X-Carve 49

[cg49me]

I have grand aspirations for continuing upgrades on my 1000 mm X-Carve, so I thought I’d start my own thread. I probably won’t be doing anything unique for awhile (if ever), but I do at least plan to document the process for anyone else who might follow.

Current Upgrades:

• 3GT3 9 mm Belt Upgrade Kit for X- and Y-Axes, from tbd cnc (https://tbdcnc.com/)
• Y-Axis 2" x 2" x 1/4" Aluminum Angle Stiffening Brackets, from metalremnantsinc via eBay (Metal Remnants Inc | eBay Stores)
• 3GT 9 mm 20-Tooth Pulleys for Z-Axis Motor and Threaded Rod, from OpenBuilds Part Store (https://openbuildspartstore.com/)
• 3GT3 9 mm 60-Tooth Closed Loop Belt for Z-Axis, from MiSUMi (https://us.misumi-ec.com/)

Planned Upgrades:

• Laser Zeroing System (need to figure out mounting)
• Y-Axis Riser End Plates (will fab my own, will have to replace stiffening brackets)
• Linear Z-Axis (after risers)

After running my X-Carve solely on the weekends for about six weeks and breaking three belts (Y, X, and Z) in that amount of time, the decision to beef them up was easy. I came across Luke’s ad in the Inventables FS&T sub-forum for the tbd cnc kit that comes with everything you need. I went poking around for all of the components myself just to compare, and the tbd cnc kit really is worth buying everything all in one place. Having now installed everything and run a few test carves (primarily for calibration), I would highly recommend the kit to anyone else seeking to perform the same upgrade. I even contacted Luke with a question shortly after placing my order, and he responded promptly and cordially.

I started with my bone-stock 1000 mm X-Carve, sans Z-Axis belt as I hadn’t yet installed the wider pulley for the threaded rod to accommodate the new 9 mm wide belt.

The kit arrived quickly and in great condition in a bubble-pack envelope. In addition to the base kit, I’d also ordered a spare set of belts.

The belts were already cut to length for each axis, and had plenty of extra length. I ended up cutting about 6" off of each one after getting them installed.

For some comparison, the old (left) and new (right) belt clips…

…the old (left/top) and new (right/bottom) belts…

…the old (right - whoops, consistency) and new (left) idlers…

…and (what remained of) the old (right) and new (left) pulleys…

First step was removal of the old belts. In doing so, I discovered a spot in the middle of one of my Y belts with goobered-up teeth. I’ve seen plenty of these type of pics from other members, so I just chalked it up to being part of the club.

I then removed the stepper motors and idlers, and cut the old pulleys off of the motor shafts. I first tried to “punch” the pulleys off using a couple of 2x4’s and a pair of thin metal bars (there’s a video of someone getting them to come off that way), but wasn’t having any success, so I resorted to the Dremel.

I started to re-assemble the X-axis components first since I knew they’d be the biggest pain (I wanted to avoid removing the carriage). This involved very careful use of needle-nose pliers for positioning the nut at the end of the bolt for the idler and spacer, then getting enough threads started to let go of the nut, switch the pliers out for a box end wrench, and finish tightening down the bolt. A healthy supply of patience and judicious application of swear words is advised if invoking a similar strategy. Unfortunately, after completing these acrobatics and getting the first idler tightened down, I discovered that the flanges of the idler were contacting the gantry. I thus removed the idler (removal is, thankfully, much easier than re-installation) and used some 220-grit sandpaper to sand the flanges down. The below pictures show an un-sanded (left) and sanded (right) idler, and the sanded idlers re-installed into the carriage (and just clearing the gantry).

Next was installation of the new pulley on the stepper motor, re-installation of the stepper onto the carriage, then threading the belt. I ground a small flat onto the stepper motor shaft for one of the pulley set screws using a polishing wheel on the Dremel (you can see where I nicked the shaft a little from cutting the pulley off). In order to get the pulley lined up with the idlers, I put the pulley onto the shaft and tightened one of the set screws just enough to hold it in place, then put the motor onto its mounting location on the carriage to eyeball the alignment. There was just enough room to loosen the set screw, slide the pulley along the shaft as needed (it didn’t take much), then re-tighten the set screw. Following that, I carefully removed the motor (with pulley) again, then permanently tightened both pulley set screws. I also used a droplet of blue Loctite on both of the set screws.

For the belt, I highly recommend installing it in the same manner as is suggested with the stock 2GT 6 mm belt in the Inventables instructions - insert the belt down through the top hole in the carriage on either side of the pulley, then thread it under and out past both idlers. I had to use a small hex key to help navigate the new belt under the the idlers.

Before buying this kit, one of the things I’d been concerned with was the lack of belt sleeves for interlocking the teeth and securing the belts at the clips. The new belts’ teeth and width are large enough, however, that a zip tie does just fine.

You can also see in the background of the above picture my professional supervisory team.

I did come to find that the size of the new belt clips eats up a small amount of working area, since the V-wheels don’t clear the clip, and the new clips stick out a little farther than the stock ones. I adjusted the hard/homing stop appropriately.

After getting everything on the X axis installed, I went ahead and fired things up, and used the motor to move the carriage back and forth to either side of the gantry to check the belt tracking on the pulley. I’d managed to pay close enough attention to line everything up, and the belt was riding in the middle of the pulley.

After all that, installing everything for the Y-axes was a cinch. The idlers cleared the Y rails without any sanding, just! A small amount of Y working area was also lost, but as with the X, it’s negligible, and a small price to pay for what I’m getting out of the new hardware. One of the belts here ended up tracking to the outside flange on the stepper motor pulley after I had everything assembled, but I was able to loosen the set screws, use a small tool (hex key) to apply leverage underneath the pulley, and manually move the gantry forward and back to get the pulley to slide out just enough to where I needed, then re-tighten the set screws, all without needing to remove the belt.

After getting all of the electrical re-connected, I used an engraving bit and a ruler to “rough set” the new X- and Y-axes cal factors ($100 and $101 values), then performed a few test carves to fine tune them. My values ended up being $100 = 26.661 and $101 = 26.683 (26.666 is the “ideal” value for both).

One thing I notice, or rather don’t notice, with the new belts, is flex. With the old belts, whenever the steppers would change direction, I could see the belt stretch/flex along with them. There is absolutely no perceptible flex to these new belts.

Using the fish scale method, I tightened all of my belts to 4-4.5 lb at 1" deflection.

I then proceeded to the Z-axis. The only pulley I needed to change out here was the one on the threaded rod. After that, it was a simple matter of loosening up the securing bolts/nuts on the stepper motor to fit the new belt onto the pulleys, then use a small tool to lever the motor out while tightening the bolts/nuts back down.

As the old and new pulleys were both 20-tooth, no re-calibration of my $102 value was necessary (I checked with a dial gauge just to be sure). However, I already have 3GT pulleys for the Z axis in hand, and am just waiting for the 3GT3 closed loop belt to arrive before installing, and that will require calibration. I also managed to strip the hex socket of one of the grub screws on the currently installed pulley, so that’s going to make for a nice speed bump in that adventure.

I’m planning on working on the stiffeners for the next couple of days, so hopefully more details will follow soon!

[xfredericox]

nice!, you’ll enjoy all of that i’m sure!

[LukeWilson]

Thanks for taking the time to write this up! I can’t wait to see what else you upgrade!

[cg49me]

Well, the stiffening brackets took me a couple of extra days to complete, but you’ll see why below…

I started with 2" x 2" x 1/4" aluminum angle brackets that were 48" long. I determined that I wanted my stiffening brackets to be 39" long (the distance between the Y rail end plates is just a little bit more, something like 39.5", so I had to cut them down (I used my miter saw, and just went nice and slow). I ordered three brackets just in case I messed one up…

A quick note on why I went with the angle brackets rather than one of the many perfectly valid stiffening plate options that are advertised around the forum (which usually are flat, connect to the inside of the Y rails at the top and the outside of the extrusion underneath the wasteboard at the bottom, and are WAY easier to install)… Those plates do a superb job of stiffening the Y rails in the vertical direction, however I don’t think they do as great a job of stiffening in the horizontal direction. The angle bracket, however, provides resistance to deflection in both directions.

As with the stiffening plates, I wanted to attach my stiffening brackets to the inside of the Y rails at the top/sides, but I wanted to attach to the top (rather than the side) of the extrusion underneath the wasteboard at the bottom of the bracket. I’ve seen at least one other person who also used angle brackets to make stiffeners, but they just used wood screws to attach their brackets to the wasteboard at the bottom. I wasn’t crazy about this method, since MDF doesn’t hold threads well at all, and I could see those screws working themselves loose over a not very long period of time.

For making all the connections, I ordered hardware from Inventables - post-assembly T-slot nuts, button head cap screws, and washers (all M5). I did some mathification based on the thickness of my brackets (1/4"), the thickness of the aluminum extrusion rails (20 mm), and the thickness of the wasteboard (3/4"), and determined that 12 mm screws would be appropriate for the sides, and 30 mm screws for the bottoms, including the washers (about 1 mm thick).

I knew that I wanted to make slots in the rails for my screws, since the likelihood of me getting holes drilled in perfect locations was pretty small. I figured this was also a perfect opportunity to get my feet wet milling aluminum on the X-Carve, so I clamped down one of the angle off-cuts, and…

I had to use a single-flute bit for length in order for the bottom of the router to clear the other side of the angle. I started with the recommended settings in Easel (5 IPM feed, 3 IPM plunge, 0.003" DPP), and it turned out pretty good, although the plunge seemed a little quick to me. My hole size was 0.205" wide and 0.410" long with fully rounded corners (the M5 screws have a diameter of about 0.193"). I did a test fit with my hardware, and it was perfect!

I did have a small problem though… I had to stay a certain distance away from the other side of the angle when milling in order to avoid colliding with the collet nut. I did another test to see how close to the angle I could mill a hole…

…and it wasn’t going to be close enough for where I needed the screw to go to thread into a nut in the extrusion under the wasteboard. I also wasn’t going to be able to mill from the other side of the angle, since I didn’t have any great way to clamp the material down in that orientation. So, I had to resort to the manual method - drill a small hole at either end of where I wanted my larger hole to be, and widen it with successively larger bits - extra fun in my case since I don’t have a drill press, and had to do it by hand. Again, using an off-cut as a test platform…

It didn’t turn out too bad! I used a 1/4" bit for my final hole size (a little on the large side, 0.25" compared to the 0.205" holes I was able to mill). For clearing the material between my initial drills, I basically wiggled the drill bits between the holes and they slowly chewed out the “bridge” in the middle. I DO NOT recommend this method to anyone else attempting the same thing, as it’s terrible for your drill bits (totally not designed for lateral cutting or loading), and not super great for your drill. However, if you are going to be as dumb as me, make sure you wear your requisite protective gear. In fact, even if you’re NOT going to be as dumb as me, make sure you wear your requisite protective gear.

Having proven to myself that I could put a satisfactory hole in the angle bracket where I wanted, I moved on to the “real” pieces. I clamped them to a 2x4 suspended between work horses and marked my hole locations…

I used a nail to put “pilot dents” where I wanted to drill.

…and then I started drilling! I stepped up through four bit sizes for the initial holes (I think I began with a 1/16" bit), then started back with the smallest bit to remove the material in the middle. I made the “real” holes slightly longer than the test one, so there was more aluminum in the middle to remove, and I had to resort to using a small metal file on a few of them.

…but, I finally got there! I used a sander with some 220 grit as a final step to knock down the burrs.

For the second piece, I started with three holes, and the whole experience went much smoother.

Unfortunately, as I was anticipating would happen from my egregious misuse of my drill bits, I suffered a casualty…

However, both angles now had all the bottom holes drilled, and were ready to have the side holes milled using the X-Carve itself.

Now here’s where I got off on a bit of a tangent, and why this effort took four days instead of two. Feeling a bit like I had a new superpower, I wanted to see what all I was capable of achieving in aluminum…

Pretty neat, right? Lest I draw your attention to the bottom of that hunk of aluminum, where you’ll see the gnarled mess of a way over-ambitious carve that resulted in some aluminum welded to the bit. For the record, 25 IPM feed and 0.02" DPP with a single-flute 1/8" bit on the X-Carve is officially too demanding. However, I was able to settle on 20 IPM feed, 0.5 IPM plunge, and 0.005" DPP with very favorable results.

At this point my supervisor turned up, and told me, in his own way, to get back to the task at hand.

I had spaced the holes in the brackets in order to avoid the screws holding the wasteboard down at the sides. Double-checking…

Good! I then measured, marked, and checked where I needed the holes for the sides…

…but I was just a little bit low. So, adjusting…

Perfect!

I designed my hole pattern in Easel - it’s great for simple cuts like that. I split it into two jobs, one for each half of the angle bracket, since the working area of the X-Carve wouldn’t quite cover the whole thing. This meant I had a total of four jobs to get all of the holes milled in both brackets.

After getting the first half of holes milled on the first bracket, I set it up again to triple-check…

For the record, I was off my mark because I was a little off on my zero position, not because my cal is off.

Awesome! Good to finish off the rest!

Now, again getting a little off on a tanget… Since I had a tested and working pattern, I decided to add a little flare to my stiffening brackets…

Alright, brackets satisfactorily fabricated, it was time to cut channels in the wasteboard to accommodate the screws that would go through the bottom of the brackets into nuts in the extrusions beneath the wasteboard. So, out came the wasteboard…

This didn’t really come as a surprise, but it was neat to see. There were small piles of wood and aluminum dust underneath where all of the threaded inserts were located.

I used a 1/4" straight router bit to cut the channels in the wasteboard. It was just a hair bigger than the slots already in the board for holding it down to the extrusions.

I used the freshly drilled-and-milled brackets themselves as guides for marking where the channels needed to go. I used a square to translate these marks to the edge of the wasteboard, then, as is my modus operandi, double-checked the marks were where they needed to be.

I then measured the distance from the center of my router to the edge of the base (about 1/32" under 3" for my DeWalt DW618), and made guide marks from all my cut marks, offset by that distance.

I used a large ruler as a depth gauge, and and a speed square as a guide. The wasteboard was clamped to the 2x4s underneath, and the ruler and square were clamped to the wasteboard.

I brought the bracket back to check my cuts before moving on to the other side…

Beautiful!

Moving things back to the X-Carve, I again used the brackets as guides for inserting the post-assembly T-slot nuts in the lower extrusions before replacing the wasteboard.

Things had been going so well up to this point, I knew that I must have made a mistake somewhere, and here’s where it finally creeped up on me. While I had made sure to locate my screws so that they avoided the wasteboard screws, I had positioned the end screw holes too far out, and there wasn’t enough extrusion underneath to properly insert a nut (the extrusion I wanted to screw into butted up against the front/back perpendicular extrusion). At this point, I wasn’t going to go through the effort of putting another set of holes into the brackets and another set of channels in the wasteboard, plus these holes were so close to the Y rail end plates that their mechanical contribution could be said to be marginal. I was still able to use the matching holes on the sides of the brackets, as the Y rails themselves run the full length of the wasteboard.

After re-installing the wasteboard, I again checked the positioning of the T-slot nuts…

Still good! Using the brackets as guides one final time, I inserted T-slot nuts into the Y rails.

At long last, I was ready to screw it all down. I first installed all the screws just to finger tightness, then, starting with the bottom screws, tightened everything down firmly in an alternating pattern. I ended up ditching the washers on the lower screws, as they were riding up on the fillet in the angle bend just a little bit, and I wasn’t getting as much threading into the T-slot nuts as I would have liked with the washers left in. I went over all the screws a third time just to make sure they were nice and snug, and finally called the project complete!

As far as results go, I’m going to be a little spoiled. I’ve only got about 100 hours on my machine, so I’m not exactly going to know what I was living without. I can say that, before, I was fairly easily capable of deflecting the Y-rails a few millimeters, especially horizontally. Now, they’re at least solid enough that I can’t move them with all my weight.

Also, some of my lessons learned on aluminum…

Set your Z = 0 position high, by 0.015" or 0.020" or so. Your first few passes will be in air, but this ensures that your first pass in material is no deeper than you want it to be. That first pass is CRITICAL - if insufficient material is cleared out, the bit will be deflected on subsequent passes, leading to all kinds of issues. The below video shows this tactic being utilized - note that the first pass in material isn’t uniform (around the 1:40 mark)!

WARNING: LOUD!! TURN DOWN THE VOLUME!!

In setting your Z = 0 high, set your depth of cut in your CAD model a little deeper to compensate. Say you wanted to cut a 0.05" pocket, and you plan on setting your Z = 0 point 0.015" high - make your pocket in your model 0.065" to adjust.

I highly recommend setting your Z = 0 point without using a probe. Just go slow, and jog by smaller and smaller increments - I go down to 0.001" when I get close to my material. I also like to use a small piece of thin printer paper - when the bit grabs the paper to where I can’t slide it around anymore, I know I’m effectively right at the top of my material.

If you’re wanting to cut a hole completely through your material, set your material thickness in your model just a little bit over what it actually is. For instance, the slots I milled in my brackets only needed to be 0.25" deep to go completely through, but I set my material thickness (and cut depth) to 0.3". Since I also started 0.015" high, it actually only punched an extra 0.035" through the other side of the bracket (I used some 1/4" plywood underneath the bracket to keep from digging into my wasteboard).

All of those tips above are actually good to follow no matter what you’re milling, but especially with aluminum since it’s likely to be the least forgiving thing most people will attempt milling on their X-Carve.

I also wanted to show the condition of my bit. I used one bit to mill all the holes in my brackets, plus all four of the logos, plus four other test patterns roughly the same size as the logo, all of them performed without any kind of cooling (other than the downdraft from the router) or lubrication. The picture below shows both the used bit, and a brand new one. Can you tell which is which?

The new bit is on the left, and the used one is on the right. As you can see, the used one is actually still in pretty good shape! Point being, look for that sweet spot (and use all the advice here on the forum to help you get there) - your bits will actually last quite awhile if you’re using them appropriately.

Next upgrade on the list is the 3GT pulleys and belt for the Z axis. The belts are supposed to arrive tomorrow, so that should be a quick one…

[AndyF]

@cg49me

I’ve seen at least one other person who also used angle brackets to make stiffeners, but they just used wood screws to attach their brackets to the wasteboard at the bottom. I wasn’t crazy about this method, since MDF doesn’t hold threads well at all, and I could see those screws working themselves loose over a not very long period of time.

Not to diminish what you’ve done, because it’s great, but to let you know that there hasn’t been any issue with screws loosening for me. Between the rail connections, the number of connections and the fact that my brackets fit snugly against the the risers, there is practically zero movement in any direction. It’s been months and nothing is the slightest bit loose despite running the machine extremely hard. And it only took me 2-3 hours. Doesn’t look near as good as yours though. I hope you’re as happy with yours as I am with mine. Such a huge improvement.

[cg49me]

To each their own! I’m glad yours is still solid, and hope it remains so. I just wanted to eliminate (reduce, anyway) the chance that my bottom mounting screws would work themselves loose.

[cg49me]

Another day, another upgrade!

Quick and easy this time, just swapped out the 2GT hardware on the Z-Axis for 3GT parts (yes, some of which I just swapped out to accommodate a 9 mm belt - I thought it was going to take longer for the 3GT closed-loop belt to arrive).

Comparing the “old” and new hardware…

As I’d done with the X and Y steppers, I Dremel’d off the pressed-on pulley from the Z stepper motor, and ground a small flat for the set screw on the new pulley.

There were just enough threads left above the new threaded rod pulley for the securing nut on top. This pulley clamps onto the rod, rather than using set screws.

I started re-mounting the stepper motor with the set screws on the pulley loose so that I’d be able to slide the pulley up and down as needed to align with the threaded rod pulley.

While I got the pulleys lined up, the belt slowly tracks to the top of both pulleys until it finally rides against both top flanges. I believe this means the “tops” of both pulleys are slightly pointed toward each other, and I’m not quite sure what I can do about that. Shim the far stepper motor mounting screws? For now I’ll just keep an eye on it - I have plenty of spare belts if this one decides to go.

I started out thinking I’d need to re-cal the Z-axis, however I came to realize that since the ratio between the pulleys is still the same (1:1, both are 20-tooth), the original $102 = 188.976 still applied.

[MarkA.Bachman]

Regarding tracking of the belt; there were some past cases of the top Z plate being bent.
You may want to check that. You might be able to bend it to get both shafts parallel.

[JustinBusby]

@cg49me

I started out thinking I’d need to re-cal the Z-axis, however I came to realize that since the ratio between the pulleys is still the same (1:1, both are 20-tooth), the original $102 = 188.976 still applied.

I’d highly recommend recalibration the Z axis anyways. Belt tension can cause a difference (it did for me at least).

i feel like changing the pulley will also change it but I don’t know and it’s friday afternoon and just too much math.

[LukeWilson]

Since the pulleys are the same number of teeth as each other, they don’t change the ratio between the motor shaft and the leadscrew as compared to stock (both stock and Dave’s new setup turn the leadscrew the same amount as the stepper shaft turns). I agree about re calibrating for tension, though.

Dave - If you happen to have some 1-2-3 blocks around, I recommend using them to get a super precise calibration: Precise Z Calibration Procedure - Tips and Tricks - Inventables Community Forum

[cg49me]

Thanks for the tips guys. I borrowed a friend’s dial gauge, and actually started the calibration process. I thought the correct number would be 188.976 * (2/3) = 283.464, since I was going from 2 mm to 3 mm pitch belt. I went ahead and plugged that number in, then started to correct, and kept zero-ing back in on 188.976. Then I had the “duh” moment.

@MarkA.Bachman

Regarding tracking of the belt; there were some past cases of the top Z plate being bent.

You may want to check that. You might be able to bend it to get both shafts parallel.

I wasn’t having tracking issues with the 2GT pulleys and belt, so I’m not sure why I’m having it now. Maybe I got some shmutz between the motor and plate on one side when I was reinstalling it?

[JustinBusby]

@cg49me

188.976 * (2/3) = 283.464

Yea that’s not right.

2/3 of 188.976 is like 126.

You were doing 1.5 (3/2) * 188.976.

On the X and Y axis, it’s definitely 2/3 * steps (another unspoken benefit of the 3mm pitch, lowering the working RPM range of the stepper therefore getting more torque) but on the Z axis, the pitch of the ACME screw is dictating the steps not the pulleys.

[LukeWilson]

on the Z axis, the pitch of the ACME screw is dictating the steps not the pulleys.

Just a note for other users, this is only true if the drive pulley (on the stepper) and the driven pulley (on the leadscrew) have the same number of teeth.

[MarkA.Bachman]

Wider belt would be more suseptable to the issue.

[cg49me]

@JustinBusby

2/3 of 188.976 is like 126.

You were doing 1.5 (3/2) * 188.976.

Erg, yeah, I invoke the Friday evening math excuse too.