Belt Stretch and Stepper Holding Measured

@gmack Yes, that’s the conclusion I seem to be coming to. Not what I was expecting - from following the forums, I had figured steel belts + linear guides was the magic combination. I do see deflection in the V-wheels, but it’s an order of magnitude less than the belt stretch… enough that I’ll want to fix it eventually, but it’s nowhere near the top of my list, now. And the extrusions + frame plates seem pretty solid.

Here are the belts I’m using:

9mm steel:
(Same ones @luc.onthego mentions below)

10mm steel:

(Plus 8mm 20-tooth GT2 pulley to mate with my new stepper, links mentioned in an earlier post… stock stepper would want a 6.35mm. Edit: to clarify, the 8mm/6.35mm refers to stepper shaft diameter. Pulley width is 10mm for 10mm belts.)

Re: spindle - I have a Jianken JGD-80/1.5R40 on order. 80mm diameter, “1.5kW”, ER16, 40k RPM. Unlike some of the other high-RPM spindles, power is flat between 30k and 40k RPM, which I think is nice - it gives me the flexibility of using either .25" or .375" end mills with good MRR without exceeding the “happy” SFM of 3000 for aluminum (or is that a myth and I can just go as fast as I want???). I don’t need peak power at peak RPM when using smaller end mills (like 1/8"), just the extra RPMs.

Thanks for the link to the Mechatron… I was looking at their spindles, but somehow missed that one (and the HFP-8022-50-ER20). The power/torque curves on those look great compared to other spindles on the market. Maybe for next time!

Oh, also, regarding how tight I tensioned the steel belts. I still got a nice improvement over where I was going into this experiment by tightening them as much as I could. For the 10mm steel belt, I have it so tight that the belt teeth are almost slipping in the retaining clip. Any tighter, and a tooth will deform enough to slip in the belt clip to the next tooth (2mm looser). I had to use a longer screw in the belt clip to start it threading, as I can’t physically pull the belt close enough to the end of the X axis to start threading the standard screw. And I had to press down on the belt clip pretty hard while tightening, to prevent the belt from slipping in the clip.

At some point in time, steel belts alone should appear in the Carbide 3D shop.

Back when the machine first came out, I advocated for tensioning using the weight of the machine — this was easier with the old sliding motor mount tension system — we’re working to update the instructions, hopefully this will become clearer.

I will note that I turned down a set of steel belts for my XL (but did install them on my SO3 at my mother-in-law’s) since I bought Gates-branded GT2 belts from SDP/SI and still feel the sting of that expense, and I still have a spare length — they’ve held up well, and I’m hoping to get my money’s worth out of them yet.

Be careful because the belt pulls on the pulley/shaft of the stepper and if you tighten too much, we are warned that we can damage the stepper shaft.


So I’ve heard. On the other hand, the belts clearly have some give in them. I wonder which would break first. My new stepper has a larger 8mm shaft - maybe more resilient, though I don’t know how well the stepper’s bearings can handle side loads.

As pointed out in my comments here, Jianken’s spindle specs aren’t very realistic. For this spindle they claim 220V * 4.4A * 1.73 = 1.675kWA input produces 1.5kW output. You’ll probably need to increase the VFD drive current to about 7A to get that level of power output. (See what Jianken and/or the motor label says?) How much did you have to pay for it and shipping/fees - to where?

According to this unit horsepower for aluminum decreases from 0.3 to 0.2 HP/(cubic inches/min) when cutting speed is increased from 3000 to 5000 SFM at 0.005 IPT/Revolution chipload. It levels off after that, but he could only test to 25,000 RPM with that 1 inch diameter 2 flute MSS endmill cutting 7075-T6 with Cimcool.

1 Like

$555, including shipping, to the US. Plus about $200 for a VFD from elsewhere. Yeah, their power numbers seem a bit optimistic, the math doesn’t entirely add up. They did offer an explanation for this, but I’m waiting until I have it in my hands to take my own measurements. I’ll post about it here once I’ve had a chance to try it.

Even if it’s under-powered, it’s still about where I want to be “for now”, for my upgraded SO3. I’m thinking about building a from-scratch CNC with what I learn from all this experimenting, and that’ll be when I consider splurging more on the spindle(s).

That’s interesting about SFM vs cutting power. I’ll have to try pushing it above 3000 and see what happens :slight_smile:

1 Like

Excellent testing and documentation. These results intrigue me and I plan to duplicate them with some nice testing equipment at work I have access to. I plan to complete the following tests:

Pull polyester and steel-cored belts with 10, 20, 30, 40, 50 pounds-force and measure strain.

Bonus test - run cyclic testing of 0 to 50 lbs-f loads and see if the belts undergo any degradation.

Anyone else have other thoughts on some useful tests to try?


Use Shapeoko stock pulleys and document belt part/lot numbers?

1 Like

I don’t have a way to mount the pulleys in the testing apparatus. I intend to clamp directly onto the belts.

If the belt part/lot numbers are on the belts, I will. Otherwise I have no information on them other than materials.

1 Like

This could be useful testing device


Right, so the machines (Instrons) I have access to are much faster and more accurate versions of what he makes in this video.


Great - I’m interested in whatever you learn. Do you have access to a dynamometer for the measurements? FYI - Piotr and I are continuing our discussion about his “2.5 kW” Jianken ATC spindle.

No, I was just going to measure actual input power to the spindle. If there is a straightforward way for me to measure output torque and power, I’d be interested in trying that, too.

@AlexN may I ask what kind of cutting forces were you seeing when cutting aggressively.

Estimated? I forget, somewhere around 6lbs according to @gmack spreadsheet, I believe. Running around 90% of the Makita’s estimated output power, 30k rpm, .25” single flute. I’d have to look up the exact DOC, etc. which I don’t have handy, but I want to say around 5mm DOC and 1.25 WOC? Going off memory.

I don’t know the actual force. I feel like there were some parts of the cut (large circular adaptive clears) where deflection from prior passes may have left enough behind that the momentary force exceeded what the X stepper could hold, also taking into account that the 18lbs max holding force will drop off at speed when a stepper starts moving fast.

So, I think the X stepper had enough holding power in theory for the Makita, but when combined with the belt stretch, I’d sometimes be biting off much more than expected.

Definitely a healthy bite. I’m guessing you could hear chatter on the heavy circular cuts as well, especially transitioning from Y to X right? I know these belts have a bit of give but imo the X wheels do not deal well with the axial loading and unloading.

Have you seen the Millright megaV with its rack and pinion drives? It should be very interesting to see how those perform when they start coming out. Also one tip on long term use for steel cores, they don’t like being over tightened and they do have a lifespan.

I think we can all agree on belts not being the golden ticket but they are very good for a hobby machine in terms of maintenance, ease of use, and backlash.

Super interesting thread
Subscribed :slightly_smiling_face:


Wow that’s a big bite, but how fast are your feed? I have a rigid machine but that much stepover i would just dream of :slight_smile:

Yes, there were definitely directions of heavier chatter… it’s been a few weeks now, so I can’t remember which part of the circle it was on. The V-wheels on mine, I couldn’t measure much deflection just pulling on the router (but maybe cutting is different). For sure belt stretch would differ a lot, transitioning X to Y.

Super interesting about the new MillRight, I hadn’t seen that… the rack and pinion would address the belt stretch and the rest of the SO3 seems pretty solid already, so that looks pretty beefy.

Steel belts do feel like a good sweet spot for most things. I guess MillRight got R&P into budget by manufacturing it in house…?

@ydrefalk 160 IPM, give or take. Normally I’ve been cutting about half that depth at that speed, but I was working on a large 24”x12”x1” piece and was getting impatient… I wouldn’t repeat those settings on my machine without additional upgrades. It vibrated my Makita to death, actually - the speed dial broke and it is stuck on min speed. Guess I’ll have to buy that spindle sooner than I was planning :slight_smile:

Also, what problems have you seen with over-tightening and lifespan of the steel belts?