One thing I will say though - if 10k+ rapids prove possible with more standard motors at 24v with more standard motors - it’s hard to see any disadvantage to using a 3:1 gear reduction. 3x more torque to resist lost steps without sacrificing rapid speeds in any meaningful way seems like a no brainer to me - especially given how cheap these belt reduction sets are.
1 Like
Wonder what’s a good enough easy to repeat test folks can do with their stock and modded builds to understand and verify the limits of accurate motion they can expect for rapids, and expected loads?
Have seen a few, and personally done, rapid motion drag tests. But I don’t remember seeing reproducible simulated resistive load tests, e.g. push-pull 6 pack of beer, or, a belt running off the table (idler assisted corner) to some hanging known weight load.
FluidNC macros support variables and control flow… GCode Parameters and Expression | Wiki.js, so thoughts on…
A test that starts from Xmin with endstop triggered then moves close to Xmax at half the target speed, then rushes back to Xmin at target speed, then measure if Xmin endstop is triggered, if so, increment target speed and repeat until we end up with a cycle where step(s) were skipped? Detect X endstop not triggered as indication of steps missed and machine’s limit for accurate motion?
Thinking a post build assembly test along these lines can help folks with regular stock or modded builds understand if they’re observing their machine are operating as expected of whether there’s things that need aligning, lubing, snag avoiding, tension adjusting , etc…
I don’t believe there is any way in gcode to determine if an endstop is triggered. I think this is only in the status string.
1 Like
It’s all good, I just stepped in to say the tone of the discussion was getting…not fun.
I don’t care what people do, I made a machine that is made to mod. I love it, with the careful use of “upgrade” of course, and once pointed out I think we all understand why.
As to why people really dig in about this sort of stuff is what works or does not work for you is very different from other users experiences. What one user values others might not care at all. The issue is when the information is presented in such a way it seems to sort of point that these things needed, I am not saying this right. So a real world example, over the last ten years this happens alllllllll the time, but here is one specific example. One person made a great looking machine, and used two belts on each axis (meshed teeth). They made a stunning video proving how much better it was and how it solved so many things for almost no added cost. The video went slightly viral in the CNC world. Everyone quickly started making the changes, even though I had my doubts and was vocal about that. Before I could even test it, the person made a follow-up video saying it actually didn’t work, made things worse, and needed to be removed. No one watched the follow-up video made just a few days later. For several years, we had to get users to stop making the dual belt change.
Obviously there is/was no ill intention. No hard feelings on either side. Now, we all just get triggered into high alert when things are said very factually, in a way that sounds like it should be done despite words like “testing”, “for my use case” “I prefer”.
So I am sure you can see the dilemma. We want testing, we want people challenging assumptions, we want the boundaries tested and found…but how the heck do we present the information is such a way a new user doesn’t feel the need to start there??
Please, don’t stop, this is not criticism, I just want to keep the tone of the conversation light and fun, and worded carefully not to encourage new users to start changing a lot of things.
8 Likes
Another good example maybe a bit more relevant, is shop vac versus dust collection. Dust collection is better in nearly all ways, why don’t I use that as the standard for a build??
I am making a machine that users should not be intimidated to build, make mistakes with, build on a shoestring budget to their exacting needs. If I tell people a dust collection system is needed then you are talking about a very expensive vacuum, ducting…and a friggin shop to have it in. Most makers already have a shop vac, if not they most certainly will when using a CNC and building fun stuff.
New people come in and ask about dust collection. Half the group jumps in and says use whatever you have and open any doors and windows and the other half says you need dust collection, an N95 dust mask, automated fire suppression system, dedicated shop router, and dust removal system on your ceiling. One gets you started, and one is more “better”.
6 Likes
You might want to get together a few mod kits, I am sure some users would appreciate that. Trading max rapids for max torque is surely valuable to some.
This part really makes me wonder what is going on. That seems pretty weak to me. What drivers, current, and steppers are you using?
I, and a few others, have milled right through steel screws. 12mm deep slotting with a 1/2" dovetail in maple instead of 1.2mm, pulled the shop vac around the floor when the hose gets snagged, and not missed a step. Actually I am not sure I have ever missed a step other than having a piece of material break it’s tabs and snag the machine, and at that point I want it to skip steps not rip the belts off the table.
4 Likes
I’m using 2209’s in standalone mode tuned via pot to 1.4A rms - I’m currently using a tinybee and the uart mod simply didn’t work for me. My solder tests correctly when tested with a multimeter.
I’ve not gone back to try to fix that.
That is likely part of the issue there. Instead of troubleshooting UART going to swap over to dm556 asap(they just arrived) so I can control my micro stepping and current more accurately again and test.
I’m not at all happy with these in standalone and believe they are a significant limitation for me right now, especially comparing my rapids to the ones EG is seeing.
How this machine performs once the drivers are swapped in will determine next steps - if I get those high speed rapids and the increased torque then I’ll look for more testers as it’s a clear advantage in all scenarios.
If I don’t get the high speed rapids then I will swap to the speedy power motors to get the performance boost, but likely will not be sharing any files as I don’t want people using this without the required hardware to get that performance and not understanding the downsides.
That would cook my steppers, and those are very large nema 17’s, I have them set to .800 or .900, some 2209 have a solder jumper you need to remove to use the pot. I have never done that so maybe the scale is different from UART numbers. I would guess at 1a my steppers would easily hit 60c and fall out of a pla mount. I say all that and it makes me think about the other way to check. Your steppers should be at a about 45-50C At that point you should not be able to skip steps unless it if a violent hit.
I have also never used a tinybee so I guess I can not be of that much help.
1 Like
Mine are running very cool. So I know something is off. These happily take 1.4A rms in my experience. But I’ve been monitoring temps consistently as I have a lot of experience with hot steppers fucking up printed parts 
And thanks, not looking for help with the tinybee it was chosen because it was $20 and I wanted to see how it would work, external stepper drivers solves the issue for me it’s a big stick approach and I normally prefer to troubleshoot and work with what I have, but not this time. Work and project load is enough at the moment where $30 in drivers to not have to troubleshoot was worth it.
1 Like
Funny enough, that’s one of the tests I did. I increased rapids by 1k and pushed around a 14kg 3 phase motor on the table. When I got to 18k, I called it quits cause it was already taking too much time and it wasn’t really necessary anymore. I was never going to run the machine at its limits anyway. I wanted to increase them so that 20% below the limits was a number I could live with.
1 Like
Also my parts are in asa. Overture Natural asa so about as stiff as ASA gets in my testing. I had some PLA for this project originally but it turned out to be trash, and I had the ASA in hand as it’s one of my favourite materials for many other things. Good PLA would be more stiff, but this one element is where ASA shines - overture ASA(polymaker) has a heat deflection temperature of 100c @ 1.8MPa, making it significantly better for motor mounts specifically.
1 Like
I would not be surprised if there is a 2x factor from voltage to amperage and your pot is at 1.4V and 700mA
I think you and I got off on the wrong foot. I thought you were criticising, not critiquing, so I felt the need to defend my reasoning.
4 Likes
I get what you’re trying to say, but this machine will probably never see any hardwood other than flattening some slabs. I have a second CNC machine for that sort of cutting with a larger Z cutting capacity and way more rigidity. I use that machine for steel and alu.
And beyond that, I live in NZ. Hardwood is as expensive as gold. Just looking at oak or walnut will drain my wallet. And native hardwood is even more expensive than the traditional ones because native species are protected and you have to be licensed to cut them.
5 Likes
I have heard reports of pet cf bring used on the the big parts and gaining rigidity over pla while also gaining heat resistance over asa. I can’t not confirm the reports personally, but other members here have hands on experience.
I wonder if it helps with the nod you mentored previously
1 Like
Pet-cf and asa-cf would both be fantastic choices for this use case. Many monies here in Aus for decent quality versions of either unfortunately.
Yes @50usd+shipping per 1kg roll… Not cheap
And the three big parts are where it would be the best used. So 1.5kg give or take.
yes, but the added heat resistance is welcomed in every part for me.
my shop temps are kind of high, and you have to sum up all those temps as i live in the carribean and not climate control in the shop, this is like hell heat 24/7