[catahoula]

I guess not. I knew it would be discolored but had no idea of this much scale build up. I don't know if they have that capability but in hindsight I def wish I'd asked.

[catahoula]

I also remembered that drilling holes into the EG is not ideal so it will be machined as is

[catahoula]

Linear guides arrived yesterday- whew! All parts accounted for and to spec. I'm like a kid at christmas. Cleaned out the shop over the weekend so now the apartment is the inspection site Gotta keep it moving forward while looking for new space.

Notes from first assembly: BST Motion ballscrew end machining is good. Snug slip fit into the bearing blocks- perfect. Pulleys were also custom machined and fit well, though slightly looser than the bearings. One rail set was 3mm short, but that matters not a bit. Their "house brand" angular contact C5 BK blocks feel tight in a good way- I don't think I'll be needing to shim the bearings apart as i've seen mentioned with cheaper blocks, but we'll see with an indicator once they're loaded. These BK blocks do have set screws on the adjusting nut, and while I'd prefer a second locknut, it should be OK as long as I don't have to adjust them too many times. BF blocks have a slip fit that matches their purpose to simply float the end. Also feeling more and more confident in the ballnut selection, these feel like a high quality bicycle hub adjusted just a shade too tight. Consistent and smooth resistance, but light.

On the Hiwin guides, ZA preload is a bit heavier than I expected, but makes sense for the added rigidity. I'm counting on the whole drive system being rigid enough that that resistance won't affect fine movements, and I'm sure they'll loosen up with use. I went with H precision class rails and C precision class carriages as H carriages weren't available. The parallelism of the rail is most important for accuracy, in my thinking, anyways. Since the beginning of this thread (specifically with the adjustments to the gantry) I modified the rail selection: they are 20mm on Y and Z with H (long) carriages, and 25mm on X with C (standard length) carriages. The standard length X carriage bolt pattern lined up better with my Z plate design, and the larger size more than offsets the rigidity difference between standard and long.

It's well too early to say overall, but I'm thinking I haven't f*cked it up too much, yet...

[spumco]

The rug really ties it all together...

[catahoula]

All the dude ever wanted was his rug back, man

[catahoula]

Machined the columns and the Z-axis plate. The z-plate photo was before I finished the linear guide shoulder, rail mounting holes, and ballscrew mounting plane, so a lot more was done but it's not quite finished. I need a better way to hold the edge as this 10" wide plate on a 9" bridgeport clone was tough to work around, had to move clamps around a lot but got it to within about +/- .0005". The backside is as flat as the mill table, however flat that is. May have it surface ground now that it's pretty close, but I'll probably give it a whirl as is.

Another question for the hive mind, particularly to other Americans: Am I insane for considering attempting to have a full metric setup in the USA? Given that I make bicycle parts, which are generally speaking metric, my modeling is metric, and I'll mostly be using Datron tooling and self-made fixtures, it would make my workflow easier. But even with the internet, metric availability for the machining equipment here can be low, and for working with machine shops/suppliers/etc imperial is very much standard. What do you think? Bite the bullet and go imperial, or fight the good fight and attempt to keep it as metric as possible? Obviously there will always be quite a bit of crossover necessary, I'm mostly thinking about how hard I should try to obtain metric dial indicators, parallel sets, and other measuring and workholding tools.

[peteeng]

Hi Cat - what bike parts do you make? Peter

[mactec54]

Machined the columns and the Z-axis plate. The z-plate photo was before I finished the linear guide shoulder, rail mounting holes, and ballscrew mounting plane, so a lot more was done but it's not quite finished. I need a better way to hold the edge as this 10" wide plate on a 9" bridgeport clone was tough to work around, had to move clamps around a lot but got it to within about +/- .0005". The backside is as flat as the mill table, however flat that is. May have it surface ground now that it's pretty close, but I'll probably give it a whirl as is.

Another question for the hive mind, particularly to other Americans: Am I insane for considering attempting to have a full metric setup in the USA? Given that I make bicycle parts, which are generally speaking metric, my modeling is metric, and I'll mostly be using Datron tooling and self-made fixtures, it would make my workflow easier. But even with the internet, metric availability for the machining equipment here can be low, and for working with machine shops/suppliers/etc imperial is very much standard. What do you think? Bite the bullet and go imperial, or fight the good fight and attempt to keep it as metric as possible? Obviously there will always be quite a bit of crossover necessary, I'm mostly thinking about how hard I should try to obtain metric dial indicators, parallel sets, and other measuring and workholding tools.

It will make no difference if it is Metric or Inch ( Imperial ) so if everything is Metric Machine and control, and you want or need to cut part that is drawn in Inch then it is just in the programing of the part G20 is Inch G21 is Metric your machine won't know any difference of what you are making so if you are confident in using Metric then go with it

[pippin88]

Doesn't really matter what units you use.

Indicators are mainly used for comparison not absolute measurement.


Imperial end mills are often useful for making metric parts.
E.g. a 3/8" for a 5mm corner radius pocket (rthaer than using a 10mm endmill which tend to chatter in the corner)

[catahoula]

Ok, I understand that it makes no difference to the machine, I'm just thinking about setting myself up for spending less time multiplying and dividing by 25.4. Maybe I justwon't worry about it.

Pete, I design and make an ergonomic bike saddle. Been chipping away at sorting out in house manufacturing for a few years now.

[catahoula]

Updated the model to reflect as-fabricated condition. Ran an FEA just for kicks with the full epoxy granite fill and the spindle box, and it's essentially the same as before, 6.4 um deflection @ 300N load, with only the center of table restrained. Of course, joint conditions do not reflect bolting etc and the linear guides are modeled solid and bonded, so real world will be lower. But 46 N/um is a comfortable starting point; I could lose a lot and be totally fine. The ZA preload linear guide carriages are rated at 500-600N/um radial stiffness, the bearing spacing is good, and bolted joints are spread wide with many fasteners, so I'm optimistic about performance. Spindle with ER20 collet is clearly the weak point- the worst part of the frame is only deflecting 2 um. I have the spindle modeled as a 5mm wall steel tube (85mm dia) with a 24mm steel shaft to represent the collet, for a loose approximation. A big fat 110 mm ATC spindle with ER32 collet would probably help quite a bit here. But for now I gotta KISS and stick with the 2.2kw 85mm spindle, so I don't totally lose my mind.

I would definitely do a few things different if I did it again, most prominently make the base ~30% deeper to guarantee stability during machining of the rail seats. Also a little adding frame and linear guide length to all axes. But overall coming together. The base was a spot where I got too excited about FEA- there's more considerations than specific performance in the cut and I optimized a little too far. The rails hanging over the front are just for extra travel to leave room for a future ATC, not used in the working envelope.

Specs:
X work area: 580 mm (600 travel)
Y work area: 460 mm (580 travel)
Z work area: 200 mm
Weight: 320 kg steel frame, complete machine probably twice that
Guides and screws:
X: HGR25CA-ZA, TBI OFUR2005 P1
Y: HGR20HA-ZA, TBI OFUR2005 P1
Z: HGR20HA-ZA, TBI OFUR1605 P1

Controller will be EdingCNC from CNC4PC with DMM DYN4 400w motors and drivers. Going DYN4 as per advice here so can switch to other drives without changing voltage.

[catahoula]

X deflection is almost identical, 6.7 um so 45 N/um. Z comes in around 2.5 um, or 120 N/um. Z isn't bending the poor collet so much which is the main driver for the big difference. It'll be interesting to tug on the spindle bottom vs collet in real life and see the difference. These numbers are really good and I'm not expecting that to hold.

[pippin88]

Watching with interest

Nice build!

[peteeng]

Hi Cat - Most FE systems you can put in a load or a forced displacement. If you put in a forced displacement it will tell you how much force was needed to move that far so if you move the tool 0.001mm you will get the stiffness in N/um directly.... saves doing the math. Your build is coming along very nicely. Peter

[catahoula]

Good tip, thanks. Quick question- anyone know about hardware sizes for mounting BK/BF style blocks? Been scouring the web and coming up empty handed. For a BK15 block, holes are 6.6mm dia- which means 4x M6 bolts will leave very little room for adjustment. But go down to M5 and that's quite a bit of slop. Holes will be positioned with a DRO on a mill so should be pretty accurate, but if not.... Thinking I'll start with M5 and some larger washers under the bolt heads. Any advice?

[jackjr-123]

You are overthinking this
Use M6 bolts. Drilling holes within 0.6mm of one another is not an issue at all, even manually.

[catahoula]

You are overthinking this
Use M6 bolts. Drilling holes within 0.6mm of one another is not an issue at all, even manually.

My specialty
The linear guides, which are meant to have adjustment, have 7mm holes for M6 bolts, so that was the context. But those holes have to line up over a long distance, whereas the blocks are mounted independently. Thanks.

[spumco]

Stick with the M6's. If you can't get enough play for alignment you can ream the bearing block holes out a little... and the M6's will hold better than M5's.

Since it seems like you have the same hole-locating abilities I do, you know wallerin' out an existing hole a wee bit is standard practice.

[catahoula]

Since it seems like you have the same hole-locating abilities I do, you know wallerin' out an existing hole a wee bit is standard practice.

Yeah, I mean half the reason to have a CNC is so you don't have to locate holes anymore

Tnx

[spumco]

Yeah, I mean half the reason to have a CNC is so you don't have to locate holes anymore

Tnx

You should see my collection of reamers, burrs, round files, and everything else needed for 'precision' hole locating. If there's a scale where a Moore jig borer is at one end... I'm at the other.

Seriously, back to the task at hand... something to consider are tapered dowl pins. If you have lots of fastener clearance and are concerned that friction between components may not be enough to keep things precisely aligned, taper pins installed after alignment are a great addition.

You mount and align the ballscrew however you can (i.e. per manufacturer, or hillbilly, your choice). Now drive the axis back and forth with the screw and check for binding. Assuming it's happy, drill two additional holes through each block and in to the mounting surface. Ream them open with a taper reamer and drive in tapered pins. The pins take up all (more or less) lateral/axial loads and the fasteners just have to keep things clamped down. You're no longer relying on friction alone to keep everything located.

If you can't do a through-hole, you can drill & tap (or thread) the end of the dowel for later extraction. Or buy them drilled & tapped. Metric taper pin reamers are painfully expensive, but 'Murican reamers are pretty cheap and the pins are very inexpensive as they're usually not hardened.

This also permits later disassembly & reassembly in exactly the same alignment - something round pins can't really do unless they're a press fit. You just reverse the order of installation - clean everything, pins first to align, and then torque fasteners. The pins keep the fastener tightening process from moving the blocks around.

Just did this on my Z axis ball screw and it worked out surprisingly well. A 1/4" drill and some #5 pins went smooth as silk.

-R