Ok so yesterday I broke my z-belt by over tightening the little screw that pushes the idler down. I did this because my z kept slipping upwards while trying to cut this file in some walnut: PCPF rough&finish xl outside v2.nc (805.3 KB)
The slipping happened twice. First time: I set the z-zero, ran the job, and on the first pass, instead of going about 12mm deep, it went about 3 or 4. I’ve been having odd behavior with my z-height, though, so I read the forum, checked that the pulleys were correct, made sure my v-wheels were not too tight or too loose, and tightened the z-belt a little bit until it seemed sort of twangy, but what the hell that means I don’t really know.
Second try: ran job, first pass was at correct depth. Hooray! Second pass seemed correct. Third pass was 3 or 4 mm too shallow. Ok, so now I felt the z-belt must certainly be too loose, since all the other stuff seems good. Tried to use a guitar tuner to tune the z-belt. That was a failure. Decided to just make it as tight as it would go, and the damn thing snapped. Learning is fun!
So, what is wrong? I’m using a downcut bit. Feeds and Speeds are from g-wizard. G-code is from Fusion 360.
And, please, for the love of g-code, can someone explain the proper tightness of a belt in terms that are somewhat less vague? Try to imagine you are talking to someone who, prior to this, has never even touched a timing belt, let alone tried to tune one. When I squeeze the z-belt, how close should I be able to get the belts? Can I use a luggage scale to pull them sideways to measure the tightness? If so, how tight should it be?
Also, should my belt have broken before my pulley bent? Suddenly I wonder if my pulley is now bent…
From your description, it would seem to me that the Z belt tension was not the problem in the first place: the “sort of twangy” should have been enough, and while a slightly loose Z-belt can explain inaccuracies in the cut, for me it cannot explain such a huge depth difference (3mm versus 12mm). It would take a completely loose belt, and you would probably have heard the belt slipping over the teeth (ugly sound): very unlikely.
I know you said you checked the pulleys (I guess you mean you double checked the set screws?), but it’s easy to consider the set screws are as tight as they could be, and not realize that they are not on the flat side of the shaft, or damaged, or something. The only really effective test is to draw a line across the shaft and pulley, run a few jobs, and see if everything is still perfectly aligned.
I suppose it’s not the endmill slipping into the collet, you would probably have noticed if it got pushed back into the collet by 9mm
Could it have been a very aggressive plunge rate ? (unlikely, but asking just in case, as plunging 12mm into walnut can be hard on the Z axis if you plunge really fast. By the way, I’m surprised by the 12mm depth per pass, if I read that correctly?)
Sorry, I know this does not answer your main question about a quantifiable way to tell if the belt is “tight enough”, but honestly if it’s tight as in “when you pinch and release it, it goes cloooong”, it should not matter if it’s 20% more or 20% less.
I don’t suppose you would post a video of how tight yours is?
I checked my set screws to make sure they were on the flat of the pulley shaft, but I will try drawing a line for confirmation.
A person on Instagram suggested the down cutting bit might be causing it to creep up - maybe it’s losing a step here or there, little by little? I don’t notice the cut getting progressively shallower, though.
I already ordered a 10 pack of replacements, I need to get the machine running on a job ASAP. Do you know if they make timing belts of this size with steel in them?
It’s my understanding that due to how they are manufactured steel core belts aren’t possible.
Rather than a 10-pack, my recommendation would be to purchase some good quality belts from a reputable vendor which has Gates-branded GT2 belts. BB Man., or SDP/SI are the usual suggestions.
I would have, but I upgraded to an HDZ a long time ago. However, I pulled out my old Z from the box of spare parts, and setup a makeshift Z-belt assembly of something that feels/looks similar to how tight I used to have my Z-belt:
Yeah, losing 9mm means losing 360 steps, that’s a LOT, and indeed you would notice the cut getting progressively shallower. What’s your plunge rate ?
Plunge rate is 762mm, ramp rate is 1200mm. Here is a very short video I shot of the cut happening, I doubt it will tell you anything, I just shot it because it looked cool:
I for one would not plunge that fast/that deep in walnut with a downcut bit (they are notorious for poor chip evac), and using the stock Z axis.
Actually, I like to do a roughing pass using an upcut (to not have to worry about this), and do a finishing pass along the walls with a downcut to get perfect top surface finish.
I see on the video you are using an adaptive cut, but are you using a straight down plunge initially ?
At that depth (~200% DOC) I would use either a slower plunge rate (say 400mm/min), or at least a helix ramp in.
Nah it’s a helix ramp in. If you check the .nc file I attached you can see how it did it: kinda nibbled its way in from the sides and then went in little adaptive sweeps. But I’ll try a slower plunge. Piece being cut is only a little bit smaller than the stock, so it’s generally not having to plunge into the material for this operation.
