that I TIG welded the end of the pinion to the motor shaft on my z-axis because it kept moving :)

[KenC]

Sean, you are a brave man! :cool:

[salewis]

what about taper pins / spring pins? They would be a pain to get in BUT a bit more stable than the screws and less drastic that TIG.

[Gerald D]

Snag is that you may want to swop motors and pinions around the various axes, where the pinion position on the shaft varies a bit. Or you simply want to replace a pinion 5 years down the line. Then a tapered pin through a hole in the shaft is a problem to line up.

The set-screws do work if you follow the basic "rules":
- bigger diameter screw
- finer pitch of thread on screw
- minimum of 2 screws, 90 degrees apart
- good steel quality of screw (wrench hole must never want to strip out)
- use good wrench and tighten until the wrench deforms
- Loctite screws and tip of shaft where it mates in the pinion (A bit of heat loosens the Loctite)

[Greg J]

I have double grubs screws (one on top / backing the other) plus locktite on all my pinions. Have never had an issue with any loose pinions.

[salewis]

greg, by double do you mean two screws 90 degrees apart or two in each hole (really 4 screws)?

[Greg J]

greg, by double do you mean two screws 90 degrees apart or two in each hole (really 4 screws)?

Two in the same threaded hole. 2 screws per pinion. Thats it.

[Gerald D]

If you have large pinions (on geared/belted motors) the screw holes are deep enough to get a pair of screws into the same hole, on top of each other. But there is no harm in a second hole, 90 deg off, with another pair of screws in there as well.

The small pinion (direct-drive) guys won't get 2 screws per hole, therefore they are obliged to go to two holes.

[salewis]

could it be that our flat spots on the shaft are too deep?

[Gerald D]

how long is a piece of string?

[jwt]

If it shifts after using high strength loctite, either you don't have enough area on the threads (too small a diameter or too thin a wall) or you haven't cleaned the threads on the pinion and screws of grease and contaminants. Bit of acetone works wonders.


Loctite 270 is strong enough. You could use 290 but it penetrates, be careful it doesn't get onto the pinion or shaft, it will wick its way between the shaft and pinion and you'll have a job removing it without overheating the motor shaft.


There is an oil tolerant threadlocker from loctite but afaik its more that once it sets it is tolerant to oil, not that you can use it on an oily surface.

John

[Gerald D]

We do wet the shaft & pinion with Loctite, in addition to the threads, and a puller has always shifted the pinion later, without applying heat. But I can't remember the specific number of the Loctite in our arsenal - I think it is "medium" strength stuff.

However, I have a personal belief that Loctite does not help that much for something that wants to start coming loose. Sure, it stops a loose screw from falling out, but it can't help for the contact metal between screw and shaft relaxing and developing a bit of slack. First prize is for a mechanical joint to work without adding Loctite.

[salewis]

as for the string comment - I walked right into that one :eek:

After a long run last night, Ted thinks the screws have backed up... When I disassembled it yesterday, the screws had not backed up/out.

Gerald, your last comment has me thinking that the back/forth action on the z (we are doing a lot of 3d, especially z) is causing the shaft to "wallow out" where the screw contacts the shaft. I have seen/heard people relief drilling into the shaft before the set screw to give the set screw a bit of a shoulder to work on.

As for loctite, I've got some red (aka permanent or bearing mount).

BTW, this is on the larger gear (72 teeth) of the belt transmission. The pinion is holding solid. This gear is also Aluminum though we have a steel gear in our stash.

Back to the length of string is that before/after knots and twists? Besides string is either too long or too short, it is never just right. That is why we have teeth. (Just made Ted aka Crashingsucks crazy on that one....:D )

[Gerald D]

Sorry, couldn't resist the string snark :o

If the metal/metal contact is "wallowing out", it can only help if you double the number of those interfaces.....a second screw hole.

A pre-machined hollow wouldn't be as good as a a hollow worn in by the screw itself. A couple of re-tightens in the early days should get the screw/s nicely bedded. Then add loctite.

[salewis]

Yeah it is better to pull the string than push it! :D

I can see the point there on just letting it wobble a bit and retighten (using new screws).

BTW, we already have two screws 90 degrees apart

[jwt]

Aha!

This is where my cunning stunt of cutting the grub screws down and having a flat surface pressing against the flat on the shaft comes into play.
*cough*

[Red_boards]

Very educational thread.
John, I'll store your trick for later use.

[jwt]

You might want to wait until I try it out in anger :) so far it's all theoretical until the pinions get a hard workout

[salewis]

on and similar but different wavelength (figure that one out guys.... :D )..

I fly R/C aircraft and we use a collar arrangement to keep (or at least supposed to keep) the wheels attached to the wire landing gear. Think small wire (0.125 or so) with brass collars and a single set screw (smaller than #2). Well a buddy and I got tired of his wheels coming off (due to collar falling off)....

So.. I took a collar rethreaded with a 4-40 and used an alloy/hardened socket head machine screw that I had cut off...

Similar to JWT's thoughts.... different application... but it might work. Especially if we ruin another $400 part. :mad:

But we are running even as I write this so I know more in about an hour....

stan

[salewis]

Just got back from the shop. Recut of the mold.... Wait for it....

V
V
V
V

PERFECT!

Better than what the guy was getting from another guy running a ****bot :D

Just got to keep an eye on the set screws before running a 6.5 hour finishing pass.

[Gerald D]

. . . . cutting the grub screws down and having a flat surface pressing against the flat on the shaft comes into play. . . .

That was the old-fashioned way and that style of grubscrew was notorious for working loose. The most popular screws today have a knurled cup at the end.

http://brijmohanandsons.com/yahoo_site_admin/assets/images/ms_grub.115233131_std.jpg

[Greg J]

Better than what the guy was getting from another guy running a ****bot :D



One of the main reasons why I went the MM route. Before the MM, I asked for and received samples of what other machines could cut. I was amazed how bad the cut quality was. And, that a company would send out something of such poor quality. Bad for them, good for me. :)

[liaoh75]

I get better cut quality than a large number of CNC machines manufactured in Taiwan in the $15,000 to $20,000 US dollar range. Of course the $50K to $100K Big Iron machines cut about the same as the Mechmate. I've seen a number of big iron machines around here sporting a lot of bells and whistles like massive servos and 55mm lead screws, but I'll have to say that I'm really not jealous. I've even seen 5X10 machines go from x and y extremes to home in under 3 seconds (Yes, I've see a few that could actually pull that off and it is really impressive) but, can you really cut at that speed? No!!!

[salewis]

Yeah, I mean come on folks... Why pay that when for less than $5-6K one can get this quality of cut....

Our silent partner in the process uttered a quiet "wow," when he saw the mold... He's pretty hard to impress....