RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
What bearings do you think are the best?
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Let me describe where I'm now:
- Rods/Bearings: Misumi rods / LMU8 bearings, greased with Divinol Lithogrease 000/150
- Y asis: Trinamic QSH4218-51-10-049 1.8° μ-stepping optimized stepper (bulky motor, but very linear)
- X axis: Trinamic QSH4218-51-10-049 1.8°, same as Y
- Bondtech BMG Prusa upgrade
- Dual-bearing toothed idlers
- LDO-42STH25-1404MAC 0.9° pancake stepper on E axis (but has retract issues)
- Gates GT3 belts
- Aluminum Y-rod holders, Y rods parallel to 0.01mm
- Aluminum X-ends, X rods parallel to 0.005mm
- Part-aluminum lightweight X-carriage, rod distance exact to 0.005mm
- Aluminum Z motor holders and Z top ends with Z rods fixed at both ends, Z rods parallel to 0.01mm
- Nylon Y bearing holders
- Z threaded rods carefully aligned with Z smooth rods and fixed in bearings on top
The prints are really nice, and any vertical artifacts are seen only under very specific lighting. But I keep wondering if this is as good as I can get, or if it still can be optimized much. 🙂
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Hey vojtech,
I'm new to this, but I'm curious about those Trinamic 1.8° steppers. I can't find much info on them elsewhere. Am I right in thinking they're bigger than the standard motors? What did you need to do to mount them? Any firmware considerations? Have you compared them to alternatives like the Moons 1.8° or 0.9° steppers, or others?
Thanks in advance!
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
They're what Trinamic uses in their expensive servo solutions and purportedly have an internal geometry tuned for microstepping. A manual is easy to find and has a lot of information on them. I bought them from Digikey. They're still NEMA17, so they can be mounted without issue, only they are longer for the same torque than the default LDOs. They have the same 1A nominal current like the LDOs, do, so no firmware changes are required.
The main difference is supposed to be the internal shape of the magnetic poles. The LDOs are said to have wide poles which allows high torque in a small volume at the expense of microstepping linearity. The Trinamics are supposed to have narrow poles, thus having a much more sinusoidal behavior, resulting in more linear microstepping. But narrower poles require a longer body for the same torque. So the Trinamics are also a bit heavier. Not a big deal on X and Y, somewhat on an issue when mounted on E.
I would love to somehow compare them to the Moons 1.8° or 0.9°, but I'm not quite ready yet to buy all the Moons' and the Sanyo's I'd like to test against just for the purpose of science. They're not cheap and after testing would just be collecting dust.
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Thanks, Vojtech! Are you in the US? I have a pair of Moons 1.8° I'd be happy to lend you if shipping isn't crazy. After all, it would help me save on deciding if I should invest in the Trinamics, and then we all benefit from the comparison. 😉(Though it sounds like the 0.9° motors are the ones to beat.)
As I've been researching steppers tonight, my brain stumbled across a couple other (undoubtedly dumb) questions:
1. Has anyone experimented with oversized steppers on the X and Y axis? For instance, Moons MS17HD6P4100/MS17HA6P4100 look like a nice bump in torque and still run at 1A. There's no mass considerations on those axes, and they have plenty of room for longer motors. Any reason to think the extra torque would help move those parts smoothly, or is it just going to fight its own added inertia?
2. Has anyone experimented with midsize steppers on the extruder like Moons MS17HD4P4100 or Trinamic QSH4218-41-10-035? I'm tempted to hook up Kuo's geared-extruder mod for the Mk3S and try it out, but no sense in replicating the experiment if someone has already tried it.
3. Along those same lines, has anyone tried the smaller Trinamic QSH4218-35-10-027 on a geared extruder?
I apologize if these have already been answered somewhere and I missed them, or if I'm just asking stupid questions!
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Thanks, Vojtech! Are you in the US?
Nah, Europe. Very close to Prusa HQ actually. 15 minutes or so.
I have a pair of Moons 1.8° I'd be happy to lend you if shipping isn't crazy.
That'd be way cool. But I expect the shipping to be crazy. Maybe printing & sending over test prints (same material, same gcode ...) could work out less expensive.
After all, it would help me save on deciding if I should invest in the Trinamics, and then we all benefit from the comparison. 😉(Though it sounds like the 0.9° motors are the ones to beat.)
That is the big question for me: Can the Trinamic be 2x better at microstep linearity? Because the 0.9° Moons will only need half the steps. On the other hand, if the microstepping is not good, the 0.9° resolution will only go so far at making it better.
As I've been researching steppers tonight, my brain stumbled across a couple other (undoubtedly dumb) questions:
1. Has anyone experimented with oversized steppers on the X and Y axis? For instance, Moons MS17HD6P4100/MS17HA6P4100 look like a nice bump in torque and still run at 1A. There's no mass considerations on those axes, and they have plenty of room for longer motors. Any reason to think the extra torque would help move those parts smoothly, or is it just going to fight its own added inertia?
Extra torque will help with microstepping. Without a closed-loop feedback, microsteps get compressed and stretched when the motor is under load, resulting in non-linearity that the E-correction cannot fix. The more torque the motor has compared to the load, the smaller the compression will be. The Trinamic controllers don't seem to mitigate that, although it could be possible by open-loop control, basically plugging CoolStep into the microstepping compensation table.
A disadvantage of bigger motors is bigger rotational inertia (needing more torque to get spinning, so you lose some of that torque advantage) and a bigger inductance (robbing you of some max speed).
2. Has anyone experimented with midsize steppers on the extruder like Moons MS17HD4P4100 or Trinamic QSH4218-41-10-035? I'm tempted to hook up Kuo's geared-extruder mod for the Mk3S and try it out, but no sense in replicating the experiment if someone has already tried it.
3. Along those same lines, has anyone tried the smaller Trinamic QSH4218-35-10-027 on a geared extruder?
I will be putting the 035 on a geared Bondtech. Not exactly what you ask about, but I hope for really good linearity there. If the inertia doesn't kill retracts again.
Oh, and I also plan to design a 1:2 reduction for X and Y using only the belt and a two extra idler pulleys - no gearbox. Let's see what that brings ...
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Nah, Europe. Very close to Prusa HQ actually. 15 minutes or so.
Oh, cool! I hope you get over for a factory tour some time.
That'd be way cool. But I expect the shipping to be crazy. Maybe printing & sending over test prints (same material, same gcode ...) could work out less expensive.
Ouch, yeah, my quick round-trip estimates are coming out at around the same cost as the steppers themselves. I may go ahead and pick up a set of the Trinamics for comparison in the next couple weeks. If I do, I'll at least print and post some comparisons. Did you ever decide on a test print? I've been using this XYZ cube, because it has both a labeled side and a smooth side. I've generally been printing in both black PET and the silver PLA that came with the printer. I use the 0.10mm "Detail" preset in Slic3r PE and generic PET/PLA filament presets, and otherwise leave the rest of the settings at defaults. I'm happy to switch to something else, though -- I'm still pretty new to this!
That is the big question for me: Can the Trinamic be 2x better at microstep linearity? Because the 0.9° Moons will only need half the steps. On the other hand, if the microstepping is not good, the 0.9° resolution will only go so far at making it better.
...
Extra torque will help with microstepping. Without a closed-loop feedback, microsteps get compressed and stretched when the motor is under load, resulting in non-linearity that the E-correction cannot fix. The more torque the motor has compared to the load, the smaller the compression will be. The Trinamic controllers don't seem to mitigate that, although it could be possible by open-loop control, basically plugging CoolStep into the microstepping compensation table.
A disadvantage of bigger motors is bigger rotational inertia (needing more torque to get spinning, so you lose some of that torque advantage) and a bigger inductance (robbing you of some max speed).
Thank you for sharing! Since I'm still in the don't-know-what-I-don't-know phase, explanations like this give me lots of leads to go read up on.
I will be putting the 035 on a geared Bondtech. Not exactly what you ask about, but I hope for really good linearity there. If the inertia doesn't kill retracts again.
I'm eager to see the results of your Bondtech + Trinamic combo. If the Trinamic stepper works there, I might just go ahead and try the -035 or even -027 on a Bunny Science Extruder Motor Gearbox.
Oh, and I also plan to design a 1:2 reduction for X and Y using only the belt and a two extra idler pulleys - no gearbox. Let's see what that brings ...
I think I finally understand where you're headed with the geared X/Y axes. I'm definitely also interested in what you find. I probably can't do much to help, but I'll happily be a guinea pig and post comparisons with whatever motors I happen to have at the time.
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
- Y asis: Trinamic QSH4218-51-10-049 1.8° μ-stepping optimized stepper (bulky motor, but very linear)
- X axis: Trinamic QSH4218-51-10-049 1.8°, same as Y
Hey Vojtech,
I'm interested in these engines.
You have been able to compare them with some others or you have some test even if it is comparing them with the stock.
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Oh, and I also plan to design a 1:2 reduction for X and Y using only the belt and a two extra idler pulleys - no gearbox. Let's see what that brings ...
I'll have to keep an eye out for that - I have idlers aplenty and relatively cheap stuff like that to improve the surface finish is always a win in my book!
I'd been contemplating trying to model a planetary gearbox with split spring-loaded gears to cut backlash/slop but my design skills are not quite there yet.
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
^^ missed the edit window, but count me in as a tester if you need one, @vojtech-p6
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
I made an exceptionally dumb mistake last night. It's also kinda funny, so I figured I'd share so you can laugh at (and learn from) my error.
I was having no luck taking comparison photos and decided to try 0.9° steppers. I have a MK3S upgraded with a BNBSX extruder using a Moons pancake, Trinamic steppers on X & Y, and new Misumi rods and bearings. I'm really happy with the improvement -- especially the BNBSX! -- but I'd ordered some 0.9° Moons earlier with a long lead time that just arrived, and thought I'd give them a shot.
They went on easily, and the firmware worked the first time. But when I started moving them, I was greeted by a chorus of loud chirping noises, like a nest full of demon birds escaped from hell and possessed my printer. Ahh!
It took about twenty minutes of troubleshooting for me to realize my mistake. I had been researching different sizes of steppers like the larger Moons MS17HA6P4100 out of curiosity. I'm certainly no expert, but I learned enough to realize that the 16mH inductance of the larger Moons would make a poor fit for our humble 3D printers. Intending to order the proven MS17HA2P4100, I figure I then copied the model number from the wrong browser tab, hit Buy It Now, and then just threw them on my printer like a fool. Copy-pasta strikes again!
Aaaanyway, anyone know a good application for some big, high-inductance, 0.9° steppers? 😋
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
I made an exceptionally dumb mistake last night. It's also kinda funny, so I figured I'd share so you can laugh at (and learn from) my error.
I was having no luck taking comparison photos and decided to try 0.9° steppers. I have a MK3S upgraded with a BNBSX extruder using a Moons pancake, Trinamic steppers on X & Y, and new Misumi rods and bearings. I'm really happy with the improvement -- especially the BNBSX! -- but I'd ordered some 0.9° Moons earlier with a long lead time that just arrived, and thought I'd give them a shot.
They went on easily, and the firmware worked the first time. But when I started moving them, I was greeted by a chorus of loud chirping noises, like a nest full of demon birds escaped from hell and possessed my printer. Ahh!
It took about twenty minutes of troubleshooting for me to realize my mistake. I had been researching different sizes of steppers like the larger Moons MS17HA6P4100 out of curiosity. I'm certainly no expert, but I learned enough to realize that the 16mH inductance of the larger Moons would make a poor fit for our humble 3D printers. Intending to order the proven MS17HA2P4100, I figure I then copied the model number from the wrong browser tab, hit Buy It Now, and then just threw them on my printer like a fool. Copy-pasta strikes again!
Aaaanyway, anyone know a good application for some big, high-inductance, 0.9° steppers? 😋
A DIY benchtop CNC milling machine? 🙂
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
I made an exceptionally dumb mistake last night. It's also kinda funny, so I figured I'd share so you can laugh at (and learn from) my error.
I was having no luck taking comparison photos and decided to try 0.9° steppers. I have a MK3S upgraded with a BNBSX extruder using a Moons pancake, Trinamic steppers on X & Y, and new Misumi rods and bearings.
How are the Trinamics working for you?
They went on easily, and the firmware worked the first time. But when I started moving them, I was greeted by a chorus of loud chirping noises, like a nest full of demon birds escaped from hell and possessed my printer. Ahh!
It took about twenty minutes of troubleshooting for me to realize my mistake. I had been researching different sizes of steppers like the larger Moons MS17HA6P4100 out of curiosity. I'm certainly no expert, but I learned enough to realize that the 16mH inductance of the larger Moons would make a poor fit for our humble 3D printers. Intending to order the proven MS17HA2P4100, I figure I then copied the model number from the wrong browser tab, hit Buy It Now, and then just threw them on my printer like a fool. Copy-pasta strikes again!
It's really hard to believe that the inductance would be a major factor. The LDO's that come from Prusa with the printer are specified at 13mH±20%, so they can have up to 15.6mH of inductance and still be in spec. Plus, inductance helps with smoothness of motion nicely at low speeds, it's only at high speeds where it starts limiting current (and by extension torque).
Are you sure you got the wiring right? Have you tried to boost the currents up? Big motors have a bigger cogging torque and need more current to start spinning.
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
I have a MK3S upgraded with a BNBSX extruder using a Moons pancake, Trinamic steppers on X & Y, and new Misumi rods and bearings. I'm really happy with the improvement -- especially the BNBSX!
Paws up! Print results on mine are SO much better than from stock configuration. They don't even look like they are from the same machine.
Good to hear someone else is enjoying the transformation.
I am particularly happy that you did each pre-requisite change that makes each subsequent upgrades meaningful.
1. Upgrade to smooth running, less shaky bearings & rods - remove random noise and smear. (That random noise hides a world of sins, but also reduce effective printer resolution)
2. 0.9 degree motors on X and Y - remove VFA's
3. Geared extruder (BNBSX) - remove extruder inconsistency
4. Change X and Y drive pulleys to BALITENSEN 10 mm wide units - reduce belt intercalation artifacts. Surprisingly, these intercalate with Gates 2GT better than even pricey E3D's drive pulleys. https://www.amazon.com/gp/product/B07BH26P2D
Each stage must be completed before the next becomes important. It's like doing auto paint. You cannot jump to a finer grade paper before finishing work with the coarser paper. Skipping the first two and going straight to the extruder is like not sanding and going straight to polishing compound.
Doing each pre-requisite, your end results are sharper prints with less surface artifacts. The Prusa becomes a totally different grade machine.
.
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Hey Vojtech,
The Trinamics are quite nice, and if I hadn't already placed orders for some 0.9° steppers, I might have just stuck with them and focused on optimizing elsewhere. They easily beat the 1.8° Moons MS17HD2P4100. The surface finish looks nice and clean, similar to some of the photos with 0.9° steppers. The one XYZ cube I printed with those over-sized Moons (and a pair of earplugs) might look just a hair nicer (until the layer shift), but man, it's close. I've mounted a pair of StepperOnline 0.9° OMC 17HM15-0904S this morning, since it sounds like they're the pair to beat, and a Benchy is printing as I type this. First impressions: I miss the quiet of the Trinamics.
Speaking of: The Trinamics are some of the quietest motors once I got my belts tuned correctly, but over-tension those belts and they'll scream. In the spirit of sharing my mistakes: I briefly made the error of chasing the numbers from the Belt Status screen before stepping back and thinking about how it works -- different motors are going to have different target numbers. I went back to tuning belts by feel and that's when I really started liking the Trinamic steppers. There's the occasional talking-robot sounds and squeaky accelerations, but I'm still pretty happy.
I also have a smaller Trinamic to try with the BNBSX extruder, and the Bunnies were kind enough to design a custom gearbox to fit, so...more to come! (and photos, once I figure that out...) It seems like my problem these days is that I have to stop experimenting with the printer in order to print parts for my next experiment!
For the inductance and those larger Moons steppers...that's interesting. I used the same cables as I did with other Moons motors, plus both X & Y steppers made the same noise. I doubt it's a wiring issue, but definitely worth a double check. I'll also break out the multimeter and make sure there's nothing off with those wiring harnesses now that I've unplugged and re-plugged them a few times.
When you mention tuning the current, are you thinking the first two entries in these arrays here? Any suggestions for target values? My day job is software (well, "Site Reliability Engineering" these days), but most of my electronics experience is with audio recording gear. I'm still coming up to speed with stepper motors and micro-controllers. Would it also be worth revisiting any of the other TMC values, like the PWM grad/ampl or toff/hysteresis/etc.? I know it's overkill, but I'd love to get these steppers working if I can, if only for the fun of learning new stuff.
Again, I really appreciate your help and patience. This is an awesome community!
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
...and hey guy-k2!
Thanks! I'm having a blast with this process. Thanks for laying it out so well!
I think you're absolutely right about how one improvement lifts the veil and let's you see the next problem. It's been fun seeing that in person, in real time. It's also cool walking through each step, and seeing exactly what each artifact looks like, what impact it has on the print, and then what we can do about it.
Oh, and I realize I didn't list them, but the Balitensen pulleys are on there, too, and I'm so incredibly glad you pointed them out. Even without improving the print quality, I love that they make it so much easier to experiment with steppers. With the extra length, I haven't needed to grind a single motor shaft!
It's funny, I look at how much the surface quality has improved through this process, and I have to make myself think back to my first print. I have a friend who's an avid 3D printing enthusiast, owns several printers, and has been tweaking for years. I showed him my first Benchy off the stock Prusa MK3s kit and his jaw dropped. I suppose part of why we're able to take improvements this far is because we're working from a pretty darn solid starting point!
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
@guy-k2
I love this thread and have been following along.
At the moment I have a slightly modified Bear MK3S. Changes I have made:
- BMG extruder
- Linear rails (12mm mounted on 20x20 vslot) installed on Y axis. Working on finalizing the design then X axis will be next.
- OMC motors on X and Y.
- Trialing vesconite bushings on X axis.
- 10mm pulleys on their way from Amazon.
With the linear rails testing I’m constantly fiddling with the firmware in order to pass self test and XYZ calibration so mine if heavily commented with my own observations and I would like to keep it. Your firmware has a host of changes - some of which I have no idea what they are and why you changed them. Would you mind giving us a short write up on the whats and whys? That would be greatly appreciated.
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Forgot to mention something else regarding the firmware, I'm seeing two issues. First one is during homing, when moving X to home it tries to keep pushing once it reaches the far left, the printer then reboots. I do recall you had the same issue, unclear as to what to change in the firmware.
Second issue is random crashes, although the the printer might be making a slow move (like on an outer layer).
Any suggestions?
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
Did someone compare the Trinamic QSH4218-51-10-049 against the Moons motors and the stock ones? Maybe even a photo comparison of print results?
Also, are there trinamic stepper motors that could replace the extruder motor or should I get the Moons one for this? (My extruder is overheating to the point that I cannot print PLA for longer than 5 hours - it definitly needs a cooler motor)
RE: Stepper Motor Upgrades to Eliminate VFA's (Vertical Fine Artifacts)
@albe-g,
I did have a few random crashes with the OMC 0.9° steppers on X & Y. Eventually, I found that my x-axis idler was jammed. It was not rotating smoothly, and was also deflecting the belt to the side, making it rub against the pulley's flange. I was planning to eventually replace it with a toothed idler anyway, so I went ahead and made the swap. That cleared it up, and I had no more crashes over a modest couple hours printing. I haven't had any issues with the homing, though -- that worked fine for me right out of the box. Ironically, it sounds like the opposite problem. In one case, the Stallguard threshold is too low, and in another, too high.
@fynn-a,
Yep, I spent a good week or so with the Trinamic QSH4218-51-10-049 on X & Y. My entirely subjective, very preliminary, totally amateur opinion is that they're better than the 1.8° Moons MS17HD2P4100 on those axes, and close to the 0.9° OMC 17HM15-0904S with linear correction set. Unfortunately, I've been struggling to take photos that I feel are honest representations of what I see in person (and also not, you know...blurry blobs...). The 3DBenchy folks put together a printable smartphone studio that I'd like to print and try, but I spent this weekend diving into TMC2130 tuning instead of printing all the stuff in my backlog. 😉
I suppose you could use the Trinamic -049 on the stock extruder, but I personally wouldn't. Those Trinamic steppers are pretty beefy. It would add 50g over stock (70g over the Moons) and extend the extruder's moment arm even further out in front of the smooth rods. The Moons MS17HD2P4100 worked for me on the stock extruder, and did make an improvement. It cleaned up the surface a bit, and I ran several 15-hour overnight prints without any issues. I found a much more significant difference when I switched to a geared extruder, though. I'm using the Bunny Science BNBSX "Bunny and Bear Short Ears" right now and I'm very happy with it.
Speaking of driver tuning: I never did find a satisfactory configuration for those over-sized MS17HA6P4100's. I checked the wires, swapped for new wires, and tried only one at a time in case there was an odd defect with one. I tried various combinations of Tbl and Toff, but the noise stayed around even with configurations that should be far into the infrasonic.
Tuning the current proved interesting, though. It made no difference at all until I went from 31 to 32, at which point the noise just disappeared. I'm guessing it has something to do with the change in Vsense from 1 to 0 at that point (`tmc2130_setup_chopper(...)` in `tmc2130.cpp`), but I don't understand why that would make a difference. At any rate, the drivers hit thermal shutdown about 40 minutes into a test print, and I wouldn't want to run them that hot over the long haul. It's all starting to make a _little_ sense, but I'm still a long, long way from an expert. I certainly don't understand how the current configuration works in the firmware.
Also: thanks for those extra M919-M924 commands behind the `TMC2130_SERVICE_CODES_M910_M918` #define, Kuo! They definitely made it easier to iterate more quickly.