Kb58's granite CNC router build

[kb58]

Picked up the fabricated granite slab that will become the gantry. Would have been disappointed if this had been an appearance item, as they didn't stack all three pieces evenly. They sawed three sides square but for some reason left the fourth untrimmed. That side will be placed downward where it's not visible, but it was a reminder that one has to really check over anything fabricated for important requirements.

Used roller guides, rails, and ball screws have been ordered off eBay but were still expensive. As I tend to do, it's a bit over-the-top, meaning heavy duty and precise, but I don't want to spend more in the future replacing poor quality parts. Still to be purchased are servos, power supplies, and controller, but that can happen later. A whole lot of holes have to first be drilled in the granite and I haven't decided how that's going to happen: handheld drill; buy a small drill press or farm the job out to a CNC business that can drill granite. I suspect that last one could be hard to find but it's worth a try if for no other reason than to know what they'd charge. That way, if I decide to drill them myself, I can at least keep reminding myself how much I"m saving, hah.

As mentioned previously, I'm pretty much building Piotr Fox Wysocki's YouTube build, but also incorporating the upgrades covered in his "Epic CNC Upgrade" series:

. I'm doing this even though his uses and mine may differ. He's running a business and uses this as such, so it's a deductible business expense, never mind that he's getting enough publicity from his videos that vendors donate some components. Me, I'm building it just because I want to, with no immediate source of income lined up once it's finished. Still, being able to machine aluminum - extremely well - does present opportunities. Cutting wood would be completely trivial as well, so who knows!

Before that all happens though, there's still a remodel to finish - I took the summer off because of the heat. Also, a fence between us and a neighbor needs replacing, and we'll probably build it ourselves. Then there's painting the house, installing new rain gutters, and on and on. Of course, if drilling the granite is farmed out, that could happen while I'm working on the house or yard, so there's that.

 

[kb58]

Ordering stuff from overseas always has higher shipping charges, but lately it seems like it has taken on a life of its own. Depending what's ordered, shipping can range from "free" to 100-200% on top of the item price. Then there are the "mistakes" made by sellers, where one of something might have $20 in shipping, but when you order five of the same, it's $100 shipping... no, no, no. Every time I bring this up with a seller, it's always, "Oh yes, we can reduce the total shipping." I suppose the platform they're using may not provide a shipping fee reduction factor based upon number ordered, but what's passed off as accidental or random, always seems to benefit the seller. Hmm.

In seller's defense, shipping fees are based upon both size and weight. If something heavy is ordered, ordering three may indeed cost nearly 3X as much. If it's a small light item, not so much, yet those are the items where they're trying to add shipping over and over for multiple pieces.

 

[kb58]

Parts are starting to arrive. I don't know what's going on at the seller's end, because their ship dates are disconnected from when stuff shows up. For example, one box "sent yesterday" showed up here the very next day. Looking at the tracking info, it showed that it had actually left their end a week earlier. Apparently their shipping department doesn't keep Sales updated, but I"m not complaining, as it's far better than stuff never showing up at all.

I've been looking at spindles and the Automatic Tool Changer (ATC) 2.2kW 220V units are really sweet. Having ATC is a necessity when running a business, where time is money, but as a hobby it's hard to justify. How hard? A high quality 2.2kw ATC spindle is around $1500. Buying the exact same spindle without the ATC option drops the price to $500 - ugh. While it would be "nice to have", that $1000 goes a long way toward buying cutters, collets, and nearly pays for upgrading from steppers to servos. As much as I'd like to just push a button to change cutters, it's too much.

 

[kb58]

The current plan is a moving-gantry design, but oriented in the "wide direction", meaning that the Y rails are on the short sides of the base, which is different than most. The reason given by the designer is that he doesn't have to reach into the machine as far, and since he has a lot more experience than I do, I figured, "okay", not feeling strongly one way or the other.

Watching other videos though, I wonder about a fixed gantry design. Advantages are (presumably) improved stiffness and less chatter, but with the consequence of the moving bed extending out beyond the machine's base both front and back, nearly doubling the machine's floor space, something I have very little of. Also, it means buying a large plate for the bed, probably 24" (600mm) square steel of 12-20mm thickness. A Blanchard-ground plate would be nice but doesn't guarantee anything other than it being flat - but not necessarily level once installed. A much cheaper hot-rolled plate could probably be used, machined flat and level in-place by the router itself.

I keep going back and forth on it. The indecision isn't helped by not knowing its uses, or even the size of the parts being machined; I doubt that it would be anything large. A big point in favor of the first design is the near-mirror-like finishes the designer is getting, so how much better do I really need, given that he's already getting micron levels of accuracy and repeatability? Here's a pretty impressive video on that

. All first-world problems.

 

[kb58]

The last two boxes successfully traveled from Korea to the FedEx hub in Memphis, Tennessee, labeled as Box #1 and #2 of 2. Unfortunately, only Box #1 is scheduled for delivery today, but for some reason, Box #2 was left in Memphis. It's supposed to be here tomorrow.

Box #1 carried the crazy-heavy 45mm rails, ones with square/sharp corners, perfect for ripping though the end of boxes. The shipper knows their stuff though, and packed them right, and even at 70lbs (30kg), came through well.

 

[kb58]

The big rails are here now, they're used, heavy duty units out of who knows what. The giant black 180mm roller blocks ride on 45mm(!) rails, and the green 160mm blocks ride on 35mm rails. I'm half way thinking about using only one black roller block at each end of the gantry, instead of the traditional two blocks. One reason is that it allows the spindle to get closer to the ends of the work surface. Another reason is that one of these big blocks is as long as two of the typically small blocks used on consumer machines. We'll see.

 

[kb58]

There is some thought that a single large block at each end supporting the gantry might be sufficient due to their size (180mm). It might work if the gantry is well balanced and the end support plates are very rigid - but how is that quantified and what's good enough; how stiff does the machine have to be to make it "acceptable?" Short answer is that I don't know. This suggests building it with one block and seeing if stiffness is sufficient (definition pending), and if not, adding the second block. Trouble with that is it means making new aluminum gantry end pieces, which is a fairly laborious and expensive process due to the heavy aluminum required. It all depends on what's deemed good enough.

A possible compromise: The Y axis ballscrew sub-assemblies bolt to the sides of the granite side pieces, and the servos stick out beyond the granite surface plate by about 200mm. That means anything else sticking out the same amount won't increase floor space requirements. Once all the parts are on-hand, they can be mocked up or modeled to see what's what. The huge side rails are plenty long, and I'm only going to cut them once(!) after being really sure about the final lengths.

 

[kb58]

For something that won't be completed for quite a while, you may be wondering why buy the parts now. It's like building a car, where it may be hard or impossible to get good dimensions of all parts up front. I work best when I have parts in-hand and can physically measure and/or mock them up. A big first question is: "how high must the granite sides be?" It's not clear at this point how all the parts will stack up (and down), meaning there's the granite side block + rail + rail block + gantry end bracket + gantry rail offset - spindle offset = ? Yes, the spindle moves up and down via the Z carriage, but how high and how low should it go to be most versatile? I don't know and want the parts to tell me.

The granite sides are 12" as-is, and the rail/block assembly is about another 2", and that may be too tall, necessitating the spindle carriage to over extend. While it's not hard to cut down the granite side pieces, I only want to do that once - and after a lot of double checking!

 

[kb58]

My mind doesn't work well with CAD*. I picture things how I want them, maybe do a few sketches to determine critical dimensions, and that's about it. I'm torn on the efficiency of CAD because I have this idea stuck in my head that it takes longer than using my method above. With CAD, there's measuring every component, creating it in CAD, and then moving the parts around (I realize that more and more new parts include CAD files, but much less so for used parts.) My approach is to use the actual parts, sitting them out on the table and determining the best placement. Okay, let me finish before you start typing.

I know that by definition, using a CNC router requires CAD and CAM to accomplish anything - I have no choice. Right now though, I'm in a self-created Catch-22 situation of: "It'll take much longer - and be much more frustrating - to learn CAD and do this, than to just figure it out using the actual parts." I'm also aware that there is no deadline for this, no pressure to have it done by some date. I did follow a series of YT training videos on Alibre CAD and replicated the cylinder head used as the examples. That was several years ago though, and it brings up another issue. It seems like many (most?) people who love CAD learned it at work on someone else's dime, and usually use it everyday, making it trivial to use at home. For me though, I never used it at work and hardly ever use it at home other than the above training. Not using it for a few months means starting all over again, leading to renewed frustration at the wasted time.

And so there they are, my lame excuses about resisting CAD. Now that I'm retired, I realize I have even fewer excuses.

*Neither Gordon Murray nor Trevor Harris design with CAD, so obviously neither do I

 

[kb58]

The Teknic servos arrived. Which models to get was based upon the expected load, balanced against price. One way to make sure they'd work is to just buy the highest-power units, but the cost suggested giving the selections a bit more thought. The two Y-axis servos share the gantry load, reducing the torque requirement by half. They X servo drives only the carriage, so the same model was selected, but with a high-resolution encoder. The X axis uses a 10mm -pitch ballscrew, while everything else is 5mm. The fine pitch screws give reasonable positional accuracy, but was judged not enough with the 10mm. The Z servo is a smaller unit due to it moving the least amount of weight.

Different topic: I'm a little concerned about whether the glued-together granite gantry beam is flat and square, and how to correct it. I'd have bought a proper block of granite but it's simply not available around here and too expensive to ship in. Hopefully my concern is unfounded, but the place that did the work is a granite countertop fabrication shop, not a surface plate manufacturer. I asked that they ensure the assembly was "flat", but what that means to me and them are probably two different things. I'm thinking of doing something that might be considered a bad idea: setting it up on my mill and facing it with a diamond grinding tool normally used on cement or maybe countertops, which I already have. The issue is granite dust, but I'm thinking that if a vacuum hose is nearby, and it's kept wet, and plastic sheet to protect the mill, it might be okay.