Z-Depth Trouble

I'm refining my last post a bit.

My machine is consistently NOT cutting through the material when it should.

My machine is a 2016 PRS Alpha ATC. I have skimmed the spoilboard. I program the file to cut through the material by 0.15". That doesn't come close. I keep changing it until I'm telling it to cut into the spoilboard by 0.080" before it does. And it's NOT cutting 0.080" into the spoilboard.

On my ancient 2001-2002 PRT machine, it worked like a dream, even if I didn't skim the board for weeks.

What on Earth could be causing this?

I don't have an ATC so I can't answer for sure...

Did you adjust your zero plate when you skimmed the table? Do the end mills in your ATC carousel have to be re-zeroed?

Yes, I ran the C73 table offset for the zero plate after skimming, and all the bits have been zeroed correctly.

Does the Z axis actually move the commanded distance? Zero it out, move 1" and use a dial indicator to verify. Try 2", etc. If it's off, your unit values may be incorrect for the Z axis.

After verifying that it is setting the z height to the correct position, run a file and then check again after the file runs to see if it moved. Bits can slip in a collet sometimes as well. Mainly like was said above, verify it is actually placing the bit at a certain height when you tell it to go there. I have seen loose spoilboards that were flycut without vac hold running and then when a sheet of material is there to cut the vac pulls it down lower. Make sure your cutting file does not have a start depth in vectric. (the box above the desired cutting depth)

Have you run the C72 on all your tools since you set up? Tis is vital to all the values being correct. When I start each day, I run C3 to align on 3 axis. If I am referancing the top of the material for my Zzero, I run C2 to zzero. If I'm going to be referencing the surface of the spoilboard as my zzero, I C73. I use Shopbot Link which throws a lot of other wrenches into the smooth running gears. After running C73, I have to MZ to the thickness of my material and then do a ZZ to set my zzero reference.

Since you just got this machine and are new to ATC operations, I would suggest running the entire course of alignment routines. Using an ATC is very different from a machine that does not have an ATC. This is something I struggled with when I added an ATC to my machine. All of a sudden zero references are very different. Zzero reference was the toughest for me to get my head around. You have to make sure all the other alignment routines are run successfully first, because all the ATC code uses these routines to determine where it is in the world. Everything runs of proximity switches with the ATC, especially the Z proximity switch. If you don't have the instructions for a full ATC alignment, ask the Shopbot folks and they can walk you thru the whole thing. I can't stress enough how vital this initial alignment is.

I ran my PRSAlpha from 2009 until last year without an ATC and then got the ATC. Took me a couple of months to get things ironed out and fully operational to where it now purrs along very nicely.

Thanks for all your input.

The guy who sold it to me (and delivered it) helped me set it up. We ran the CN74 program to calibrate the clip locations. We loaded the tools into the collets and ran the CN72 routine. We then ran the CN73 routine to set the table offset.

"When I start each day, I run C3 to align on 3 axis. If I am referancing the top of the material for my Zzero, I run C2 to zzero." - exactly what I do.

"run a file and then check again after the file runs to see if it moved. Bits can slip in a collet sometimes as well." - If the bit slipped, it would cut too deep. Mine is consistently not cutting deep enough.

"Make sure your cutting file does not have a start depth in vectric. (the box above the desired cutting depth)" - I don't fully understand this. I'm using the same post processor that I used for years on my old PRT. Why would it behave different on this machine?


Today, I re-skimmed the spoilboard so it was fresh. I re-ran CN73. Then I ran a file that basically cut out 3 circles with some interior detail. I initially programmed it to cut 0.015" into the spoilboard, but it did not cut through. So I modified the program to cut into the spoilboard 0.035" into the spoilboard. In the two corners (see image) it cut into the table approximately 0.025", and in the middle, RIGHT WHERE I SET Z-ZERO (red x), it not only didn't cut into the spoilboard, but left material that I had to manually remove. This is right after a fresh skim and a fresh CN73.

Attachment 34555

Attachment 34556

So, I turn the vacuum system on when skimming, which I learned with my last vacuum table. Now, this seems preposterous to me, but I'll go ahead and say it.....

Is it remotely possible that the actual table for my machine, becomes concave once the vacuum is on? In that the center portion sinks down and the outer portions pull up? Making, in effect, a 96" x 48" bowl, which allows the edges to cut through but not the center?

You should not be using the same processor for the atc machine that worked on the old prt. Not the post processor anyway. And yes the bit can slip both directions. I'll post of pic of the start depth box

I tried to use my Micrometer to measure the Z travel. With the imperfect method of measurement it seems to be accurate in terms of the stepper values. It certainly does not explain the pattern from the above picture.

http://www.talkshopbot.com/forum/att...id=34557&stc=1

If it is consistant about cutting on the edge and not in the center like your last picture, the spoilboard moving has to be the issue.

Thanks for the pic. I've verified with Vectric that the post processor they have for V-Carve PRO is, in fact, the correct one for the ATC. It has all the proper call-outs for the ATC. Anyway, none of that explains the specific depth problems I'm experiencing I don't think.

Thanks Ken. Yea, that start depth is always zero for me.

How stable is your plenum? I've heard of plenums flexing when the vacuum is on. As far as the spoilboard, as long as the plenum is stable and can't move, once you surface the spoilboard, it should be perfectly aligned with the CNC. Since 2009 I've not had a problem with my spoilboard moving. But being in GA, I suspect humidity could be a problem over time causing an MDF spoilboard to swell and shrink.

Is your Z-assist cylinder working properly? The ATC spindle is a WHOLE LOT heavier then the smaller spindles. The Z-assist (pneumatic) helps the Z motor handle this extra weight. I had that issue when I first setup my ATC. Once I got the Z-assist cylinder securely attached a lot of problems seemed to fade away.

Are you using the pressure plate for ZZero or the aluminum bar? I found the pressure plate can be a little off. I use the pressure plate for my CN72 routine but not for zeroing on the spoilboard or top of material. I have a pressure plate for Zzeroing on the spoilboard and material but I found it to not be accurate enough. Good old C2 on the aluminum plate seems to be the best.

The pressure plate enables you to Zzero without the alligator clip being attached to the bit. Since the ATC spindle has ceramic bearings, you need the alligator clip or the pressure plate. On my 2.2hp spindle I never had to use an alligator clip on the bit.

I have the pressure plate on the ATC, but I use the Z-Zero plate on the table.

I am leaning towards a plenum problem. I just re-zeroed to the table and ran this pattern on the blank table with the vacuum on.

Attachment 34558

This was the result, both on a piece of 1/8" hardboard AND on the table itself. The odd part is that where it does not touch the table is exactly where I zeroed it.

Attachment 34559

Attachment 34560

One possible way to test if the plenum is at fault:

1. Cover a zone with thin plastic.
2. Place a piece of paper in the center of the zone.
3. Zero an end mill to that piece of paper so that the paper slides with resistance.
4. Turn on the vacuum.
5. Check to see if the gap between the end mill and the piece of paper has changed.

My bet is that when you are flattening your table, the vacuum does not have as much downforce on the open table as it does when it is covered with masonite. So the middle section is sitting higher when it is being surfaced and then pulled down when nearly sealed with the top sheet. Do you have a gauge on the table side of your vacuum? I'm not sure if mine measures psi or inches of mercury, but my table only draws 2-3 when it is just the spoilboard, but I can get it upwards of 15 when I have it completely covered with material. Also sealing the edges of my table made a big difference

Have you tried pushing down on the table when the vacuum is off, to see if the layers are not bonded together and there's some spring in the lamination?

Questions and points not yet addressed:

Did the previous owner have these issues ? Sounds like they are actively helping you, assuming they're decent enough to help = they'd let you know, right(?)
If you're confident it wasn't a previous issue, you should completely review your assembly steps at current location.
Is this the same plenum, or did you make a new one? If it's bending in your shop, it must have in previous owner's. If no = look somewhere else.

You've verified steppers are responding correctly, post-processor is correct, yet we're still having issues. Indeed, spoil board with vacuum is not the same as loaded table, but we all do this without experiencing these alarming amounts of variation.

Not yet discussed; V-rollers and gantry mechanical contacts with rails. If the V-rollers are not tight consistently all along the rack, some twist may be happening. Both your images suggest table is thin and flexing, or V-rollers may have a few areas of loose contact, as in gantry may have a slight twist. This can look good in a cursory check in one place, so you need to pull covers and examine travel full distance - disengage steppers for this so it may freewheel and reveal any variations in contact. Look for light and carefully try spinning v-rollers by hand to see if one is popping a wheelie somewhere along the line. Gantry end carriages typically need to be re-squared to cure this, and a recent move certainly could have an effect on this.

Footnote: while re-surfacing is being trusted to generate 100% flat surface, it cannot be trusted as so if the V-rollers are loose. Two different sized bits with different geometry and a loose wheel = chaos.

Good luck - we're racking our brains and hoping for victory with you!

Jeff

Hey Red. That very well may be the case. My vacuum numbers are very similar covered vs. uncovered. Also, I just noticed that the v-shaped support under that area is missing all the bolts on one side.

I have not done that, Bill, but will try it. Either way, I'm planning on completely re-doing the table. Until then I'll just cut everything deeper than is necessary.