Hi guys, after reading comments here are two things: 1. Plenty of people saying they don't have layer shifts, plenty saying they do. The ones I showed at the start were genuine for me, not manufactured for this video. If you don't need these, that is great news, but it doesn't mean they are useless for everyone. 2. I was confused with the price by the '4PCS' in the description, but I see now they are referring to the mounting hardware. They are in fact $25 each for the units shown in this video, my apologies for the mistake in the summary. I will see if UA-cam let's me crop that part out as to not confuse people. 3. The endstop lever went back on without issue. My printer is proud to have 'taken one for the team' in the pursuit of testing and thanks you for your concern.
@@ColinWatters My upgraded A8 sometimes skip steps when I'm printing materials like ABS and print start to lift corners. I didn't have this problem with PLA or PETG.
Today I received two of these. The DIP switches are all set to on by default - the inverse to what your drivers looked like. Do you know where I can find a manual on what they do and how they should be set?
So much anxiety watching you whack the printer with a mallet. Amazed to see the Closed Loop system in action, very impressive. Interested to see what the maximum print speed is now.
So glad you did this video. I have been waiting on a proper review of these closed loop stepper units for 6 months. Seems like every printer should come with this type of control of the steppers and maybe they will going forward. Thank you again for putting the time so we can all see what is involved with using these units.
While I am unlikely to use these on my 3d printer, it might be really great for a CNC mill where the forces are much higher and the chance for skipping steps greater. The price for motor and driver circuit is really impressive!
@@jothain True, but servo motors are normally very expensive as are the drivers. Bigtreetech is selling a complete solution for the price of the motor.
Yesterday I successfully upgraded my home-build H-bot printer following these instructions. Just want to share couple of things I learned during the upgrade: - At 8:29 Michael talks about removing jumpers for micro stepping. This is optional because cable adaptor is not using these pins. The only 6 pins it connects to are VMOT, GND, VDD, EN, STEP and DIR. - Pin connectivity allows me to answer my own question asked earlier here - there should be no problem connecting 2 motors in parallel. Obviously you either need a board with 2 stepper sockets wired in parallel or buy/make a cable splitter. - At 13:45 Michael mentions that he increased steps per mm from 80 to 100. In my case I had to go from 320 to 410.214. So these values are printer specific. - It worth mentioning that once motor is configured, you don’t need LCD and can remove and store it. Once again, many thanks to Teaching Tech for yet another extremely helpful video.
Oh my freaking god as a side note at 9:12 the way you removed the old motor while keeping the belts routed in tact was absolutely genius! I could use that same method with initially routing the belts to begin with! That is always such a pain in the ass to route the belts through that spot so doing it your way will make it so much easier. Just wanted to let you and everyone know there are so many things you can learn from watching these videos!
I'm so glad I found this video, I've built my own corexy printer 300x300x400, linear rails on all axis and a MKS SGEN L V2.0 main board. I was given a pair of BTT S42B steppers, no matter what I did I couldn't get them to move until I watched this video, those 2 small changes in the firmware solved everything. Thanks.
@npgoalkeeper _ TMC2209s aren't closed loop. They will not auto correct for any movement. All they do is give you tighter control over movement and quieter motor movement. The TMC2209 is the easier upgrade, but it isn't better.
As someone who has battled a lot with layer shifts in the past, it's amazing that closed loop motors have taken so long to hit the consumer printers. With the Creality K2, it looks like it's finally happening. My original "fix" was changing stepper drivers and running them at close to the maximum my board and driver are capable of, so the print head can power through a lot more resistance. It's of course not comparable to closed loop, but it does minimize the risk of layer shifts (for me). But I'd be a lot more comfortable having closed loop steppers - but alas, my printer's mainboard is closed source and doesn't allow for such customizations.
I like the idea of using an observer to read the motor coil current and inductance to determine if it stepped or stalled. It's an all electronic solution with no moving parts or metrics. Thanks for these great videos!
Excellent Michael. Good stuff... the only thing worse than the magic smoke is that deep fried ‘scent’. Back in the day I was calibrating /troubleshooting a HP DC p/s and slipped with the probe... then all you hear is OK ! who let the smoke out!! Good times LOL
3 changes.... Removable drivers. Spi or uart connections. Nema17 and Nema23 sizing. Tmc 5000 series drives with closed loop, 20amp feedback for cnc, cheap and easy, love it. Load your 2208's or 2209's that you already have, love it. Get rid of the screen and use spi or uart.... better cooling, better control....
I saw these on the BTT website and was hoping someone would do a piece on how well they actually work. I'll keep these in mind for whenever I get around to building that coreXY machine.
I too have seen them on shopping sites and wondered what they did but after 2.5 years as a 3d printer hobbyist and 4 printers I am yet to experience any layer shifts so I won't be upgrading just yet
Closed loop tech has been around for a LONG time. It is about time 3D Printer makers 'discovered' this. I was totally surprised when I first started reading about 3D Printer tech and how the motors could skip? What!?!?!
dear Michael, it is kind to have done this test, especially on a CORE XY printer. I have tested BTT previous version of the closed loop steppers and I have been very disappointed. 1/ the board has some electrical problems that cause the steppers to stop randomly (my guess the connexion to the RESET pin is not good and BTT denies but others have the issue) but I guess the new BTT version has fixed the issue 2/ the stepper can get stronger current and torque when needed, that's why the stepper will not loose steps BUT if you have a non CORE XY like most of the 3D printers (my FLSUN printer is one), you will loose steps because the toothed belt will skip if there is too much torque. So I guess I would need to change the kind of belt I am using but currently I have no idea what sort of belt I need. So in the meantime, I have put TMC2209 along my old SKR Pro 1.1, it is less noisy, less random to use than the closed loop steppers. Also while BTT has done a good job to industrialize the product from a DIY system, last time I have seen they contributed ZERO to the software. But maybe they will do better and contribute because they have talents at BTT they're not Chinese copiers, they do improve products. edit 1: 14:31 thanks for the vampire Michael meme again today :) :) :) :) :) :) :) :) you're so cool edit 2: if you have the option to produce a video about the types of belt pro/con please please please
I'd also like to see a video comparing results with different belts and printing at high speed. If a different belt and/or toothed pulley would help things, especially on a big printer like a CR10 S4 or S5, I'd really like to know.
I'm relatively new to 3D printing, but always wondered why closed-loop stepper control wasn't more common these days. It's such a great way to improve reliability. Even if it's only needed 1 in 10 prints that seems worth it to me if a print is saved.
It's because print failures like this are even less common than 1 in 10 - it's relatively difficult to make a stepper skip steps if you give it a little overhead for its current. I'm using TMC2130s for XY in an i3 configuration, with the current limit of 800 mA through SPI - and they're perfectly capable of powering through obstructions. Thus, a layer shift failure is extremely rare and often caused by issues that cannot be rectified by a closed-loop system, such as belt tensioner failures or skipping belts. The downside of this system, besides cost, is the driver it uses - A4950 has lower microstepping, the system has no feedback to the controller like TMCs, and it costs a fair bit more. I'm not certain if it's going to be as quiet as StealthChop mode, which is a considerable boon of TMC drivers. Its primary advantage is the closed-loop system it offers, but TMC2130 can detect feedback from a stepper stall, which can be rectified by re-homing the machine or just the X and Y, before resuming printing. All in all, closed loop doesn't have as many applications in everyday hobby 3D printing as it may seem, but would be useful when you're printing at speeds and accelerations where step loss is almost inevitable, noise is not a concern, and open-loop systems no longer able to work reliably enough.
@@EgorKaskader Yeah true the most problem that the most people give their motors to less mA and thats why they skip sometimes just give them a little more and youre fine
I had been looking at this before but was unsure. Thanks for testing and sharing this video! My Gigabot 2.0 printer still has this exact problem even though I've done all I can to reduce the layer shift occurrence. I ordered my units today! I'm surprised this isn't standard equipment for 3D printers.
Very interesting. Glad to see these sort of solutions coming to consumer level printers. I would like to try one of these on the Ender 3 y-axis. I will have to get a socketed board though.
HI Thanks for the useful info regarding fitting S42C steppers, the BTT info is a little weak! I'm building a printing using SKR3 with S42c on X &Y with E3d V6 hotend all on a Ender 3 V2 frame. Thanks again for the good UA-cam stuff.
I carefully adjust voltages under typical load (printed a square vase and cylinder, tested normal and fast) and made sure acceleration/jerk was reasonable as to not cause a skip from sudden momentum shifts or built up resonance in multiple small movements. After doing that I haven't had a layer shift in years. With that said, I only print about 20-60 hours a month and mostly models around 100mm cubed. I used to have a layer shift about once every other month or maybe twice a month depending. Noticed sometimes it was because of slicer was far too aggressive and using a different slicer worked fine, but that's wasn't fixing the root cause which is why i adjusted voltages. In the end, that's looks like a great and relatively cheap step without getting a servo. And at that point you should also get a UPS that can last the typical time the light goes out in your area.
Thank you for the great video. (Or rather, great videos). I experience this issue with my MK3 quite often, even though the Prusa MK3 has the crash detection it still will layer shift. I will definitely add these to my Hypercube Evolution. After building that printer I couldn't really get the Marlin firmware to work right with the SKR 1.3 or SKR 1.4 board so I've just left the printer abandoned. With your amazing videos, I think it's time I get back to getting it working.
Amazing upgrade, especially for the low price. I don't get very many layer shift problems but this may be a good upgrade just for the peace of mind. I believe BTT offers a version of this kit that includes a smaller leveling knob for the Y axis, although I don't really need it since I have auto leveling on my Ender 3.
I am about to order 2 for my Ender 3 which will be running with SKR 1.4 Turbo. I am printing ABS so I do need something very reliable which when it hits something on our complex prints don't get all screwed up like it did in the past. Which on 24h long prints really set us back a lot.
I am implementing on my Geeetech A30T right now, at power cycle and initial homing I notice the close loop is in action, but still don't have movement. I think my last step is to change firmware values for min and max pulse and I'll be good to go!! Biggest tip I can give is that the Geeetech driver board pinout is opposite to what bigtreetech sent me, that means blue to black not blue to blue black to black like on bigtreetech boards Thanks for the video, without you I wouldn't have gotten so far, last step is figuring out which version of Geeetech open source code to use.
I rarely have layer shifts so I wouldn't incur the time and cost to switch, but I can see this becoming a standard feature on future machines. They'll probably have an integrated unit with serial communication back to the main board in not too long.
I can imagine it being a Marlin feature where you can specific pins for an encoder to feed back the position. Corrections could then be made regardless of stepper drivers. Hardware wise it could be hard as you would likely need interrupt pins.
Would you be able to do a second high speed motion and acceleration test? I'm interested to see if travel speeds and / or travel accelerations could be meaningfully increased. 300mm/s movement speed means little when the bed is 315mm wide. Lots of time for oozing
Good question. The original firmware this was copied from has a lot of features you can play with, specifically phase prediction, which could definitely increase max RPM. Of course torque will still drop off. Google nano zero stepper if you are interested.
I'm still modifying this printer towards it's specific purpose. Much of my time for this video actually went into that: SKR V1.3 conversion, printed case to match, all metal hot end swap, designing new part cooling duct. What I'm trying to say that once everything else is sorted I'll be willing to push it a bit more.
Have you continued to use these since you created the video? If so, have you had any issues? Also, have you updated the firmware on the closed loop stepper drivers?
Love this video and all your videos. I have an ender 3 pro, with the following mods: yellow bed springs, skr mini e3 v2.0, and a blue ptfe tube. The printer has been running constantly for the last 5 days, every single print that it has produced in that time has been absolutely perfect, with no defects, no stringing, no layer shifts and absolutely no problems. I have the latest firmware and use cura has my slicer. I am watching the video for curiousity to learn, but i must admit i have no intention of doing this upgrade, but nevertheless this upgrade is absolutely impression, and thank you for the videos, i have learnt a hell of a lot from you!
Interesting Gizmos. Dunno IF I will be buying some though. Only had One layer shift issue in 3 1/2 years of printing . Very close to being a solution for a small to nonexistent problem :-) Also.... without doubt there will be a flurry of Clones available in a month or 2 .. IF.. these prove any kind of sales success.
Very cool. I have a Duet in my printer and hope Duet offers a similar solution at some point in the future. I'm impressed with BigTreeTech though, they do seem to be innovating and producing good quality products.
What amazes me the most is no matter how thoroughly the principles and operation of the closed loop system is explained, people still go "WAAAAAAAH I won't have one until it has an open loop driver".
"Yoda's splitting headache" I love it. _There are no mistakes, just happy little accidents_ Definitely not the print you wanted, but now it's arguably a bit better
Just saying, I'm so glad your intros don't have some dumb and overly emotional intro line like, "COME LEARN WITH ME!!!!!" Just what you are covering, and then intro scene. Good. :)
FWIW - these are really affordable for this kind of tech. I use something similar on my DIY CNC Bridgeport VMC, Gecko G320 brushed servo drives. But, they accept step/dir signaling and ensure the steps are returned via a shaft encoder. They error out if they get out of sync with the encoder signaling after 128 steps +/-. The error pin has to be fed back to the control software so the program stops else they reset and keep going which actually tends to make the issue worse. These have been popular devices. Gecko upgraded them fairly recently to the G320X which has PID. They support 80V @ 20A max. Max. They are very expensive for the 3DP world @ $115 ea. Then you have to buy the motors, encoders, & run all the wire. Each motor needs four encoder wires and two power wires. I just looked and noticed that Marlin has M860 - M869 set aside for reporting back from I2C position encoders. Not sure if that project is stalled or what. Didn't look into what pins they are using or if the data is being used for PID regulation of moves within Marlin. So someone is looking at it there as well. As may have been mentioned, the Linux CNC project has support closed loop feedback for a while. I'm in the early stages of moving from Mach 3 to Linux CNC. I've seen more than a few people stuff Linux CNC on Rasberry Pi. Linux CNC has plenty of axes so it could run a printer. Should be interesting to see where this goes.
I had the problem, I solved it by slowing the acceleration(the parameter in mm/s²)from infinite because when it command to move it wasn't taking in account that the motor can't turn that fast between steps + add it the weight of the bed(40x40cm aluminium +glass) and it happend only on the bed motor never on the head. The print are a little slower but quality improved mostly in corners
Many printers use a4988 drivers. Use tb6550 or tb6600 instead. With these You can increase stepper driving current and you never see layer shifts. it is especially useful for extruder stepper
Wow this is really cool now the only thing I would have suggested that they do is that they should have made the buttons that control the S42B be attached to the screen that way they can come off with it then you could simply get longer cables so that you can have more clearance for the motors and be more compatible
Finally got my shipment from BTT. Adding skr 1.4, closed loop, and bl touch to my Chiron that doesn't print. Should be a fun, just because, type project.
I tried them before, noise is much louder than with TMC2209 drivers, and because the position is controlled dynamically they produce a whistling sound even when they should not be moving. The principle works quite well.
Great video as always very clear and concise. It's something I had been considering but the choice of stepper driver is dissapointing and puts me off; layer shifting isn't a big problem (for me) so it seems a backward step from the 2209s I am using.
Will not use thoe on my printer BUT I will use one on my Magnetic loop antenna for Ham radio as it use a stepper motor to control a variable vaccum capacitor. As I am building the software to control the antenna. Thnaks for the video. this will also come in handy the day I wish to switch to core x-y printer
Finally, they arrived. The first problem - the driver didn't work with 0.9° motors. Also calculate steps use different formula, for my printer it's 8192 steps / (2mm * 16 tooth) = 256 steps per mm.
I just added a pair to my x and y axis on my tevo tornado. The x axis works really well. Unfortunately the y axis has severe vibration and it will not home in the right direction. According to the installation manual you can change pid settings to remedy the vibration along the y axis, the problem for me is that the board has no USB port so I’m not sure how I would update the firmware. I’ll most likely have to return them. Solid idea, just need to workout the bugs.
On the Sidewinder X1 Y-Stepper, it goes thru the whole case which has a cutout at the bottom for the stepper, so fitting a CLS won't be much of an issue space-wise but you won't be able to see its screen.
The Yoda was actually printed on my old Solidoodle 2 back in the day. It came with a regular imperial threaded rod and nut. Z backlash and inaccuracy was a constant issue and z hop hop wasn't an option. More modern printers are less sensitive but z hop isn't always the answer.
It should be noted that you do need extra space but the screen seem to be removable so you can take it off and use a small length of flatband cable to move it out of the way and anyone who has ever held a soldering iron could break out the buttons too.
I think it's easier to understand if you got behind the general logic before. Stepper motors are "rasterized" motors, so to speak. They are available in various subdivisions, from coarse to fine. In 3D printing, motors with 200 steps are the most common ones. That means that one rotation is divided by 200. So, if you tell the motor to rotate by 800 steps clockwise, it will do 4 full turns clockwise, and they can rotate in both directions. Because of the fine graduation, the number of steps can be translated fairly precisely into linear travel distance. You just have to calculate or measure the distance that your object moves when the motor is doing a specific number of steps. (depending on the transmission ratio that is used) For example, 100 mm of movement could be 1832 steps. When this ratio is known, the system knows how many steps are required for any distance. Regular stepper motors don't measure their current position/step. That's why you need to use end-stops with them. Anytime when starting a print, the motors need to be brought to their starting points, from which they just "play back" their steps. If the nozzle now hits something - or the mechanics are jamming at some point - and a motor is skipping some steps, there is no "error message" sent back. It will continue as if nothing happened. But the position will be shifted from that point onwards, and there is no way to restore the original position because there is no record . That's what's meant by "being blind". You have to restart the entire path, by bringing the motors back to the end-stops from which they "blindly" do their steps again. Closed-loop steppers simply have this "missing" sensor built in to read out the current position/step, eliminating this issue. It's called a rotary encoder. (some more info on this in the reply to this post) If, for example, the movement is blocked, the printer can recognize that a motor isn't at the position/step where it's supposed to be and go into pause immediately. After fixing the issue, it's able to continue from where it stopped. Even if you can't save the print, it can still save you from damage to the printer, or from hours of dry printing, wasting lots of expensive filament. Another benefit - since the controller knows how much time was needed to reach a position, it can also react to mass inertia and other machine-specific things, and calculate patterns to optimize acceleration and speed to run smoother and more efficiently.
You didn't go easy with that rubber mallet, you knocked a bolt loose. You really jarred something loose tiger! I probably show my age with that last joke, but hopefully the kingpins amongst you get the reference :)
Cool technology, More processing power in the motors then your main board ;-) Larger machines with 2 Z motors would benefit from this technology to keep the motors in synchronization. Now that is a great application where people are having problems.
They really should produce a version that has the "UI" on a short cable connector and then offer a versions with & without it. It would help with the added height too!
Quick qestion is it worth it to switch from TMC2130 to TMC2209 ? I really like the tmc2130 but they get hot fast and are expensive with around 8-10 bucks a pop
The short answer is "Yes". Trinamic even makes a couple of drivers specifically designed for closed loop systems, such as the TMC4361 and TMC4671 (I have no idea how they compare operationally to the 22xx chips). As for people making them? I've only found the uStepper: ustepper.com/store/3-ustepper-boards Note that this isn't an endorsement for the uStepper. I've never used them, and due to cost I probably never will (especially when all I have is an Ender-2). But they use an Arduino processor (not a fan but it should still work) and either a TMC2208 (Lite) or TMC5130 (S) driver chip. They do seem to be more fiddly to install, though, what with the requirement of a dedicated DC power source to the control board.
@@ameliabuns4058 it is too a common issue.... i've probably had 40 layer shifts with my one and only FDM printer since i bought it 4 years ago. belts loosen over time, steppers overheat, speed to high, etc. try to cut down on time and everything might seem fine for 3-12 hours until it layer shifts. my friends tronxy will layer shift if you forget to turn the external fan he has for the drivers on. Its a common issue. its a common issue.....
@@darkshadowsx5949 I mean it's probably one the least common ones. Unless there's something seriously wrong with the printer. In which case you usually fix the issue instead of spending money on a servo. But don't take this the wrong way people have different opinions it's completely valid If you buy one or disagree!
Late to the conversation, but I'd love to see a followup to how the machine that was upgraded went with various tuning parameters. As far as I understand, the processor on the board counts the pulses sent to the motor and compares that to the count it has from the real motor position. The CPU on the board generates pulses of it's own to minimise the difference between the two.
Is this for earthquake prone areas or is there something I'm missing? The only layer shift issues i've ever had have been due to loss of bed adhesion, not random mallet whacks.
last time i had an earthquake in my area was 4-5 years ago and i barley felt it. just sounded like a low flying plane. and that was the last one in 40+ years i have layer shift problems frequently. a number of reasons can cause it. Over heating steppers, loose belts, speed, etc. its a more common issue than you think. my printer is 4+ years old and best i can do is upgrade it. dont want 2, dont need 2, dont have room for 2. you could be 3+ hours into a big print then all of a sudden LAYER SHIFT BITCH, and your fucked.
Never had layer shifts either, but the concept is pretty neat still. Just not something I'm gonna run out and buy. Not needed for my use case it would seem.
I wouldn't assume these are A4988 performance just because they are controlled via standard A4988 STEP/DIR interface. The A4950 on these S42Bs are basically fat MOSFET transistors (no STEP/DIR interface front-end like A4988) so they can operate spread-spectrum or other modes like TMC drivers if they are properly controlled with a nice processor. Specifically, the STM32 Cortex M0 chip found on the S42 does have hardware pulse generation with spread spectrum. I'm not saying the S42 firmware uses these STM features, but it looks like it from your review.
Cool that this tech is appearing. Now, let's do a speed test using Klipper and see what it takes to get a bad surface finish. There shouldn't be layer shifting during a print unless a vref is too low, speeds/accel are too high or it gets bumped. Curious minds want to know.
So, from what I understand, those ST arm chips on those tiny boards is what directly controls the A4950 drivers and that's why they are working quieter. Wonder if there is a cheaper way to have an encoder for each motor with TMC2209 drivers.
I was wondering since he said you could edit the firmware, how long before someone built a firmware for TMC drivers to make use of their features along with this.
that would be a dream, with skew compensation for twisted frames and missed steps , and back lash compensation....that's the printer i dream off... give it 2 years and some will be selling one for less that a 1000 unlike my bcn3d sigma which is nice, but I'm relying on the frame being setup perfectly in factory and sending it back to the for maintenance costs the same as buying a whole brand new cr10s. even though bcn3d say its open source you cant buy nozzles just whole hot ends starting at 89.99 for the cheapest option. rant over
@@ericspooner It's not quite as simple as that I think. This module, from the looks of it, is a fully contained system of it's own. So the printer doesn't even know that it missed some steps or what's happening behind the scenes. Printer thinks it's controlling A988 drivers. You can't really reprogram TMC chip like you can the ARM chip on this board. But I am not an expert in this though.
@@stranger7968 I get that, I'm saying someone programs STM on these daughter boards to use TCM drivers. I would love to see something like 5160's or something like that as a way to work with any stepstick setup(grbl smoothie) without having to program into those the TMC driver libs.
Same thought as Larry Johnson. Why can't the board be remote mounted from the motor itself? Found the answer, Magnetic encoder. However I don't see any reason why the screen couldn't be remote mounted. As for the Y motor with shafts out both ends: Buy the board alone kit, glue magnet to end of shaft where accessible, print a mounting bracket for the board, make an extension cable to get back to the motor and away we go for more DIY projects.
I have a question for you Michael… I just recently picked up these beauties for my printers. I love how absolutely quiet they are and print quality is excellent. The question is: since you can set the micro steps on the driver boards, do I set the micro steps in my firmware (klipper) to the same as what I set the boards to, or do I set 1 micro step in the firmware and X on the boards? Or do I do the opposite, set the boards to 1 and the steps to X in my firmware? Thanks in advance.
Excellent video,,BUT what was missing was the calculation of steps /mm ....before I installed this stepper..I had 80 steps/mm...(1.8 degree stepper (which this is ) 16 micro steps with 20 tooth sprocket )..with this unit for whatever reason I had to reduce steps down to 70 steps/mm
I have installed this on my printer. I have an Azteeg X3 board and dual y axis steppers. Motion is very jerky. this is no RPM reading on the OLED displays when running a print. In this video apparently everything works perfect first try out of the box but my result is less impressive. There is almost no troubleshooting information on the web. Does anybody have some information about what to do if you are not getting the desired result? I have made sure the magnet is installed as close to center as I could and can detect no visible runout when the shaft turns.
Noooo nonononono, they make more sound than the TMC2209, they're not on par, and the surface finish is clearly worse. Also, you're saying it's $25 for FOUR of the steppermotor+driver combination!??? Are you sure it's for four pieces?
primary reason I stopped using the originals (these are china clones of smart steppers) on my machine -- there is surface artifacting, and they're significantly noisier. They use a full h-bridge, which needs A LOT of tuning. And correction algorithm needs A LOT OF tuning. You have to pid tune them, BTT is not gonna tell you how -- you're on your own there. Honestly if you have high enough current steppers and things are not running at the edge of functional, these are not necessary unless you plan to do something with the encoder data.
Thanks for doing this review Michael. I've been curious about picking these up and understanding the noise and improvements available. If you had 2209's and were running sensorless homing, it looks like you'll have to go back to having limit switches, right?
I recomed use z-hop to avoid leyer shiffting.Layer shifting -in my case- was produce because the nozzle hit a part of the print, and then fall down in a shitfed position. Using zhop in cura resolved that
Very interesting Michael. Not sure I'm ready to give up on my 2209s though. It looks like the option 3 was for one S42B for $24.33, but you said you got 4 for that price? What did you break off the printer when you hit the hotend with the mallet? I saw something drop onto the bed, nothing serious I hope.
Hi guys, after reading comments here are two things:
1. Plenty of people saying they don't have layer shifts, plenty saying they do. The ones I showed at the start were genuine for me, not manufactured for this video. If you don't need these, that is great news, but it doesn't mean they are useless for everyone.
2. I was confused with the price by the '4PCS' in the description, but I see now they are referring to the mounting hardware. They are in fact $25 each for the units shown in this video, my apologies for the mistake in the summary. I will see if UA-cam let's me crop that part out as to not confuse people.
3. The endstop lever went back on without issue. My printer is proud to have 'taken one for the team' in the pursuit of testing and thanks you for your concern.
My cheap A8 based printer will skip steps if I push it much over 100mm/sec but print size also matters obviously.
@@ColinWatters My upgraded A8 sometimes skip steps when I'm printing materials like ABS and print start to lift corners. I didn't have this problem with PLA or PETG.
@@GrimmChristopher is your link broken?
@@ColinWatters please try again
Today I received two of these.
The DIP switches are all set to on by default - the inverse to what your drivers looked like.
Do you know where I can find a manual on what they do and how they should be set?
So much anxiety watching you whack the printer with a mallet. Amazed to see the Closed Loop system in action, very impressive. Interested to see what the maximum print speed is now.
When you hit that and it snapped back that's impressive wow
So glad you did this video. I have been waiting on a proper review of these closed loop stepper units for 6 months. Seems like every printer should come with this type of control of the steppers and maybe they will going forward. Thank you again for putting the time so we can all see what is involved with using these units.
While I am unlikely to use these on my 3d printer, it might be really great for a CNC mill where the forces are much higher and the chance for skipping steps greater. The price for motor and driver circuit is really impressive!
Are you running nema 17s on your CNC? Maybe you could add the encoder only kits to a nema 23 for more torque?
@@joshhyyym My CNC is a Sherline mill using Nema 23 and grblshield/Arduino controller. I would love for Bigtreetech to make a Nema 23 version.
not only motor & system, but also the driver... its game changing at this price
Well you just need feedback to controller. These kinds of things have been around for tens of years.
@@jothain True, but servo motors are normally very expensive as are the drivers. Bigtreetech is selling a complete solution for the price of the motor.
Yesterday I successfully upgraded my home-build H-bot printer following these instructions. Just want to share couple of things I learned during the upgrade:
- At 8:29 Michael talks about removing jumpers for micro stepping. This is optional because cable adaptor is not using these pins. The only 6 pins it connects to are VMOT, GND, VDD, EN, STEP and DIR.
- Pin connectivity allows me to answer my own question asked earlier here - there should be no problem connecting 2 motors in parallel. Obviously you either need a board with 2 stepper sockets wired in parallel or buy/make a cable splitter.
- At 13:45 Michael mentions that he increased steps per mm from 80 to 100. In my case I had to go from 320 to 410.214. So these values are printer specific.
- It worth mentioning that once motor is configured, you don’t need LCD and can remove and store it.
Once again, many thanks to Teaching Tech for yet another extremely helpful video.
Oh my freaking god as a side note at 9:12 the way you removed the old motor while keeping the belts routed in tact was absolutely genius! I could use that same method with initially routing the belts to begin with! That is always such a pain in the ass to route the belts through that spot so doing it your way will make it so much easier. Just wanted to let you and everyone know there are so many things you can learn from watching these videos!
I'm so glad I found this video, I've built my own corexy printer 300x300x400, linear rails on all axis and a MKS SGEN L V2.0 main board. I was given a pair of BTT S42B steppers, no matter what I did I couldn't get them to move until I watched this video, those 2 small changes in the firmware solved everything. Thanks.
What I'm really like about this is it makes setting up the steppers alot easier
@npgoalkeeper _ TMC2209s aren't closed loop. They will not auto correct for any movement. All they do is give you tighter control over movement and quieter motor movement.
The TMC2209 is the easier upgrade, but it isn't better.
This is fantastic! Thank you for this video. In particular, I was very impressed with your 'hammer test'.
As someone who has battled a lot with layer shifts in the past, it's amazing that closed loop motors have taken so long to hit the consumer printers. With the Creality K2, it looks like it's finally happening. My original "fix" was changing stepper drivers and running them at close to the maximum my board and driver are capable of, so the print head can power through a lot more resistance. It's of course not comparable to closed loop, but it does minimize the risk of layer shifts (for me). But I'd be a lot more comfortable having closed loop steppers - but alas, my printer's mainboard is closed source and doesn't allow for such customizations.
I like the idea of using an observer to read the motor coil current and inductance to determine if it stepped or stalled. It's an all electronic solution with no moving parts or metrics. Thanks for these great videos!
That's what the TMC2208 and TMC2209 do for both variable decay as well as StallGuard 😀
Seriously, seriously impressive. All printers should have this tech!
Excellent Michael. Good stuff... the only thing worse than the magic smoke is that deep fried ‘scent’. Back in the day I was calibrating /troubleshooting a HP DC p/s and slipped with the probe... then all you hear is OK ! who let the smoke out!! Good times LOL
This was/is a great video, even though I didn't understand any of the technical jargon.
you hit it with a hammer to show your test!!!! You are amazing, and i love your channel
Excellent, crystal clear, well-scripted presentation. thanks
Super film. S42B potrzebny do frezarek i robienia płytek PCB. Dziękuje i pozdrowienia z Polski.
3 changes....
Removable drivers.
Spi or uart connections.
Nema17 and Nema23 sizing.
Tmc 5000 series drives with closed loop, 20amp feedback for cnc, cheap and easy, love it.
Load your 2208's or 2209's that you already have, love it. Get rid of the screen and use spi or uart.... better cooling, better control....
I'll be getting a set after watching this video. Your videos are my favorite of all the 3D printing videos on UA-cam
I saw these on the BTT website and was hoping someone would do a piece on how well they actually work. I'll keep these in mind for whenever I get around to building that coreXY machine.
I always used to wonder what they do when I seen them on BigTreeTech AliExpress. As always Teaching Tech to rescue!
I too have seen them on shopping sites and wondered what they did but after 2.5 years as a 3d printer hobbyist and 4 printers I am yet to experience any layer shifts so I won't be upgrading just yet
People seem to forget I'm an enthusiast just like all of you. I was curious, purchased them and tested them out.
I hope this will be a standard feature on future printers.
Closed loop tech has been around for a LONG time. It is about time 3D Printer makers 'discovered' this. I was totally surprised when I first started reading about 3D Printer tech and how the motors could skip? What!?!?!
Well... engineering is typically finding the best minimum solution. Layers shifts are rare enough that open loop works fine most of the time.
dear Michael, it is kind to have done this test, especially on a CORE XY printer. I have tested BTT previous version of the closed loop steppers and I have been very disappointed. 1/ the board has some electrical problems that cause the steppers to stop randomly (my guess the connexion to the RESET pin is not good and BTT denies but others have the issue) but I guess the new BTT version has fixed the issue 2/ the stepper can get stronger current and torque when needed, that's why the stepper will not loose steps BUT if you have a non CORE XY like most of the 3D printers (my FLSUN printer is one), you will loose steps because the toothed belt will skip if there is too much torque. So I guess I would need to change the kind of belt I am using but currently I have no idea what sort of belt I need. So in the meantime, I have put TMC2209 along my old SKR Pro 1.1, it is less noisy, less random to use than the closed loop steppers.
Also while BTT has done a good job to industrialize the product from a DIY system, last time I have seen they contributed ZERO to the software. But maybe they will do better and contribute because they have talents at BTT they're not Chinese copiers, they do improve products.
edit 1: 14:31 thanks for the vampire Michael meme again today :) :) :) :) :) :) :) :) you're so cool
edit 2: if you have the option to produce a video about the types of belt pro/con please please please
I'd also like to see a video comparing results with different belts and printing at high speed. If a different belt and/or toothed pulley would help things, especially on a big printer like a CR10 S4 or S5, I'd really like to know.
I had doubts about this product due to the quiet operation of the engine, but now I found out that it is worth buying.
They are great, I used closed loop system on my 3kw CNC mill - amazing. Recommend.
I'm relatively new to 3D printing, but always wondered why closed-loop stepper control wasn't more common these days. It's such a great way to improve reliability. Even if it's only needed 1 in 10 prints that seems worth it to me if a print is saved.
It's because print failures like this are even less common than 1 in 10 - it's relatively difficult to make a stepper skip steps if you give it a little overhead for its current. I'm using TMC2130s for XY in an i3 configuration, with the current limit of 800 mA through SPI - and they're perfectly capable of powering through obstructions. Thus, a layer shift failure is extremely rare and often caused by issues that cannot be rectified by a closed-loop system, such as belt tensioner failures or skipping belts.
The downside of this system, besides cost, is the driver it uses - A4950 has lower microstepping, the system has no feedback to the controller like TMCs, and it costs a fair bit more. I'm not certain if it's going to be as quiet as StealthChop mode, which is a considerable boon of TMC drivers. Its primary advantage is the closed-loop system it offers, but TMC2130 can detect feedback from a stepper stall, which can be rectified by re-homing the machine or just the X and Y, before resuming printing. All in all, closed loop doesn't have as many applications in everyday hobby 3D printing as it may seem, but would be useful when you're printing at speeds and accelerations where step loss is almost inevitable, noise is not a concern, and open-loop systems no longer able to work reliably enough.
@@EgorKaskader Yeah true the most problem that the most people give their motors to less mA and thats why they skip sometimes just give them a little more and youre fine
Great update. Was ackley talking about close loop steppers in the job today
Thanks for sharing :-)
honestly, that recovered print is rly cool that it recovered so good
What a fantastic price for this! Thanks @TeachingTech. I just ordered some for my 400x400 moving portal printer(CNC style) and my ender 3! Great find.
I had been looking at this before but was unsure. Thanks for testing and sharing this video! My Gigabot 2.0 printer still has this exact problem even though I've done all I can to reduce the layer shift occurrence. I ordered my units today! I'm surprised this isn't standard equipment for 3D printers.
Very interesting. Glad to see these sort of solutions coming to consumer level printers. I would like to try one of these on the Ender 3 y-axis. I will have to get a socketed board though.
HI Thanks for the useful info regarding fitting S42C steppers, the BTT info is a little weak! I'm building a printing using SKR3 with S42c on X &Y with E3d V6 hotend all on a Ender 3 V2 frame. Thanks again for the good UA-cam stuff.
Dear,Thank you for your professional video, thank you for your support to BIGTREETECH, and love your video very much.
Omg it corrects angular error.. i can actually see this improving print quality as well.
I carefully adjust voltages under typical load (printed a square vase and cylinder, tested normal and fast) and made sure acceleration/jerk was reasonable as to not cause a skip from sudden momentum shifts or built up resonance in multiple small movements. After doing that I haven't had a layer shift in years.
With that said, I only print about 20-60 hours a month and mostly models around 100mm cubed. I used to have a layer shift about once every other month or maybe twice a month depending. Noticed sometimes it was because of slicer was far too aggressive and using a different slicer worked fine, but that's wasn't fixing the root cause which is why i adjusted voltages.
In the end, that's looks like a great and relatively cheap step without getting a servo. And at that point you should also get a UPS that can last the typical time the light goes out in your area.
Thank you for the great video. (Or rather, great videos).
I experience this issue with my MK3 quite often, even though the Prusa MK3 has the crash detection it still will layer shift.
I will definitely add these to my Hypercube Evolution. After building that printer I couldn't really get the Marlin firmware to work right with the SKR 1.3 or SKR 1.4 board so I've just left the printer abandoned. With your amazing videos, I think it's time I get back to getting it working.
Amazing upgrade, especially for the low price. I don't get very many layer shift problems but this may be a good upgrade just for the peace of mind. I believe BTT offers a version of this kit that includes a smaller leveling knob for the Y axis, although I don't really need it since I have auto leveling on my Ender 3.
I am about to order 2 for my Ender 3 which will be running with SKR 1.4 Turbo. I am printing ABS so I do need something very reliable which when it hits something on our complex prints don't get all screwed up like it did in the past. Which on 24h long prints really set us back a lot.
I am implementing on my Geeetech A30T right now, at power cycle and initial homing I notice the close loop is in action, but still don't have movement. I think my last step is to change firmware values for min and max pulse and I'll be good to go!!
Biggest tip I can give is that the Geeetech driver board pinout is opposite to what bigtreetech sent me, that means blue to black not blue to blue black to black like on bigtreetech boards
Thanks for the video, without you I wouldn't have gotten so far, last step is figuring out which version of Geeetech open source code to use.
I rarely have layer shifts so I wouldn't incur the time and cost to switch, but I can see this becoming a standard feature on future machines. They'll probably have an integrated unit with serial communication back to the main board in not too long.
That would be a perfect solution while still letting you use your TMC drivers.
I can imagine it being a Marlin feature where you can specific pins for an encoder to feed back the position. Corrections could then be made regardless of stepper drivers. Hardware wise it could be hard as you would likely need interrupt pins.
Would you be able to do a second high speed motion and acceleration test? I'm interested to see if travel speeds and / or travel accelerations could be meaningfully increased.
300mm/s movement speed means little when the bed is 315mm wide. Lots of time for oozing
Good question. The original firmware this was copied from has a lot of features you can play with, specifically phase prediction, which could definitely increase max RPM. Of course torque will still drop off. Google nano zero stepper if you are interested.
I'm still modifying this printer towards it's specific purpose. Much of my time for this video actually went into that: SKR V1.3 conversion, printed case to match, all metal hot end swap, designing new part cooling duct. What I'm trying to say that once everything else is sorted I'll be willing to push it a bit more.
Have you continued to use these since you created the video? If so, have you had any issues? Also, have you updated the firmware on the closed loop stepper drivers?
I’m definitely interested in giving these a try, thanks for another awesome and informative video!
Love this video and all your videos. I have an ender 3 pro, with the following mods: yellow bed springs, skr mini e3 v2.0, and a blue ptfe tube. The printer has been running constantly for the last 5 days, every single print that it has produced in that time has been absolutely perfect, with no defects, no stringing, no layer shifts and absolutely no problems. I have the latest firmware and use cura has my slicer. I am watching the video for curiousity to learn, but i must admit i have no intention of doing this upgrade, but nevertheless this upgrade is absolutely impression, and thank you for the videos, i have learnt a hell of a lot from you!
Interesting Gizmos. Dunno IF I will be buying some though.
Only had One layer shift issue in 3 1/2 years of printing . Very close to being a solution for a small to nonexistent problem :-)
Also.... without doubt there will be a flurry of Clones available in a month or 2 .. IF.. these prove any kind of sales success.
lol you hit it so hard that filament detector pin broke off ! lol
i know this is an old video, but i appreciate the info.
With all these btt upgrades, a future btt printer will look great
Very cool. I have a Duet in my printer and hope Duet offers a similar solution at some point in the future. I'm impressed with BigTreeTech though, they do seem to be innovating and producing good quality products.
What amazes me the most is no matter how thoroughly the principles and operation of the closed loop system is explained, people still go "WAAAAAAAH I won't have one until it has an open loop driver".
"Yoda's splitting headache" I love it.
_There are no mistakes, just happy little accidents_ Definitely not the print you wanted, but now it's arguably a bit better
These are definitely useful. Anyone that says they dont get layer shift, just hasnt yet. They will
The invert logic pin is probably useful for upgrading to Moons' Motors.
Nice features with that.
Just saying, I'm so glad your intros don't have some dumb and overly emotional intro line like, "COME LEARN WITH ME!!!!!" Just what you are covering, and then intro scene. Good. :)
When you hit your printer with the mallet, I felt it.
FWIW - these are really affordable for this kind of tech. I use something similar on my DIY CNC Bridgeport VMC, Gecko G320 brushed servo drives. But, they accept step/dir signaling and ensure the steps are returned via a shaft encoder. They error out if they get out of sync with the encoder signaling after 128 steps +/-. The error pin has to be fed back to the control software so the program stops else they reset and keep going which actually tends to make the issue worse.
These have been popular devices. Gecko upgraded them fairly recently to the G320X which has PID. They support 80V @ 20A max. Max. They are very expensive for the 3DP world @ $115 ea. Then you have to buy the motors, encoders, & run all the wire. Each motor needs four encoder wires and two power wires.
I just looked and noticed that Marlin has M860 - M869 set aside for reporting back from I2C position encoders. Not sure if that project is stalled or what. Didn't look into what pins they are using or if the data is being used for PID regulation of moves within Marlin. So someone is looking at it there as well.
As may have been mentioned, the Linux CNC project has support closed loop feedback for a while. I'm in the early stages of moving from Mach 3 to Linux CNC. I've seen more than a few people stuff Linux CNC on Rasberry Pi. Linux CNC has plenty of axes so it could run a printer. Should be interesting to see where this goes.
I had the problem, I solved it by slowing the acceleration(the parameter in mm/s²)from infinite because when it command to move it wasn't taking in account that the motor can't turn that fast between steps + add it the weight of the bed(40x40cm aluminium +glass) and it happend only on the bed motor never on the head. The print are a little slower but quality improved mostly in corners
Exactly this. Remember F=ma. Too much acceleration means too much force.
Thanks a lot for the video. That may be a solution for my Ender5 losing steps on the X-axis.
Many printers use a4988 drivers. Use tb6550 or tb6600 instead. With these You can increase stepper driving current and you never see layer shifts. it is especially useful for extruder stepper
Wow this is really cool now the only thing I would have suggested that they do is that they should have made the buttons that control the S42B be attached to the screen that way they can come off with it then you could simply get longer cables so that you can have more clearance for the motors and be more compatible
I dont understand why this doesnt come standard for all 3d printers. Its so cheap, and adds so much quality.
Finally got my shipment from BTT. Adding skr 1.4, closed loop, and bl touch to my Chiron that doesn't print. Should be a fun, just because, type project.
I really would hope they make a version that fit your own drivers like the TMC
I tried them before, noise is much louder than with TMC2209 drivers, and because the position is controlled dynamically they produce a whistling sound even when they should not be moving. The principle works quite well.
The previous models where much louder than the current ones apparently.
Very worthwhile examining
Great video as always very clear and concise. It's something I had been considering but the choice of stepper driver is dissapointing and puts me off; layer shifting isn't a big problem (for me) so it seems a backward step from the 2209s I am using.
Will not use thoe on my printer BUT I will use one on my Magnetic loop antenna for Ham radio as it use a stepper motor to control a variable vaccum capacitor. As I am building the software to control the antenna. Thnaks for the video. this will also come in handy the day I wish to switch to core x-y printer
This sounds interesting do you have a link with more info?
@@vaughancahill5870 Magnetic loop are researched a lot lately.. ua-cam.com/video/UH8BXD0oQfs/v-deo.html
Really cool steppers the link is no longer available threw Aliexpress
Superb video as usual Michael 👍👍👍 I'm ordering them for sure for my monster Ender 3
Finally, they arrived. The first problem - the driver didn't work with 0.9° motors. Also calculate steps use different formula, for my printer it's 8192 steps / (2mm * 16 tooth) = 256 steps per mm.
I am excited to see someone implement these from factory. Either Prusa or MatterHackers on their house brand.
Definitely on my upgrade list now!
I just added a pair to my x and y axis on my tevo tornado. The x axis works really well. Unfortunately the y axis has severe vibration and it will not home in the right direction. According to the installation manual you can change pid settings to remedy the vibration along the y axis, the problem for me is that the board has no USB port so I’m not sure how I would update the firmware. I’ll most likely have to return them. Solid idea, just need to workout the bugs.
Thanks for the video. Have been thinking of testing these on a new build project.
On the Sidewinder X1 Y-Stepper, it goes thru the whole case which has a cutout at the bottom for the stepper, so fitting a CLS won't be much of an issue space-wise but you won't be able to see its screen.
"Yoda's splitting headache" hahaha
Would not have happened with Z hop enabled.
The Yoda was actually printed on my old Solidoodle 2 back in the day. It came with a regular imperial threaded rod and nut. Z backlash and inaccuracy was a constant issue and z hop hop wasn't an option. More modern printers are less sensitive but z hop isn't always the answer.
The VLC player example was nice
It should be noted that you do need extra space but the screen seem to be removable so you can take it off and use a small length of flatband cable to move it out of the way and anyone who has ever held a soldering iron could break out the buttons too.
I think it's easier to understand if you got behind the general logic before.
Stepper motors are "rasterized" motors, so to speak. They are available in various subdivisions, from coarse to fine. In 3D printing, motors with 200 steps are the most common ones. That means that one rotation is divided by 200. So, if you tell the motor to rotate by 800 steps clockwise, it will do 4 full turns clockwise, and they can rotate in both directions.
Because of the fine graduation, the number of steps can be translated fairly precisely into linear travel distance. You just have to calculate or measure the distance that your object moves when the motor is doing a specific number of steps. (depending on the transmission ratio that is used) For example, 100 mm of movement could be 1832 steps. When this ratio is known, the system knows how many steps are required for any distance.
Regular stepper motors don't measure their current position/step. That's why you need to use end-stops with them. Anytime when starting a print, the motors need to be brought to their starting points, from which they just "play back" their steps.
If the nozzle now hits something - or the mechanics are jamming at some point - and a motor is skipping some steps, there is no "error message" sent back. It will continue as if nothing happened.
But the position will be shifted from that point onwards, and there is no way to restore the original position because there is no record . That's what's meant by "being blind". You have to restart the entire path, by bringing the motors back to the end-stops from which they "blindly" do their steps again.
Closed-loop steppers simply have this "missing" sensor built in to read out the current position/step, eliminating this issue.
It's called a rotary encoder.
(some more info on this in the reply to this post)
If, for example, the movement is blocked, the printer can recognize that a motor isn't at the position/step where it's supposed to be and go into pause immediately. After fixing the issue, it's able to continue from where it stopped. Even if you can't save the print, it can still save you from damage to the printer, or from hours of dry printing, wasting lots of expensive filament.
Another benefit - since the controller knows how much time was needed to reach a position, it can also react to mass inertia and other machine-specific things, and calculate patterns to optimize acceleration and speed to run smoother and more efficiently.
You didn't go easy with that rubber mallet, you knocked a bolt loose. You really jarred something loose tiger!
I probably show my age with that last joke, but hopefully the kingpins amongst you get the reference :)
Cool technology, More processing power in the motors then your main board ;-) Larger machines with 2 Z motors would benefit from this technology to keep the motors in synchronization. Now that is a great application where people are having problems.
They really should produce a version that has the "UI" on a short cable connector and then offer a versions with & without it. It would help with the added height too!
Agreed. The SKR 1.3 and 1.4 have a serial connection for every driver, why not use that to configure the servo parameters?
I wish there was a way to use TMC2209s with a closed loop system.
#metoo
Quick qestion is it worth it to switch from TMC2130 to TMC2209 ? I really like the tmc2130 but they get hot fast and are expensive with around 8-10 bucks a pop
The short answer is "Yes". Trinamic even makes a couple of drivers specifically designed for closed loop systems, such as the TMC4361 and TMC4671 (I have no idea how they compare operationally to the 22xx chips).
As for people making them? I've only found the uStepper: ustepper.com/store/3-ustepper-boards
Note that this isn't an endorsement for the uStepper. I've never used them, and due to cost I probably never will (especially when all I have is an Ender-2). But they use an Arduino processor (not a fan but it should still work) and either a TMC2208 (Lite) or TMC5130 (S) driver chip. They do seem to be more fiddly to install, though, what with the requirement of a dedicated DC power source to the control board.
They make them, they're just a lot more expensive
Here: www.trinamic.com/company/news/news-detail/plug-and-play-smart-stepper-motor-for-closed-loop-operation/
I feel really lucky since I only had one layer shift in my three and a half years of printing.
Great tech though.
Only layer shift I've ever gotten is when I first got my printer, one of the belts wasn't tight enough.
@@Krytern Same thing with mine.
It's not a common issue
@@ameliabuns4058 it is too a common issue.... i've probably had 40 layer shifts with my one and only FDM printer since i bought it 4 years ago.
belts loosen over time, steppers overheat, speed to high, etc.
try to cut down on time and everything might seem fine for 3-12 hours until it layer shifts.
my friends tronxy will layer shift if you forget to turn the external fan he has for the drivers on.
Its a common issue.
its a common issue.....
@@darkshadowsx5949 I mean it's probably one the least common ones. Unless there's something seriously wrong with the printer. In which case you usually fix the issue instead of spending money on a servo. But don't take this the wrong way people have different opinions it's completely valid If you buy one or disagree!
Late to the conversation, but I'd love to see a followup to how the machine that was upgraded went with various tuning parameters.
As far as I understand, the processor on the board counts the pulses sent to the motor and compares that to the count it has from the real motor position. The CPU on the board generates pulses of it's own to minimise the difference between the two.
Is this for earthquake prone areas or is there something I'm missing? The only layer shift issues i've ever had have been due to loss of bed adhesion, not random mallet whacks.
It's a thing for step loss. That mainly happens when you go too fast, or right at the edge of the capabilities of your stepper/drivers.
last time i had an earthquake in my area was 4-5 years ago and i barley felt it. just sounded like a low flying plane. and that was the last one in 40+ years
i have layer shift problems frequently. a number of reasons can cause it. Over heating steppers, loose belts, speed, etc.
its a more common issue than you think. my printer is 4+ years old and best i can do is upgrade it. dont want 2, dont need 2, dont have room for 2.
you could be 3+ hours into a big print then all of a sudden LAYER SHIFT BITCH, and your fucked.
@@darkshadowsx5949 you might try upping your motor current
It is more likely to happen on large prints over 500g of plastic, the added weight on the bed can cause shifts randomly.
Never had layer shifts either, but the concept is pretty neat still. Just not something I'm gonna run out and buy. Not needed for my use case it would seem.
I wouldn't assume these are A4988 performance just because they are controlled via standard A4988 STEP/DIR interface. The A4950 on these S42Bs are basically fat MOSFET transistors (no STEP/DIR interface front-end like A4988) so they can operate spread-spectrum or other modes like TMC drivers if they are properly controlled with a nice processor. Specifically, the STM32 Cortex M0 chip found on the S42 does have hardware pulse generation with spread spectrum. I'm not saying the S42 firmware uses these STM features, but it looks like it from your review.
Cool that this tech is appearing. Now, let's do a speed test using Klipper and see what it takes to get a bad surface finish. There shouldn't be layer shifting during a print unless a vref is too low, speeds/accel are too high or it gets bumped. Curious minds want to know.
So, from what I understand, those ST arm chips on those tiny boards is what directly controls the A4950 drivers and that's why they are working quieter. Wonder if there is a cheaper way to have an encoder for each motor with TMC2209 drivers.
I was wondering since he said you could edit the firmware, how long before someone built a firmware for TMC drivers to make use of their features along with this.
that would be a dream, with skew compensation for twisted frames and missed steps , and back lash compensation....that's the printer i dream off... give it 2 years and some will be selling one for less that a 1000 unlike my bcn3d sigma which is nice, but I'm relying on the frame being setup perfectly in factory and sending it back to the for maintenance costs the same as buying a whole brand new cr10s. even though bcn3d say its open source you cant buy nozzles just whole hot ends starting at 89.99 for the cheapest option. rant over
@@ericspooner It's not quite as simple as that I think. This module, from the looks of it, is a fully contained system of it's own. So the printer doesn't even know that it missed some steps or what's happening behind the scenes. Printer thinks it's controlling A988 drivers. You can't really reprogram TMC chip like you can the ARM chip on this board. But I am not an expert in this though.
@@stranger7968 I get that, I'm saying someone programs STM on these daughter boards to use TCM drivers. I would love to see something like 5160's or something like that as a way to work with any stepstick setup(grbl smoothie) without having to program into those the TMC driver libs.
@@ericspooner ooh. I see what you mean.
Same thought as Larry Johnson. Why can't the board be remote mounted from the motor itself? Found the answer, Magnetic encoder. However I don't see any reason why the screen couldn't be remote mounted. As for the Y motor with shafts out both ends: Buy the board alone kit, glue magnet to end of shaft where accessible, print a mounting bracket for the board, make an extension cable to get back to the motor and away we go for more DIY projects.
I have a question for you Michael…
I just recently picked up these beauties for my printers. I love how absolutely quiet they are and print quality is excellent. The question is: since you can set the micro steps on the driver boards, do I set the micro steps in my firmware (klipper) to the same as what I set the boards to, or do I set 1 micro step in the firmware and X on the boards? Or do I do the opposite, set the boards to 1 and the steps to X in my firmware? Thanks in advance.
Excellent video,,BUT what was missing was the calculation of steps /mm ....before I installed this stepper..I had 80 steps/mm...(1.8 degree stepper (which this is ) 16 micro steps with 20 tooth sprocket )..with this unit for whatever reason I had to reduce steps down to 70 steps/mm
I have installed this on my printer. I have an Azteeg X3 board and dual y axis steppers. Motion is very jerky. this is no RPM reading on the OLED displays when running a print. In this video apparently everything works perfect first try out of the box but my result is less impressive. There is almost no troubleshooting information on the web. Does anybody have some information about what to do if you are not getting the desired result? I have made sure the magnet is installed as close to center as I could and can detect no visible runout when the shaft turns.
Would be interesting to know what peoples thoughts are on this vs something like stallguard on TMC drivers
Noooo nonononono, they make more sound than the TMC2209, they're not on par, and the surface finish is clearly worse. Also, you're saying it's $25 for FOUR of the steppermotor+driver combination!??? Are you sure it's for four pieces?
I wonder if TT was high or otherwise intoxicated while making those comparisons
I’m sure he means $25 ea and $100 for a set of 4..
Although just looked Aliexpress has the for $16
@@rambokd $16 only for the one without stepper motor :)
primary reason I stopped using the originals (these are china clones of smart steppers) on my machine -- there is surface artifacting, and they're significantly noisier. They use a full h-bridge, which needs A LOT of tuning. And correction algorithm needs A LOT OF tuning. You have to pid tune them, BTT is not gonna tell you how -- you're on your own there. Honestly if you have high enough current steppers and things are not running at the edge of functional, these are not necessary unless you plan to do something with the encoder data.
Thanks for doing this review Michael. I've been curious about picking these up and understanding the noise and improvements available. If you had 2209's and were running sensorless homing, it looks like you'll have to go back to having limit switches, right?
I recomed use z-hop to avoid leyer shiffting.Layer shifting -in my case- was produce because the nozzle hit a part of the print, and then fall down in a shitfed position.
Using zhop in cura resolved
that
Very interesting Michael. Not sure I'm ready to give up on my 2209s though.
It looks like the option 3 was for one S42B for $24.33, but you said you got 4 for that price? What did you break off the printer when you hit the hotend with the mallet? I saw something drop onto the bed, nothing serious I hope.
I made a mistake with the price unfortunately, they are $25 a pop. Just a lever arm from the endstop, it went right back on.
@@TeachingTech That's still a dam good price. Ah cool, glad it was nothing serious.