Hello. Yes, I did check the tightness of the pulley and that is not the source of the play. Good thought though. I can see visually that it comes from the zip tie mounts. I have the belt already really tight but can see that it flexes at that point. I measured it with a caliper to 1.26 mm (that’s in either direction) before the stepper is overpowered. I dialed down the stepper power to give me a little idea that I’m not putting a lot of pressure with my one finger when I press against the stepper holding to look for the source of play. Right now on my machine that is by far the biggest source on the x axis. I’d need to get a device to measure accurately how much play there is at a specific weight, say 1Kg
BTW, Ryan. The 1/8" 1 flute upcut bit I got from you has by far done the best in cedar and is lasting a lot longer than any of the other bits I have tried from other sources.
Okay, lets just get all the details in one place. you are giving one tidbit of information per post.
Gcode?? and router setting.
1 hour 10 days?
We all use zip tie mounts they can hold 40lbs of tension, this is not the issue, this is the side effect of teh issue (dull bit, bad gcode).
Vague, tells us nothing 150mm/s or 2mm/s?
This proves 100% it is the bit, why do you think it is the zip ties? A more rigid machine will just break the bit. Busy machine shops track the time on there endmills and replace them before they get dull for this very reason.
Sounds like the gcode is bad, that is why I have asked several times for it, as I said Kevin gets days in aluminum and I get 6-7 hours in MDF, cedar is nothing comparatively.
In the last cut picture what is that? Is it packed with saw dust on one side or is your cut actually fuzzy?
That’s sawdust at the bottom. There is a channel in the waste board underneath from the same cut being repeated. The cut is good.
GCode is attached, Router is at max (6)
3 hours
Sure, that doesn’t mean there is flex there. That is the question this whole post was about, how much flex other people have in theirs on the X axis. Forget if anyone things it is relevant or not, that is what I was asking and I’d love to know what other people have compared to me.
I’m cutting at 10-11mm/s right now. More than that and it’s too much load for the steppers at the moment as well as the rigidity. Cutting at more like 30mm/s is where I would like to be at, which can give good chip loading for 1/4" bits that can handle that speed at deep cuts. (10mm+). I’m going to get backlash on that, I know. that’s where I would like to be at for cutting out pieces.
Yes, the bit is what caused it, I never said otherwise. HOWEVER, it does show where there is play on my machine. I’d like to improve that play. That’s what this post was about and is getting pushed in other directions.
Actually, this cedar is pretty hard on bits, even though the wood is soft. It has a lot of oils in it as well as silica, so it’s very abbrasive
The gcode will show 9.75mm DOC at 11mm/s I believe. Two passes making 19.5mm depth. That is not actualy what is being cut in the wood for the most part, there is extra there to leave room for where there is some warpage. 8’ boards aren’t perfectly flat. The actual DOC is about 8 - 8.5mm per pass.
arch.gcode (3.97 KB)
That is really fast, faster than a video where I was not slotting, I was adaptive clearing on a tiny build and completely shocked that the machine could handle it.
So you want to add more than 2x the load (up to 4X) with a larger endmill and more than double the speed (almost 3x).
What you are wanting is extremely far out of the range of the machine, and honestly not sure what machine can do this.
A typical slotting cut can be no deeper that the diameter of the bit, for chip clearing, and 8-10mm/s keeps accuracy really well on my machines. In my opinion you are looking for a machine with many more zeros in the price tag.
It’s definitely a lot faster than what the machine in it’s current form can handle. Those are ideal speeds for bits. Machines like the shapeoko 3 can do that and faster. There is a lot more cost there.
You have however come up with quite an awesome design. It’s a very capable machine for the minimal cost and work required to build it. (Basically the table…)
I think it’s possible with some additional work to move at least in the direction of those speeds. A lot of the fun is in trying, but I think the potential is there to certainly move it in that direction, and that’s where my questions come from and where I will be experimenting.
The belts clamping on one side, with an adjustable tension on the other end I believe would be highly beneficial in that goal. The zip ties are strong, but require too much tension before it removes the flex. Like I said, great design and machine, I just want to see if I can take it some steps further down the road. Yes, more cost would be involved, but a fraction of other machines I believe still.
A shapeoko is not 8’ long an 5’ wide, not even remotely comparable, and I highly doubt that those speeds in a 1/4" at that depth is honestly possible with any degree of accuracy and repeatability.
Please stop with the belt flex…it is not the belts. You are going to propagate a misconception I feel very strongly about, and I had to fight for years. The belts are not the issue unless you are pushing them harder than 150lbs, which you are not. The machine is designed for a few lbs of load.
Bits have a range, dependent on many things. I designed a machine to max out a 1/8" bit, YOU want a 1/4". Why use a 1/4" bit for slotting, decrease your load by 2X using an 1/8".
At this point it is great you want to try and do more but you are asking for some very unrealistic numbers in my opinion. I am often wrong but, before you start telling people you are improving things take a very close look at the loads you are expecting at those speeds.
Look at big machines and the figures they advertise and the prices they want. I am not sure how to convey what you are asking and how out of reach it seems. To make something 4x as rigid you need to make it ~4x smaller.
Optimizing your CAM and finding the endmill that best suits your purposes is the best way to increase the speed of your cuts. But expecting to make some tiny changes and increase your load by 4X is kinda far fetched.
I didn’t say the belt was flexing, I said the connection is flexing. It is, I have measured it with a caliper repeatedly. Are you suggesting that I am incorrect that the joint can flex or misunderstanding what I am saying? This is your forum, if you would prefer I don’t have discussion about this here I am happy to oblige, please feel free to let me know if that is the case. I enjoy discussing, but not at the cost of others.
I did not say I expect those numbers. That is where I would love to be at with a machine, I do not expect this machine to do that. But I’d be happy with a 20% increase. Thrilled if more was available. I believe it’s possible to increase it enough that it’s worth the effort, I never meant to imply with a few small changes the load could be quadrupled.
We are going in circles. I feel you will have (and said so yourself) no issues with a single flute 1/8" SHARP end mill. You have room to optimize your Gcode and get a slight cut time improvement.
I gave you a suggestion on how your belts are mounted. Although I feel it is a waste of your time. I actually have all sorts of little zip tie tweaks we have tried over the years. If you could post a very clear picture of both sides of your machine with a clear view of you zipties. If you look at the zip tie section of the LowRider page you can tell I spent the most time here, with extremely clear pictures of how I use them. If you do it the suggested way there is such extremely small room for any flex variables you might see why I am not very receptive to changes with them.
As for the rest I am just trying to be informative, not agumentative.
I am going to redo the belts, but I need longer ones to do so. I just placed an order, but I actually meant to order 2x the 10mm belts of 10M each. I included a couple other items. Is it possible to change the order? Or can we cancel and redo it?
I had posted about flex in my x axis for my Lowrider v2 recently. I could see a lot of it was coming from the way my belt was zip tied on. Ryan pointed out the gap I had. I purchased more GT2 belt to be able to make it long enough to go right snug against the 3d printed part as was recommended.
I did some testing with a digital scale and dial indicator. What I found was with the steppers locked. I could easily flex the 611 plate on the x axis.
I tested with the existing setup, and then with the newly installed gt2. The results were essentially the same. Using an scale I could see that to either direction I had 1.3mm of play with only 1kg of force for a total of 2.6mm. There was still visible plat at the end of the belt where secured with the zip tie very snugly right against the plastic mount.
So I tried removing that connection and clamped the belt down under the aluminum angle bracket. That reduced the play to .5mm. To me that seems like a significant difference. Increasing the tension to 8kg only changed the play to .6mm so I’m really pleased with that. I would like to redesign the one part to be able to clamp the belt. I’ll attach a picture of what I did for a quick test.
The remainder of the play is actually coming from the stepper rotating a bit when the 1kg load was applied. The stepper does that have enough holding power to prevent any movement. I have some 92oz steppers so I tried one of those and had the same result. Looks like a NEMA 17 can’t hold firm enough to prevent rotation, but the small belt change removed a large percentage of the play.
Thoughts?
1 Like
This doesn’t seem right. 0.5mm is 50 1/16th microsteps or about 3 whole steps. If the stator moved that much, it wouldn’t return. It would be skipping steps. Something must be flexing.
Before writing off nema 17s, you could consider changing the gear ratio. If you moved things around to add in a 3:1 reduction, you’d lose top speed but the flex would be reduced and the torque increased. Nema 23s are a huge pain.
1.3mm or 0.5mm just doesn’t seem right to me. I haven’t gone through the experiment, but I’m sure I’m getting tighter tolerances than that.
I hear you. I had a friend over who is an engineer and very knowledgeable in these areas help me measure. He milled me some beautiful y plates out of aluminum we installed at the same time. We were surprised to see the gear turning without the stepper letting go and returning. Both the stock nema17 from Ryan and my bigger 92oz one did the same thing.
I’m curious to know why you think nema23s are a huge pain. You mean all the changes required to switch to them or are you generalizing about that format specifically for some reason?
However. With power set to 1400 on the nema17, their static holding power measured at 19lbs. We were thinking gearing also, or perhaps a worm gear, which would not allow any force from the bit on the gantry to push back on the stepper at all.
Well the 1.3mm in each direction was because of zip tie mounting. Even really tightly zipped right against the gantry has flex. Clamping the belt reduced it to .5mm
The two of us measured several times with the same result. .5mm flex at a couple pounds pressure and then up to 19 pounds before the stepper let go it would go to .6mm so very little more with a lot of force. When removing the force it would return almost back to where it was. (Off but about .03mm) which would be from some other little bit of flex.
I’m feeling very confident the zip tie change makes a big difference. I don’t understand the stepper rotation and return though. We thought maybe the mount was flexing, but we could both clearly see the rotation of the amount required for .5mm
I’d really like to know if anyone else can just grab the x tube with the stepper locked on and press with your thumb on the 611 plate and see if your stepper has some give also without letting go. .5mm doesn’t mean a lot of rotation but two of us checked and saw the same thing.
I wonder if a quick swap to these might help:
I’m not sure if they’d fit through the existing zip tie channel. I can test this if you’d like as I have some of those steel ties.
1- Are you using my little tiny plastic inserts on the belts? These were designed to not let the belt collapse on itself.
2-How about a video of your zipties flexing? There are some tricks to make it better but mine in no way moves that much with 1kg at the endmills tip. That force is divided between both sides so each belt/ziptie only sees 0.5kg. I have a scale and have done testing on my smaller belt and tiny zip ties on my MPCNC and I had to pull a significant amount of weight to deflect it (there is a post here with real numbers somewhere).
In use my Lowrider with very little concern for perfection in assembly (trying to replicate a new user) is giving me un-measurable tolerances over 13" (LowRider parts) with a tape measure, most certainly less than 2.6mm.
3-Where is your other thread with all your pictures so I don;t have to ask for them again? Your numbers seems very extreme to me. If this is true I need to stop selling machines until this is fixed. So I am taking this seriously, 2.6mm is extremely unusable.
I merged and moved this so we can keep all the details in one place.
I see the yellow zipties you have but not the Y axis ones. What rating are the ones you are using? I see one side has a littel slack is the other completely snug?
Yes, I am using the plastic insert. That was a brilliant idea. Without it would not work nearly as well as it does I’m sure.
Since the first pictures I ordered new belts and zipties from you so I was using only your parts as I had run out of your zip ties. I had space but when I redid it I had the one side tight right against with no space tightened as much as I could with pliers without snapping it. The tension is not evenly devided when there is force from the bit .the bit force if going straight in the x axis + or - is applied to the one side.
I could see this because I initially left one side zip tied really tight with no space and then clamped the other side simply by putting it under the aluminum angle bracket and tightening the bolt down as in the pictures then when I tested pulling away from zip tied side it moved 1.3mm (dial indicator accurate to .ooo5") and when pulling the other direction with a strain guage to measure weight vs movement it moved only .5mm away from the clamped side. And when we tried to find it we so zero movement of belt moving, only the stepper rotating slightly. But the belt was rock solid.
My table is full sheet. Although I don’t think that matters since there is no belt stretch in the measurement I’m taking. I’m happy to do any testing you want. I was thinking of modeling a new top piece to clamp the belt better as what I did would not last. I can get a pic of the one side that is still zip tied so you can see what I did. It’s using your zip ties and belt and the plastic bit to put in the zip tie. But the 30 second clasmp I did totally removed all that play. I would not say by any means you need to stop selling it. But to me it just seemed like a good modification that is not costly or a lot of work.
That’s why I’m hoping someone else can check this on their Lowrider and see if mine is somehow acting differently for some reason I’m not seeing.
I appreciate the insights and feedback!