Saturn 2 4x4 - Nema 23

[GME]

wmgeorge

Apparently, I offended you. It's not clear to me what touched a nerve, but it certainly was not my intent. Good to hear your're content with your setup.

In the future, please feel free to simply ignore any suggestions I might make that don't interest you. Please also keep in mind that there are others participating in this and other threads who may find the information useful (or not), or at least interesting in an academic sort of way. Just trying to be helpful.

[wmgeorge]

wmgeorge

Apparently, I offended you. It's not clear to me what touched a nerve, but it certainly was not my intent. Good to hear your're content with your setup.

In the future, please feel free to simply ignore any suggestions I might make that don't interest you. Please also keep in mind that there are others participating in this and other threads who may find the information useful (or not), or at least interesting in an academic sort of way. Just trying to be helpful.

I am not offended in the least. But it’s a router for wood. Where in wood working do you need .001 accuracy?

[PLJack]

Thanks for the effort Robert. That makes a lot more sense now.
I think I must have came to the wrong conclusion regarding the Y and B sensors. I figured that if I was able to dial both in to reflect a squared gantry that I would turn on both and it would auto square the gantry each time I did a homing operation. Seems that is not the case. Especially if it went out of square with all the bolts tight moving that B stepper would just stress against it wanting to be where it is.
My thought was that it would clean up and backlash / slop introduced from running a job. Don't know. I might be overthinking that.
In any case I'll deal with it when I get there. Seems I should get my platform bed in place before moving too much further.


Thanks all.
Jack

[PLJack]

See my post #19 in David Falkner's "Calibration Question" thread for a discussion of an alternate method for determining whether your gantry is perpendicular to the Y axis, and if it's out, quantifying exactly how much adjustment is needed. It's called the 5 Cut Method. It is a bit time consuming, but works better than any other method I've found. On the Saturn 2, you'll have to be able to adjust the sensor blocks. You'll also need to loosen the gantry and move it the amount necessary to correct the error. I can say from experience that having to loosen both sides of the gantry to adjust it makes the whole process very fussy. Leaving one bolt snug as a pivot point helps but it's not perfect when trying to achieve movements in the thousands of an inch range. Having a very stiff machine is both a blessing and curse. It's great for cutting, but a bear for adjusting making adjustments. With extrusion-based machines, forcing the gantry into square works OK. Not so much for machines with linear rails where binding can be a problem.

Gary

Thanks Gary. Can't have too many methods to choose from .
Also thanks for linking to CNCNutz in that post. Some good information on that channel it seems.

[GME]

Good question. I agree that you don't usually need that level of accuracy in woodworking. However, there are circumstances where that level of accuracy is desirable enough to make it worth the effort. I use a lot of Festool products including the Festool MFT table. It has precisely spaced and sized 20mm holes in the table top. The tops are CNC cut and are within far less than .001 tolerance, according to a fellow at Boeing who took one into a Boeing shop to check it out with their precision measuring equipment. Used in conjunction with their guide rails, Parf Dogs (a type of bench dog) and saw, it's possible to cut sheet goods with surprising accuracy. Unfortunately, the mdf tops are consumables and obscenely expensive to replace (like all Festool products, which are German made).

Achieving a high degree of accuracy allows one to make new tops that are as good as the originals. It's also possible to make custom tops that are much larger, to breakdown sheet goods. I'm making some large torsion boxes for just that purpose. My largest guiderail is 3000mm long, which allows a full length cut. Much easier, safer and more accurate than wrestling with a 4 X 8 sheet of plywood or MDF on my table saw. The internet has much on the subject. If you or anyone else is interested, you can google the Ron Paulk workbench, Tomothy Wilmont's site for his MFTC and other variations, and the New Brit Workshop. Peter Parfit at New Brit has developed products specifically designed to achieve high levels of accuracy around the Festool MFT concept. It's a niche area in woodworking, but there is a strong US and European following.

Errors tend to be cumulative over the span of processing wood parts. IMHO, it's best to start with the most accurate cutting tools possible. I know many woodworkers who will devote substantial time getting their tablesaw tables and fences square to their blades. Most use dial indicators in the miter slots to get their setup as close to perfect as possible, since cutting errors tend to multiply (think assembling cabinet parts cut out of square). They make indicator mounts specifically for that purpose.

It all depends on what you're making. I cut 3d rosettes for Victorian style door trim on my CNC. No way you would see a few thousands error. However, MFT tops are another matter.

Gary

[difalkner]

I am not offended in the least. But it’s a router for wood. Where in wood working do you need .001 accuracy?

Cutting fret slots for guitar fingerboards and saddle slots on bridges, Abalone inlay on delicate pieces, etc.

I may not every achieve that tolerance on my CNC but I'll certainly shoot for it.

David

[wmgeorge]

Ok I see where you would need, and why not make it the most accurate machine you can? I was wrong. to butt in.

[ger21]

The tops are CNC cut and are within far less than .001 tolerance, according to a fellow at Boeing who took one into a Boeing shop to check it out with their precision measuring equipment

MDF moves far more than .001" with changes in humidity.

[GME]

Hello Gerry,

I agree with you. MDF, like wood, moves with changes in humidity. I've been a woodworker for over 40 years and have unfortunately seen firsthand how failing to take movement into account can ruin a piece.

The stainless Parf dogs I mentioned are 19.9 mm (about .004 undersized) and fit in the holes without any perceptible slop. During times of high humidity, one might expect that they would get tight to very tight. Over years of use, I have never found that to be the case. I cant' perceive any difference, no matter the weather or humidity. Go figure.

As far as the Boeing guy is concerned, I suppose the numbers were dependent upon the moisture content when the top was made versus the moisture content when it was later measured. Alternatively, I was fed a load of bull excrement. Wouldn't be the first time.

Gary

[wmgeorge]

Hello Gerry,

I agree with you. MDF, like wood, moves with changes in humidity. I've been a woodworker for over 40 years and have unfortunately seen firsthand how failing to take movement into account can ruin a piece.

As far as the Boeing guy is concerned, I suppose the numbers were dependent upon the moisture content when the top was made versus the moisture content when it was later measured. Alternatively, I was fed a load of bull excrement. Wouldn't be the first time.

Gary

I am trying to figure out how you got a 48 x 30 inch sheet of MDF into and out of the Boeing Plant and had a tool and die maker on company time do your measurements for you. I will admire you fussing with your machine to get .002 inch over 58 inches. Marveling that while David needs the accuracy at close tolerances for his guitar why building furniture needs +/- .001? I have seen and so have you what close fitting does when the humidity changes a fine piece of furniture. I will work with my machine to get it as close as I can within reason.

[GME]

I am trying to figure out how you got a 48 x 30 inch sheet of MDF into and out of the Boeing Plant and had a tool and die maker on company time do your measurements for you. I will admire you fussing with your machine to get .002 inch over 58 inches. Marveling that while David needs the accuracy at close tolerances for his guitar why building furniture needs +/- .001? I have seen and so have you what close fitting does when the humidity changes a fine piece of furniture. I will work with my machine to get it as close as I can within reason.

Maybe the Boeing thing is an urban legend among the Festool crowd. I heard about it from a co-owner of a woodworking store who supposedly knows the guy. He may have been blowing smoke, or the guy from Boeing fed him a load. Regardless, the MFT is much beloved, despite the obscene price tag.

Getting to a super level of accuracy is sometimes more a question of "want" than "need," depending upon what you're making and when it's "good enough" for any given person. For much of the work I do on the machine, several thousands one way or the other makes no practical difference. I plan to use the Saturn 2 for some dovetail work (Gerry's Jointcam). I've used the Leigh Jig for dovetails, and a few thousands can be the difference of a sloppy, too tight or a good fit (I gave up on hand cut dovetails years ago). It's just good to know its there when I feel like I need it.

There is also another dimension to it. I enjoy the challenge and am willing to put in the effort - even if sometimes it frustrates the heck out of me. If I was dependent upon the machine for making a living (I'm retired), I would probably have a different perspective. I've thought about trying to pick some paying jobs, but I keep busy with remodeling projects, and am not hot to reenter the working world. Having to worry too much about the money, customer service, stress and the myriad of other problems attendant to running even a small business could easily take all the fun out it - at least for me.

Gary

[ger21]

Regardless, the MFT is much beloved, despite the obscene price tag.

As are most Festool tools.

I plan to use the Saturn 2 for some dovetail work (Gerry's Jointcam). I've used the Leigh Jig for dovetails, and a few thousands can be the difference of a sloppy, too tight or a good fit (I gave up on hand cut dovetails years ago)

Yes, dovetails are very intolerant of errors.
0.003" in depth is the difference between sloppy and too tight

[difalkner]

I'll post this over in my thread on Calibration but I finally got it as close to perfect as I can see with 3.0 reading glasses.

David

[PLJack]

More progress.

I bought two sheets of MDF and cut them to length. Also picked up a few metric rulers. With the boards just sitting on the frame I started to tweak the "steps per" settings.
Up to now I had just used a dial indicator with a 20mm travel to train the motors. I made a pointy wood dowel to help with the measurements.



That worked out really well. They were off by about 2mm over 600mm travel. Got it to within a 1/4 mm for now. Then I started thinking about what else I could measure just to see where things were. So I attached a dial indicator to a dowel and inserted that into the spindle. I was pretty shocked by what I saw. Then I pushed lightly on the MDF and was able to move the dial by 80 thou! So, I need to secure the MDF before moving forward.

That go me thinking about machining the holes to attach the MDF to the frame. After a quick measure of the holes I remembered you all saying they holes we not spaced evenly. Which is fine.
Staring at the pointy dowel I got the bright idea of having the machine tell me where the holes are. That turned into a bit of a job.
I would move the dowel to each hole...



Then bring it down slowly and move it back and forth until I got it to about .5mm spacing.



Then recorded what Mach3 listed as X and Y with my cam software. Did that for all 60 holes. The back 20 were measured by hand and milled. The bad news is that was a lot of work. Good news is that I learned a good deal about Mach3 fixtures and how repeatable my Saturn 2 is, even half calibrated. Over several nights I ran the homing procedure with the fixture (G58) selected before starting to measure another batch of holes.
When I finished that job it dawned on me that I just painted myself into a corner. I can't make any changes, software or otherwise, until I mill the holes for the MDF. If I did all the numbers I just collected would not be valid any longer.

The holes in the frame are 6mm. So I went with an optimistic MDF hole of 7mm and a 11mm recess for the screw heads.
I was amazed that I hit 77 of the 80 holes within the 1mm.



1 in the back where I placed the spindle over the wrong mark before milling. The other two seem to be a bad measure.The rest look great.




I have four screws per rail installed for now. So I ran the test that started me down this hole. Again with the dial indicator.



Again I was concerned about how the dial was moving. Until I did that math. That dial moves a long way to cover 1 mm.

With front left at zero it travels 1.27mm up. So my gantry is low to the right. I've since noticed that the left of the gantry is shimmed. Maybe I replaced the shim on the wrong side? In any case I still need to loosen the gantry to square it so I can shim then.

From front to back I'm seeing less than 1mm difference. I understand this is MDF and not a 10 ton machine for milling steel so those numbers are pretty good given the size of the machine.
Still I like to think the better I can make the numbers at this stage it will pay off later. To night I'm going to draw some right angles on the MDF with the machine and get a rough idea where I am with a squareness of the gantry.

I will wait until I get the second layer of MDF installed and faced before dialing in the rest. With that in mind I have a couple of questions.

I bought a bunch of T-nuts to install into the back of the bottom MDF so that I have something mechanical to attach the top layer with. Is that a good idea? If so what spacing should I use.

[PLJack]

For the squaring test I attached the bottom of a mechanical pencil to the dial indicator.



Then drew some 600mm lines at right angles and measured.



If I measure the long side of the triangle in the cam software it comes out to 848.528mm

I ran the test twice and got 849mm!! I was not expecting that. I'm still hearing some creaking every now and then and assumed it was because my gantry was not square.
Not sure what to make of that.

The best I can do with this method is 1mm so I guess I'm done with that for now.

Jack.

[wmgeorge]

I guess I thought when you milled the top of the spoil board all would be fine? Don't you want the gantry and the tool to be parallel with the spoil board. I would assume the cross braces are welded in place on a jig but that can allow a few thousands of an inch or a few mm out of alignment with the one next to it or at the end would not be out of the ordinary. I would not shim the gantry based on one cross support.

I plan on screwing and gluing the top spoil board to the bottom base board. David used plastic screws on his.

[ger21]

I bought a bunch of T-nuts to install into the back of the bottom MDF so that I have something mechanical to attach the top layer with. Is that a good idea? If so what spacing should I use.

I wouldn't bother, as putting T nuts in is a lot of work.

What I would do, is surface the first layer. Then glue the second layer down to the first. I'd pre-drill the top layer, and use wood screws every 6 inches or so. Then remove the screws when the glue dries.

[difalkner]

For the squaring test I attached the bottom of a mechanical pencil to the dial indicator.



Then drew some 600mm lines at right angles and measured.



If I measure the long side of the triangle in the cam software it comes out to 848.528mm

I ran the test twice and got 849mm!! I was not expecting that. I'm still hearing some creaking every now and then and assumed it was because my gantry was not square.
Not sure what to make of that.

The best I can do with this method is 1mm so I guess I'm done with that for now.

Jack.

That is definitely looking good, Jack. I started out with a triangle and switched to large squares and rectangles so I could measure the diagonals both ways and then compare. It took some tweaking but I finally got it down to as close to dead perfect as my eyes and measuring capabilities allow over a 52" diagonal.

David

[GME]

I'm still hearing some creaking every now and then and assumed it was because my gantry was not square.

Just an outside shot, but could the creaking your hearing be a cable chain? I was hearing odd noises, and that's what it turned out to be.

Gary

[PLJack]

I guess I thought when you milled the top of the spoil board all would be fine?

Agreed. Should be. My concern is with the material being faced. If I use the layered version of MDF and cut through a layer halfway across the board, because it's a few mm lower to the right, I may have a delamination issue. Besides, OCD is real

I would not shim the gantry based on one cross support.

Good advice. I've decided to wait until I have a stable face before worrying about shimming.

I plan on screwing and gluing the top spoil board to the bottom base board. David used plastic screws on his.

I like both those ideas but gluing sounds like a nightmare if I have to replace the top layer. Plastic screws sounds great. I looked into to that but I was looking a nylon, not plastic, and they were way too spendy.So I bought a bunch of 14/20 hex drive bolts.

Thanks.
Jack.