Method for accurate Z zero between tools

[blackhawk]

I know a lot of people (including myself) struggle in keeping an accurate Z zero position between multiple bits. The biggest issue for me is when I am doing 2.5D text in a pocket. Usually, I will have a 1/4" EM doing the pocket clear, then I use a V bit for the letters. I can never seem to get the EM and V bit to cut exactly perfect to get a clean pocket bottom. So, I came up with the idea of using a dial indicator as seen in the picture. I mounted the dial indicator to a linear rail slide to give me smooth and accurate travel. Then I simply hold it in place with a very strong neodymium magnet.

First, I zero to the top of my material with the Z zero plate as normal. After I finish the cut with the first bit, I bring it over to the dial indicator and do an MZ=0 command. Then I slide up the indicator until the dial is set to zero, holding it in place with my magnet. I then move the router over and change bits. Then I move the router back over the dial indicator and slowly jog the router down with the "K" command active, until I get to the zero reading on the dial. Then I do an ZZ=0 command to set the zero point of the new bit. Now, I have an accurate zero point between each bit.

I have only gotten to try this out once on a 2.5D pocket so far, but my first trial worked great. I had a very clean and even bottom in my pocket.

[carlosgmarroquin]

Thanks. I will try this technique for sure.

[michael_schwartz]

Smart and simple.

Another use for this technique could be for zeroing bits when the top of the surface has been machined away. (If you didn't want to zero to the bed)

Using multiple bits within a 3D file could be an example

[dana_swift]

Brad- Thanks for the technique! I switch bits often enough to appreciate the simplicity of the idea.

Not only that, I have a Dial indicator I used for other setup work that I could use to try it.

D

[ATX Poly Products]

Nice little tip! I also have seen a bit of variance when doing a tool change and re-zeroing. Nothing super big but just a hair.

Thanks!

[bleeth]

You could also rig up a fixed position Hex nut with a bolt that had the head machined flat in it and after you zeroed to top the first time move your bit above it and screw the bolt up to touch the bit. Then use it to set the next bit. Or set it at bed height. Or use Gary's (and others) fixed z-zero method with offsets.

[Gary Campbell]

Another method I have used on multiple bit files is:

1) Zero to top of material with bit 1
2) Affix a machinable block out of the cutting area
3) With bit 1 at z zero, machine the top flat
4) Change subsequent bits over the block with z at zero
5) raise and then lower bit to zero to double check

This eliminates rezero to a material that may move due to machining. A loose zero plate and OEM files will give results plus or minus .010. Thats ten thousandths up or down. In worse case scenario, 2 bits could be off as much as .020

[dana_swift]

Being the curious sort that I am, I got to wondering how repeatable is the ZZero routine? So.. I wrote a SBC file to measure it.

In Summary:

1) I found zeroing at the same point and the same bit reliably returns to the same value of ZZero, +/- 0.000" (Almost always, and +/- 0.001" when it wasn't.)

2) Just removing and replacing the ZZero plate could cause a +/- 0.002" variation.

3) Use the K command to select a nearby point to measure on the material surface, then significant variations were seen over +/- 0.010".

This is just a quick look, but I will be paying more attention to it now.

Anybody else wanting to measure their ZZero repeatability is welcome to a copy of my code. In exchange you must publish your experience.

What my routine does is pull up exactly 1.000 inch from the measured material surface without changing the zero value in Z, so if the Z readout reads 1.003" the variation was +0.003". If the Z readout is 0.996" the variation was -0.004". Etc.

[[ Also note the file should be ".sbc" however the forum will not not allow uploading anything but ".sbp". So you are advised to rename it back! ]]

[blackhawk]

I ran a 3D file last night that used 3 bits. Between each tool change I used the zero plate and zzero routine. I then ran the bit over to my indicator to check the accuracy of what you get using the zero plate. One bit was off .004" and the other .001", which is actually pretty good. I placed my zero plate in a different spot everytime, but each spot was probably within a 4 inch circle. My part was pretty flat too, because I had surfaced both sides on the bot beforehand.

[applik]

We have tried all sorts of methods for re-zeroing without much luck. The bolt is the latest method, it comes closer than any other we've used, but still not acceptable when doing the 3D. The Z just will not go back to "dead on". In the second picture, the ridge measures .004 higher on one side than on the other. Sometimes it's lower, never the same outcome. Would be great if it would be exact, but I seriously doubt that will ever happen. Funny thing is, we can get the X and Y right on, but never the Z.

Will be watching this thread with great interest.

SL

[bleeth]

The problem is that the smaller the carving the more difficult it is as the needs are for virtually no difference and when you are doing a small carving the small differences are exaggerated. Even the tiniest variance in table or material flatness to begin with can really throw a monkey wrench into things. If you have followed the advice of "The Professor" Dr. Crumley you note that he recommends running texture backgrounds, finishes, and pockets for critical elements to be cut separately. This is so you don't get nailed by too much hand finish labor to make the parts look pristine.
I've learned that at times I am better off letting the machine work longer and use the detail bit for the whole project than go nuts because switching from a roughing bit or area clear bit to a fine carving bit blows precise z no matter what. That may seem contradictory to previous post but that is only because I neglected to mention earlier that there is no "perfect" when changing bits, only darned close.
When a difference of .004 on a small part is unacceptable then one needs to understand that to get small carvings really sweet folks with jewelers mills expect their machine to work for many hours to cut one wax.

[applik]

Dave,
I'm not changing bits here. The same bit is still in the router. I'm just trying to recover from power outages or turning the machine off at nite and then trying to pick up where we left off the next morning. We have experimented with temperature variances, time for router warm up, etc. After 4.5 years of manipulating and correcting for every conceivable machine ailment and wood peculiarity, we have far exhausted any and all capabilities of the bot. In order to minimize the original "ridges issue" we have slowed the machine to 1.5 in/sec and that makes for extremely long carves (20 - 36 hours for large slabs of wood). I have to admit, I think I started out with too high of expectations of what the machine could do with 3D. I have been jolted into reality. I always end up using the micro carvers to even out the bump. I've hand removed everything from a 1/16" inch to the .004. Takes an incredible amount of time when it's a 4 foot long bump in an intricately designed slab. That's time I could be sculpting in Artcam.
And of lately we have tested several machines on the market that are capable of being stopped and started, power interrupted, as well as undergoing various bit changes and still manage to relocate zero accurately without any "bumps" regardless of the size of the carvings. :) Life is gonna get easier.

[bleeth]

Shari: Life never gets easier. The challenges just change!
You have my e-mail. I'd be glad to hear more about howyour new machine works out and the new challenges (or is it opportunities?) you face.

[Brady Watson]

Thicker material, like a 3X3X1/2" block of aluminum or brass is more dimensionally stable and reliable than bit-gouged 1/8" plate that gets twisted around when it is placed back in the holder. Those plates are easily off by .005" from one side to the other, even when they are brand new. I'm not going to mention repeatability of the rack & pinion, or backlash...


Shari,
Don't be fooled. No machine out there is going to give you perfect repeatability when stopped & re-started overnight on a long 3D file. The variable or culprit is not the machine.

What most don't realize that the material absorbs heat as it is being cut, in addition to internal stresses being released, on some materials. This small amount of expansion increases the overall size of the material - and leaves you with a ridge when you restart the tool cold the next morning. I've cut enough large 3D files to make this observation.

The trick here is to learn how to tweak the Z after you restart the file to get it to match up exactly where you left off. It isn't hard to do.

-B

[applik]

Brady,
If tweaking the Z is not hard to do, would you mind sharing the process. Hey, even if it is hard to do, I'm willing to give it a try. :D The strange thing is, when I only pause the machine for a few minutes, then resume, I still get the "bump". We have been over this issue with Shopbot and was told that the variance of the thickness of a sheet of paper is acceptable and within the tolerances of the machine.

I agree about the aluminum plate. That's why we zero to the bolt. I have an electronic caliper on a heavy base we use to find the height of the wood, then set the bolt to match that. Works great. Actually, we have two bolts, one med size, and one large size for the really tall projects.

Yes, we are testing out new machines. I'm not a traitor; just looking for alternatives to up my carving production. The servos on ball screws run faster and that means we should be able to get the slabs carved in a shorter time. So maybe we won't have to turn off the machine at night and then resume again in the morning. I'm still running time tests to figure out our percentage gain. Like was stated, lots of variables going on here.

If there are other "options" for me to try, I'm open to suggestions.

[Brady Watson]

Shari,
When you are ready to restart the machine, and the machine is sitting where it left off in XY, with Z up in the air, use the MZ command to move the Z in the negative about .01 ABOVE the intended cut depth, as indicated on the M3 line of code when you stopped it.

Using the keyboard command and the Fixed mode set to .005", 'sneak up' on the intended depth that the line in your SBP is supposed to be cutting. Get right down there and observe where the tip of the ball is...Can you see light between it & the material? Were you not able to sneak up on it before bottoming out the bit into the material? Turning the spindle by hand can verify if it is too deep.

When you have the Z set to where you know it is on track again, use the VA command and enter in the correct negative Z value for where the tool is sitting. It should be the same value that was commanded in the SBP on the line you left off of.

Raise the Z, and add a J3 command above the line you left off of with the same X,Y coordinates, your desired MS, SO commands if you need to turn on the spindle and then use the FG command to execute the SBP a few lines before where you left off (wherever you have added the lines).

Make sense?

Now...this sort of thing should NOT be happening when you stop the tool for a 'couple minutes'. If it does, a thorough inspection of the Z axis rack and pinion are in order. If you have never replaced the Z pinion, I'd say you are overdue with all the 3D you cut...Knowing how particular you are, you probably have already done this. The other important area is the pinion to rack lash. It can't be too loose or too tight, or you'll have some squirrelly things going on in the Z department. In an ideal world, the rack and pinion should be raked out of debris, lubed and checked for slop often. If you are running a PRT, those V-rollers should be checked as well, since they set the lash on the gears.

No doubt a screw Z would add resolution in the Z...

-B

[jerry_stanek]

Have you thought about the material thickness changing over night. I have had sintra change with in 1 hour of my cut also corian has changed as the temp changes.

[Brady Watson]

Jerry - up a few posts:



What most don't realize that the material absorbs heat as it is being cut, in addition to internal stresses being released, on some materials. This small amount of expansion increases the overall size of the material - and leaves you with a ridge when you restart the tool cold the next morning. I've cut enough large 3D files to make this observation.

The trick here is to learn how to tweak the Z after you restart the file to get it to match up exactly where you left off. It isn't hard to do.

-B

[tuck]

The great artist Salvador Dali once exclaimed to his students; "Forget about perfection! You will never achieve it!"

http://i50.photobucket.com/albums/f332/Superfan52/Salvador_Dali.jpg

[gundog]

I was cutting my vacuum plenum the other day and I noticed on the second zone it was way deep about .300" too deep. I stopped the file checked the bit for tightness and it was tight. I re-zeroed the z and cut the rest of that zone and the other 3 zones with no issue. I wish I would have noticed the indication of z before I hit the E stop but I did not so I am not sure what happened. This is a first but I have never had great Z repeatability.

When I surfaced the spoilboard I ran the bit to 6", 6" and zeroed the file was supposed to cut .020" when I started it, it did not touch the material @ 0, 0 I figured it was just not flat so I remade the flattening file to remove .030" and started it, it still did not make contact. For the hell of it I zeroed the bit again and then chose the same file and it cut the .030" deep and surfaced the whole table. I have also recently had the z zero routine stall on the first plunge and had to exit out of the control program to get the machine to go again.

These are all fairly new so I am wondering what I have going on.

Mike

[joe]

I'm with Shari.

Every one who sets up one of these machines has different results. I tweaked the one I had till I threw in the towel.

I have a new VW Bug. I love it but I'm quiet sure it's not a RR.

Tweak away.

Joe Crumley
www.normansignco.com (http://www.normansignco.com)

[Brady Watson]

Mike,
A few things come to mind...

1) If you are running a PRS, you need to check the bottom rollers on the YZ car to make sure they are adjusted. Every one I've looked at that was installed by the owner has play in the rollers to rail. This results in ridges when machining the spoilboard and inconsistencies in Z because there's enough slop there to move in +Z when the tool hits the material and lifts the YZ car off of the top rails.

2) Unless the vacuum bleeder board is 100% secure & being held down by both mechanical fasteners/countersunk screws around the perimeter & vacuum in the center, then there is without a doubt going to be movement of the bleeder board in Z.

Like Joe said, "Every one who sets up one of these machines has different results." - It's often true.

There is no doubt in my mind that you can get these machines to run 'tight' - but you have to eliminate all the places where you can get accumulative slop. Even with a *perfect* machine, you can still have your Z be off by inconsistencies in the machine bed, material and other places. This holds true whether your machine is painted blue, green or whatever. Investigate & go after the facts of why your machine is doing what it is doing. Get down in there and investigate the why/what & how of your machine. You'll figure it out if you look hard enough. Things as stupid as the machining tolerance in ArtCAM could make your Z off by a few thousands of an inch.


-B

[jerry_stanek]

I just added a zeroing point on my PRS. It is an aluminum rod that is just off the X Y zero corner and I machined it to the table top. Now all I have to do is move to -.5 X and run my z zeroing program. Like I said I just installed it and only tried it this afternoon I seems to work very good and it should give the same zeroing each time. I will run the surfacing program and let the cutter mill off the top to keep it at the same height as the spoil board.

[joe]

Jerry,

Your soluction may work perfect at first but the real test comes in a few weeks or months down the trail. On light weight machinery like this it's hard to maintain tight accuracies unless you work at it.

Joe Crumley
www.normansignco.com (http://www.normansignco.com)

[lrheimpel]

This has been an educational thread for me. I had solved the problem by zeroing to the table bed. Also, I use an "L" boundary fixed to the table and each time I start working (if I left the shop) I set the XY zero to the corner of the "L".

[Gary Campbell]

I want to give a bump to Brady's post #23 above. I have walked a good number of guys with cutting inaccuracies thru a few steps to make sure that their machines mechanical adjustments were set properly.

In virtually every case the operator was sure his machine was "tight". In each case, proper adjustment cured the problem. Except in extreme cases, an operator cannot "feel" the .010-.015 that results in unacceptable cuts, especially in the Z. This almost invisble amount, added to accumulated slop inherent in all motion parts such as bearings, easily adds up to some horrible cut quality if the adjustments are not checked regularly.

Remember, there are a good bunch of guys out there cutting fast and hard and in some very dense materials, even aluminum. Their machines can do this on a regular basis, why cant yours? (the operator is the only intelligent part in a CNC machine)

[applik]

Rest assured that there are other operators out there that have experience well beyond what is assumed as your level of expertise. I am Shari's husband, and have over forty years experience with engineering, designing, purchasing, installing, and operating sophisticated paper mill machinery involving everything from thousand-ton, city-block long, paper machines to space-age micro-machinery smaller than a pencil eraser. I know my way around dozens of forms of computerized distributive control systems, automated instrumentation and specialized electro-mechanical control systems, and have decades of experience in preventative maintenance principles, electrical engineering design and control, as well as optical and laser allignment principles and applications.
I spent decades working with some of the best manufacturers, engineers and mechanics the industrial spectrum has to offer in the areas of bearing applications, lubrication technology, failure analysis, computer technology, power grid systems, equipment optimization programs, etc.etc.
I have an electrical engineering degree, as well as a chemistry degree. I have spent literally tens of thousands of hours managing maintenance crews, construction crews, and engineers involving well over a billion dollars in greenfield projects as well as years of managing routine maintenance programs and leading teams during crisis management and catastrophic machine failures. I knew what was going on in my world because I was always right there beside the workers with a flashlight and calipers and not hanging out with the lounge lizards.
I have been fortunate in having attended thousands of hours of training programs at the facilities of the some of the world's most sophisticated, leading-edge manufactures of complex electrical and mechanical equipment made on the planet.
In short, the CNC machine in my shop is a toy erector set from my viewpoint. I know how to run dial indicators and have seen more than a few sets of calipers and micrometers in my time. We have tweaked the Bot in every conceivable direction and we now know beyond a shadow of a doubt what it is capable of and what it is not.
Like all machines, it has its good and bad points. Be assured that we are in control of those limitations as much as humanly possible.
As far as a person not being able to feel, or see the difference of ten or fifteen thousands of an inch...either they aren't looking or don't know what to look for. It is impossible not to notice an offset of just several thousands if you know how and what to look for. In my experience, five thousands of an inch tolerance is not even in the ballpark when speaking of accuracy.
Shari can tweak the Bot down to about a four thousandths variation in return zero height but the return height is never consistent. She is the best practical engineer I know, and I know a lot of them! She is also a perfectionist and that is reflected in her designs. She is one of the best in the use of Artcam (call them if you don't believe me). She simply wants the machine to perform consistently.
Between Shari and myself, we have investigated every performance angle that the Bot operates under. There are no loose or worn parts, no signs of end-play or excessive run-out. The complete system is grid-grounded with no signs of unbalanced voltages, points of static generation or dispersal, no spurious equipment noises or induced AC effects. There are no computer issues on the dedicated system.
I could go on for another ten pages about what is not electrically or mechanically "wrong" with the machine, but will end it here: If anyone thinks they have the "Lone Ranger" answer to the unrepeatable Z question, please contact me and we will discuss it together. I would be most happy to tame the Z monster but I am confident that we are working against both the design and mechanical function limits of the machine. If there is one thing I have learned in my 60 plus years of life, it is that I try not to offer a definitive diagnosis of any problem without seeing the problem firsthand. We are all in the same Bot boat, hopefully all paddling in the same direction. Reese Loveless. (Bot co-owner!!)

[Gary Campbell]

Reese...
Welcome to this forum. Please understand that my comments were not directed so much at you and/or your machine as to make a general statement of what my personal experiences with a good number of machines has taught me.

I agree, the ShopBot is an entry level, bolt together kit CNC. It is very complete, and partially assembled, but a kit none the less. Fortunately there is a good number of ingenious operators that find ways to workaround the inherent limitaions of the machine and actually make them perform above their pay grade. Hopefully I am thought of by others as one of them. I am obviously not as qualified as yourself, however, I do find a way to get problems solved.

It is very possible that you or your needs exceed the capabilities of the machine and need to move up. Many, including my friend Joe have done this and have had success. There is a good amount of tinkering involved with keeping a machine cutting above its abilities. I enjoy this (most of the time), others, like Joe, prefer to work and make money. If this is the course that you select, please dont forget to stop back here and tell this group how you are doing.

There are many specific instances where the ShopBot machine and/or its software do not have the best solution. Nor would I expect them to. Again, this is an entry level machine. I would expect that there are a number of users, myself included, that could reduce or eliminate your problem by either adjustment of mechanical or software items, or initiate a workaround that gets the Z accuracy to at least the advertised level of repeatability.

There are many instances in my shop where the ShopBot is not the best tool for the job. The same may be also true in yours. I do stand by my words, as I have faced that same issue in the past and had it compounded by the addition of multiple (8 at the time) tools. The only thing consistant was the inconsistancy. It did take a long weekend, but I now have 12 tools that I can trust.

You are correct, It takes a keen eye, firsthand at the machine to solve these problems. That is one reason I didnt offer up a long distance diagnosis to Shari's specific problems. The other is that Brady is as good as they get and there is no need to pile on. I did however, want to reiterate to others following this thread, the importance of frequent maintenence adjustments to keep the machine tight.

On another note, I personally would never pause or shut down the machine. I would prefer to leave it on..... for days if needed.

[gundog]

Mike,
A few things come to mind...

1) If you are running a PRS, you need to check the bottom rollers on the YZ car to make sure they are adjusted. Every one I've looked at that was installed by the owner has play in the rollers to rail. This results in ridges when machining the spoilboard and inconsistencies in Z because there's enough slop there to move in +Z when the tool hits the material and lifts the YZ car off of the top rails.

2) Unless the vacuum bleeder board is 100% secure & being held down by both mechanical fasteners/countersunk screws around the perimeter & vacuum in the center, then there is without a doubt going to be movement of the bleeder board in Z.

Like Joe said, "Every one who sets up one of these machines has different results." - It's often true.

There is no doubt in my mind that you can get these machines to run 'tight' - but you have to eliminate all the places where you can get accumulative slop. Even with a *perfect* machine, you can still have your Z be off by inconsistencies in the machine bed, material and other places. This holds true whether your machine is painted blue, green or whatever. Investigate & go after the facts of why your machine is doing what it is doing. Get down in there and investigate the why/what & how of your machine. You'll figure it out if you look hard enough. Things as stupid as the machining tolerance in ArtCAM could make your Z off by a few thousands of an inch.


-B

Thanks Brady,
I checked my rollers on the bottom of my Z car and they are tight, at least the way I remember the manua said to adjust them when I built it 3 years ago. The bottom rollers roll against the V rail but have a noticeable drag and do not free wheel when you turn them by hand and the machine is turned off. I will check my rack and pinoin to make sure they are tight I need to do that.

The problem I had the other day though was a first and was not the result of any bad adjustment. The machine flat cut a section of the plenum .300" too deep when I saw that I hit the E stop. The only thing I did was re-zero the Z and start the file from where I stopped it and it returned to the normal height. I have no clue why it cut that one section of the plenum that deep and had I not stopped it, it would have cut the rest too deep. I checked the collet and the bit was tight nothing changed except I stopped and ran the zero routine again. and started it again where I stopped it.

Mike

[Jerome]

I am still very new to the shopbot but I did have a thought on the Z zero problem. I have a widebelt sander and planer that both have digital controls to raise and lower the table. Any time they are raised or lowered, the table will go slightly lower than the measurment called for and then raise to the desired height.

This is to eliminate backlash. As near as I can tell the shopbot can't use this system since it is constantly raising and lowering while working.

I am amazed they can be as accurate as they are.

Jerome

[Brady Watson]

As Gary pointed out, my post was more of a generic one, than one specifically for your machine. A lot of users read the forum & never post. I agree 100% - remote diagnostics are difficult at best.

I think ShopBot advertises something along the lines of +/- 0.015" accuracy on these machines (may be tighter now) - since they are after all, geared mainly for larger machining jobs and woodworking - where material can easily move that much from morning to afternoon from thermal and humidity driven expansion. I am sure you are perfectly aware of the limitations of your CNC.

It sounds to me, with your background of expertise, that you could easily put a ballscrew on the Z and immediately eliminate all of your woes...provided you keep it clean :D

Have you given this any thought?

-B

[magic]

Sometimes I measure the distance from the collar nut to the bottom of the bit then change bits and tighten the nut with the distance equalling the same as the measurement.

This usually requires me to put the first bit in the collar by a fixed amount. I use various "L" brackets on both bits.

I have 3 "L" brackets. one is 1 inch the second is 2 and the third is 2.5. - The possibility exists to use a short bit first then to change to a longer bit but pull up the Z by .5 and rezero - then both bits cut at the same depth.

So, it does not matter what height the program stops the Z at, when it resumes, it resumes at the same depth. I got this idea from using a collar on a drill bit and assumed if I had collars on every bit I could control how far they cut

[applik]

Brady, Thanks for the followup. When I unpacked the machine years ago, my first thought was to install slide bearings and worm drives all around. After using the Bot extensively, I think it would be an unwise move, even for the Z axis. There are too many cumulative issues at work here.
Last week, we saw a 5 year-old, extensively used, CNC machine that had just plowed an inch-wide design through inch-thick bullet resistant Lexan at a speed that would make you cringe. We then carved a delicate leaf pattern, complete with miniature veins and patterns on that same machine. After a series of pausing the machine or shutting it down completely and restarting, we saw absolutely no variance in the Z height; even with an optical collimator. This is what we are after.
The Bot is simply what it is...nothing better or worse. It is a serviceable tool for cutting sheet goods most of the time. It still has a mind of its own when inputting MZ commands. Considering its design and construction, I am amazed it accomplishes for us what it does.
Our substrate expansion due to temp and humidity in Arizona is nowhere near what ya'll suffer with in other locations. Our relative humidity has been about 2 to 3 percent all winter...the driest in 75 years of record keeping. Still, we are very much aware of this problem.
As far as my stating my expertise, I did so only as a matter of information on the subject and not bragging rights. We have talked with other Bot owners around the U.S. that are incredibly gifted and talented about the common ills plaguing the machine. I am always interested in other owner's opinions on the subject, no matter their expertise level. I always learn something from every conversation in life, even if it is not always positive.

[bleeth]

After years of owning a bot and also investigating the plusses and minuses of other equipment I have reached the following conclusions that are based on my own observation:
1. Out of all the "entry level" machines out there I am convinced that the Shop-bot is the best of the bunch.
2. It is not now, and under current construction, control software, and motor configurations the best machine available regardless of cost.
3. There are superior designs and engineering out there.
4. If one is looking for a serious upgrade to the SB for reasons of resolution, accuracy, and repeatability, todays economy allows one to purchase a lightly used machine superior to a Bot if one is comparing to the current top of the line Alpha with toolchanging machine. This is only due to the severely depressed economy in general and the large amount of really fine equipment available on the used market.
5. Don't even consider trying to find that superior peiece of equipment unless you are ready to part with 40-50K (Although you never know-you may get lucky)
6. The vast majority of Bot users do not need to run a 16 hour 3-d program with multiple pauses and complete halts during their execution and therefore:
7. Out of all the "entry level" machines out there I am convinced that the Shop-bot is the best of the bunch.

Having been in the habit of using the auto analogy for some time now I see no reason to stop at this time.

You just can't compare a Cadillac to a Volkswagon.

Good luck on your new machine Reese and Shari.

[jkaras2000]

Just read the Bolt/Zero thread. Very good idea. I do 3D Wood Carvings presently. Depending on material availability, as well as how it comes off my planer, my material thicknesses can vary. The bolt interests me, because I am thinking I could bolt it through the table to make an adjustable height to match the material I am carving. I have been trying to Z Zero off a piece of the same material thickness when possible. Mostly a problem if I get a missing Com message and have to Z Zero again. Any Thoughts? Merry Xmas Joe :cool:

[rb99]

Oh now I read this thread... after I buy a 12" new Z and motor. I thought it was the ballscrew Z that was the problem and the new Z fixed the problem. Apparently not?

[propellers]

Nearly all of our work is 3D on a PRT. We have experienced all of the issues I read in this thread but have found means of getting around them with the understanding the machine does not have the mass or rigid enough for extreeme accuracy. Two things that has made my cutting better, one is never turn off the machine if the part is not completed at the end of the work day..gravity will change the Z position when the power is off. The second is the machine has to do the measurement of zeroing the axis and certainly will never be acurate using the eye.

I use the K command for all movements. To keep explanation simple and since the question is about the Z axis and a caliper has been firmly mounted to the machine then by using this as a measuring point will be simple to zero you current cutter at a Z position.

Use aligator clips on a length of wire. Clip on your zero plate and to the metal caliper. Move your cutter above the caliper so the same point on the cutter will touch the caliper. Use the "K" command then hit "D" on the keyboard. This will allow you to enter a value in the step box. Use a very low value to obtain accuracy. Then repeatably keep tapping the page down key until the light on you #1 switch activates, this tells you that the cutter is contacting the caliper, ZZ and you have repeatbility of confirming a location as acurate as one can possibly achieve. I use this proceedure for all axis..even when offsetting to other locations for multiple part cutting.

[rb99]

Nearly all of our work is 3D on a PRT. We have experienced all of the issues I read in this thread but have found means of getting around them with the understanding the machine does not have the mass or rigid enough for extreeme accuracy. Two things that has made my cutting better, one is never turn off the machine if the part is not completed at the end of the work day..gravity will change the Z position when the power is off. The second is the machine has to do the measurement of zeroing the axis and certainly will never be acurate using the eye.

I use the K command for all movements. To keep explanation simple and since the question is about the Z axis and a caliper has been firmly mounted to the machine then by using this as a measuring point will be simple to zero you current cutter at a Z position.

Use aligator clips on a length of wire. Clip on your zero plate and to the metal caliper. Move your cutter above the caliper so the same point on the cutter will touch the caliper. Use the "K" command then hit "D" on the keyboard. This will allow you to enter a value in the step box. Use a very low value to obtain accuracy. Then repeatably keep tapping the page down key until the light on you #1 switch activates, this tells you that the cutter is contacting the caliper, ZZ and you have repeatbility of confirming a location as acurate as one can possibly achieve. I use this proceedure for all axis..even when offsetting to other locations for multiple part cutting.

Interesting... Could you post a photo of this caliper? How do you zero X & Y and account for the caliper?

What step down number do you use?

Thanks

[tomwillis]

Ok - I'll chime in here too....as I've been keeping track on this subject as well. What I've found is my prox switches on my PRT based on global acceleration rates are not accurate. Depending upon what my acceleration rate is set - depends upon how fast the gantry hits and then moves off the prox switches....including the Z. I have been coding in a sep G4 (pause move) in 2 step increments - typically start and pause at 0.1000 off the zero work surface and then a G1 down to zero another G4 to make sure it's at zero and then G1 up and start the program and run. It seems like this has been my only solution of accurate repeatability.

[propellers]

Interesting... Could you post a photo of this caliper? How do you zero X & Y and account for the caliper?

What step down number do you use?

Thanks

I’m not sure if the picture uploaded, however, I do not use a caliper, I mention it because the author of this thread had a caliper already firmly attached so this will provide a non-movable point to be used each time for ZZ. Unless you have a fixed point to return too, reset zero readings will vary. For instance, if one uses the zero plate to zero Z then no matter where it is placed on the table it will be a different height. I use the zero plate because my Z height is non-critical but my XY is very critical for alignment, but, the procedure is the same one has to have a fixed locating metal surface for either X,Y,or Z. The caliper has nothing to do with this process; it’s just a fixed position.
I set the “Fixed” setting at .001, use any setting that will provide moving to the zero device without ramming into it. This thread is about repeatability of Z zeroing process so I will use that process but X, Y is identical with the exception that if a hold down bolt/device is used then both sides is measured, divided by two, move the axis to that point an ZY, or ZX.
Process Used. 1. Must have a fixed non-removable metal location. 2. Attach a wire/alligator clip on a wire works good, to the Z Zero Plate and to the fixed metal location. 3. Move cutting bit over this location..use the same X, Y, location each time..do not use the eyball method..we do not want to have variables here. 4. Move cutter Z height close to the fixed metal object; do not touch the metal object with the bit. 5. Enter “K”, Enter “D” or click on “Fixed”. 6. Enter a value that will move the Z Axis in very small increments. 7. Tap the Pg.Dn key multiple times until the bit touches the metal location..the touching of this location will be known when the #1 switch light illuminates on the monitor..PgUp until the light goes out then PgDn until the light illuminates then ZZ. This can be even put into the beginning of a program that will make it automatic.
Lonnie Prince
419-877-5557

[jerry_stanek]

We have tried all sorts of methods for re-zeroing without much luck. The bolt is the latest method, it comes closer than any other we've used, but still not acceptable when doing the 3D. The Z just will not go back to "dead on". In the second picture, the ridge measures .004 higher on one side than on the other. Sometimes it's lower, never the same outcome. Would be great if it would be exact, but I seriously doubt that will ever happen. Funny thing is, we can get the X and Y right on, but never the Z.

Will be watching this thread with great interest.

SL


After you re zeroed the file and ran it could you try again to re zero and run a third time. I wonder if the bit is bending just a little on the first cut but stays straight on the second. You are taking off a lot of material the first time.

[bleeth]

Jerry:
I doubt if Shari will answer-She sold her Bot and bought another machine.
Apparently she is happy but as I pointed out early on in the article, If you want a better z repeatability it will cost you. Her current machine costs around 50K with a router. It is servos and ball screws, linear rails and a heavy duty welded steel frame.

[donclifton]

If you have a fixed plate you can zero you bit as usual then do a probe on the fixed plate and note the reading. Then after you change a bit reprobe the fix plate and change the reading to the original reading.
Don