Beltless I3 project, feedback?

Maybe your leadscrews and the acrylic frame are too noisy for an underbed Piezo? I had to adjust my sensor to 'almost def' and then got into the same missed trigger issues as you did.

Quote
o_lampe
Maybe your leadscrews and the acrylic frame are too noisy for an underbed Piezo?

I can't discount the possibility, though the X and Y screws aren't moving at all during probing, and most people use Z screws that are similar. It doesn't seem unusually noisy when the false triggers happen, in fact it's quieter than the I3 ever was. The pitch of the noise that it does make is different though.

I can do something with the machine base to try to dampen noise transmitted through the frame, and maybe I'll add some neoprene washers between the Z motors and the towers. The nuts on the Z screws are spring-loaded to hold them in place, and sometimes the springs make noise. I should improve the way that they're held in place so that they can never touch the screws. I want to be careful though, since they're what provides protection for the frame of the printer if it tries to (As it often is now) crash the nozzle into the bed.

If this keeps up, I might change my setup, and use a couple of drilled discs over the 5mm Z rods detecting those nuts being forced downwards, but if I were to go that way, I could probably just use microswitches.

Ah the issues never seem to end.

So last night, during a 5mm Z travel, I got this wonderful screeching noise, and it seems that the right side Z motor has just stopped moving.

It is true that I'm trying to get Z motion much faster than I ever did from the I3 running Marlin on the Mega 2560/RAMPS. I think that the current is set about the same (The duet's configuration defaults to 800mA. and I've left that alone for now.) I think I had Z current set to about 1.5A with the 2 motors in parallel on the RAMPS, with X and Y at about 750mA each.

I have replacement motors, so I ought to be able to handle replacing the one that died. I wish I knew for sure that the cause of it dying was. I'm thinking that I should maybe change out the Z axis from using the 5mm 0.8mm pitch screws to using 8mm ones. I could probably use the 1.25mm lead all-thread, or the 8mm lead rods, which may be the better choice, since 400 steps/mm seems to be more than good enough for almost any layer resolution that I'm ever going to use.

In the meantime, I guess I'll replace both Z motors to get things back to functional. I'm pretty sure that the replacement motors that I have are better then the ones that I was using.

Edited 1 time(s). Last edit at 09/24/2018 03:25PM by SupraGuy.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

Are you sure the motor(s) are dead? It's nearly impossible to kill them using an IC driver. Driver chips, OTOH, are very easy to kill.
Motors are just iron, coils of wire and magnets. There isn't much to go wrong. If the coils are good- check with an ohm meter- the motor is 99.9% OK. If the motor doesn't smell burnt, it's 99.9% OK. If you can turn the shaft with your fingers, it's the other 0.1% OK.

Have you ever taken the motor apart? If a foreign object got in there (magnets will attract all sorts of stuff n a workshop) it might be jamming it up. Tolerances are very tight inside a motor...

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

no, I'm not 100% sure. I don't think that anything has happened to the Duet board though. I't s mounted in a reasonably safe location.

It's possible that the threaded rod just jammed and that there's a problem with the coupler. It's just a 3D printed coupler, so it's not particularly strong, and has pretty wide tolerances for a machine part, after all. (I've been meaning to get better ones.)

I can turn the motor shafts with my fingers, though the one felt kind of chunky. (Printer power was off.) I've never opened the case on those motors. I have not taken a multimeter to the wiring to check it yet., I've been kind of busy.

So I need to do some diagnosis. I'll just pull the shafts off of the motors, and see if the motors spin under no load.

I'm a little concerned about the spinning speed of the motors anyway. I had most of the work done to convert them to a T8 lead screw instead of the 5mm. I had aborted it it because the smooth rod shaft and the lead screw centres are a little too close. (18mm centre to centre) so the 15mm diameter LM8UU bearings and the 22mm diameter nut normally dictate a minimum separation of 20mm, so with those shafts spaced 18mm apart, there's a little interference between where the lead screw nut has to be and where the bearing would need.

So I can resolve that by placing the nut so that the flange surface is in between the bearings, and be super careful about how I handle the cutout for it. It would require some redesign of my gantry pieces... Naw. I decided on a simpler option. Take a dremel to the nut to cut a flat side into it. It's a reasonably simple brass piece after all, and I have lots of spares, since I don't need anti-backlash on the Z axis (Gravity will provide all of the anti-backlash that I should need.) It will make the parts a little harder to replace later, but it will mead a 10X reduction in needed RPM for the same travel speed. This should be easier on the motors.

Besides, I'm apparently getting a little Z wobble after all, so I need to alter this a little to uncouple the gantry from the screws a little better.

There are ridges that exactly match the 0.8mm pitch of the lead screws in this piece, you can see them on the left side of the picture quite clearly. I am VERY happy with the lack of ringing and overshoot though, For the record, this is what I would call my first successful print in PETg ever. (80mm height, 50mm across the faces of the top hexagon structure, 0.1mm layers, 240 deg C nozzle, 60mm bed.) Since I had hoped/intended to use this printer primarily for PETg filament, I'm quite pleased that this actually worked.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

looks ok... do you have a pic of the parts where the nut is, maybe there's another way to do that, be also interesting to know if the Z patter is also on X & Y?

Edited 1 time(s). Last edit at 09/28/2018 01:09PM by MechaBits.

Thanks for the feedback!

Well, I've completed the switch from the 0.8mm lead 5mm threaded rod to the T8x8 lead screw. I haven't re-printed anything tall enough to diagnose Z wobble with the new setup yet. I've got a looser fit with the nut this time, so it's not likely to exhibit the same problems. I hope. Anyway, I'll see if I can snap some pictures of the old mechanism, but it had a square nut popped into the gantry piece, and retained with a "C" clip that spring-loaded the nut into place against the support. The spring loaded mount was intended to allow the nut to come out the bottom in the event that the head was driven into the bed. (This has happened several times in testing, so it probably saved the glass.) I have lost that capability with the new mounts.

As for the motors:

I did try them under no load. They were configured to allow 780mm/s with a 0.8mm lead, meaning 975 RPM. They seem to be fine spinning at that RPM if you can get it there, but accelerating to and decelerating from that has about a 75% chance producing a noise that makes a dentist's drill sound pleasant and calming by comparison. The one motor is definitely worse than the other, but both are doing it. I pulled the motors apart and the support bushings at the output shaft are spalled. I replaced them with spare parts from the RC car bin, but installed the new motors with the T8x8 lead screws anyway. These will manage the same speed at 97.5 RPM, which is definitely more sustainable. With a full-step margin of 0.04mm, this shouldn't post a problem when I typically print with 0.2mm layer heights, which is 5 full steps. I guess I could change my "fine" resolution from 0.1mm to 0.12mm (3 full steps) but I'm guessing that it staying at a half step will probably be okay.

Anyway, it was indeed the motors causing the problem. I'm not sure if I got something in them or if it's just because they were probably the cheapest motors that the kit provider could find and they were crappy. One of the other motors in that set had already failed, which is why I had the replacement motors on-hand.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

How dead are these motors ? Open coil, seized bearings or something got loose inside and is interfering with rotation ? The coils wire can take a lot more current than what the driver can supply. Now, shitty connection could come loose too eg cold solder.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

The bearings weren't siezed, but had some spalling on the shaft surface. The shaft was a little scratched where it was. I think that it was probably "crap got in there" that shouldn't. They would probably be okay if they were kept to low RPM, but they were relatively high RPM with the 5mm 0.8mm lead screws, They appear to work okay with the replacement bearings, but I haven't driven them past 100 RPM, and haven't put them under any load. The coils seem okay, and the insides seem clean enough. I did clean out some old grease, and put a bit of fresh lithium grease on the bearing surfaces when I replaced the bushing/bearings.

Edit: The motor that had failed earlier had a coil wire break loose and short circuit, which also took the motor driver on the RAMPS board with it.

Edited 1 time(s). Last edit at 10/01/2018 11:26AM by SupraGuy.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

Should I change course before wasting time & plastic?
I'm not sure if I should continue with a completely beltless machine as I expect some issues, but it is fun playing with the design,
I need to order more leadscrews as the one I have on X while fine...the motor probably wouldnt be upto the job, will try a better/larger lead one(but then have to sort backlash).

I'm holding back on the designs of thee final parts, long enough to think of some permutations...I dont want to waste filament..and would like the final parts to at least have 2 uses. I want the option to replace the X axis drive arrangement from leadscrew to belt. So I have a couple of parts taking shape. I have to print a new cart...and I'm playing with a lever idea for the leadscrew nut trap(like on the power feed of a lathe) But also with attachment points for the belt...trouble is I'm close but it's tricky to have the belt where I want it, it's competing for space with the screw...and i'd probably have to remove the screw and just go belt....now I can see this coming...but really want to see how it goes with the screw because it is simple looking with screw.

I'm hoping to power both z leadscrews from one with a closed loop belt...distance between screws 445mm and because they are in line with the 12mm rods I'll need to use 36 or 32? pulleys on the leadscrew...might have to figure some belt tensioner.
I'm playing with a method for isolation any z wobble will have to see how that turns out, looks promising




Edited 1 time(s). Last edit at 10/15/2018 10:26AM by MechaBits.

Well, I'm quite happy with the print quality that I'm getting from my leadscrew driven printer. Using the Duet electronics, print head travel speed isn't a problem, I'm able to get up to 100mm/s in either X or Y directions, and actual print speed is fast enough to be limited by the extruder, and not the leadscrews.

I had a Z wobble issue, which I resolved. When I got that taken care of, to my surprise, I found that I had a wobble on the X axis. So yeah... |that was a thing. I've got it taken care of for now by putting a flexible coupling on the X motor in place of the solid one. I thought that this might result in some ringing artifacts, but that does not appear to be the case.

One thing that I like with the leadscrews is the high accuracy. I'm having much better luck making parts that fit together than I had with my belt driven printers, even when they are quite well dialed in.

Throughout this whole project I've been treating filament as entirely expendable. I've gone through a fair amount of revision by print parts, test fit them, stress test them, change the design, rinse and repeat. I have many versions of the drive parts wile I was still trying to preserve the general I3 frame (Which I later completely abandoned, so I still have a bunch of printer parts around, so maybe I'll just build another I3 and put a 1mm volcano nozzle on it for rapid/low res printing.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

Well that's good news, though I wasn't planning on a 32bit for this...might have to the fine leadscrew was for the mill but as the design went on with that I wish I had a smaller 200mm one for the mill, but if I carry on with the beltless and forget abut the belts, I save a load of hassle(multifuntion parts create a lot of revisions I dont want to print them all) & plastic, though I may add a belted gantry at the back of current one. I'm not too worried about Z wobble but X is a different issue, my z nut can move a little as it spins so I hope its enough of a decouple. A part for the ends of X to join Z has been through a number of revisions but the one I just printed ended up reduced from probably over 100g to 17g...not sure if it will do the trick, cant really test until I have 2 leadscrews, seems ok will post pic when i have one.

I'm pretty sure that you can reach reasonable speeds on a RAMPS type board, like say an MKS type. Probably 24V would be a safer supply voltage. I don't know about printing at 100mm/s that way, but then I don't do that as is.

Solving wobble isn't hard to manage with a decent design. Oldham couplers and the like will do their parts, but I didn't find it actually necessary, though I did spend some time getting things aligned precisely. I only had the problem after making a change to the X axis, and then not spending quite enough time to make sure that the coupling was straight.

Yeah, I went through a number of iterations. This printer started out as replacement parts on an I3 clone frame, and ended up as this. The gantry parts were initially meant to hold 5mm Z nuts in with just gravity (To allow them to come out the bottom rather than drive the print head into the bed) They were modified to decrease their height, then again to increase their strength, then again to spring-load the nuts into place, but still allwo them to come free, and then finally to use the 8mm lead screws instead of the 5mm/0.8mm lead threaded rod. I'm still thinking about the idea of modifying them again to restore the spring-loaded anti head crash properties (Though that did stop being a problem after I got other parts of the motion system dialed in.) I'd also like to make the motor mounts a little more rigid, and add some cable management points to it.

My head carriage was initially designed to hold a J-head and use a Bowden extruder, but I ended up with a direct drive, so I never did print the J-head holder portion.

Speaking frankly, I don't think that I'd do it again, but I definitely don't regret designing this printer the way that I did, and I plan on continuing to use it with no major changes for the forseeable future.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

I would like to get some 24v bits as I might need them for cnc...maybe more for that though,
and at some point I want 32bit controller to tinker with, just holding out so i dont make an expensive mistake,
as my leadscrews are 400mm they are more susceptible to arrive bent, the ones I have seem ok though maybe a slight bend, and I think I need at least 2 more maybe 3?

Aligning this printer is pretty easy, lots of straight lines to help out, only a few plastic parts,
besides the hotend cart, there are 2x 12mm rod ends, 3x 8mm bearing holders, and 2 small parts to help with z lift.
maybe 150g total, 190g with the 4545 nuts i printed to save a fortune.

My X gantry sits on the lower bearings, the long ones, so they really arent doing much except lift, at the moment the gantry rides on them so could seperate if touching the bed...I have a feeling I might need to fix that. The slots on the motor mount just wheren't wide enough for me to use with the longer bearings, I might make something in the future so I can make better use of them,

The linear rail is 400 but I have another 450...l could have made it a little wider but I think things line up better this way, though was thinking of adding the other rail behind the other one, on the other side of the L plates, just to put a direct extruder on.


How my Z evolved so I only printed a tiny part, I really wasnt looking forward to a big print.

One drive system you could consider is the so called "servobelt". It uses a fixed, glued to a flat surface length of belt, and then another belt on top of that running in the omega drive configuration, with the teeth of the two belts interlocking. This limits belt stretch to the short length between the two idlers on either side of the drive pulley, since past that the drive belt is meshed with the glued down belt which stops it from stretching (assuming you glue it down to something solid). It's a cheap-ish alternative to rack and pinion, keeping mass low and avoiding the binding and whipping issues of leadscrews, and the belt stretch issues of belts.

wouldnt be beltless then, what's the benefits over just standard omega without the extra belt, I imagine the one with the extra belt is better for moving the profile(if it was glued to profile).

Edited 1 time(s). Last edit at 10/18/2018 12:18AM by MechaBits.

Well, some of the things that I was looking at as advantages to going beltless aren't just avoiding stretch.

one example: 400 microsteps/mm is 25 FULL steps/mm, which gives pretty decent resolution where I don't have to worry about full-steps. 92 microsteps/mm is less than 6 full steps/mm, which means that something knocking the print head to it's nearest full step could result in a problematic layer shift. I have had this happen on my I3. this also means that I can print with very high dimensional accuracy, which is sometimes important to me.

For another,it's an unusual configuration, and I kind of like that.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

Quote
MechaBits
wouldnt be beltless then, what's the benefits over just standard omega without the extra belt, I imagine the one with the extra belt is better for moving the profile(if it was glued to profile).

The advantage is as stated, much lower belt stretch since only a small length of belt (between the two idlers) can stretch. Past that point the belt locks teeth with the one glued to the profile, which wont stretch unless whatever it's glued too stretches (yikes!) or the adhesive shears (also yikes! or use better adhesive).

Supraguy is right that it doesn't come with the extra resolution, but you can do what a lot of rack and pinion drives do and gear down the motor before it is coupled to the drive gear.

here's a video which might help explain it:
[www.youtube.com]

Using the belt as rack&pinion is interesting for CNC, but I wouldn't like the moving mass of the stepper on a 3D printer.
Any chance to drive it with a geared remote drive like Zesty nimble or Flex3drive? Or would their rotating shafts cause more problems? ( torsional backlash or whatever)

BTW: Why do they need the second belt anyway? A drivepulley without collar would lock into the glued belt as well. right?

Edited 1 time(s). Last edit at 10/19/2018 02:51AM by o_lampe.

Yeah but it would have more teeth to engage with 2 belt, nice & wide too.

Quote
o_lampe
Using the belt as rack&pinion is interesting for CNC, but I wouldn't like the moving mass of the stepper on a 3D printer.
Any chance to drive it with a geared remote drive like Zesty nimble or Flex3drive? Or would their rotating shafts cause more problems? ( torsional backlash or whatever)

BTW: Why do they need the second belt anyway? A drivepulley without collar would lock into the glued belt as well. right?

And what about the dust that will be squeezed between these belts ?
When rack and pinion are used, better have the rack up and the pinion under.
Note the belt tooth profile is not fit to work as a true rack and pinion.
Anyway, there is more, just read the comment by Kevin Delaney.

Edited 1 time(s). Last edit at 10/20/2018 04:49AM by MKSA.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

If it helps I am using an MKS GEN-L in a custom built belt driven 300mm cube and it is happily running at 120mm/sec, which is the limit of the current extruder setup.

Quote
MKSA
And what about the dust that will be squeezed between these belts ?
When rack and pinion are used, better have the rack up and the pinion under.
Note the belt tooth profile is not fit to work as a true rack and pinion.
Anyway, there is more, just read the comment by Kevin Delaney.

You don't need to tension the top belt so it permanently locks against the bottom belt, you can leave the parts on either side of the idler completely slack and free floating so dust can easily be blown or dusted off, check the video. Regardless, even a plain old belt drive or screw drive can have dust and debris get stuck between the pulley and belt, or nut and leadscrew, so I don't see how it's an issue unique to this setup. As for the profile not matching up, the belts lock together pretty well (if not perfectly, I'm not an expert on the tooth geometry but I'm sure there is a tooth standard that fits with itself perfectly) and the actual actuation between the drive pulley and the driven belt is exactly the same as any other belt drive.

The reason you'd do this over just using a belt as a rack and a pulley as a pinion is tooth engagement. A pulley running along a belt as a rack and pinion is no different to how the belt in a regular belt drive comes in tangentially.

Quote

The reason you'd do this over just using a belt as a rack and a pulley as a pinion is tooth engagement. A pulley running along a belt as a rack and pinion is no different to how the belt in a regular belt drive comes in tangentially.

OK, I take back my former comment about 'the second belt is obsolete'.
Now I rather think of using a closed loop belt around idlers placed on the stepper motor fastener screws. ( like my 'cloverleaf cable drive', but with a belt )
The idlers in contact with the glued belt are fixed in place, the other two idlers rest on excenters like in the video to tension the belt.
But still the stepper would move with the gantry

I fail to understand why you try to replace timing belts in a 3D printer by lead screws and lead screws by belts in a router (aka CNC here).
And frankly the "rack and pinion" seen here [www.youtube.com]. is amusing but not convincing Again the belt tooth profile is not fit for it and you don't have that many tooth engaged. The dust is an issue.
A large toothed pulley, may be two would do a better job.
OK to machine Styrofoam, ash wood
If you look at real machines, leadscrews are protected or have wipers. (in some cases Igus leadscrews are OK without) or rack are mounted with the pinion below.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Well, my leadscrew 3D printer is an exercise in design implementation for building a CNC router, at least in part. It was conceived because I had a useless I3, or at least most of an I3 around that I didn't have much use for, and I had another 3D printer that needed stuff that it wasn't quite up to doing for itself. Plus, I had almost all of the parts before I started building. Some of those, I bought when I was thinking about the idea of using screws, but before I bought my second printer. That I3 left a bit of a bad taste with how bad those belts were. I was constantly trying to fix belt tension, fighting with keeping the alignment, and I still had problems with 1cm over here was different from 1cm over there. Even though the X and Y belts were supposedly the same, the steps/mm were different. I could never get prints that were supposed to interlock or be able to be fit together to actually fit.

The MakerBot clone printer was able to do that, and X and Y steps/mm were consistent with each other and with the reality of the printed part, but by that time, I had pretty much committed to this being how it is.

I am also (albeit very slowly) building a CNC router. I have many of the parts purchased. Some assemblies are being prototyped with the 3D printer, but the plan is to replace many of those with routed pieces which I believe will be stronger and/or more rigid once the router is up and running. I have learned a few things through the design and building process of this printer which I plant to take forward into the design and building of the router.

In the meantime, as I've said, I'm quite pleased with the performance of this model. This thread was started very early in the build process, and before a near complete redesign which resulted in the printer that exists now. the pieces that I had made at that point were all intended to be replaced, though I had hoped to be able to get those up and running. Some of the reason for the redesign was assumptions that I had made that I'd be able to find parts that matched certain specifications. When I later discovered that the specifications that I had were one-of parts that were impossible to find "in the wild" I started looking at what I could so with the parts that I could get. I still ended up with some non-standard pieces that required me to do some unusual fabrication, which I'd like to change so that another such printer could be made without the need. Well, I suppose that it's no weirder than the laser-cut acrylic pieces.

I can assure you that I have no intention whatsoever of using belts for a CNC router. For the router, I will want the best precision that I can get, and standard 8mm threaded rod (1.25mm pitch/lead) will do nicely. I do not plan to run a spindle through material much faster than a few mm/s.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts