Squaring mystery...Solved?

[jTr]

Have experienced a variation in squareness of cuts with my PRS Alpha 9660. Not fatal, but puzzling for a number of reasons. Decided to run Gary Campbell's gantry squaring routine again this week.

Puzzle Pieces:
1> Gary's routine came out perfect. I mean super- tight on square.:cool:
2> Next day I fired up and made cuts, I was out of square roughly .028. Precisely the same variable on each piece cut, regardless the size or location of cut on the table.
3> Parked gantry @ 96/60 that night in anticipation of running dovetails the next day- I use the negative symbol in front of X/Y Unit values to invert 0,0 home for processes run on the far end of the machine.
4> Very similar measure of .03 ish out, but in the opposite direction from the previous day. :confused:
5> Fully check legs, rails for level/true and all is very good - the concrete floor it rests upon has been quite stable for the 3 yrs it's been there, and I've not needed to adjust at all.

Questions pondered:
- Is gantry twisting somehow depending on where it's parked? Leveling checks indicate this is not the case.
- Is there something loose in the carrier systems? -all is snug and good there.

Then I realized:
- The day I ran Gary's squaring routine, It had been parked at 24,24.
- The day it was measuring long diagonally from 0,0 towards 96,60, it had been parked @ 0,0
- Each time I experience the opposite problem for end-work routines such as dovetails, it had been parked/reset at 96,60

Sounds like someone's pulling one end of the gantry towards center of the machine....
Could it be the cables tethered to that end of the gantry?

Dropped motors to disengage gears from rack and began free-wheeling. Sure enough- when parked at 0, had a friend hold up the cables to alleviate their weight from the gantry. When gently released, weight of cables pulled the gantry towards center of table by an inch or so. Even greater movement when applying this procedure to the 96" end. When clamping a block to the opposite side of the gantry, the cables appear to pull the gantry approximately .020 - .030 .

Theory: Cables are introducing enough tension to pull gantry out of square prior to resetting drivers at start up. Once reset, motors are locked into this mode for duration of the day, and cable weight is not enough to affect normal operation. Dust collection hose is centered above table, and incidentally contributes to the issue as well, most acutely on the far end of machine where things are pulled most out of square. Furthermore, if this theory holds true, the extra-wide gantry for a 60" machine is more likely to display this phenomenon than a 48" machine.

Short term remedy/test of theory: always park roughly centered on table. Diligently monitor parts for squareness. If proven true, long term solution would be to purchase e-chain to alleviate the variable pull caused by the hank of cable currently swinging on that side of the machine. At very least, allow more slack, but concerned about entanglement issues if too slack...

Thank you for any guidance as I seek enlightenment.:o

jeff

[Brady Watson]

Jeff,
Having assembled my fair share of PRS tools, it is most likely that one of the end plates is not clocked properly, splayed out or the v-rollers are not meshing with the rail along the full travel. I've seen/adjusted some that were pigeon toed where the rollers are slightly splayed out. You have to be careful making any adjustments as you can rock the beam where the Z is now cutting on an angle. Adjustments need to be made with some long bar clamps and a block of wood to clamp the beam down while adjusting. Don't do this until you are sure it needs to be done - and know what needs to be done(!).

For now the focus should be on finding out the culprit. Drop both X motors out of mesh with the rack with YZ parked @ Y30. Remove the 4 X axis v-roller covers. Then go from one end to the other 'free-wheeling' (slowly - like 4"/sec by hand) with your eye looking at one wheel at a time and nowhere else. Do not press down on the beam while doing this. Just nudge the gantry lightly to get it going and look at each wheel per side noting any deviation from it sitting perfectly registered to the rail. Walk along with the gantry to keep your eye focused. Do the other side. Repeat the procedure again.

See if that turns up anything.

If one or more of the wheels ride up a rail, then this can cause all kinds of 'variable' issues. The key is to get the gantry perfectly riding on those rails - even before you put the YZ car on (as this adds weight). If one of the 4 wheels is up or opposite wheels diagonally, then some adjustment is in order.

You also want each X motor to mesh with the rack at the same pinion 'clock' - SB recommends pulling against the axis stops, then turn on the control box to assure it is square, but I prefer to get the tool as perfect as I can first. I never pull it against the stops to square.

-B

[jTr]

Brady,
Thanks for the guidance. Have tried "test 1" you described. Several trips up and down the rails, focusing on 1 wheel at a time from different angles - side, above, and (as best I could) observe the light piercing the gap between top of rail and point of the V of the wheel. No signs of trouble there.

(Kind of fun though- what a nice, smooth carriage motion - like new after 3 years full time service!- yeah, I'm infatuated with the 'bot....;))

Jeff

[curtiss]

I made a simple hardwood "T-square" when setting up /with a digital caliper taped to the far end/ low end of the T

A shallow saw groove was cut on the top of the T / which then sits on the first / straight & level rail....

The "depth gauge" works as a "feeler gauge" to get / make sure the rails are exactly parallel.

If both rails are parallel and straight, the y car about has to run true.

[jTr]

Curtiss,
That sounds like a good idea, which I'll employ once I get a chance. Though level and visually straight, maybe one of the rails is curling in or out on the ends. Meanwhile, as I've committed to parking in the middle of the table, my cuts are incredibly square, so I feel I'm on to something. You'd think Brady's method of testing would show some sign of a curved rail, but just could not get any suspicious behavior there. Probably because Z car is in place yet and it's weight is keeping the wheels tight as Brady suggested - maybe I'd better pull the Z car and try again- just crazy busy right now, and short term solution is keeping things very accurate.

I appreciate the input - any more advise is welcome.

jeff

[jTr]

Found It. At least part of the issue. Must have had a minor collision on far end of the X rail. Used a straightedge to inspect and found it was out about .030. Will now monitor/test when time permits, but think I got it now. Brady's method probably would have revealed it if the Z car had been removed - glad I didn't need to.

jeff