Gut Check

[chunkstyle]

Hello everybody,

I've finished my table and vac hold down and have done some surfacing of the table tops along the way. My table was constructed using two layes of .75" Medex MDF and a .5" uldf spoil board.

The second, upper layer of MDF had the plenum pattern cut into it and was finished with 3 coats of laquer.

The plenum grid was surfaced using the shop bot program and the supplied 1.25" diameter onsrud bit. The surfacing to the plenum was done to remove the laquer and surface of the MDF to allow for good glue bonding. There was definately tooling path marks on the top of the plenum but I didn't mind since it was only for gluing down and the assembly manual mentioned that the first pass was going to be a bit rougher than final.

The ULDF was loaded and the vac hold down was turned on (heaps of thanks to all who shared their ideas and designs, what a HUGE help in my being able to build mine!!!). ULDF was then surfaced and the tooling pattern was not as bad.

I removed the ULDF from the machine bed and proceeded to roll on some Titebond 111 with a foam roller. After covering the entire surface of the plenum I flipped the newly surfaced ULDF face down on the plenum, covered with plastic, turned on the vac hold down (another big shout out to the forum!!) and waited for about 45 minutes for the glue to set up.

After this I did a final surfacing of the table top's newly glued down surface.

The tooling marks are still apparant but no where near what they were like on the first surfacing made to the plenum. The upper left corner of the table appears to have a bit of a 'hump" at the corner whereas the kitty corner (near right) has the opposite or slight depression at the corners.

My first question is how good is good nuff. I will primarily be cutting out cabinets. I don't know yet what is realistic for these machines and what can be expected for the tolerance of flatness for the bed surfacing.

I searched the past posts and "tramelled" my spindle. With the spindle moved about 18" in the X direction and 24" in the Y I tested the plumb of the z. It was done using a peice of .5" thick glass and a trammel with a dial indicator at the end of a 5.5" R arc (or 11" circle). The results were 20 thou out on the y axis and 13 thou out on the x. Would this be within the range of expectation?

Again, let me say how much help the forum has been in the design of my shop bot. It has taking alot of the mystery out of getting set up. I just wanted to do a gut check with the community to see what else can be done to get the best level of performance out of the machine.

Thanks for any advice,
Tim

[chunkstyle]

Whoops,
I forgot to mention that I'm running a PRS standard with a 2.2 hp spindle.

[bleeth]

For cabinets the real test comes in when you find out how square your parts are. To me the test for z level squaring is to cut out a part and put a machinists square on it. If my edge is square for that and it run's through the edgebander clean and when I butt join two parts together I get the same clean join whether they are flipped or not then I am happy. The funny thing is although we all know cabinets are "only boxes" (That was a joke) they are the most demanding when it comes to size and squareness requirements. The 3-d guys and signmakers can usually get away with comparitive murder in that area.
OKAY-Let's hear all you 3-d and signmakers out there SCREAM!! :eek:

[gc3]

For cabinets the real test comes in when you find out how square your parts are. To me the test for z level squaring is to cut out a part and put a machinists square on it. If my edge is square for that and it run's through the edgebander clean and when I butt join two parts together I get the same clean join whether they are flipped or not then I am happy. The funny thing is although we all know cabinets are "only boxes" (That was a joke) they are the most demanding when it comes to size and squareness requirements. The 3-d guys and signmakers can usually get away with comparitive murder in that area.
OKAY-Let's hear all you 3-d and signmakers out there SCREAM!! :eek:

ok on the edge bander and squareness....if you use dado for box construction who cares if you are out a few thous...

on another note I have started to use the vintage machine I have with a soon to be upgraded y carriage which I will purchase from... ahh, bleeth...:eek: for cutting dadoes in face frame parts to accept plys, works well.

I also butcher out 3d stuff and it looks much better when the z is out about ohhh .250" :rolleyes: gives it a real "rustic" look

BTY I am still looking for used prt, pr parts to create another bot.

Gene

[michael_schwartz]

on another note I have started to use the vintage machine I have with a soon to be upgraded y carriage which I will purchase from... ahh, bleeth...:eek: for cutting dadoes in face frame parts to accept plys, works well.

I also butcher out 3d stuff and it looks much better when the z is out about ohhh .250" :rolleyes: gives it a real "rustic" look

BTY I am still looking for used prt, pr parts to create another bot.

Gene

The stuff you crank out with that vintage PR is pretty impressive. Gives us guys with PRS machines no excuses.

[dlcw]

Tim,

Your X & Y numbers should be fine for cabinetry. A single sheet of plywood will vary in thickness, over it's surface, by .015 (on a good sheet). A crappy sheet might be upwards of .03 - .05 over the surface. I've even had some import versions vary by .06 across the surface. :(

.0012 shouldn't be a problem. My dado clearance is set for .007 with a depth clearance of .01 for M&T joinery cabinets. Now if you were cutting steel then this would be a little sloppy.

I think you are ready to try a cabinet and see how it all fits together.

'Bot on! :D

[chunkstyle]

Thanks guys,

Not knowing what is considered "tight" for tolerance had me wondering. I worked some more on the spindle mount and got the x deviation down to about 7 thou and the y deviation at 13 thousandths. I was able to shim the top of the spindle out from the mounting plate and get the X run out figure reduced. For the y direction I loosened all the allen head bolts that mount the spindle to the z-plate and tried to move the spindle in the proper direction to lower the deviation. I was only able to get a few thousandths out of it.
I'm glad to hear it's really not a cause for concern. I'll make a few cut outs and check the edges for square as suggested. I"m hoping to be able to take parts directly from machine to bander without problems. I have a laminate countertop scheduled to do this week and I think it would be an ideal starter project for my bot.

Thanks for the advice, it's really appreciated.
Tim

[chunkstyle]

Just a quick update,

I cut out a long rectangles that was then cut in half to form two rectangles.
As Dave suggested, I checked the edges for square and did the butted up joint of the two short sides that had a common cut and flipped one side over. The results were not to my liking on either the squareness of the edges check our the gap of the butted up joint. I can get a better result cutting parts on my cabinet saw that cost one tenth what this cnc is costing me. Is anyone getting square edges as good as a decent cabinet saw can cut with their shopbots or are my expectations to high for this machine?


Looks like some more time spent getting this cnc dialed in some more.I build frameless so it would be nice to have the edges of the cab parts square to the surfaces.

I also put a length of .5" drill rod in the spindle and slipped a peice of float glass under the rod to create a larger flat reference plane. A machinist set up block was then slid up next to the rod and the same run out could be visibly seen as my previous dial indicator trammel test was telling me.
I realize that these machines are not the most robust and there is some slop that's always going to be there but shouldn't it start out as close to being perpendicular to the bed as possible so that the slop won't be further magnified?

I'm currently reading Gary Campbell's article on tuning up the PRS gantry. I'm thinking that this might be my next step to getting it cutting close to square. I get how you can get the spindle square to the table in the X direction. Not so sure I understand how you can get it square in the Y, as it's mounted to the z-carriage. Have to re-read it again.

Sorry if my posts comes off abrasive. Just getting annoyed at the amount of work that is needed to get this up and running at a level of accuracy I had hoped for. Time is something I don't have alot extra lying around these days.

[michael_schwartz]

I'm currently reading Gary Campbell's article on tuning up the PRS gantry. I'm thinking that this might be my next step to getting it cutting close to square. I get how you can get the spindle square to the table in the X direction. Not so sure I understand how you can get it square in the Y, as it's mounted to the z-carriage. Have to re-read it again.

Sorry if my posts comes off abrasive. Just getting annoyed at the amount of work that is needed to get this up and running at a level of accuracy I had hoped for. Time is something I don't have alot extra lying around these days.

Make sure to adjust the end stops so you can push the gantry up against them to square the machine on power up.

If you need particularly tight tolerances use an offset roughing tool-path, and climb cut to flex the bit/machine away from the cut line. Then run a finishing path to take off the slight bit of remanning material.

1/4" bits have a tendency to flex as well.

[chunkstyle]

Here's some images of what I'm getting from the table surfacing and the drill rod/set up block relationship in the Y axis.
Is getting your spindle any more perpendicular to the table feasible or is this whats to be expected?
I would not be happy id my blade was set up to make this cut in my table saw but maybe I'm being overly critical?

[chunkstyle]

Thanks Mike,

I will keep the tool path advice in mind for future cutting.

[michael_schwartz]

I would not be happy id my blade was set up to make this cut in my table saw but maybe I'm being overly critical?

You should be able to get the spindle square up perfectly.

[dlcw]

Tim,

It took me two days to get my spindle perfectly trued up but boy is it worth it. It took me a weekend to get my automatic edgebander to perfectly trim the top and bottom of the edgebanding after it was applied, but boy has that saved me tons of hours on production jobs.

Heck - two hours to get perfectly trued up planer and jointer blades (after each change).

You just need to take time to get your spindle perfectly plumb. Yes, it takes time but once it's done, it's done. Just check nuts and for tightness on your regular maintenance cycle and you should be good to go.

[chunkstyle]

Thanks Mike and Don,

Did either of you guys use Gary's PRS tweaking suggestion? One of the questions that I have is how I'm going to get the spindle squared to the table in the Y axis. I've loosened all the mounting plate bolts that tighten the Spindle's mounting plate to the Z slide and tried to get the spindle plate rotated (in my case, counterclockwise) but there doesn't seem to be enough play in the bolts holes. Any suggestion on how you were able to take care of this, if you had to, on your machine?

Thanks again,
Tim

P.S. Funny you should mention it Don, but I have to dive into my newly acquired, used edgebander after I get this shopbot squared away. Picked up a SCMI k201 and just arrived today.

[bleeth]

Learning how to get your CNC squared up to your satisfaction the first three or four times is painful and time consuming. Then you learn the tricks and when you need to do it again it is not so bad. But it is a learning experience and some of it you will just have to handle with patience and intuition as well as the good advice offered.
If you think getting the CNC tweaked is a pain you will really hate tweaking an edgebander!!! They are beyond a doubt the biggest pains in the world. Or at least the shop. And they need it much more often.

[Gary Campbell]

Tim...
As you can tell from the above posts, you can and need to get the spindle perfectly square in both directions. The spindle mounting bolts take care of one direction and the 16 gantry to Y extrusion bolts (along with the gusset bolts) will do the other. Make sure the rails are dead on first.

Like Don says, its a pain and doesnt go quickly. My article was short, but it took me most of a Saturday and a good part of Sunday to get the spindle set to my liking.

I would suggest you get a surfacing routine that uses just the x direction for cutting. The SB OEM file exaggerates any error by flexing the tool in all 4 directions. We use a 2 1/2" bit to surface the table, buying it was what mandated rethinking how perpendicular my spindle was.

All machines require retuning on a regular basis. Most of them will allow you to gain more experience before resquaring than the ShopBot, which is not factory assembled.

If its any consolation, I am doing this again on my ATC machine. The electricain finally got the 3 phase box in and I ran it for the first time last nite. I should be ready for cutting in a week or so.

[chunkstyle]

Thanks Guys,

I'll be starting in on it today. I'll probably have a few more questions about it along the way. Thanks for the help and perspective on it.

Tim

[michael_schwartz]

Thanks Guys,

I'll be starting in on it today. I'll probably have a few more questions about it along the way. Thanks for the help and perspective on it.

Tim

I remember when I was setting up my machine I had to wait 3 days for an electrician before I could fire it up. Those 3 days felt like a month, since I had to sit there and stare at my machine wondering if it would even work.

In a few weeks you should have your machine tuned up nicely, and you will be past the initial learning curve. My advice is to start drawing stuff in partworks when you have spare time since that will be the most important thing after your machine is squared up.

If you want to practice cutting stuff get a few sheets of pink foam from home depot. I started out cutting a few shop projects out of cheap plywood.

One last thing, a cheap digital caliper is a must have. Just keep extra batteries around.

[nat_wheatley]

Thanks Mike and Don,
One of the questions that I have is how I'm going to get the spindle squared to the table in the Y axis. I've loosened all the mounting plate bolts that tighten the Spindle's mounting plate to the Z slide and tried to get the spindle plate rotated (in my case, counterclockwise) but there doesn't seem to be enough play in the bolts holes.

I have the same question. How do you do this? I, too, tried to loosen the mounting plate bolts, and there isn't enough play with those.

[dlcw]

Your Y plate might be a little slanted where it mounts to the gantry. It could be one of your concentric rollers is a little higher/lower then the other one cause the alignment issue.

As far as adjusting the spindle itself, I had just enough play in mine to get it lined up. That in conjunction with the Y plate brought my spindle into perfect alignment (like the moon in the seventh house and Jupiter aligned with Mars- :D - dating myself there aren't I :eek:).

You might want to contact ShopBot tech support to see what you should do for this.

[tracys]

I had the same problem as you showed in your pictures. I talked to shopbot and they suggest putting steel shims behind the v bearing on the gantry. It took a few different trys but I have it really close now. My table looked like a drive in movie for mathbox cars, now it looks good with light shining across it.

[Gary Campbell]

Let me add a little to the mix for the YZ car. My first (single Z) was assembled slightly out of square with the axis of the upper V rollers. I disassembled it and using a steel straight edge on the rollers, squared the Z extrusion mounting block to that straight edge. I then cut a hardwood block and made sure the two blocks were parallel. I was then able to get it all aligned the way I wanted. Checked the distance between with digital calipers. If I had needed, I would have clearanced the top and bottom set of holes, as to me, plumb is more important than a perfectly round hole.

On the pink foam... you will have microscopic particles in every thing you own! Its safe, but very messy.

On the calipers... buy Mitutoyo. More money up front, but batteries last for over a year, and they dont seem to go out of calibration like the cheapos. I have gone thru 3 or 4 a year til I got the good ones.

[michael_schwartz]

On the calipers... buy Mitutoyo. More money up front, but batteries last for over a year, and they dont seem to go out of calibration like the cheapos. I have gone thru 3 or 4 a year til I got the good ones.

I suppose I have had better luck, but of the two I have the older one has been going on for 3 years now, and the newer one is about 6 months old. I have replaced batteries a few times. I suppose these will eventually get replaced with a Mitutoyo, but as long as they still work. I use both on a daily basis.

the older caliper is from Lee valley, and then I have another bought at Home Depot under the general brand. The OEM batteries that came installed died pretty fast. The replacements I bought seem to last quite a bit longer.

[chunkstyle]

Got the Y-alignment down to 2.5 thousandths on a 11" diameter trammel with dial indicator. We'll see how long the allen head bolts hold it there in the slightly enlarged holes.

Got the X axis alignment down to 8 thousandths on the same 11"D trammel. I will pull the z-car off the beam for a fifth and hopefully final time to make the necessary shim adjustment between the upper V-rollers and the z-car's plate steel.

Somehow I can't avoid the feeling that a little built in adjustment would have been nice for this critical point of design. Were getting there, though. It will be interesting to see how long the machine can hold the tolerances here and further back up or discount my whine about the need for some better adjustment in the design. Time will tell.

[chunkstyle]

O.K.,

Here's the results of the squaring effort on my bot. The one picture shows the spindle with the trammel and dial indicator set up. There's a small cresent wrench sitting on the peice of glass that the trammel will revolve around in a clockwise direction. The head of the cresent wrench will be 12 o-clock and the starting point of the dial indicator is at the 6 o-clock position.

Dial was zeroed out in at the starting point (6 o-clock position).

Moved dial indicator to the 9 o-clock position, then the 12 and finally the three o-clock position. The cresent wrench helps to orient the position of the dial relative to the start position in the pictures.

That's as good as I'm going to get it, I think, and I'll get the bot re-connected and put back together.

I will use this as my base line and trammel the spindle as a routine monthly chore, as others have suggested. I hope that the X-alignment stays put. The thought of having to pull the Z-car again.....

Thanks for all the help with this and getting me squared away with the rest of the world.

Tim