I do love how quiet the motors are. My thinking about racking: plain bearings always have clearance, so... The maximum angle the bearing can sit at is where the two opposite ends hit the rod. If racking is occurring (the bearing is at an angle relative to the motion axis), there is only contact at the two ends. So rather than having one long bearing, it would be just as effective to have two extremely short bearings but placed very far apart (assuming they are sufficiently coaxial). It is lighter and since there is less contact area when sitting flat, there is less friction too. Incidentally I think the same thing applies to linear rails and everyone always using a single long "H" style carriage is a bit weird. Something else minor I noticed, it looks like your umbilical has a short bend point right where the wire goes into the toolhead. We use a short section of large diameter spring there, just large enough ID for the wire bundle.
That makes sense, During wrap you are contacting two outside edges. I Have two bushings spaced apart on the Y axis toolhead. Maybe those heaver Bronze bushing on the outside rails would act like a vibration damper? What did you think my thoughts on the capacitors? I've seen some pretty massive capacitors to buffer the servos on large CNC machines. I have done a few large CNC electronics retrofits for customers.
@@fintechrepairshop I've actually not used a commercially made control board since my first printer which used a RAMPS 1.4 (of course making your own is a lot more fun...). These days with TMC2209 I use 50uF of capacitance per driver, connected with large traces with no vias, directly to the drivers. I think it is going to depend a lot on the board design (trace resistance and inductance, also cap ESR, etc.) and the small single-driver boards may just be better because the simple requirements allow them to put the cap right on top of the driver. But I am not doing 35k accelerations either, and there may be a need for more capacitance too to merely capture deceleration energy. But there is no need to guess here! You have an oscilloscope, right? You should scope the driver supply pads (maybe across the h-bridge for a discrete bridge driver) during a fast acceleration and deceleration and see what happens!
@@jessicav2031 Yeah, I have an oscilloscope. I originally got the scope to test PWM circuits. I had a few ideas for caps. Might want to look a drivers with external mosfets and large capacitors. Another thought was to solder some caps on the drivers power source pins. Not sure how much more current I can send through the TMC2240's. I might be able to push them to 2.0A if I have better cooling. They are heating up at 1.5A. Maybe machine a better heatsink or upgrade to external. The bigtreetech TMC5160 Plus work pretty good in my Orcinus printer.
I will agree the Mini cuts some corners, and the extruder and runout sensor setup is a disappointment. But I am not sold that dual small gears are the best solution. I am quickly becoming a fan of large drive gears, even in a single gear configuration. The single large gear that Prusa is now using on the MK4 Nextruder has performed flawless for me over the last 9 months. I have just started to build a dual large gear Clockwork2 extruder for a Voron using HGX gears. I have high hopes for it! Do you use the Integrated Filament Sensor version of the Bondtech? My understanding is that it still is a basic sensor, and doesn't do any loading or unloading, like the MK3/MK4 would, correct?
0:48 no that's no the problem, the real problem is that the motor pulls the filament so fast from the hotend that it leave mini traces of filament inside the extruder. The only solution is pull the filament in a lower temp and speed, also to change the bowden tube for a capricorn one. I preferred to have a gcode to remove the filament at 185 C for PLA and 200 C for almost everything else.
I've fixed 10+ of these printers. They all come in with the same problem. The only reliable fix is convert to BMG extruder gears and ditch automatic feeding. Load the filament manually.
BLtouch was disconnected and wires were a mess. I fished the bltouch wires in the sleeve. I made sure everything works and did a test print. I never know how I'm going to get them.
That aluminium body maybe it's worth it for cooling purposes, connect a v5 or v6 throat to it maybe via another alu plate, slap a heatsink-fan on the side? Make the whole assembly and post extruder filament path way shorter. What do you think?
@@fintechrepairshop Yeah but you're just using e3d-v6-like for your hotend. Remember how 3d Hemera reduced the height of the toolhead by removing the heatsink dangling below the extruder. Instead the throat gets screwed straight into the extruder body, which has a heatsink integrated. I'm thinking that same trick might just work with the aluminium HGX extruder. Maybe with an extra 40mm heatsink-fan leaning straight against it and a bottom plate. Maybe fan blowing through extruder body is enough.
"Tool post grinder" and if you can get a textbook from a vocational machine shop it will help you connect the dots and get the terminology right. Check community College bookstores that sell used books...
I've used it, probably good for this customer because domain controller and roaming profiles. I'm not a fan of windows and prefer Linux. All my servers are running Linux.
That vibration sounds similar to one of my AWD printers before "syncing" the motors. You should try engaging the motors and loosening the grubscrews on one of the pulleys before tightening them again. I would do this for both pairs of motors if you haven't already. Vez3D has a video on mechanical motor syncing as well.
Good info, I'm going to look into it. The printer is Cartesian. Makes sense, motors need to stepping at the exact same time. I might try running two motors off one driver so I know they are both being command to step on the exact same time. I also made sure the motors wire length were all the same length, gauge and brand.
@@fintechrepairshop I would be weary of running both motors off one drive as it will cause extra/double the back EMF and usually results in worse performance. I would highly recommend looking at Vez3D's video for motor syncing, more about making sure the pulleys on motor pairs are exactly aligned so that they don't fight each other during some moves.
I like that 45 degree idea, that bowden looks pretty great for bowden! That resonance you have is the same thing I was talking about that we encountered in our croxy printer which limited our top speed. Our conclusion was that it has to be interference between the steps of opposite motors. Have you done input shaping yet? It is surprising that it wouldn't prevent it, since it seems very frequency-dependent?
Yeah, I have a USB accelerometer. It is really bad at low speeds. You are actually making really good points. I didn't even think about the steps on opposite motors. I might try running two motors one driver. Right now I'm running a TMC2240 on each motor. I wish I could find the source of bambu labs carbon rods. Those things are perfect.
@@fintechrepairshop Using the same driver seems to defeat the purpose of having two motors though? Would that even work very well? A driver works by having a comparator, the mosfet closes and the coil current rises until the comparator fires (it rises slowly due to inductance of course) and then it switches to PWM. But the two motors won't have identical inductance, so initial current will not be the same. Are you synchronizing steps now between opposite motors? Or homing each side individually? I think that having the steppers mirrored would be a better approach here because then the resonance would always be the same (and thus able to be canceled each time without recalibrating input shaper). If you independently home both motors, the motor angle difference will be random, so resonant behavior will be different each time. Just a theory. When we built ours, input shaper didn't exist 😀
I had another comment about syncing the motors. I'm not sure of that applies to Cartesian printers. I'm going to double check all my axis for squareness and alignment. I might mess with hole spacing on the toolhead.
Neat. I've been wanting to see inside one of these. It looks extremely hypercube-esque. I love how they buried the extruder motor at the very center of rotation. I bet they carefully check the Z screws too to prevent wobble, because that Z design looks prone to it. You should send them a GPL request for the firmware source 😅
Yeah, its incredibly smooth. I've worked allot of printers and this is by far the smoothest and quietest. The owner found the hotend and I got it printing. I didn't video of it.
Since the 3 bed screws are driven by one motor I'm wondering what happens to the Lidar if one the screws skips the belt. There is now way for it to compensate. Thanks for the comment.
@@fintechrepairshop I doubt a belt skip is likely. The printer I designed and built in 2018 has a very similar arrangement with a large closed loop belt driven by a single motor (but running on rails instead). It has never skipped a tooth in this entire time. I've crashed the machine plenty.
@@jessicav2031 Well I guess if you have the Z motor current dialed in. Did you ever upload your design? I always like seeing new things. I'm still trying to figure out how there Lidar works. There was 6 wires and a ribbon cable going to it. Definitely Not like your typical probe.
@@fintechrepairshop Nope. It is basically a Hypercube Evo clone except I used multiple Z screws on a single motor with rails on Z. And the printhead has gone through a million revisions of course. Back then we were experimenting with extruders driven by 20:1 worms on a typical round style motor (with 36V) but friction was too much so I switched to belt reduction, then an Orbiter recently. Our more modern printers all use 3-point Z but the one I am talking about was from the Marlin/ATMega/etc days and I never changed it since it has just been so reliable 🙂
Thank you! Of all the forum posts and videos I’ve looked at, this is the only one that helped at all. Nothing else was even close. Luckily, I have another broken B1 I can grab spare parts from, so I didn’t have to order anything. I threw the new fuse in it, and it was good to go.
I loved my Simple Metal, but I couldn't deal with Brooke. Tried pointing out that they're using a bad thermistor that single strand wire was prone to breakage do to the bend and he got so mad I Tweeted about it to let other owners know that he blocked me on ALL social media!
I still have mine, it running klipper now with 7" touchscreen. I've done thousands of prints with it. I still print with it today. Its not my primary printer but has got me out situations when my primary went down.
I've been using ebb42 on 2 of my printers for quite a while and haven't had any of those issues. Your comment about USB working better makes me think your canbus umbilical may not have a good twisted pair for data communication. The only tuning I've had to do is figuring out the best baud rate to set the can interfaces. 500k seems to be the sweet spot. Running at 1m, weird things tend to happen.
Very odd that you were able to get your Novosun cards to work without glitches in Linux but not in Windows. I have the Nvem 6 axis ethernet controller which I bought several years ago but never got around to installing it into my existing Windows XP parallel port setup, for fear of breaking a very reliable Mach3 machine. However I am so fearful of my Win XP dedicated CNC rig eventually dying due to either a failed PC or OS that I am now thinking of biting the bullet and doing the upgrade, which is how I came across your video. Could you please elaborate on your issues with the controller and can you give me any tips. Thanking you in anticipation.
probing was unreliable crashing destroying tips. Random lockups. I went to a smoothstepper. I spent too many hours on those boards. The older ones were more reliable.
I have also gone through same thing. Got good results with plastic bushing initially. I used drill with smooth linear rod run it until no binding while bushing were installed in the part. Then tried graphite long bushing which are too heavy 8x12x45. Then changed design again to use 8x10x20 and used two while saving 9 grams per bushing. However perfectly aligning two bushing in ABS part is difficult. I had to used soldering iron which is 8 mm in diameter and aligned them at around 280c as soon as bushing got hot while installed in parts i turned off iron and let it cool to be remain aligned. However corexy put too much load on Y axis bushing when if i want to run system stiffer for high accell, on low belt tension input shaper is all over the place. So no real good alternate for me on corexy. I can make a short video explaining all my failures if you want.
You will need slomoguys to be able to catch movement at that speed :)) Now all the air is going first on the bed and is getting hotter, maybe to blow somehow from top to bed ? i'm struggling to get my printers fix each 2 days, not an expert so it is just an ideea. Also, like you said, maybe they cancel each other in the center, try to print on the side the test cube. You can adjust the flow for testing with someething to partially block or adjust the internal oppenings.
I forgot to say this Why not just one long bushing for the hotend? then you could make the top plate one piece with a hook clip on one side and a screw on the other to cut a screw out.
For the cooling add one more fan to each and make in the air nozzle three air chambers ( like three straws next to each other) one for the back, middle and front, If you just want to keep the two fans you got you have to make a nozzle that has three air chambers one for the back, middle and front. rotating it will also help one less bend it has to make.
Hey there, just curious if you tried chopper-resonance-tuner for your setup. Should reduce motor vibrations a lot too. I need to try that one one day for all my motors aswell. Should be easy to do, just takes a few hours to measure the data.
use equal parts baking soda and vinegar after shop vac-ing the water out, let sit for 10 minutes and run a rinse cycle. this should clean all that crap out
Last try to post if this doesn't go through o well but let do this. You can try using a rubber sleeve over the bushings and you can also try changing the steel rods to carbon fiber rods to be able to use the heavier bushings and just to cut weight.
Nice project. I gave it a go on the graphite bushings but unfortunately they had too much play on the carbon rods. Such a shame, they seemed to slide really well.
@@fintechrepairshop I spoke briefly with a guy from Easycomposites. He suspects Bambulabs machines their carbon rods, because it's not only about the tolerance, but more like keeping the tolerant consistent along the whole rod. Hope that makes sense.
Subbed. verry intresting. I am 54 just bought an Ender 7 (the one everyone hates and no one bought) for 250$ :) Already run's Klipper with an CM4 4Gb/32GB, and a BTT Microprobe, printed my first bunchy, was nearly perfect. Got 2 LDO's for my x y but the damn thing comes with two 2160's boards for the x y and i can't "tweak" them and i dont think they do UART or SPI So yea came here from your "Tmc 5160T Plus" video :P Thinking of putting an BTT Kraken in my ender 7.it has four 2160''s that do 8amp, and four that do 3amp
The extremal 5160 with the LDO Super HT steppers are awesome. I wish they ,made an external driver board with large capacitors. One that would run any driver. I might design one.
thank you so much for this video. no BS about it, just straight to the point. no 15 minute rant about your life story and how you chose this car, and your uncle billy wasn't too happy with your decisions, blah blah blah, all the other youtubers rant about. Just straight up replacing the battery. very helpful
Wouldn't the probe that stays on allow you to walk back to the tone generator and clip it onto different wires until you hear the probe go off? Just curious
