As you may have guessed from the title, I'm looking for recommendations for a suitable material to use for an extremely temporary spindle mount on a new short-gantry Lead Machine. Criteria: Soft enough to machine using a cheap, underpowered, runoutariffic Harbor Freight trim router as a spindle - so domestic hardwoods are ok, but aluminum ain't. Strong enough to hold an 80x188mm, 4kg water-cooled spindle while cutting .5" 6061 aluminum... ...but not for very long. It only needs to last long enough to cut some proper mounts, and then it's going in the trash and/or scrap pile Cheap. See "going in the trash" above. Ideally no more than 3/4" thick, due to endmill length constraints. Local availability is a plus - something I can run out and grab at a big orange box, hardware store, lumber mill, etc, is preferable to Amazon, and Amazon (Prime, specifically) is vastly preferable to anything slower. McMaster-Carr would probably be a close tie with Prime, if it weren't a weekend. Because the spindle's so **** big (and because I'm using .5" aluminum instead of .85"), my plan is to use a pair of mounts - one by the the bottom of the axis, another towards the top - rather than the single Openbuilds-style (for the temporary, I'm even considering a set of three mounts). The final ones will be similar in design to the Openbuilds ones, while the temporary version could be made significantly thicker around the three outside edges. I'm not planning anything overly aggressive - I fully expect I'll be babying it all the way through (extremely shallow passes, low acceleration, and chasing the tool around with compressed air) - but even the with the gentlest passes, cutting aluminum's not cutting MDF. At the moment I'm thinking about trying to get away with some 3/4" oak that's sitting in the scrap bin, but I'd appreciate any feedback - especially from anyone experienced in cutting aluminum on Openbuilds machines, and the stresses involved. I'm also open to tips & tool recommendations for machining aluminum, too (my previous machine struggled with even engraving it), but first things first. And probably second things too. I'm not sure which of those is the mount, and which is wiring up the VFD. -Bats (maybe I should've put the mill together before the CNC... oops.)
Check out Acetal/Delrin. I use it a lot as a temporary model material when making bits and pieces for CNC designs. It is strong without flex when used within sensible expectations. As a Spindle support, I would look into maybe beefing up the model to hedge your bets for the short time it will be needed. It is an excellent material and can be found on FleaBay as small offcuts and the such.
Could do it like this, temporary mounts .. really didn't see any improvement with aluminum vs the hose clamp mounts.. the black parts are hdpe but could be plywood or oak cut on bandsaw. Later I used this setup and machined 1/2" aluminum mounts. Cheers Gary
Delrin was my first thought too, but it failed both the 'cheap' and 'local' tests... The best prices I could find (domestically) looked like it would be $25-30 for that much of the stuff (a 1.25" 5"x6" block that I could hopefully bandsaw in half to make two >.5" blanks, which is still a little thinner than I'd hoped for), and take at least a week or two to get here. Any thoughts on HDPE? It obviously doesn't compare to Delrin, but cutting boards are a lot easier to source. That was the idea. Widening the mounts starts cutting into travel, which the big spindle already eats into, and which was ridiculously tight to begin with (only something like 9" travel, due to the fact that bats can't math), but that's an easy sacrifice to make, since it only needs to work once, and since the final mount shouldn't require much travel at all... as long as I can get big enough chunks of material to cut it out of. -Bats (hrm... if Delrin works, and HDPE is kinda like a wimpy Delrin, and HDPE is used to make cutting boards, and stuff also gets cut on vinyl mats, and records are made of vinyl... *looks up "lounge LPs" on ebay*)
Now that's something I didn't consider... And I really should have, considering my first inclination was to use hose clamps to hold the HF router on in the first place. My only worry would be that, while your Makita flares out at the top & provides a nice lip to sit securely on top of the hose clamp, my spindle is a smooth cylinder all the way up and weighs more than twice as much... although I suppose lining the hose clamps with something soft & grippier (grippyer? grabby?) probably wouldn't introduce too much flex to be tolerated. Either way, I think it's something I've got all the parts for on hand, so I might as well toss something together & see how it feels. Thanks for the idea! -Bats (and by "see how it feels" I naturally mean "strap it to the machine, crank it to full power, and see if I end up with an endmill embedded any inconvenient body parts")
Two-hole conduit clamps or pipe strapping with long enough bolts to not bottom out and get a really good torque on it. Rear supported by poplar turned with an ID of the spindle's OD.
The router is not resting on the lip at the end of the router, the clamping force is very secure. I put heat shrink around the hose clamps to keep the stainless from scratching the aluminum. Cheers Gary
Forgot you were using this to hold the HF router, hdpe should work for that. If cutting boards aren't thick enough and you want to be ultra thrifty, lookup some YouTube videos about melting milk jugs. It only takes a toaster oven and a mold, then you can make the blank whatever size you want.
Nope - that would be too easy. The HF router is on there right now, being shimmed all to hell in the OB router mount. It's being used to cut this temporary mount, but there's not a snowflake's chance in [questionable political analogy] that it could ever cut the aluminum for the final one. The temporary mount has to hold this beast: ...which is slightly larger than the trim router: The thought actually had occurred to me (it's an idea I've been toying with for not-quite-wood turning for a while now), but, aside from the whole melting/molding/pressing process being a bigger tangent than I really want to take from the build right now, I came to the conclusion it'd take a lot more milk jugs than I have on hand. -Bats (I may be a packbat, but... ok,I'm definitely a packbat, but empty milk jugs are one thing I don't have stockpiled. scavenged thermoform blister packs, on the other hand?)
Why nor just buy the spindle mount that fits the OD of the spindle. They are cheap. My Makita is 65mm so I bought a 65mm spindle mount from China. 3D printed shims did not seem ideal to me. Not all of them sold have predrilled holes for mounting on a plate, but the one linked appears to have the holes for mounting set at 100mm which is what mine was.
I considered it when I ordered the spindle, but, aside from seeing generally poor reviews for the things, they're about twice as wide as the C-beam and add another kilo to the weight on the gantry. I haven't entirely ruled out ordering one & making an adapter plate to mount it to the C-beam, but that's at least another $20 and another week of waiting that I'm hoping to avoid for something that'd still only be temporary while I cut a mount that was a better fit to the machine. Of course, at the rate I'm botching the soldering on the spindle connector, a week might not be long enough. -Bats (and why buy a miserable poor quality spindle mount when I can make a miserable poor quality spindle mount?)
I see... The lead machine's spindle mounts direct to the cbeam. I was thinking it mounted to a plate that went up and down the cbeam.
Thank you again, Gary, you're a genius. Have I mentioned that I want to incorporate your genetic material when I clone my army of genetically engineered super-bats to take over the universe? The real test will come tomorrow, when I actually try to stick it on the actual machine, but so far it looks like the idea is a huge win. I tried it out on a spare length of C-beam tonight, and it's not going anywhere. Not "it seems pretty stable and and it probably won't come flying loose and send angry spinning death dancing around the room" like the shimmed HF trim router in the Openbuilds mount. Not "yes, that's a fairly confidence-inspiring connection" like the Openbuilds mount presumably would be with a router it was actually designed to handle. This felt like it might as well have been part of the C-beam - like a really huge, bulbous, manufacturing error that for no apparent reason is composed of a random assortment of entirely different materials. Including wires. And water. Or like it was just really, really securely attached. ( woo! check out that baby in the background - she's got huuuuuge tracks o' T! ) I must have at least a couple hundred feet of heat shrink in a dozen different sizes... but none that would fit the hose clamps. So I used masking tape. It prevents scratches, but unfortunately doesn't provide much grip. I still couldn't get it to twist or slide (or shout), but, since the spindle can generate a lot more torque than I can, I'm thinking about possible lining it with something like a slice of bike inner tube The version I've got on there right now is just out of 3/4" MDF, since I didn't want to waste anything better (ie, pretty much anything) until I was was sure the shape was going to work right... but it occurs to me that maybe it doesn't need to be made out of anything better. It's essentially just padding between the spindle, C-beam, hose clamps, and the vertical bolts that hold the corner brackets together. The C-beam and hose clamps keep it all constrained, so there's no chance of anything breaking off - the only worry would be that over time it might compress and squish out of the gaps, the way tiramisu squishes out between your fingers when you slowly crush it in your fist. That's not to say I want to stick with this long enough to test the accuracy of the tiramisu analogy - the idea of using MDF as anything more than a spoilboard on this machine makes me feel even dirtier than that nasty little HF trim router already did. It does, however, have me rethinking how the permanent mount should work, and whether a scaled up Openbuilds style is really the right way to go. One big advantage to the hoseclamp design is that the mounted spindle is no wider (or no more than twice the thickness of the hose clamp) than the rail it's mounted on, while the Openbuilds design would inherently hang over the sides and result in lost travel. -Bats ( tl;dnr : "It works! Probably. I think. Better duck." )
Looks fine for now? I mean, it's concerning in that all that's holding the thing on is those four puny screws and the only thing stopping it from getting violent enough to obliterate them is a piece of glorified cardboard, but other than that, it looks great! Get it mounted up, cut out the same arch-bridge shapes in 1/2" aluminum, everything's sorted!
That's two pieces of glorified cardboard, thank you very much. But, yeah, the fact that it all rides on those four long screws worries me almost as much as the cardboard. Short M5 screws feel nice and stocky, but at 35mm they start looking a lot skinnier. Ok, so maybe there are more reasons than just my deeply offended aesthetic sensibilities to upgrade to something sturdier for the long term. Maybe. Perfect! Then all I have to do is figure out how to anodize it to match the rest of the machine. Easy! Oh, wait, no... I never did quite get the hang of anodizing... I also still need to do a little tweaking to make sure I can use the LED ring, as the hole is (not surprisingly) too narrow to fit around the body of the spindle. -Bats (I know! Maybe I'll just sand the black off the rest of the machine to match the mounts!)
Having just fired it up for the first time & jogged a few zigzags across the glorified cardb MDF still clamped to the table, all I can says is "Wow". Well, actually I can say quite a lot of things... but the vast majority of them could be reduced down to "Wow". ( don't mind the tattered and torn edges of the failed cuts from the old "spindle" ) Suddenly instead of feeling like an out of tune old power tool shrieking away on top of a rube goldberg contraption of aluminum bits & just waiting for a momentary lapse in focus to destroy both the bit any workpiece, it feels like one solid machine, powerful but completely under control. I'm sure some of that can be attributed to the near silence (the fans on the PSU & VFD are louder than the tool at idle, and only slightly quieter while it's cutting) and lack of howling wind blowing clouds of dust into the air, and the fact that I was running at a sedate 10krpm instead of the trim router's 35k doesn't hurt either (10k just felt about right for jogging around on MDF - it can run up to 24k, and doesn't seem to lose torque until down around 300rpm), but the entire feel of the machine is different. The different in power is amazing, too - pushing a high feed with a low spindle speed rips off messy layers of MDF, but not only does it not bog down, it doesn't even flinch. So, yeah, again: Wow. Of course, the Newton's Third Law of Awesome says things can't be this great without an equal and opposite minor annoyance, and there is still one problem that I'm not precisely sure how to solve yet - 'precision' being the key, as it happens... Ignoring the horseshoe and focusing on the surface (or as much as my phone was willing to focus) you can see fairly dramatic peaks & valleys between the passes, which I've generally found to be a sign of a crookedly mounted spindle (or... erm... occasionally a crookedly mounted Z-axis). It's not terrible, considering I've so far made zero effort to get it aligned, but it's a long way from ideal. In the past, my usual practice was to chuck up a 4" long 1/4" end mill or drill blank (depending on which was available and/or less bent) and then wail on the axis with a fine adjustment sledgehammer carefully tweak it until it visually aligned with a try square/setup block on the spoilboard, and then eventually give up after remembering that the X-Y table had a nasty habit of shifting at various points in its travel, making the whole exercise futile. I'm interested to hear how other people manage it, though - especially how you manage it with any degree of precision. Also, how you (Gary) tweak the alignment with a dual-hoseclamp (or dual-mount in general) setup - just loosen the top mount, wiggle it around and try to get it tight again when everything lines up properly, then swear & start over when it squirms out of the way? Also, am I the only one who's constantly frustrated by the complete lack of available ferrous metal on these machines when it comes time to mount an indicator? -Bats ( maybe if I pour molten iron into the slots in the v-rail... )
Personally I'd probably use M8, but I do like to over-build things, for the most part. Seems like the less flex in the screws you have, the more precise your z axis is going to be, as well. Uh-oh, then I guess you're gonna have to paint the angle brackets and the spindle body itself, too! Having cracked MDF under almost purely compressive load before, and this being only superficially compressive, but realistically heavily shear, load, with two horses trying to pull on it... I'd definitely be changing it out. The pipe clamp is pulling it "through" the screw against the rounded body of the spindle, which then applies even more force in rapidly changing and completely uneven directions. It's begging to rip that MDF in half using the screw as leverage. Even a simple round aluminum spacer around the screws to seat the spindle against would, I suspect, be a better option in terms of both rigidity and ultimate strength. But that's just me! From the scalloping but general lack of striations, it looks like it's straight in the B axis but tilted in the A axis. Could be 1) spindle (probably), 2) spindle mount rigidity (probably), 3) overall z axis rigidity (maybe), 4) tilted x axis rail (perhaps, depending on mounting), or 5) your side columns/y-axis wheels (probably not) This is how I trammed my z axis, but it needs something that'll act as a surface plate (or something you can shim square to your machine so that it can be a reference for your z): For the other axes, I clamped a piece of cold roll to the V-Slot so the mag base could mount: With a piece of alu sheet on the other side to protect the rail. Worked a treat overall, my entire machine is +/-0.002" over entire travel distance on x and y, and +/-0.0005 over the 2" or whatever on z. Plenty good for a laser cutter, so that's where I stopped. (Also check out that sexy black and silver colour scheme.... Just sayin' )
I would've liked to go with something a bit beefier, but I was limited by what I had on hand that was long enough, and would fit in the cast corners. I'm still hoping to pick up something larger for the final version, but M5s were the largest that would fit the corners without widening the holes (slots, actually), and it looks a little dicey as to whether there'd even be room for the heads of M8s. It looks like M8 heads are supposed to be 12.7-13mm dia and the model shows exactly 13mm space in the cast corners... but those corners don't seem to be made to very tight tolerances & I'm less than enamored with the thought of filing out the inside edges. Well, the spindle body is aluminum, so I can just anodize that too... Oh... right... anodizing... I guess I did a pretty good job accidentally anodizing titanium... maybe if I just replace all the Al with Ti? Maybe I just decide to like the deliberate contrast between the silver spindle and the probably-not-actually-black mounts? Maybe if I just turn off all the lights, and... how does the old adage go? "At night all aluminum extrusion-based light machine tools are anodized grey?" Those four long screws & the material they're running through definitely seem to be a weak point, but, while I'm not much of an engineer, I'm starting to think things may actually a little more complicated. If the hose clamps held the spindle to the MDF mounts, and the screws held the mounts to the rail, then yes, you've got two horses worth of spindle torque plus, potentially, X&Y movement working to rip the screws. In this case, though, the hose clamps are acting to hold the mount together, with the screws inside it, and the screws are holding not just the mounts to the rail, but the mounts and hose clamps. The more we talk about this, the more I wish I'd spent more time learning to use (and/or understand) Fusion's simulation & stress modeling... Yep, the issue definitely seems to be A and not B, as the ridges don't appear if the passes are running on the Y axis. My suspicion was primarily #1. #2 & #3 seemed unlikely, since, while neither is especially rigid, these were slow (or at least leisurely) passes through maybe .05" of MDF - practically cutting air. I hadn't given much thought to #4 or #5, but it seems like either one would cause a smooth surface with a varying thickness, unless the spindle was also misaligned with the axis... wouldn't it? I'll have to play with it a bit... it's that "something square to the machine" that may be a little tricky. I don't really trust the MDF spoilboard to be square to the machine without surfacing it (which, obviously, is difficult here)... but I suppose if I yank the spoilboard off & tram it relative to the frame underneath, it should at least give me a decent starting point. (you also seem to have a variety of attachments that I don't recognize on your indicator, but they don't seem to be relevant to to the task at hand) Showoff Ok, black & silver it is! -Bats ( I suppose I'd better like black & silver, since those distinctly-not-aluminum hose clamps wouldn't take very well to anodizing anyhow )
A hand drill and an angle grinder would make short work of those screw heads (I use this technique for bevelling, diameter reducing, radiusing... You name it). And you could get away with some wider straps too, if you wanted, because you could use longer screws. Ti seems easy (blue, green and orange machines?!), Al seems a lot of dying, which I have minimal interest in. Although I think hard anodising might be a little nicer, albeit longer to do. I'll have to remember that one for the ol' Space Tombstone! The compressive support of the strap would definitely help in the case of catastrophic failure, but you've gotta remember that the spindle is mounted to- and inflicting forces upon- the screws, not the MDF. Imagine how the spindle would move if it would magically float in mid air with the pipe clamp magically holding its trilobed shape- ie. the MDF is empty space. Now as the spindle moves, it pulls the clamp through the air toward the screw, and opens up new space on the other side of the screw. Except now remember that as you pull the strap toward the screw, it wants to flex and wrap around the screw, "pinching" the MDF around the razors of the screw thread and opening up space on the inside to eliminate the compressive support it had been receiving. It's almost certainly not long for this world, but if it lets you cut an aluminum duplicate of itself with a reasonable DoC, then it's done its job heroically. Spend 30 minutes just messing about with random shapes and it all comes together surprisingly easily. You can get meaningful numbers out of it in a very short space of time, I was shocked (I'd also heard the stories about FEA...). One mounting bracket being closer to the viewer from the front is definitely highly likely- it doesn't take much angle to get scalloping, though it does seem like a lot of scalloping for the size of tool, which lead me to consider the other options as well. Mount rigidity is a possibility inasmuch as Y+ movements apply a torque moment, likely flexing the hose clamps to some extent, but with such a light cut, that seems less likely now. Axis rigidity would fall into the same boat here too, you're right. Seems like perhaps it actually is 4 or 5 (or 3.5, actual z axis mounting, if it's possible to mount it tilted about A). It's an odd one, given that it seems to be a dramatic enough error that simple visual inspection should be enough to diagnose it, but I'm assuming there's nothing wildly obvious that's the issue. I suppose, if you could rely on the peaks being consistently the same height relative to an arbitrary X-Y plane, you could just surface the spoilboard, but otherwise, yeah, it's time to dig out the shimstock. Not sure about attachments, it's just a test indicator held onto the machine with or without a mag base using a Kant-Twist clamp. Very simple setup (though for that z setup, you have to be very careful about how hard you squeeze the indicator in that direction... The innards lock up if you flex the sidewalls too much!). The z is against a 1-2-3 block on a granite 12x18 surface plate (there's a reason I built the laser cutter that size...). A little. Though technically I can't back up those numbers, since I don't know the natural variation in the surface of the V-slot, nor the squareness of the 1-2-3 block. It's possible that they just so happened to fluctuate in the same direction as the machine's motion, and it's out, like, 20 thou! Metrology is a pain. Victory!
Oops... Twice, oops, even: Oops One oops : It occurred to me that I have no idea whether we're even talking about the same axes here, as my machine is both narrow and installed sideways - X being my long axis, with a short gantry as Y. The ridges appear when the passes are parallel to the X (long, stationary) axis. When running parallel to the Y (short, gantry) axis, they either don't appear, or aren't prominent enough to notice on the fuzzy surface of milled MDF. So, as I see it, an A axis twist would suggest that upper & lower spindle mounts are misaligned on the Z rail. Oops Two oops:): It took until I walked over to the machine to realize that this works for tramming the Z axis, but not for aligning the spindle. Considering it would be much easier for the spindle to have been mounted crooked than for the Z (or Y) axis to have been built crooked, I feel like that's probably where I should focus for the moment. Putting too much effort into tramming the rest of the machine is probably premature optimization - especially if I'm still thinking about rebuilding it with a longer Y (and with the #$%^&%^ gantry eccentrics mounted on top where they can be adjusted after everything's built without elevating the whole. ****. machine. Thanks, Mark). -Bats ( now maybe if I zip tie a setup block to the side of the spindle to stick a magnetic indicator mount... don't zip ties just scream "precision"? )
But that... that... that's.... sensible! Or at least that's the only reason I can come up with that it never occurred to me to try it. Or even to stick it in the lathe. Well, not really... The 6061 plate I have is only .5" thick, which is more of a limiting factor than the screw length. Finding wider hose clamps looks like a challenge, too - even McMaster doesn't offer anything over .625". Also, funny story about screw diameter & length: Last time I went looking for 35mm metric screws at the Ace Hardware down the road, I discovered that they stock them in M4, but for M5 and larger they only carry up to 30mm. Length was more important than diameter for whatever I was working on, so I figured, "fine, I'll just go down to M4 and pick up some new nuts to match"... except that they only carry metric nuts for sizes M5 and up. "Ace is the place" indeed. Ti is easy, and turns all sorts of pretty colors - you can get nice deep shades, too. I learned this by accident when all the Ti I was using to support my Al in the bath came out looking beautiful, while my Al failed to do much of anything. And I was trying for colored anodizing on the Al, but mostly because I didn't have any other decent indicator as to whether or not it was working... which, of course, means that I don't necessarily know whether I was failing at anodizing, or just at the dying. I think I ruled out hard anodizing at the time because it was going to be significantly more involved process - it looked like it would need constantly circulated, refrigerated bath, and probably more amperage than the PSU I was using could put out. Space tombstone? Is that where you mount your stock on a rocket-powered horizontal mill? I hadn't even considered the threads cutting into the MDF - which the strap tension would be encouraging even without any torque from the spindle... But, either way, yeah, I think I'm going to be glad once the MDF is out of the picture - and I don't much care if the DoC I need to use for the aluminum isn't reasonable. This alignment issue is bugging me, though, and I keep wishing I could come up with some way to make the final design either easier to fine-adjust after mounting, or less likely to be misaligned in the first place. All I could come up with was making the back edge slot into the C-beam, but A) it would mean that instead of a continuous strap wrapping around the spindle & mount, the straps would have to screw into the sides of the mount, which seems like a significantly weaker design, and B) I don't know how much it would even help. Maybe I'll give it another shot when I get fed up with tweaking the spindle alignment. I've already got the spindle & mount in there, which are straightforward enough, but trying to model a strap under tension might take some fiddling. Not sure I understand what you're describing... You mean one mount being farther out from the Z rail than the other (as might happen if the strap on one was pressed a little less flat between mount & rail), or both being flat on the rail, but one skewed to the side of the other? I've been assuming the latter, since the former would manifest ridges/scallops between passes on the other axis. Right. If this were happening this much with a big >1" tool, I wouldn't worry so much (at least not until it came time to surface something with it... which, of course, is probably the only reason I'd have had it chucked in the first place, so... err... nevermind...), but on a .25" end mill, it's a bit dramatic. Exactly. If this were hogging, a hard material, an aggressive feed, or anything other than a glorified skim cut on less-glorified cardboard, rigidity would be a likely culprit. But for it to have this dramatic an effect under the circumstances, Openbuilds would have to have replaced my regular aluminum extrusion with rich, delicious Folgers wet noodles (let's see if I notice!). The ridges aren't particularly deep - probably less than half a degree skew between the mounts could cause it, and it's hard to visually align the spindle too much more accurately than that without being able to sight straight down the X axis - which is made more challenging than it probably should be by the drill press that so inconsiderately decided to stand right at the far end of the machine. Did I mention things are a little cramped here? Just like scraping in ways or surface plates - it's all about the peaks averaging out and... err... hoping that whatever workpiece you clamp down doesn't fall between the much-wider-than-a-scraped-surface-and-inconveniently-parallel peaks. (and... err... not using cardboard for your surface plate or ways) The clamp's what I was looking at - apparently I haven't spent enough time staring at them from the backside to recognize it. And to think - they say that's all some guys look at. I think I've got some spare granite floor tiles... that's almost the same thing, right? Right? Rob....? Hello...? What's that sound... is that laughing or crying? A proper reference surface is something that's made it close enough to the top of the list that I've ended up shopping for one a half dozen times, but never quite close enough to actually buy. So I have to settle for making things I generally just treat my 1-2-3 blocks as being ideal geometric solids, because if they're not actually square, at least they're square-er than anything else I have, and anything that's verifiably squarer than them has a pricetag that makes me want to cry. -Bats ( sniffle.... sniffle... I always cry at metrology )
So, in the course of this conversation (literally! I've been typing this all with my toes!) I have made some progress. I bolted a 123 block to the gantry, stuck a dial indicator to it, and indicated on the side of the spindle (on the probably vain assumption that the outside of the spindle would be at least somewhat close to parallel to the chuck) while running the Z-axis up & down. This didn't occur to me before because I'm a bat of very little brain. The mounts were indeed misaligned - showing at least a .03" difference between the top & bottom limits of travel. With a certain amount of struggling and swearing I managed to get that down to ~.001" - still not good, but at least when I cut the mount I don't have to worry about the sides being dramatically sloped. I only did this for the A axis misalignment - now I'm getting about the same surface finish along each axis. It's still not as smooth or level as I'd like, but, again, passable for a temporary mount that won't do any surfacing. With the final mount on, presumably the next step would be to repeat the process with the higher sensitivity of a dial test indicator (probably indicating on a drill blank in the collet instead of the outer surface), but that's where I run into a problem. The way the mounts are fastened to the Z axis (M5 screws through cast corners into drop-in T-nuts) is loose enough to leave plenty of room for adjustment, but it's also terrible for making any sort of fine adjustments. The difference in mount position between the point where it's loose enough to adjust and the point where the screw (just one, to hold it in place) is snugged down can be .005" or more - so even getting down to the current .001" between mounts was largely a matter of repeatedly loosening & retightening a screw, each time wondering "will it move the right way this time?". I imagine I should (once I have a mount with a little more structural integrity and a little less structural squish) be able to tweak B-axis misalignment fairly predictably by leaving the mounts in place and shimming between mount and spindle, but are shims practical for the A axis without a solid frame around the sides of the spindle? -Bats ( o/~ shim shimminy, shim shimminy, shim shim sher...line mini mills? o/~ )
Since you're taking pictures from, and presumably loading stock from, the right side short end, I've been using that as the de facto Y- "front" of the machine. I always use the open side as the front, it just works better in the grand scheme of things. So for me, the short rail is the X. Sure, not very aesthetically pleasing compared to a mill table, but more practical for the purposes of trouble-shooting. I mean, it's a difference of four wires either way, so... So when I say A axis, I mean twisted front-to-back around the X axis line. "Pitch", in aviation terminology. B would be roll, and indeed would be the culprit in your coordinate system because the spindle is mounted to your right- the spindle is swept in some form of arc through the vertical plane that runs parallel to the scallops. This is hard in words. Aligning the spindle to the z axis is much easier- just stick a magbase on a piece of steel plate and dump it on your spoilerboard. If the spindle is tilted, regardless of the angle of z, it'll appear to be a ramp to a stationary indicator. Yeah, at 8mm, throwing it on the lathe is more doable (small screws tend to just bend under all but the lightest cutting pressure, I've found). Doubling up the plate and the straps is perfectly doable for a larger clamping area though, if it's worthwhile (and it might be... Might not be, either). Yeah, that sounds about right. "Specialist" (metric) hardware locally is a nightmare. I realised the double meaning in this context after I posted it! Who doesn't want to be shoved inside a nice hardened, ground and tapped steel box forever?! I think it's a plan. I'd aim for adjustability over perfect alignment, personally. The reason my Z axis is on a single bearing block is so that that block could be shimmed against the gantry carriage plates if necessary (it came out perfectly, as it happens, but it's safer to assume that it won't). I'm not sure there is a non-bulky way of securely mounting a spindle like that, honestly. Ideally you need something that's a flat plate at the back, and then the straps catch onto some sort of ledge or screw on the sides of it, though the worm drive would still be obvious. What you could do, actually, now I think about it, is to absolutely solidly mount the spindle to the Z axis extrusion, and then turn precision spacers for the wheels that that extrusion is riding on. If you added a couple thou (are shim washers a thing on McMaster? I guess you could holepunch some feeler gauges) to the top wheels so they came outward a bit, or vice versa with the bottom wheels, you can adjust the angle of tilt about the A (your B) axis, without having to get fancy with anything else. Being skewed about the extrusion would result in hard triangular ridges in the long-axis direction of travel, and scallops in the short axis travel. Being tilted up from the extrusion would scallop. If you run a toolpath as it's set up now, you should get those sharp ridges (like a factory roof) running along your short axis (and they should tell you which direction the spindle's tilted, too). I actually have some granite floor tiles that I use for sanding and things (abrasive dust and surface plates, I'll never understand some people). They seem remarkably flat, might be good enough to at least get the machine somewhat settled. I should measure them against the plate sometime. I was at a point on some project or other that I was measuring parts near-constantly, and the lack of a reference surface started to significantly outweigh the $100 that plates that size can go for. Exactly. When I get a surface grinder, I'll have a conversation with these things. Meanwhile... The issue with that is that your A axis is parallel to the scallops, and therefore a twist about it shouldn't be able to affect them?! I mean, progress is good, though! Let's not look a gift horse in the mouth. (That phrase makes no sense to me) But yeah, the scallops should be caused by (to you) a B axis twist. I had that too. M5 screws are just bad for adjusting with. Maybe add washers at the relevant points so it's at least not grabbing, and you have more of a chance? I found that there's also a very precise point at which the screw is tight enough to tap in without flopping about, but not quite tight enough to be immobile. It's a very narrow range, though. Shimming against the spindle is theoretically only viable if the spindle itself is perfectly cylindrical. If so, go for it! If not, you're gonna need to scribe in some permanent reference alignment marks.
Aesthetics don't enter into it... Setting up the machine with the open end out would mean completely blocking the path from one end of the shop to the other (the whole reason I went with a narrower build to begin with), and keeping those axes after rotating the machine 90° into the current orientation would require also rotating my brain 90°, which would almost certainly break some of the important thinky wires that make it work. It would probably also be really messy. But now I know which way your description is oriented, which is what matters. We need better words. I think I'd somehow I'd been stuck on the idea of needing some sort of flat/square reference surface rather than just a fixed one - which is likely why I bolted the setup block to the gantry rather than just clamping on the chuck of slightly-bent CRS I initially grabbed & then discarded. I'm going to put it down to insufficient caffeination. True, but unless I'm overlooking something (which - based on recent history - is highly probable) I don't see much mechanical advantage to bolting up multiple thicknesses vs. using multiple single-thickness mounts spaced out. Although I suppose the former would require less cast corners. Titanium box. Can't anodize steel. Then again, if you do go with steel, you could have the entire surface hand-scraped, which would be pretty classy. That's not the axis I'm worried about - that one can be adjusted by shimming between the mounts & spindle. It's the other axis that I've been puzzling over. It looks like our words have yet again proven deficient. I was starting to suspect that your "scallops" and my "ridges" were describing entirely different phenomena, and it looks like they were. Now that we've achieved linguistic synchronization, the problem I was focused on is definitely ridges. I actually haven't noticed any scalloping in either direction of travel - if it is there (which it probably is), it'll likely require a material that gives a smoother finish and/or a better endmill. Surfacing MDF with a $2 1/4" 4-flute of uncertain vintage leaves a fuzzy enough surface to disguise a lot of detail. Actually, pretty much everything cutting MDF leaves a fuzzy surface. I may try cutting a couple limit switch mounting plates tonight to see how it works on Aluminum and/or not-cardboard, but after last night's almost-but-not-much success I think it's good enough for government and/or mount-cutting work right now - assuming nothing explodes and I can figure out a design that allows for my-A axis adjustability on the final one. I generally use glass when I need a flat surface for sanding, since the powers that be have expressed some vague threats about dire consequences if anything happens to the granite tiles and they then turn out to be needed, so at the moment they're only being used to raise up/weigh down the little wood lathe. Mind you, about the only thing I can see that might ever necessitate their use is if I drag several tons of cast iron machine tool in the front door & across the entryway, smashing tiles as I go, only to then face... the stairs! No, I can't see the tiles ever getting called back into official service... maybe I'll dig some out from under the lathe and see how flat they are. Was that before or after the shipping to transport a big ol' block-o-rock? (I'll leave the "granite", "outweigh", and postal scale puns to someone else... can't take all the low-hanging fruit) I know a guy with a surface grinder. One of these days maybe I'll go visit him and make absolutely everything I own absolutely flat! My setup blocks will be surface plates. My coffee mug will be a surface plate. MY BEAGLE WILL BE A SURFACE PLATE! HOLD STILL, YOU! Ridges, scallops... need better words. Stupid president has all the best words. *grumble* The big problem, from the beginning, was ridges, not scallops (I think?), being caused by a my-A axis twist. Based on the fact that I now seem to see more ridges from my-Y axis travel than I used to, I think I probably introduced some my-B twist in the process (although I may just not have tested it enough before to notice them), but it's significantly less dramatic in both directions than it used to be in X, so... yeah... hopefully good enough to get the job done. And the reason you don't look a gift horse in the mouth is because the Arabian Gift Horse (as well as the related European and American breeds) is known for having teeth that grow unevenly, making them a lousy reference surface for metrology and causing you to resent your horse for its inadequacies. I think the problem was less the M5 screws and more the combination of drop-in T-nuts with brackets that aren't fixed in position. As the drop-in nut tightens and rotates into place, it shifts around and drags the screw & corner bracket along with it. If the little bumps on the cast corners were perfectly fitted to the slots, it wouldn't be an issue (and, of course, the price of the corners would shift the decimal by a place or two). I've thought about trying to do something with eccentric nuts, such that the alignment could be fixed after the T-nuts have locked into the (inevitably wrong) position, but putting eccentrics between the corners & rails would mean sacrificing all the stability of the cast corners' nice large bearing surfaces and hanging the spindle on just eight skinny posts, there isn't much space for something like that inside the corners - never mind space to stick a wrench to adjust it. I suppose I could try sticking coupling nuts in a 4-jaw chuck and turning extra-long eccentrics, but it doesn't look like there's enough material there for much adjustment. Making some from hex stock might be an option, although the idea of hand-tapping that many deep holes makes me cringe a little. Good point. And I suspect the mounts will make enough marks of their own on the spindle that I'm not going to flinch at scribing a few of my own. -Bats ( most breeds of gift horse also have non-ferrous teeth, making their mouths difficult places to mount indicators )
That's what you get for buying mills and drill presses and things! In a machine shop?! Not so much about mechanical advantage as higher total clamping force at a lower clamping pressure- spreading out the force. Really depends on how thick those spindle walls are and how well they can take the clamping under load over a period of time. I have no idea. Drifting through space forever in a hand-scraped drilled and tapped box would be the bee's knees! I'm sure all kinds of things could be conveniently attached, too. This is true. But for the record, this is what the lighting in your photo is showing: Now whether that's what's ACTUALLY there, I can't know. Lighting can lie just as readily as Photoshop. But in whatever direction you get ridges, in the other direction you should get scallops. It's just geometry! (Really nasty 3D geometry with tilted circles, though) It's clearly incredibly unlikely that you'd ever do that, so tile away! That was after shipping, I believe. It's a very nice rock, for sure. These days it might be $150-200, but still quite inexpensive for a 12x18x3" A grade plate (because all those tree-necrophiliacs use 'em for sharpening chisels, it's a supply and demand thing... I think). It's not Starrett pink or anything, but it definitely does the job. This sounds exactly like what I'll do upon receipt of a surface grinder. I don't have a small furry creature to attempt grinding though, so I'll have to try and lure something in from the back yard. My house will have the flattest squirrels around! (I love the idea of coasters that are +/-0.00015 to +/-0.00005 though, I might have to remember that one) I think this is the important takeaway here, waffling and pontification on referential frames aside. I'm gonna need a new horse for the next project then, I suspect. Could you clamp the brackets in first? Or clamp the brackets to a 1-2-3 block that itself is clamped to the side of the V-Slot, so that the brackets are perfectly flush and square with the edge? MOAR CLAMPS
My feeling was that 3-4 separate .5" mounts would tend to offer more of an advantage than two bolted up 1" mounts. That feeling has absolutely no grounds in engineering or physics, though. You'll need dovetails! Lots and lots of dovetails! Ok, now I finally understand what you've been getting at (yeah, I know I said that last time... but this time I mean it!). I think those actually are more ridged than scalloped, but between the MDF's soft fuzzy edges and the camera (one of the reasons I stopped posting progress shots), I can see how they'd look otherwise. Either way, we'll know once I get something harder under there. I wanted to get some aluminum in there last night, but got distracted trying to get too mathy with my feeds & speeds, and then had a long text dump I've been needing to get off my chest that kept me busy past bedtime. Not a bad price, then... Although I've always suspected the need for a high-grade surface plate to sharpen chisels and plane irons comes less from necessity or any advantage over float glass and more from a nagging sense of inadequacy at the size of your sharpening setup's pricetag. Might I recommend a rented steam roller? Said guy-with-a-surface-grinder used to brag that every measurement on the garage he built was accurate to tenths. At least I think he was bragging... those might not have been tears of pride. Yes, a waffle iron will make the squirrels flatter, but I don't think they'd technically be fla... we changed topics while I was distracted by squirrels, didn't we? Probably. It won't even fit in the waffle iron. You'd think so, but I did align & tighten down the brackets to the rail first, since it was the only way I'm confident all the drop-in t-nuts have engaged. Even assuming corners themselves are reasonably square, though, the fit between the long bolts and the corners (between any of the screws and the corners) isn't what you'd call "precision". In the case of the temp setup, that was pretty much a wasted effort though, since - between the HF trim router, the 1/8" or so of shimming to fit it in the OB mount, and the squishy nature of MDF (and at least one getting burned in a fire started by said HF trim router) - the two mounts weren't reliably regular to begin with, and the most careful possible pre-alignment would inevitably have needed correction after the spindle went on. I'm hoping the aluminum version will avoid a lot of those problems, but I'm still not sure if I'll be able to trust aligning the outside of the mounts squarely when I put them on to square the spindle when it goes on. After all, when phrases like "only viable if the spindle itself is perfectly cylindrical" are serious parts of the conversation, mounts cut with an MDF-mounted spindle on a largely untested machine by an unskilled operator are going to be at least a tier lower in reliability. This is why I'm hung up on finding a way to design in adjustability. Any thoughts on the idea of some sort of eccentrics around the corner-to-rail mounting screws? Even though the holes in the corners are vertical slots, it would probably still result in some vertical movement for any horizontal adjustment, but, being spaced out as they are, it seems like the A/B axis alignment of the two mounts will have drastically more impact than a little potential my-A axial twist to an individual mount. It seems like a lot to fit into the available space in the corners though, and I suspect there's more that I'm overlooking, so I've been reluctant to just go and order a bunch of hex stock. I can't help but feel like adding an extra inch to the mounting screws (outside of the corners, mind you, and with a spacer taking up the slack) is a bad idea, but I can't come up with a reason why it's a bad idea, or any worse than an especially fat washer. -Bats ( there's also that unfortunate business with the 4-jaw chuck and its rather awkward lack of a backplate... but that's a minor inconvenience )
You've got too many variables. Squishy MDF mounts (which can split when the force is parallel to the top surface, I've had it happen in other projects as it's compressive strength is higher in the Z-axis than the X or Y axis due to the way it's made) and squishy rubber under the hose clamp mounts. Turf the rubber and accept that your working machine will get a few scratches on it, or polish them out when you're done if it bothers you. Or if you feel you're ready to cut aluminum the way it is, cut version 1 of your permanent mounts. Then use them to cut a final version 2 once you've figured out what changes you'd like to have after using version 1 (which seems to be the way in my shop, regardless of how much thinking and design planning I do). I'd personally try cutting Version 1 out of Acetal (Delrin). You mentioned it was expensive, but I've been able to find it cheaper in local shops that sell acrylic than online, especially if they have some small offcuts to get rid of. Or a piece of hardwood or furniture/marine solid core plywood if you can't find any reasonably priced Acetal.
What I would personally do, actually, is design two different mounts. One temporary set that look like the ones you have now, just to take out the slop and get you running without a lot of guesswork, and one permanent mount which is one single brick with a radiused channel running longways on one side and lots of shallow 1/2" slots for pipe clamps running shortways on the other with V-Slot-sized blocks sticking up from the surface for orientation: Easy to design and machine, and would permanently solve both the mounting and adjusting problem. (Just maybe not with the screw holes behind the spindle... That would get frustrating) I wouldn't add even more spacers and overhang and all that. Eccentrics would (probably) work, but at the expense of even more leverage on your gantry and wheels. The closer to your z axis the spindle axis is, the better. In an ideal world, your spindle would be inside your z axis and your z would be inside a dual gantry axis. But most of the time that's not necessary (when it is, usually the solution is just to go horizontal). I concur!
