Struggling to identify problem.

So I'm new to 3d printers. In a pretty poor choice, I decided to buy a cheap Chinese (afinibot) delta kit as my first printer. It's more like a rostock design I believe with the 8mm rods and bearings, and acrylic plates for the frame instead of wood.

Anyway the quality of the printed parts is abysmal, I had to get replacements for most of them, and had a friend print a new effector because both that came with their kit have 4mm holes for the rods to screw into on 3mm bolts, along with being warped like all of their parts which suffer from the edges lifting off the bed when they were printed.

So all that said and done, I've put a ton of time into trying to get it dimensionally accurate, watched a lot of guides, but I'm stuck and not sure if it's just too many bad parts adding up to a mess or something I'm missing.

I've checked that the towers are on the right angles, the distance between the towers is the same at the top and bottom, the bed is squared to the towers. Because they printed the holes wrong for the endstop screws also, only 2 out of 3 actually bite enough to use, so I've just screwed them all the way in and tried to rely on software offsets rather than mechanical alignment with the screws. The bed is also not perfectly flat, and made of cheap aluminum, but it's not super far out, It dips maybe 0.1mm here and there, I do have a glass plate coming to replace it. I measured the arms, I don't have a big enough caliper yet to get a great measurement so I had to use a ruler, I do have a 300mm caliper coming for better measurements, but It appears 4 of the arms are equal and the remaining 2 might be 0.5 mm longer than the rest.

The problem I'm having right now is the radius settings. I can complete an endstop calibration without too much trouble but I cannot find consistency on adjusting the horizontal radius. This is using repetier 0.91 which is the firmware they preloaded on it. Once I level my endstops, my build surface appears convex, I can move the horizontal radius and it appears to flatten the build surface until I have the same distance with a piece of paper at all the towers, and the center. All guides seem to stop here and indicate this is the procedure, and this should be good.

However, It's not flat. while its flat between the towers and center. If I move the effector from center along the X and Y planes it's completely out of whack, contacting the build surface on one side, lifting away on the other, Contacting again and lifting on the Y plane as well moving up and down the z tower. I did manage to pull off a calibration object print last night and it wasn't even that far out according to the object. 60.6 on the X, 60.4 on the Z 60.4 on the y for a 60mm object. I Adjusted the rod length offsets as per the guide I found on thingiverse for the calibration object hoping that would correct inconsistency in rod length and start moving toward correcting it out, but I couldn't even get it level enough to lay down the first layer after that.

The endstops seem to creep and act strange as well compared to guides i've seen. Particularly the Z. I go around the cricle from x to z adjusting the offset. and Z just keeps creeping and creeping smaller and smaller than the x and y offsets each pass. Eventually I get it to where it looks good then it does unpredictable things like suddenly I need to go back the other direction a huge amount. Like last night I had been adjusting them, I got to the point where it was just about the same all around, and my Z offset was 3 steps. I checked them all again one more time, X and Y were good, suddenly z is super tight, I had to move it back up to 30 steps to correct.

I have noticed some strangeness in the movement as well that leads me to question if I could have issues with the board itself. Sometimes after I home, and send it to one of the tower calibration positions, example g1 x0 y0 z80, it will randomly surge, like it makes 3 jumping movements like it's struggling to move or something. If this happens I usually re-home it and try again and it doesn't do it. Just once in a while it randomly surges like that. Also I've only seen it do it twice, but I was at the Y tower calibration position, and I clicked the script button I have set to go to the Z tower position and it just flung over at way, way too fast of a rate compared to what was set for movement speed, like the surging though, this is rare and inconsistent and only happened twice.

Make sure your endstops work. I.e. the motors will stop when they go up against the endstops.

Set proper current on stepper drivers for your motors. Set the current to the maximum motor current but not more than 1.4 A. If your motors require more than 2.5 A then you may need to get different motors (ideally with the rated current around 1.5 A).

Measure precisely the diagonal rods. You should find a way how to do this with at least 0.1 mm precision. I do not really know how to do it without a jig in which the rods were built. Pair your rods so that the rods in one pair have equal length. If you cannot pair them this way then throw them into a garbage bin and build a new set of rods in a jig. Make all the rods in the same jig so they have precisely the same length. If you do not have rods of equal length on one tower you will not get anywhere.

Put one pair of equal length rods on one tower.

Check that your printer does not have any play.

Get a z-probe.

Update the firmware so that you have set proper rod length for each tower.

Find the dc42's web calibration page and check whether it supports rods of different length. If yes then calibrate using the page. If not (and you have a bit of clue about math) then modify this calibration procedure so that it supports different diagonal rod sizes. The calibration web page (maxima notebook) will compute for you the proper calibration parameters (your tower positions (delta radius) and endstop adjustment).

Looks like your printer is in a pretty bad shape. Good luck.

Edited 1 time(s). Last edit at 09/29/2016 05:14PM by hercek.

I know the rods are less than perfect. I guess I have to wait for my new caliper to arrive to measure them to that degree of accuracy. I did pair them as best I could for now just measuring with the steel rule from my square. I'm kind of stuck with this kit now and sorry I wasted money on it. Live and learn I guess.

I did order a set of 2020 Kossel XL parts from a better source to make another from scratch, as I did enjoy the process. I also got a set of replacement rods and ends with those parts when I get them in the hopes of getting a precise printer going to print parts to fix this one, then either give it away to family or sell it when it's working right. I'm debating whether it's worth spending any more time and money on it or just cutting the loss and calling it.

I saw a calibration tool that looks promising in this forum further down, possibly the one you mean. I'm going to try that one tonight and see if it can at least get things moving forward. I'm at a loss for things like motor current because being a cheap chinese kit, they tell you basically nothing about anything spec wise so it's all a mystery what you even have, and provide the most barely comprehensible user guide with it. I assume they're nema 17 motors, but not being knowledgeable about this stuff yet, and not wanting to blow shit up I don't want to make assumptions.

That does lead to one question though. The replacement rods I got are 220s and this one has 185s, Is it ok to use longer rods without changing the rest of the geometry completely to accommodate them? I'd rather not cut them down if possible because then I at least have a more standard set of mini rods to salvage should I want to scrap this one to make a different mini in the future.

Edited 2 time(s). Last edit at 09/29/2016 05:59PM by feistylemur.

You can use longer rods. Just set the correct length in the firmware.

Pulling motor current high is only to make sure you do not skip steps. 1.4 A is about the maximum A4988 will handle without thermal shutdowns (if it has a small heat sink and some air movement around). If your motors are driven near their maximum current you will hear very noticeable "dum" sound when they skip a step. To hear how does it sound drive the stepper motor and hold it with hand so that you hear it and know what it is. If your machine is not repeatable then skipping steps or loose pulleys are probably the most common reasons. If you believe that your steppers are not skipping steps then you can skip increase of the current.

Edited 1 time(s). Last edit at 09/29/2016 08:16PM by hercek.

I've been playing around with it tonight. I tried that eschertek calculator and it was disastrous. I don't know what I was doing wrong but it was giving me values for the Z offset in the completely wrong direction.

But the behavior issue is consistent for the most part beyond the two scenarios with the surging and the 2 times it went completely the wrong speed. I did a full reset of all values and a basic calibration. So I set my Z height. Set all my endstop offsets. Reset my Z height. Now I adjusted my horizontal radius to bring the ends up as they were concave at the towers. So now the center is on, and the 3 endpoints under each tower are on. however, If I swing to the opposite end of that point. For example X0 Y80 Z80 is just right, center is just right, X0 Y80 Z-80 is an entire milimeter high. This is the same for all 3 towers. The opposing end is a milimeter high.

Hi FeistyLemur,

Welcome to the frustrating world of Deltas. Hopefully you can get past the 'hump' and get some nice prints. The part you're at now is the most difficult, by far. Maybe these tips can help.

1) To verify that the length of your rods is the same, you need a jig. I made one using a couple of bits of aluminium angle, and set it up so that there was a 3mm screw sticking out. These angle pieces were held to an upright on the printer with the normal T nuts you use to assemble the printer. Stick one part near the bottom of the printer, and the other near the top, and adjust the length between them until a diagonal rod can easily slip on and off both ends. Now put all your rods onto this jig, one at a time, and make sure that all of them can easily slip on and off.
2) To measure this distance with a small caliper, stick another 3mm bolt into a T nut and screw this into the middle of the jig. You can now measure from each end bolt to the centre bolt, and also measure the width of the bolts. This means that you can add some measurements to get the total distance from the centres of the end bolts to each other, with an accuracy of about 0.1mm.
3) It's also very important to ensure that the width of the effector between rod ends is exactly the same as the width between rod ends on the carriages. I find this part extremely difficult to measure. Maybe you could use a washer between the bolt and the rod end, and measure the distance between the washers.

It sounds to me like your measurement of the diagonal rods might be out a bit. It also sounds like you might have some slop in one of your endstop switches (or the set screws on your motor pulleys).

I've found that the commissioning process on a Delta is a process of trying to eliminate every possible source of inaccuracy, going around and around in circles, trying loads of different things, getting frustrated, and finally watching the problem disappear after throwing loads of time at it.

Good luck.

I've been doing a lot of experimenting tonight.

I put the original effector, crappy as it is, back in place for troubleshooting, aligning all the screws for the rods to the same part of the hole (which has half a mm of slop around the screw) the top side of the effector. I remembered during my setup at some point i changed the steps per mm, and that had got set back to 100, the firmware default. So I measured the travel of the carriages over 100mm at the default, and found that it came up short at 95 mm. So I measured and adjusted and ended up at 100.18 steps per mm. This got me a great deal closer. My 1mm out on the axis is now .20 out.

So in addition to that I bumped 1.5mm of rod offset on each diagnal offset. This put it pretty level all the way down the axis. with a few ups and downs that could honestly be the warped aluminum.

So now i'm once again laying down the first layer of a calibration asterisk to continue to fine tuning and I'm pretty happy to see that actually going down right at this point.

When my new glass bed arrives and I get my caliper to measure stuff more effectively hopefully I can finally get it dimensionally accurate.

Edited 1 time(s). Last edit at 09/29/2016 10:50PM by feistylemur.

First you must be sure the printer is repeatable. It may be imprecise and badly calibrated but it must behave the same (when exposed to the same input) all the time. The most common reason for printer not being repeatable is lose pulley on a stepper motor, lose belts, or skipping stepper, loose endstop, or slop somewhere in the mechanics. Other improbable options are bad firmware or bad electronics. You must get rid of this (make the machine repeatable) otherwise it does not make sense to do anything else.

100.18 steps per mm is a strange value, I would almost say that you have somehow faulty belts.

If the only errors in the machine are incorrect (virtual) tower positions or endstop adjustments then you typically get at most one dip (or one raise) when traveling over a straight line (projected to XY-plane) just above bed from tower base to the opposite side of the tower. I.e. only one "hump" which may or may not return back to the original z-level (or overshoot the original z-level). Sometimes you can get two "humps". If you get more than two "humps" then you should check whether you do not have eccentric pulleys or non-planar bed.

The top pulleys that came with the kit are complete garbage. It's basically a single bearing, a 5mm bolt, and a bunch of washers stacked up to resemble a pulley and inserted through an oval hole which is meant for adjustment in the idler end and bolted down. They don't sit at a perfect 90 degrees for sure because the one on the X axis is eating away at the belt very slightly as it chews against the washer, I can see the belt wearing there. I did also order some of those nice plastic wheels with dual bearings and nylon bushings to replace them but I won't have them for a while. At the same time I ordered a bunch of belt just to have extra, so It's something I could replace. The belt is basically riding right on the surface of the bearing at the top.

I can guarantee the bed's not even close to planar. I put a straight edge down it and I see light under it in places or center rises in others.

The endstops are very solid. They screw right into the face of the idler end with two screws and they don't move at all. The biggest potential for repeatability error I can see is the effector and the carriages, which are awful. The holes in the effector are way oversize inexplicably. Even the holes for the nut on the reverse side are too big. They look as though they were printed for a 4mm bolt instead of 3, so it would be extremely easy for them to move if for example you pulled on an arm hard. It's extremely hard to tighten or center them because the nuts just spin in the recessed hole they go in. Basically I had to hold the shaft of the bolt against the top of the hole and try to hold it just so while tightening and jamming the nut against the top of its socket to make it tighten. Then wedge a screwdriver in the other side to hold it while snugging it up. This is why I had a friend print a new effector. But I think there may be some dimensional error in what he printed throwing it off worse, so I had to switch to the original for troubleshooting.

As for the carriages. They're terrible too. They're short and very wide. The quality of the print is again abysmal and the edges are warped very slightly. I had to request replacements because the originals were warped upward to the tune of about 2mm. The faces where the rods connect are the same size and level though, so they work. But they don't feel super smooth. The linear bearings are pretty grungy feeling and cheap.

If I could get just enough dimensional accuracy and quality to print new parts I would reprint everything. I considered going onto 3dhubs and getting someone to do that. But I haven't decided how much more money I want to waste on trying to make this thing work before calling it an abomination and focusing on the XL instead.

Edited 2 time(s). Last edit at 09/30/2016 10:46AM by feistylemur.

Quote
hercek

100.18 steps per mm is a strange value, I would almost say that you have somehow faulty belts.

I've measured similar on both my printers. If you haven't measured the belts then you might be in for a surprise. I'm convinced that normal GT2 belt is designed for much higher tensions than we normally run it at, so that it stretches to the 2mm per tooth value that everyone thinks it is. I've consistently measured it to be around 0.2 - 0.3% undersize.

Quote
feistylemur
I've been playing around with it tonight. I tried that eschertek calculator and it was disastrous. I don't know what I was doing wrong but it was giving me values for the Z offset in the completely wrong direction.

Most likely you entered the height errors with the wrong sign. I state in the instructions that you need to enter a positive height error where the nozzle is too high and a negative height error where it is too low, but still some people seem to enter them the other way round and then wonder why the results diverge instead of converging.

Edited 1 time(s). Last edit at 09/30/2016 11:19AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Hmm, I did read that in the instructions, but I didn't really get the wording for some reason. I'll try it again this evening.

Quote
nebbian

Quote
hercek
100.18 steps per mm is a strange value, I would almost say that you have somehow faulty belts.

I've measured similar on both my printers. If you haven't measured the belts then you might be in for a surprise. I'm convinced that normal GT2 belt is designed for much higher tensions than we normally run it at, so that it stretches to the 2mm per tooth value that everyone thinks it is. I've consistently measured it to be around 0.2 - 0.3% undersize.

Interesting. I use T2.5 steel core belt and its length is right. It is actually quite easy to measure the belt length precisely when you can already print (the bed is leveled with possibly bad steps per mm). Print something and measure with calipers whether the height is right.
Maybe you need to tighten them a bit more ... hopefully the idler will not tear away. The belts are supposed to produce a nice bass tone when plucked

I do not intend to buy any other belts than steel core ones because steel core is the most stiff. And it is harder to get GT2 steel core belts. Well carbon fiber belts are better than steel but I have not seen them in any shop yet.

Although some people claim steel core is not good since it wears out more quickly. But my belts still work fine so I'm not concerned about that.

@feistylemur: You do not need to try to center 3 mm bolts in 4 mm holes. You can use e.g. a piece of wood to push 3 mm bolts to one side of the hole. Stick it into the hole together with the bolt. Just make sure that you push all the bolts in the same direction. E.g. on the platform, push them all outside of the platform center; on the carriages, push them all nearer to the rods.

Quote
dc42

Quote
feistylemur
I've been playing around with it tonight. I tried that eschertek calculator and it was disastrous. I don't know what I was doing wrong but it was giving me values for the Z offset in the completely wrong direction.

Most likely you entered the height errors with the wrong sign. I state in the instructions that you need to enter a positive height error where the nozzle is too high and a negative height error where it is too low, but still some people seem to enter them the other way round and then wonder why the results diverge instead of converging.

That calculator is a thing of joy and beauty.

Something I found that was useful was to use a piece of .1mm shim stock (or paper if you don't have the shim stock on hand...) to check/set the nozzle offset from the build plate, and use feeler gauges to get the height off the bed if the nozzle was too high. At first I tried using the .1mm feeler gauge from my set; but that led to (very stupid) math errors, which then led to useless iterations on dc42's calculator. Once I eliminated the human error from the measurements I was feeding into the calculator, I had my printer dialed in very quickly. I had less good results when I calibrated off the surface an aluminum heated build plate (which was not very flat) than when I calibrated off the surface of a polished granite tile. (which is VERY flat)

I also reversed the signs on my measurements once. That went pear shaped immediately.

Good luck!

Well I put the calculator to use tonight and it brought me a long way, it prints way more flat than I ever managed to achieve. But it can't seem to calibrate out all the problems. Basically i ran it through about 4 times and it's at the stage where it cant quite get the height error out along the axis.

Meaning along each tower axis the nozzle rises from 0.1 to 0.25 by the time it reaches the farthest point from the tower (X and Z are .1 y is .25), no matter how many times I run it. I have to guess this could be due to rod length error, and dimensional error perhaps in the effector itself given how poorly printed it is. I notice changing the rod length offset along the axis affects this if I play around with it, but that just throws everything else off and I have no way to measure them with fine accuracy just yet. Sadly they came pre-assembled so I'm stuck with what the kit had at the moment.

I guess for now I'll have to live with it sucking and hope I can get my XL built more properly when I get the stuff for it then maybe replace the shoddily made parts and hopefully construct a properly matched rod set for this one as well.

Edited 3 time(s). Last edit at 10/01/2016 12:32AM by feistylemur.

Quote
feistylemur
Meaning along each tower axis the nozzle rises from 0.1 to 0.25 by the time it reaches the farthest point from the tower (X and Z are .1 y is .25), no matter how many times I run it. I have to guess this could be due to rod length error, and dimensional error perhaps in the effector itself given how poorly printed it is.

I looked at dc42's calibration web page. It does not seem to support different lengths for the diagonal rods. Therefore you cannot expect it to properly calibrate your printer when one tower has about 0.5 mm different diagonal rods from the other two towers. You have two options:

• Remake your diagonal rods so that they are all the same.
• Take my maxima notebook and modify the equations so that they do not use one r (for length of all the diagonal rods) but 3 variables r1, r2, and r3 (which would specify diagonal rod length for tower A, B, and C respectively). Let maxima derive the equations for ZFnsSansR or FAll (but without pulling out r as an optimization parameter). The procedure to derive them is the same as when there is only one common r. Let maxima compute RVr calibration results.

Quote
hercek
It is actually quite easy to measure the belt length precisely when you can already print (the bed is leveled with possibly bad steps per mm). Print something and measure with calipers whether the height is right.
Maybe you need to tighten them a bit more ... hopefully the idler will not tear away. The belts are supposed to produce a nice bass tone when plucked

Printing something won't give an exact height, because:
1) Your first layer might be off.
2) It doesn't isolate each tower's different belt stretch values
3) 0.2% is difficult to measure in plastic, because the plastic often shrinks more than that amount.

The way to do it is to use a caliper, and use the carriage to push your caliper closed. I use a computer interface to push the effector down 100mm in steps of 10mm. This gives a good idea of whether or not your steps per mm is correct, and also adjusts for different steps per mm on the different towers. I find this difficult to get precise readings, but it's within around 0.1% repeatability, and shows that the belts are consistently 0.2-0.3% undersize.

I would guess that you think it's perfect, but haven't actually measured it.

My belts all sound like a bass string when plucked (I tuned them to the same note with tensioners).

Edited 1 time(s). Last edit at 10/01/2016 11:14AM by nebbian.

Interesting, I did not measure belts with calipers on carriages. But it sounds like a better idea than measuring the object height. At least it is more direct. Which is good.

I always only measured the height (and width) of the objects printed. They are always bigger by about 0.1 mm (regardless of the object size is about 2 cm or 15 cm). I think it is because of the surface of the plastic is not perfectly flat and probably extruder not being calibrated well enough. So maybe the belts are not correct but when the object size is fine I'm not going to fiddle with it. I think it would not be better. The reason I think so is because I believe that the numerical calibration will tie Z dimension to the XY dimension if it converges. (Provided you do not have tilting problem on platform/carriages). In other words. If you correctly calibrate to get a leveled bed using the numerical optimization then the object will be uniformly bigger/smaller if belt steps/mm is not correct. That assumes the belts are stretched the same on all towers which is not unreasonable if you use the same belts everywhere. In other words, I quite believe my belts are precise enough.

Did you find the belt length error also on steel core belts? Because glass is about 2 to 4 times as stretchy as steel.

I am not particularly good at math, so I don't think attempting to modify a mathematical formula is going to go well for me, plus I don't have a way to measure them precisely yet. I am ok at building and fixing things so I think I'll have to wait till I get the other rods and parts I ordered and make a new set. They and my new caliper should arrive soon anyway with any luck.

Hercek,

The belts on my printers are the cheapest, nastiest, default options that are available at Robotdigg. I don't even know what the core material is. It's quite possible that steel core belts are a lot more accurate than glass.

Just putting it out there that not all GT2 belts are actually 2mm between teeth.

@nebbian:
OK, I measured my belts and they seem to be about 0.05% shorter. That was measured on 10 cm move and the measured distance was 99.95 mm. This indicated that my 10 cm big objects should be printed smaller by about 0.05 mm. This error grows/shrinks linearly with the object size. But that is such a small error that it does not make much sense to play with it because my calipers precision is 0.03 mm (repeat-ability 0.01 mm). And it was measured against a 3d printed plastic part on one side. Still this indicates that I should tension my belts by another 0.75 mm (over the whole belt length of about 1.5 m). This data point is for a steel core T2.5 belt.

I think that GT2 tooth profile is better for precision than T2.5. But most common GT2 belts in 3dPrinter shops use glass core. It is possible to buy steel core GT2 belt. You can look at one end of your belt. If it has steel core you will see the metal wires in it quite clearly.
I posted some info about belts here: [forums.reprap.org]

Well I got my big caliper and measured the arms.

I have

184.82
184.82
184.82
184.97
185
185.14

So I can't form perfectly matched pairs with this set. Is the difference unworkable without tossing them and making a new set?

I do not really know. I did not try to derive equations for this because I never had any noticeable platform tilting/rotation problems.

The math I and dc42 uses is not valid any more but it should approximately work. It is probably still better than trying to guess it manually.

I may give you some hints. It looks like the platform will move slightly and mostly rotate around the points which are at the intersection of axes where each axe is created by connecting centers of platform ball ends going to one tower. The rotation axes seem to be going through these points and are perpendicular to the plane of diagonal rod ends of the tower which has imprecise rod. And the most significant order terms of the dependency of the hotend 3D position error should be something like r * d / D.
Where D is distance between two parallel diagonal rods; d is the diagonal rod length error; r is hotend tip distance from the axe of platform tilting/rotation.
The exact formula will have more terms than the ones mentioned above which will add dependency on the complete printer geometry and on the position of hotend over the bed. The error will be mostly bigger at the bed edges than in the center because r is bigger at the edges. Most of the hotned 3D position error will contribute to XY dimensions. The smaller part will contribute to the Z dimension error.

E.g. for my printer the 3D position error of hotend tip would be somewhere in the range of 10 * 0.32 / 75 = 0.04 mm.
10 mm is very approximate distance of my hotend tip from the axe of rotation (the rotation would happen because of your 0.32 mm error in the diagonal rod length). 75 mm is my distance between angle ball joints (of one tower) on the platform. At the edge of the bed it would be about 3 times bigger.

There are other geometric problems with the frame causing issues and making it tougher to work with. One problem is the base and top are acrylic sheets. They flex quite a bit and are not terribly sturdy, so the printer could conceivably shift based on the surface it's on. In addition to this, I tore it down again to check everything, and the holes don't even line up perfectly on the acrylic sheets. Because of the absolutely terrible quality of the 3d printed parts (they are all warped to some degree, some worse than others), the y tower is quite a bit out of square to the base just sitting with the top removed and measuring how the rods sit on their own in the base mounts. So once the top goes on, this has the effect of making the rods bow on that tower as well. It's quite a bit because if I slack all the bolts and just sit the top on there without pulling it together tight it differs 3mm on the y z face from the top to the bottom measuring the length between towers rod to rod. So once that gets tightened up the rod is going to be bowing considerably in the center on the y tower. This is because the motor mount on that side has more warp in it than the others, though they all have peeled up corners that lifted off the bed when they were made.

The thing is basically a complete abortion. I'm not spending 70 bucks to get the motor mounts and idler ends reprinted on 3d hubs at this point, which is about the cheapest estimate I can find. So I guess I'll leave it stripped down and wait until I hopefully build my 2020 kossel xl successfully and can tinker with trying to get this one to run by replacing all the printed parts myself. At this stage I would probably be better off to cut new plates as well, or just scrap out the whole thing and salvage the electronics and hot end to use on another project. Maybe build a 2020 mini or something.

Edited 1 time(s). Last edit at 10/11/2016 11:42AM by feistylemur.

Yes, looks like you will need to rebuilt it or scrap it. I do not know about any good way to work around bent rods. I believe that a math can be derived to work around tilted towers but as far as I know it was not done yet.

They aren't really bent per se. The base motor mount is just warped and makes it sit on an angle. A straight edge on the rod shows it's straight, just leaning out because of the printed part. Still it's probably less work and makes more sense to just print standard kossel parts and get a 2020 frame later.