Catahoula's router/mill build - welded steel with EG

[catahoula]

I'm revisiting the concept of dual-loop control with linear scales closed at the servo drive. I found that the Delta ASDA-A2-M drives have this functionality, and while twice the price of the B2 version, aren't outrageously priced considering the potential accuracy. However, I also read on the Haas website that they don't recommend linear scales for 3d surfacing. (https://www.haascnc.com/productivity...ions/ls-2.html) That's the primary purpose of my machine. Any ideas on why that might be? Is it something to do more with how Haas is implementing linear scales that impairs smooth motion? Or something fundamental to linear scales? Seems like many high-end machines suitable for mold work use them.

I'm a little confused as I thought one of the main differences between the standard VF series and the VM mold series is that the VM's are thermally compensated through linear scales, but digging a little deeper I'm realizing I may have incorrectly assumed that, as linear scales don't seem to be standard on those. Maybe they use other methods of thermal compensation.

The C5 ground TBI ballscrews I have are rated at 18 um/300mm, which is fine for the tolerance I need as the molds will be roughly 300mm in the longest axis, and getting below 25 um in general would be a pretty lofty goal for me and this first DIY machine. Realistically I hope to get it around or under 50 um. But I'm concerned about thermal growth of the screws on relatively long 3d run times. Thoughts appreciated

[davebaldwin]

Here's some pictures (sorry for the poor quality) of my gantry being assembled, lying on its back. Interior was epoxy-granite filled with a tube in the center for cables. Top rail mounting surface was epoxy molded with a granite straight edge.
Bottom rail is a piece of steel 50x15 glued (epoxy injection after rough alignment with grub screws) and bolted to the tube, then scraped for flatness. The small pockets at both ends for mounting the ballscrew supports were milled in-place with a temporary Y/Z axis setup...

I know in a another post to this thread you said you glued and screwed the back plate to the structural steel tube, but from the photo it looks like it is part of the tube so I am curious how you achieved this. A detailed explanation of how you fabricated the gantry would be most welcome.

Also what were the steps you took to align the two rails. Having them at right angles looks like it make the alignment a nightmare.

Thanks,

Dave.

[jackjr-123]

I know in a another post to this thread you said you glued and screwed the back plate to the structural steel tube, but from the photo it looks like it is part of the tube so I am curious how you achieved this. A detailed explanation of how you fabricated the gantry would be most welcome.

Also what were the steps you took to align the two rails. Having them at right angles looks like it make the alignment a nightmare.

Just a bit of epoxy on the ends to make it look like one piece.

The angle between the 2 rails doesn't matter. A small gap (1mm) is left between the top plate and the side plate and then filled with epoxy.
Top rail mounting surface was molded with epoxy and a granite straight edge.
Bottom is a piece of 50x15mm steel that was roughly aligned with the top surface, then epoxy glued, and finally scraped for flatness.

[mactec54]

Just a bit of epoxy on the ends to make it look like one piece.

The angle between the 2 rails doesn't matter. A small gap (1mm) is left between the top plate and the side plate and then filled with epoxy.
Top rail mounting surface was molded with epoxy and a granite straight edge.
Bottom is a piece of 50x15mm steel that was roughly aligned with the top surface, then epoxy glued, and finally scraped for flatness.

The angle between the 2 rails does matter they have to be precisely aligned, using epoxy is soft like plastic and over time will move, so not a good choice of material to use to get parts aligned

A different why to make a beam but unless you stress relieved the tube before you did all this work it will never be a stable beam and the epoxy granite fill only needed to be around 12mm thick around the inside to dampen vibrations filling like you have serves no purpose but to add a lot of unneeded weight

[klaas123]

Hi Matec, there is a special epoxy mix suitable for this task:

Klaas

[mactec54]

Hi Matec, there is a special epoxy mix suitable for this task

Klaas

Dreams are free too, I have used all the machine epoxies that are available and for slide way rebuilding as well, which they are not using here, they are still soft compared to metal there is no real substitute

You should check out what the max hardness is of the best epoxies and do some math that will change your mind as to what they can be used for, many people try all kinds of things with epoxy but it is not very successful

[klaas123]

https://diamant-polymer.de/wp-conten...-P-0442-EN.pdf

https://diamant-polymer.de/en/produkte/dwh/

[jackjr-123]

I'm revisiting the concept of dual-loop control with linear scales closed at the servo drive. I found that the Delta ASDA-A2-M drives have this functionality, and while twice the price of the B2 version, aren't outrageously priced considering the potential accuracy. However, I also read on the Haas website that they don't recommend linear scales for 3d surfacing. (https://www.haascnc.com/productivity...ions/ls-2.html) That's the primary purpose of my machine. Any ideas on why that might be? Is it something to do more with how Haas is implementing linear scales that impairs smooth motion? Or something fundamental to linear scales? Seems like many high-end machines suitable for mold work use them.

I'm a little confused as I thought one of the main differences between the standard VF series and the VM mold series is that the VM's are thermally compensated through linear scales, but digging a little deeper I'm realizing I may have incorrectly assumed that, as linear scales don't seem to be standard on those. Maybe they use other methods of thermal compensation.

The C5 ground TBI ballscrews I have are rated at 18 um/300mm, which is fine for the tolerance I need as the molds will be roughly 300mm in the longest axis, and getting below 25 um in general would be a pretty lofty goal for me and this first DIY machine. Realistically I hope to get it around or under 50 um. But I'm concerned about thermal growth of the screws on relatively long 3d run times. Thoughts appreciated

No idea why Haas doesn't recommend them for 3D machining, but I would keep it simple first and finish building it without linear scales. You can always add them later.

For the ballscrews you can always let the machine run some time with a typical workload to warm it up then map the ballscrew to compensate for average thermal growth and linearity.

[mactec54]

No idea why Haas doesn't recommend them for 3D machining, but I would keep it simple first and finish building it without linear scales. You can always add them later.

For the ballscrews you can always let the machine run some time with a typical workload to warm it up then map the ballscrew to compensate for average thermal growth and linearity.

There are very few machines that have or use linear scales and if they do the linear scale normally feed back is to the control not the Servo Drive, this is quite new that servo drives have this feature, and it is common with all main name brand servo drives Yaskawa being the first to have this and then the others followed

For the normal mill / Router user, this is not going to be an issue, you would have to have very bad alignment to have a thermal problem with a Ballscrew on a hobby built machine

[mactec54]

https://diamant-polymer.de/wp-conten...-P-0442-EN.pdf

https://diamant-polymer.de/en/produkte/dwh/

These products are everywhere you could fill many posts with links like this

Have you used it and you are supplying information from your experiences, or just another armchair poster with zero experience with any of these epoxies

[klaas123]

Never mind matec

[spumco]

...this is quite new that servo drives have this feature, and it is common with all main name brand servo drives Yaskawa being the first to have this and then the others followed

I'm not sure how 'quite new' dual-loop to the drive is.

My Copley Xenus with dual-loop is from 2012, and Ultra3000's go back before 2005. I've no idea when Yaskawa started doing it, but I've seen other U3K's on ebay with date codes of 2001.

If there are very few machines that use linear scales, and those that do normally feedback to the control, why is it now common with all name brand servo drives? One would think that if it's a feature customers aren't using it would have been dropped.

Perhaps this feature is mostly being used in high-speed point-to-point machines (think pick and place) to reach a very precise target position without having to have a very expensive main control?

[spumco]

These products are everywhere you could fill many posts with links like this

Have you used it and you are supplying information from your experiences, or just another armchair poster with zero experience with any of these epoxies

Yes, as a matter of fact - it's curing right now. I'll report in the Milli thread when I have stiffness data, including the column-to-bed joint.

[catahoula]

I would keep it simple first and finish building it without linear scales. You can always add them later.

let the machine run some time with a typical workload to warm it up then map the ballscrew

100% agree, just thinking about whether to get the drives with that functionality upfront. I suppose selling a lightly used set of DMM 400w drives and servos down the road wouldn't be too hard though. Good idea about doing the mapping at operating temp. There must be some point of equilibrium where it's not gaining much more heat; I wonder where that point is. Now am glad I went with the lightest zero-backlash preload on the ballnuts to avoid extra friction.

Picked up the frame at the heat treater today. 700lbs! Fills the better part of a mini-truck bed. Only 1 hour of the dark, mountainous drive home was in a blizzard, not bad for around here. No need for sandbags with this in the back

Debating whether to fill with EG before or after machining. Best to do it before I think, to minimize any remote chance of shrinkage torque (the chance of which I believe is essentially zero given the wall thickness of the steel and the relative modulus of epoxy), but will be (even) more of a hassle to move around if filled in advance. I'll ask the machinist if he cares one way or the other.

Edit to add: pleased that the parts didn't seem to warp much during the heat-treat. I was particularly worried about the z plate, which came from a big piece of flat bar. Seems like the heavy wall thicknesses worked out. Of course, the z plate is 1" thick, with 2.5 x 2.5 x 1/2" angle iron welded on the front. It was a bit bent in cross section from the supplier, as would be expected. I used the weld sequence to pull that out and it seems to have stayed in it's new flatter position. See what happens once it's on the mill. The gantry and base are going to the machinist but I'll give the z plate and a columns a whirl on a friend's bridgeport. The z plate might have to go to a machine shop to get it truly flat, but can at least get it started.

[catahoula]

one more

[peteeng]

Hi cat - Think about your hernia! Peter

[catahoula]

ha, yeah it's engine hoists and forklifts from here on out, filled or not

[mactec54]

I'm not sure how 'quite new' dual-loop to the drive is.

My Copley Xenus with dual-loop is from 2012, and Ultra3000's go back before 2005. I've no idea when Yaskawa started doing it, but I've seen other U3K's on ebay with date codes of 2001.

If there are very few machines that use linear scales, and those that do normally feedback to the control, why is it now common with all name brand servo drives? One would think that if it's a feature customers aren't using it would have been dropped.

Perhaps this feature is mostly being used in high-speed point-to-point machines (think pick and place) to reach a very precise target position without having to have a very expensive main control?

What Copley had back then, is not the same as what they have now, which they have improved over the years, I have used the Copley drives since 2000 they are still not where these others are and I doubt they ever will be they are a good reliable drive which was ahead back in 2005 in some areas, but has lost ground in resent years

There are many machines that use linear scales for positioning that have cheap drive systems, like belt drive, rack and pinion drives this is where this is used and of cause on high end machining centers Grinders Etc.

[mactec54]

Yes, as a matter of fact - it's curing right now. I'll report in the Milli thread when I have stiffness data, including the column-to-bed joint.

Yes we have seen what you have been doing to make your machine more damp

You are only filling the gaps you still have metal to metal contact in the important places so this does not count Moglice is one of the hardest epoxies so you are on the right track but you will find it does not bond well to the base metal and is quite easy to break away like on that edge you show a photo of, that will just come right off with a scraper behind it from the top down not that that would matter in your case as it is not part of the joint

[mactec54]

one more

Well done, did the heat treater not have blasting to clean it up after