How Fast Can Your Printer Go? Stepper Motor 101
HTML-код
- Опубликовано: 29 сен 2024
- Thank you to David from Voron for help with the equations, @JoshMurrah for help testing and validating the model, and @JaredC01 for testing pulley sizes!
Follow me on Twitter:
/ eddietheengr
Excel Database Here:
github.com/edd...
Jared's RUclips Channel:
/ @jared01
Josh's RUclips Channel:
/ joshmurrah
That's so awesome man! Thanks a LOT! Very well explained so people like me who are not knowledgeable in this area can really understand all of it. Nice video. I'm subscribing
Awesome description. Now we need nonlinear acceleration in klipper (PWL model?) so that for lower speeds you get high accel but acceleration tapers off to stay below the curve to get ultimate top speed. 1/2mv^2 but power (delta energy) remains constant.
We really need that feature !
you might be able to do that with a macro
Awesome!
Thank you. I'm just sourcing my parts for a V2.4, and this helps explain a lot. Keep up the good work.
Hello, everyone.
This is your "Daily Dose of Internet".
That's all I could think of when the intro started 😂
I running the LDO 42STH48-2504AC with 48v and 1.768a with external fly 5160 drivers
And I must say the ldo motors goes very hot, can you tell me how I can get cooler? 1.5a maybe? Or is not problem for the motors? My printer stays in my print farm and they must work for many hours the day
Love this stuff. So many tutorials, and discussions just gloss over why parts are picked. It's nice to understand more about what the specs mean of a motor, and how it actually effects the performance. Then we can make informed decisions on what motors to get for what we want to accomplish. Nice.
Thank you SO much!!! I'm completely new to the 3d printing/performance stepper motor game. In 23 minutes, you explained everything I've spent all day for the last 4 full days piecing together on my own (why do I always find the good stuff after I did it the hard way?) AND provided a spreadsheet that's going to keep me amused for days! I do have one question: What is the number beside the pulley size (sorry if I missed it in the video)? Pulley diameter? Not travel distance, because I have a 32mm pulley and a 2mm pitch belt, so it travels 64mm per revolution, correct?... it is saying 10.186 (rounded).
Now I just have to take my printer apart to way the gantry and belt so I have an accurate mass #. Thanks again!.
Glad you found it useful! You're correct, the distance traveled per revolution would be 32*2 for your case with a 32T pulley and 2mm pitch. The ~10.186mm value would be the radius of the pulley, but I don't actually think that value is used and should probably be just deleted to avoid confusion!
Great video! I dont even own a 3D printer but loved the whole video on my journey to a 3D printer. Thank you.
Thanks a million for putting in this work. I'm designing an IDEX printer, and this will help greatly with selecting motors and optimizing my design overall.
Ay 6:00 you showed a chart which has a flat curvu up to let's say 260 mm/s
Actually typical unductance of the coil will start reducing the current right away. At 250 mm/s you have 800 us for a full step and 1/6...1/4 of this time you will have a current growing from zero to the given current. So the mean current will be 90% of the steady stayt by 250 mm/s or so.
Awesome work as always. You have a way of boiling things down to the simplest language, and the most basic concepts, so even I can digest it. Keep the videos coming.
I've tested quite a few 0.9° steppers and all the ones that scored good on the calculator performed worse than the lowest scoring one. The only thing I can think of is the permanent magnets and air gap are different and pay a large role in the performance.
AWESOME video Eddie! Thanks for doing this!
Just a small correction! With active cooling those 2209's can easily saturate the LDO-42STH47-2504AC's that ship as default on our V-Core 3's. They'll do 2A RMS (2.8A peak) when properly cooled and i've successfully run mine at 1.77A RMS for months now without any hickups whatsoever (right around the rated 2.5A peak for the 2504AC's). 5160's are not required :)
Contrary to popular belief it does not take much effort to cool those drivers, a fan blowing in their general direction is plenty :)
What about in an actively heated chamber? I actually had my motors on my Ender 3 Pro give up and throw an error at me when I got the chamber up to 100F / 38C.
I am now designing a water cooling system, and working on splicing wire extensions to get my Pi and MCU out of the enclosure.
Is this wild overkill?
@@BigHonkinGoose was talking about drivers, not steppers. Get some HT steppers and move your electronics out of your chamber as planned. It’s cheaper and easier than watercooling :) not as cool though! 😂
@@BigHonkinGoose I have an Ender 3 V2 that's running inside a 54° max chamber for ASA printing. My mainboard is cooled by a 6025 PC fan and my PSU has a 9225 PC fan installed (and doesn't even run all the time, indicating the PSU isn't even sweating). My steppers run at a lower Vref than usual but still manage move the axis at 100mm/s, and they also have passive heatsinks installed. The printer has been doing great for quite some hot hours so far. Watercooling for a sub 40°C chamber is wild overkill.
Hi, Eddie!
What does "N-cm" mean (at 0:30)?
Looks like it's not on engineering language :)))
Thanks for the explanation and sharing the spreadsheet Eddie! I've been looking to make something like this for a while
So that's why I could do 20k at 200mm/s but barely 8k at 300 and turning up the current did not help. I was really puzzled by that.
Definitely! Once you get in the sloped "inductance limited" region, there's not much you can settings wise to improve the performance other than reduce required torque, increase supply voltage, or find another motor
thanks for the explanation and for the documents
Great stuff! Do you plan on do an experiment to see where actual skipping starts vs the theoretical one?
Yes! I have done some of this testing over the last ~6 months or so. The numbers generally matched up with the theoretical models, though now that I know more I realize getting the backemf/Vgen calculation correct is critical. Maybe in the future with some help I can get measured Vgen vs. speed to improve the accuracy a bit more.
Awesome video👏🏻 Thx for sharing all valuable stuffs too!
That's some great content! Love the work and thanks for sharing!
Thank you! Glad you found it useful.
@@eddietheengineer I wonder if 2 motors double the speed.. My reading is 1350 but sheet say 700
Will 2 motors double the speed? I have readings saying 1350 but your calculations say 700 ish
Can you share what motors, stepper settings, and driver input voltage you are using?
in regards to pulley size ... tooth count has no real bearing on pulley size .... I can have a pulley six feet in diameter with 20 teeth ... how are you translating that to actual pulley diameter ??
Great question! Since GT2 belts are pretty much the standard in 3d printing I am assuming a 2mm tooth pitch, so 20T would be 40mm per rotation!
damn, cant get those working on Excel or Sheets :/
Error
Unknown function: 'SingleCoilTorque'.
Unknown function: 'StepperPower'.
Unfortunately you need to enable the macros for them to work 😢 I wish there was an easy fix for it!
Wouldn't it make sense, to make the acceleration be adjusted dynamically to the actual speed, instead of always being constant? I mean lowering it in the region, where the turque begins to fall off.
Or does Klipper already do that?
That's an interesting proposal! It would help to taper off acceleration as you reached your peak speed, but it would have to be a custom value per motor. You can see how Klipper implements kinematics here: github.com/KevinOConnor/klipper/blob/master/docs/Kinematics.md
It looks like most of the time it assumes a constant acceleration
This is really great content and really good explanation. Time to ramp up the amps in my config and see how it goes :D
Is there something wrong with the file or doesn't Sheets just understand some of the formulas? All of the fields just say "Error, Unknown function: SingleCoilTorque.
Have the same problem. :(
Wow, great vid Eddie! Love your recent content, please keep it coming!
Thank you!
As always good. Nice work.
I Know this is an old video But I really appreciate all this information. The spreadsheet is amazing btw. So LDO is releasing the LDO-42STH48-2804AC this month or beginning of the next, I found the datasheet and added the correct info to you database. Everything worked as it should. But then I went deeper and figured out we are not capable of running 60v to our steppers.. And well it made all kinds of things different. The ldo-42sth48-2504ac just flat lines and the graph don't show the end or at least where it falls off :D Do you mind updating to push the graph to 1500mms?
Have you tried tuning the new register in the tmc2160/5160 that's supposed to help with mid-range resonances? It's called TPFD.
I'm asking because I saw some of your answers regarding motor noise and besides the extra current and voltages allowed this is one of the extra features of those newer drivers and I was wondering how effective it was.
If I turn up the hybrid stepping limits? It can go pretty damn fast. That said... I prefer to run around 140mm/s travel with 400 accel for accuracy's sake. The machine is quite stiff, using all linear block slides but im picky about print quality.
If back EMF is causing problems like your LCD coming on that is a board design issue. I would throw out any board that is designed so badly they couldn't even put in basic safety mitigations. Ie. a resistor and diode for < 5 cents.
Agreed! I think this largely has been resolved with newer boards, but I seem to recall it being an issue a few years back, possibly with RAMPS boards?
@@eddietheengineer sadly alot of boards will still do this i even can make my duet board do this lol
Great work as usual. Thanks!
Really interesting stuff 👍
Eddie
Could you take a look at the Moons-MS14HS5P4200 (HS not HA). While it is a Nema 14, it would seem to me to be a viable candidate for AB use on a "normal" printer.
Ok, i have a SKR MIni E3 2.0. It uses 2209 tmc stepper drivers and my sys is 24 v. right now, i run creality steppers rated at 1A. This stepper 17HS19-2004S is rated at 2 A. The SKR doc mentions a power chip that can deliver up to 2.5A. Just wondering, is that the limit for the stepper drivers too or is that power chip used to drive lights and maybe external boards like Raspberry pi. Botttom line: Will that board handle this stepper? I want to do corexy and there would be 2 of them driving x and y
TB6600 can go ~50V as well, got a few from MKS years ago on big individual boards and heatsinks. drive voltage and stepping instruction are separate so very easy to have otherwise 12 or 24V and give the TB6600 36V, 48V whatever :)
Very good explanation!!! Thanks bro! You are very underrated. Keep making awesome content like this :)
Excellent video!
Servo's maybe better
You are correct that Back-EMF has a large effect on the performance of the motor. The inductance (and total resistance) also affects the performance of the motor. It causes a (speed dependent) decrease in the current produced by a given voltage! There is also a lag between the voltage applied and the resultant current. In fact, at high speeds, the motor can lose stability. Bottom line - the motor is a fourth-order non-linear system whose response may be quite different than what you are hoping for. This is why people do computer simulations of these things.
thanks again for the education Eddie Ive scored myself LDO motors that feature in the video from RatRig as shipping cost & Tax's to the UK from OMC or nay where else for that matter is more than the motors, so these seem the best option for myself sourced build at this present time.
can you talk about acceleration in a slicer or firmware and maximum torc of the motor and is there a calculation to determine the max acceleration a particular motor can handle. another way of saying it there is max torc vs Vref and then max work vs acceleration settings.
Is there an impact on the graph if we are using SpreadCycle or StealthChop?
You should check out my video that shows the chopper timing and I have the spreadsheets linked in description. You can change the pwm freq and hysteresis plus decrement values etc to really optimize how the coils energize and reduce heat output
Brilliant, I've been looking into building a high speed printer and this is very useful for understanding what equipment to buy to have a chance at achieving he crazy print speeds
excel is bugging can't show motor calculations only names of the motors.
awesome info, very usefull! I got one question tho, while looking at the excel sheet I can't find microstepping and interpolation settings. I was wondering how microstepping and stepper interpolation impacts torque for a given speed, would you be able to tell?
Just crazy good tutorial!!! Love it, I'm subscribing
This is awesome Eddie!
Wow, this is an absolutely priceless video. Thanks so much for putting this together.
Awesome content!
Wow I modded my printer to use 48 volts, but never knew that it gives me such a huge overhead! I will try out my 0.9 steppers now, because I can‘t imagine going that fast!
Hanzhen harmonic drive gear , industrial robot arm gear reducer,
This is some brilliant & informative work, thank you.
I have opened your Excel file & there appears to be an issue with the formula for the stepper torque & power cells. I'm afraid I still don't understant it enough to try to fix them.
Yes! Unfortunately the excel chart needs to have macros enabled since there are custom functions in the file. Eventually I want to port that over to a more user friendly interface so it doesn’t require this!
This cleared up some questions on speed/accell tuning. Really well done video and spreadsheet. Building a RR Vcore3.
I appreciate the great work and explanation. It really brings insight into the stepper and motion systems.
Thanks for this. I was originally going to go with a 60V system, but after using your spreadsheet with a Nema23 I was looking at as well as throwing some values at Cura to gauge the point of diminishing returns, I'd just do a 24V 1.8A RMS system instead at 16K accel. I doubt I'd want to / I could accelerate 2kg of mass any more than that anyway without serious shaking.
For sure! There's definitely a limit where while you can accelerate faster--it's maybe not all that useful long term. I had a set of 48V 5160s and 48V PSU, but then I ended up selling them and going back to 24V because I really don't go faster than ~400mm/s.
thank you for this video! ive been working on designing my own printer and, without a formal electronics education its always so nice to find such well put together resources like this! off topic, but i wanted to ask; what mod is that for your afterburner that has the 5015 fan?
Hello! This is the Voron Stealthburner toolhead :)
Could you briefly explain what to do with the Max Power (W) field? I didn't see it in your video but it is on your current sheet version.
That's a great question! In general, this is a rough approximate power consumption assuming the stepper motor windings are energized and holding the motor at a fixed position with a constant current, so power = current^2 * phase resistance * 2 phases. It's mostly useful for really small motors (like the NEMA14 pancake or NEMA17 pancake motors) that can have a tendency to overheat if their generated power/heat is too high. For the small NEMA14 pancake motors that ends up being approximately a 3.2W max power. This means that in general, you want to find a motor that has the highest torque for a given resistance--higher resistance means higher power and more heat for a given current level.
For larger motors with larger surface area, that number can be substantially higher--I haven't done any real limit testing though!
Great video! Really helped it all make sense
I grabbed the excel file but it doesn't seem to work, is there something I need to do? I get no graph and I do get a grid with a bunch of errors listed.
Check to make sure to click “enable macros”, since most of the calculations are custom functions I defined to simplify everything 👍🏼
How do you size your 48v psu for two steppers with 2 A max? Mine is 150w, which I calculated with some formula from the Internet, but I didn‘t really understand how I got to this value
150W should be safe for most cases-at 48V I still recommend running pretty low current just because we don’t need higher current. I’m not exactly sure the “official” means of calculating power, but I’ve seen the same numbers you have and I think they are a bit excessive. I used a 100W PSU for a bit, and am trying a 75W for this next test to see if that’s good enough.
@@eddietheengineer Most of the formulas are using the rated current whereas we are running less than half in practice. That might be the factor
This is super interesting and useful! However using the spreadsheet to do some quick math it seems different from my experience: I should be able to go as fast as 400 mm/s with an acceleration of 20K, but in reality I had some layer shifts at 200-250 mm/s with 6K acceleration. Is it more likely that I did the math wrong, or that there are problems on my V2.4-300?
Which motors do you have and what current are you running at? Depending on that, plus assembly/alignment/rail lubrication it’s possible there may be some extra resistance that may make the model under predict the torque required
@@eddietheengineer I have two LDO-42STH40-2004MAC, running at 24V and 0,8 A (it was the default config for the 2.4), and that is what I set on the spreadsheet. This particular stepper is not present in the database, I copied and pasted the 2004MAH line and I changed the values that I have found for the MAC, even if there doesn't seem to be much difference between the two. I just had a couple of minor layer shifts in several prints at 6K, so I went back to the default 4K just for the peace of mind. Oh, and if I remember correctly both times it was with some tight gyroid infill (40-45%). Other than that it has always been flawless!
Oh...My...Yikes. Haha... oh... no...
This is great 101 on steppers, probably worth having it in the title, so that it's more searchable and can help people who don't know you yet
Wandering if it's worth switching to 32v and ab motors to 0.9 on V2.4
Do 0.9 produce less vibration at same speeds?
The title is tricky, I don't post enough videos to know how to maximize views per the RUclips algorithm :D but I agree, adding a bit more information would be helpful. 32V definitely does help your top end speed if you're really pushing it, but in general unless you're trying to exceed 600mm/s it's not super critical. The 0.9 degree motors can produce less vibration, but it's almost that the natural resonance of the motor changes to a different speed range. Some 0.9 motors are actually pretty loud, so it's not a given that they are all quieter than 1.8 degree motors.
@@eddietheengineer I agree the ldo 60mm .9s are loudest I have ever heard
sorry if it doesn't have as much interrest, but the most quite printer is more what I would be after (unless the speed drop to unreasonable values). Or the printer set-up for something of good quality, quiet and as much speed (i.e. overnight for a full v0 volume?) as possible without compromising on the rest.
This is a great question! I also value quiet operation, but I’ve found that there are so many variables to what makes a motor quiet or not-and to be honest I have no idea what parameters go into that. I’ve ended up with 17HM19-2004S1 motors on two of my printers running at 0.6A because they are quiet-but some other people say they are loud on their printers. This may be printer structure/size/belt tension-who knows
@@eddietheengineer Thanks Eddie for the feedback. I was building a 1.8. Love the new Trident. It solve the issues I was worried about with the 1.8, while keeping the spirit (printing on the top part of the printer so room temp of 50 for ABS is reached faster and maintained more easily than on a 2.x). Great design. My plan with the 1.8 was to make is a Wall-E printer (in term of the look), not sure I'll be able to pull that off with the trident. Thanks a lot for all you give to the community.
Great info, thanks Eddie!
extremely well explained!
how Voltage work on stepers? On spreedsheet say work at 2,5V and on ratrig page 3,1V but are nothing close to that 48V. Is not working like led light? you can put driver of 56V like 5160 and he work at that voltage?
That’s a great question! The input voltage to the stepper driver is somewhat independent from the output voltage. You can think of the stepper driver as converting the voltage source to a current source-it will modulate then output to control to a specific current (what you specify in the configuration for the driver). Higher input voltage doesn’t change the steady state current when the motor is stopped, but can help transient behavior while the motor is spinning, especially at very high speeds
So the motor get 3V or 24V from the stepper driver like 2209? or i am lost?
The 24V input is essentially “Pwmed” down to 3V-it’s a bit more complex than that but it’s a good analogy
do drivers like tmc5560 or say higher voltages like 35v meaningfully reduce noise?
I don’t think noise will be reduced, it’s more likely it will be increased due to faster speeds!
@@eddietheengineer I'm talking about the same speed/accel. What about quality of the motors? Say 8$ stepper online ones vs Chinese generic ones vs 20$-17$ LDO ones
Would love to watch a short video like this about Z movements, regarding V2.4
To get a better understanding of it
Definitely! The speeds and math are basically the same, and in general with the 5:1 gear ratio there's a lot of extra headroom at "normal" Z speeds. The one unique part for V2.4 is that you want to pay attention to the detent torque listed on the spec sheet--if you go to a motor with lower detent torque, the gantry will fall when the steppers aren't powered.
@@eddietheengineer Is it still something you want to do or nah? :p
So, noob question here: can any of the backEMF be reduced or even eliminated, or is it required for some reason, such as being sensed back at the driver on the board?
That’s a great question! Unfortunately the backemf is due essentially to the inductance of the motor windings-we are switching the current through each winding on and off really fast! So the inductance tries to resist those transient changes. If we could make a motor with 0 inductance that would be great, but just due to physics there will always be some!
@@eddietheengineer I guess I should have worded my question a bit differently. I understand backEMF is naturally generated in the motor. What I was specifically wondering is if something like a diode could be added to the wiring to avoid backEMF feedback to the board if moving something for maintenance while the printer is off without interfering with functions like position sensing or StallGuard.
Great video! Thanks! 🙂
Which properties are important if you want to find a motor that is as quiet as possible?
This is a huge one! I haven't figured this out yet but if you or someone else reading this has pointers I'm all ears. I used to think 0.9 degree steppers were the answer to a quiet printer--but now I realize that sometimes they are just as loud or louder than their 1.8 degree counterparts!
@@eddietheengineer Yes, I definitely have some loud 0.9° steppers. 😅
I will try tuning the drivers when I have some time.
This is perfect and exactly what I was looking for, fantastic video! But it seems to be missing a very important motor that most people seem to use, which is the E3D High Torque motor?
That should be easy for me to add! Any other motors you’d like to see?
@@eddietheengineer brilliant that was fast. Not really at the moment, those are the two motors I'm looking at getting.
@@benclimo461 I looked up the High Torque motor and the datasheet looks like it's the Motech motor that's in the database! the MT-1704HSM168RE
@@eddietheengineer hi just wondering if there's a way to work out the speeds with a delta printer? As all motors are supporting the weight of the effector and moving at the same time.
any thoughts ont he 2 amp moons 60mm steppers they seem to be much better to have on this thant he 1 amp high inductance version
Do you have a specific part number for me to look at? Curious which Moons motors you’re talking about 👍🏼
I tried posting them a few times it keeps blocking my comment. Also I downloaded some torque curves from manufacturer and they don't match at all with the curves this plots . Can you take a look at the omc and moons curves and see what you think cause they are very diff
very well done!
I tried using this and noticed some strange things when I tried to put in the parameters for the LDO 42STH40-1684AC motors that I have. It seems that they have a very flat torque curve and can outperform most of the other motors in the database at higher speeds. I am not quite sure of the reasoning for this. I find it a little hard to believe considering many of the larger 60 mm LDO motors seem to look better on paper in almost every respect.
Thanks for pointing this out! I agree something looks funky there--I'm looking into it to make sure there's not a bug in the equations somewhere
I've done a bit more digging, it looks like the V_Gen estimate is from this website: www.oyostepper.com/article-1102-Back-emf-due-to-rotation-of-stepper-motor.html
=2 * pi * rev/s * (Holding Torque/(100 *sqrt(2)/Rated Current
This means that the holding torque and rated current both factor into the Vgen calculation. If you change the rated current to 2A and the holding torque to 59Ncm, you end up with a slope very similar to the rest of the motors, which implies that this component/estimate is why it looks different. I'm not positive if there is a better way of calculating this term from the data available on a datasheet, if someone knows I'd be really interested to find out!
@@eddietheengineer I am an electrical engineer, but I do not work much with motors. Although, it does make some sense that more current = more back emf assuming that the motor is not being driven to saturation. Torque with motor is Amp * turns. Apparently, the LDO 42STH40-1684AC must have more turns on the winding but pulls less current. The higher inductance of the 42STH40-1684AC also seems to indicate this. I am confused as to why the inductance is not being used in that equation somewhere.
@@dakotaadra1044 I am confused as well! I think if we knew more about the motor construction it would be possible to calculate a more accurate Vgen, but since most of that data is not readily available possibly this estimate is okay. I’d be curious to see what kinds of speeds can be achieved using that motor since it has a rather unique slope vs the rest-if that slope is just an artifact of the estimate being wrong, or if the estimate matches up with reality.
I’ve considered building an experimental setup to measure a bunch of steppers using the more common 3d printer control boards and various driver settings, but that’s a pretty in depth project and would take a lot of time!
@@eddietheengineer It is kind of amusing that I stumbled across this. I ordered those motors on accident for my 2.4 build. I am still a little worried about the holding torque being too low for the z-axis and the fact that the shafts are only 19 mm, but now I am considering keeping them. Do you think either of these things will be an issue?
Edit: It looks like I can turn the pullies around on the z-drives, but I am unsure about the A and B drives.
google can't read the sheet
I apologize about that! I’ll have to check and see if Google sheets allows custom functions
@@eddietheengineer i see thanks
TBH in real life use, it's really the flex in your parts and rigidity of your system giving up before current/stepper power AFAIk.
unless your motor isn't really bad or really small (pancake on xy!) it'll not be the bottleneck.
I can do 10-15k accel at 500mm/s moves very easily on my printer (used to tool change at those speeds, but the noise became hearable 2 walls away with the door closed lol. but the ghosting I get is a bit too much