Jump to content
Ultimaker Community of 3D Printing Experts
Sign in to follow this  
productdesign

New Pre assembled machine calibrated properly by Ultimaker?

Recommended Posts

Re-posted by request.:

Hi All.

 

Were just a few prints in to using our new pre assembled machine. Love the machine, but not sure its set up that well. The reason we were happy to pay an extra £500 for the machine to be pre-built and calibrated was in the hope it would be the best it could be. the test prints that were on the bed when it arrived didn't look that true. see pic.

I assume that the purpose of this part is to calibrate -seeing shapes wonky means its not calibrated hence I presume it shouldn't have passed QC ? see circle and hex )

 

So we tightened belts with wave parts. checked print head xy friction and noticed one of the axles seems eccentric and gives a slight intermittent increase in friction as it goes round eccentrically ( poss +/- 2mm out )

Dont know how to tackle that as it appears to be bourn out of the grub screw by the belt rack ( grub screw by design pushes you off centre )

As I dont know how to fix, we printed something last night which turned out ok for the most except for a poor patch see pic.

 

see all images here http://umforum.ultimaker.com/index.php?/gallery/album/139-troubleshootng/

 

Not yet contacted support but wonder if anyone can help improve the quality and consistency of the prints.

 

Cheers

 

Chris

 

Share this post


Link to post
Share on other sites

Yes, the white test piece does seem to show some signs of under-extrusion (solid surfaces that aren't closed up), and backlash resulting causing circles to come out off round. I'm a little surprised UM didn't do a better job of fixing that before shipping. That said, tightening the belts should help with that.

Regarding the orange piece - I think the main problem that I see is the droopy lines on the overhanging bulb. I suspect that is due to a lack of cooling. Make sure the fan is on, and set the minimum layer time to at least 7 seconds. This will slow down the smaller layers, so the plastic has time to cool before more is laid on top.

 

Share this post


Link to post
Share on other sites

I'm not sure of ideal settings for their PLA - it can vary by manufacturer, batch, and color. But in general, PLA is pretty forgiving. I usually print in the 220-230 temperature range, but you might go cooler if you have particularly tricky overhangs and need the plastic to be less runny, and cool quicker.

Generally speaking, you will need higher temps to print at higher volumetric speeds (i.e., the volume of plastic extruded per second = layer height x bead width x linear speed), since higher volumes mean higher pressure in the nozzle to force out more plastic per second. As a rule of thumb, anything around or over 10mm³/s may cause problems with the stock nozzle and extruder drive (under-extrusion, head blockages, grinding in the extruder drive) so you need to balance your layer height and linear speed. For very high throughputs, you'll want the temp to be 230 or 240, to lower the viscosity of the plastic as much as possible; where the throughput is lower, temp is less critical - you'll want to experiment with different temperatures to see how it affects surface texture and gloss.

Minimum layer time is on the 'Advanced' tab in Cura. The amount of time needed depends on the layer height, and extrusion temp. For layers around 0.2mm, and 220º or more, you may need 7 seconds or more. Less time is needed for lower temps, and thinner layers. When printing with 0.06mm layers, I find that 3.5 seconds is more than enough.

 

Share this post


Link to post
Share on other sites

Okayyy that's quite a bit to take in thanks. Cura says PLA is around 210 on the tooltip. On the print were discussing it was 215.

BTW I just ran this bed leveling part and noticed a problem i saw very first print. It looks like there are parts of the bed where there is under extrusion and conversely areas with over extrusion.

this photo shows the amount its put down in the first minute or so.

Is this a separate issue and how the hell do I fix it?

not that impressed that I this was'nt sorted before shipping :???:

http://umforum.ultimaker.com/index.php?/gallery/image/848-imag1844/

 

Share this post


Link to post
Share on other sites

This print isn't showing any of what you describe where one or both axes are non-linear.

To test for non-linearity you could print a grid pattern where the spacing is equal to 1/4 of the pulley diameter. It's probably visually difficult to see. It could possibly effect things like hole diameters but hole diameters are a problem regardless (they tend to be too small - easily fixed by increasing hole sizes in cad).

 

Share this post


Link to post
Share on other sites

The point of the bed leveling wizard is to allow you to adjust the screws to get the bed level; it's part of the routine maintenance that you need to do every now and then. Since the printer will have been shaken during shipping, it's not surprising that you need to adjust it a little now.

The differences in extrusion amount in the lines reflect the distance from the nozzle to the bed at each point - where the bed is lower, the plastic can flow more easily, so you get stronger lines. Where the bed is higher, you get thinner extrusion because the plastic can't get out as easily. Mainly, it seems that the bottom right corner of your bed (as viewed in the photo) is too high. So try adjusting the leveling screw in that corner - screwing it in maybe a half turn, and see how that affects things.

Generally the way that I level the bed is just to use a sheet of paper. Start by heating the nozzle, and auto-homing the head. Then in the UltiController's Prepare menu, disable the steppers, so you can move the head by hand. Move it to near the front left screw, and slide a sheet of paper under the nozzle. Tighten the screw until the paper slides freely without catching the nozzle. Then slowly undo the screw, allowing the bed to rise, until the paper just starts to scrape on the tip of the nozzle. Loosen the srew a further 1/8 turn, to account for the thickness of the paper. Then move the head to the back left screw, and repeat the process there. Continue around, adjusting as you go, and then repeat with the first corner again (in case it has been moved out of line by the other adjustments). Carry on until you get it level at all the corners. Its a little tricky at first, but you soon get to be good at it.

To be honest, the initial bed level isn't too critical. If the bed is too high, then you're first layer might not stick well (or at all), but after the first layer, it matters less. Any slight variations in level can easily be absorbed by setting the first layer height in Cura to be a little higher. Say 0.2 or 0.3mm. That will allow the plastic to flow over any slight differences in height, and give you a flat surface for all the subsequent layers to build on top of.

 

Share this post


Link to post
Share on other sites

Depending on the exact formulation, PLA will usually print anywhere in the 180 to 240 degree range. It behaves differently at different temperatures - and you can't print as fast at lower temperatures, because the plastic has a much thicker consistency, and the printer can't extrude it as easily. Hotter temperatures also usually give a glossier finish, but that is also affected by layer height and print speed.

But don't worry over much about temperature at this point. Anywhere in the 210 to 230 degree range is going to work just fine; it's something you can finesse later, once you have a feel for all of the other variables of speed, layer height, nozzle size, wall thickness, and model geometry.

 

Share this post


Link to post
Share on other sites

I have to re level pretty much every day. I have a heavy aluminum bed and the wooden cantelivered design of the UM causes it to droop tiny amounts from day to day (just as a guitar string stretches quite a bit the first few weeks). My droop is less than when I first got the aluminum bed but humidity probably messes things up as well.

I'm pretty lazy about leveling so I usually print the first layer .3mm thick and occasionally will just grab the z axis by the coupler (at the bottom) and twist it enough to force the bed closer to the nozzle. Or I'll level as it is printing the first layer with a screwdriver. But if I need a good quality bottom surface I do a good job of levelling ahead of time.

 

Share this post


Link to post
Share on other sites

Thanks for the leveling tips. I ran the leveling wizard in cura a few times. Prob every other day since I got the printer and will follow your tips with paper etc when I get the hang of things. However I am confident the bed leveling isnt what your seeing there. If you saw it operating you may understand better. I would say all 4 corners of the bed are level +/- 0.2mm.

When you watch the nozzle, it seriously only extrudes in the areas you see have more material. this is not leveling in my opinion. Im new to the UM but am an engineer. Clearly the extruder apperatus is effected by the x/y coirdinates in physical space. After all the sleve that runs from the extrude motor changes shape with x/y. Anyone heard of this happening?

 

Share this post


Link to post
Share on other sites

@ GR5. I don't think I could photo the pulley so it would show. You need to watch it spin to see the eccentricity. you can certainly feel the increase in resistance to the print head on this one axis. Should these things be sorted by UM before shipping. do you think I should contact support regarding lack of QC?

 

Share this post


Link to post
Share on other sites

It's unlikely that it's extruding less in the corner due to X/Y position issues. The extruder drive generates quite a lot of force, compared to the effects of slight changes in the shape of the Bowden tube. Furthermore, the bead is still pressed down into the tape in that corner, which means it is filling the gap between the nozzle and the tape with plastic. If the head was the same height from the tape as in the stronger-colored areas, but it was simply extruding less, then there plastic wouldn't be pressed onto the tape - it would just form a thin bead that was mostly not stuck to the tape.

The reason the bed is closer to the nozzle in that corner is either because the bed isn't level in the z plane, or because the bed isn't totally flat. It may be a little bit of both, but it's rarely a problem in practice - I think that simply lowering the screw in that corner will fix a lot of what you are seeing.

You say you think that it's level +/- 0.2mm? Well, if that corner is +0.2mm, and the layer you are printing is less than 0.2mm, then you are going to be trying to print with the nozzle pressed hard against the bed. You'd expect to get much lower or non-existent extrusion then.

You can test your theory about x/y-dependent extrusion by printing a large thin walled box thats is the same size as the extrusion test pattern, and maybe a couple of cm high. If it's consistently under-extruding in that corner, then that corner of the box will fall apart due to massive under-extrusion. I suspect however that what you will see is that after the first layer or two, once the head is no longer pressing into the bed in that corner, the print carries on just fine. :-)

 

Thanks for the leveling tips. I ran the leveling wizard in cura a few times. Prob every other day since I got the printer and will follow your tips with paper etc when I get the hang of things. However I am confident the bed leveling isnt what your seeing there. If you saw it operating you may understand better. I would say all 4 corners of the bed are level +/- 0.2mm.

When you watch the nozzle, it seriously only extrudes in the areas you see have more material. this is not leveling in my opinion. Im new to the UM but am an engineer. Clearly the extruder apperatus is effected by the x/y coirdinates in physical space. After all the sleve that runs from the extrude motor changes shape with x/y. Anyone heard of this happening?

 

Share this post


Link to post
Share on other sites

I printed a 100mm circle and got this result

http://umforum.ultimaker.com/index.php?/gallery/image/850-p1010135/

Not exactly true, and more material in places, what causes this?

however its better at squares!

i just printed an a 190x190x10mm square as you suggested and it built well. So doffing of hat ...even though you could physically see in parts of the bed more material was extruding. Overall in practice the print turned out well. So it must just be a bed issue. Actually did occur to me althogh i would be shocked if it were the case, that the bed could be warped. were talking pretty high tolerances in that first few layers. In practice thats not a prob then.

I re printed the part in question at the start of the thread ( well the prob area of that part ) re printed with no changes other than i did have to retape and remove bed and altered the min cool time from 5 s to 7s as iluminarti suggested. the results are good. So I guess thats sorted! Thanks.

Just leaving my eccentric pulley to sort out. Can i correct this without Fing everything else up? I didnt build this so its a rubiks cube to me in what order I would tackle that. http://umforum.ultimaker.com/index.php?/gallery/image/849-imag1845/

Cheers for your help

Chris

 

Share this post


Link to post
Share on other sites

that the bed could be warped

 

That's actually common. Or the rods that set the print head height can be warped. The warpage is usually no more than about .1mm but if your first layer is .1mm thick...

Less expensive printers usually print a raft which hides the bed warpage. Cura defaults I beleive to a .3mm first layer and I usually go with that for most prints. I replaced my bed with a flatter bed but it's not really that big a deal.

 

Share this post


Link to post
Share on other sites

Just leaving my eccentric pulley to sort out. Can i correct this without Fing everything else up?

 

I would probably leave it alone as I've never heard of the pulley being out of line causing nonlinearity issues but if you want to take apart anything in the ultimaker start here (you can get here by going to the main ultimaker website and clicking assembly instructions on the left side):

http://wiki.ultimaker.com/Ultimaker_rev.4_assembly:_X-Y_axes

 

Share this post


Link to post
Share on other sites

I printed a 100mm circle and got this result

 

It looks okay - sorry - I think it's the picture - I can't quite see what the problem is. Or is this considered good? Maybe you need to show at an angle also?

Anyway, I'm sure we will hear more from you as there are still things you will have problems with and we will have solutions for you now that you are mostly past "bed levelling". Next is probably "damn it - my parts are warping and lifting off the bed". Or "my part gets knocked off the bed before printing is done". Or underextrusion issues. Or stringing issues. Or overhang issues.

 

Share this post


Link to post
Share on other sites

The circle does seem like something odd is happening at the bottom/lower left, in that the wall changes shape. This seems to be related to your issue with the circles being off square in the calibration piece. My first guess is that it's related to belt tension still. Did you tension the short belts as well as the long ones?

As I suggested in my PM, try printing a cylinder - 10cm outside diameter, 5cm inside diameter, with a solid base. Try a 0.2mm layer height, and print just the first layer (or make the STL 0.2mm high). Seeing the shape of the circle, and how the infill lines up with it will give us some insights into the belt tension and backlash issues.

 

Share this post


Link to post
Share on other sites

Hey Guys, I didnt realize you had made these replies, so not sure why im not being notified by the system.

Anyway. I appreciate the input. I have watched the nozzle like a hawk as it does its first layer and i defer to all of your experience with the technology. To my eye it still looks like under and over extrusion caused by the stress inside the bowden. But practically speaking it does prints well and therefore i must be wrong. Though I don't discount the bed being warped by perhaps 0.5mm. What im observing are areas where the nozzle almost squishes the last layer it put down in parts of the bed and then in other areas its like a smeer of material going down, to the extent the layer is translucent. This may all be a mute point however as I dont think my prints are being effected, but that splooge effect im talking about perhaps is what caused that slight bulge in the circle discussed above.

For the record only mods made are belt tightners we put on thingyverse http://www.thingiverse.com/thing:118917

The short belts feel tight. But as discussed at the start I note one axis has eccentricity, see new vid

This cant be good! The issue I have, is I dont want to take the Ultimaker appart , this was the reason we opted for pre-build and tested. I assume they could be poorly machined. As I didnt put the unit together, Im concerned about damaging other bits if I dis assemble. Looking at it now, pulling out an axel will leave other bits under stress from belt tension. Or is this ok ?

I assume the procedure is wooden end caps off > grubs on pully loose >then slide out of one end of the cabinet.

The only observable issue I have seen in printing our work, is like the bed is shifting in as it winds up the z. so the layers are not going directly ontop of each other, would i be right to call this horizontal banding? I ran two identical prints more or less back ( second on 0.1layer . First was 0.18 Layer ) to back and one had really bad banding well 1mm shift, which ruined the look of it. If I can dial this issue in, I think were good.

Fingers crossed I have not had warping or parts lifting from bed. Infact I have been finding parts stick like ***t to a blanket! Hard work getting them off :D

Be good to get some advice on the pulley and or the horizontal shifting issue /horiz banding.

 

Share this post


Link to post
Share on other sites

The eccentric pulley isn't great... but mine is as bad or worse, and I can't say I really see any major side effects in the print. You could just take it apart as you describe - there's not that much tension on the belts, but I'd probably loosen the motor first to make it a little easier to move that first pulley.

In the photos, it looks like the horizontal banding corresponds to the edges of parts of the letters as well? If so, it's probably all the same effect. This 'ghosting' is generally referred to as 'ringing'. It is the result of rapid direction changes in the head as it goes around corners. It's not entirely clear whether it is because the head assembly resonates, or because the bed does (or both).

Dampening the bed by adding weight to it may help, but first, try reducing your acceleration setting to 3000 (via the Ulticontroller, or using gcode), or even less. That should reduce the forces that are being applied, and so help to reduce the severity of the ringing effect.

 

Share this post


Link to post
Share on other sites

Thanks for the reply.

will try taking a look at the ringing. what's surprising is just how many ghosts or rings were produced. I suspected this was a harmonics thing, especially as i think i ran one of them at about 60 print speed. I could imagine the bowden being a spring could introduce a harmonic. But you think the harmonic is in the bed? I would have thought it was a harmonic in the print head under the rubberised belts. Not sure how I could damp that other than use reinforced belts and maybe look at the bowden. Could attach a weight to the top of the bowden for example. will first try your accel setting.

The ringing is probably less of a mystery than the random horizontal banding, what causes this? Has the bed shifted during the climb?

I spotted similar posts where the z screw has been mentioned. mine has a slight wobble and if i handle the bed the whole thing does have some play. I cant see a way to remove that play however.

Regarding the wobbly pulleys. you can hear from the whirring sound in the vid that you have inconsistent resistance in that axis. Eric De Bruijn posted this

on your troubleshoot page. Presumably because there its undesirable for accurate prints?

 

Share this post


Link to post
Share on other sites

Sometimes you will see faint bands in a print that seem to have a 3mm vertical pitch. Those seem to be related to the behavior of the z-screw, which also has a 3mm pitch to the thread. In the photos you posted, however, it was a little hard to tell because of the parts that were blanked out, but it looked like all of the most pronounced bands lined up with the edges of the letters. So I think that the ringing is to blame for that too, somehow.

As you say, the ringing effect may be due to either the head moving, or the bed; while the head is the most obvious choice, the bed also tends to move because the bead of plastic that is being extruded tends to pull on it as the head moves, in addition to any resonances that are transmitted through the frame. There's not really a consensus yet about what the most significant causes are, nor how to fix them, but it does seem that lowering acceleration is a good way to minimize the cornering forces that drive the process.

One thing to check is how tightly the large z-drive nut fits into the back of the platform. It is supposed to fit loosely so that it has room to move in x and y as the bed moves up and down. Does yours have room to move? If its tight, that might cause more x-y movement in your bed position as the height changes - including perhaps some sudden shifts when friction is overcome, and the bed suddenly shifts.

Another thing you might try is rotating the part 45º on the bed, and see if that changes/improves the ringing, since you'll no longer be bringing one axis to an unpowered dead stop on the corners.

Regarding the belt tension, the video you linked to is addressing a different issue. The belts need to be fairly tight in order to prevent 'backlash' - when small movements in the motors get absorbed by slop in the loose belts, rather than being translated into head movement. The main visible effect of this is that infill lines no longer touch the perimeter loops around an object (because the last little bit of movement towards the perimeter gets absorbed, and the head starts moving back in the other direction before it's finished moving as far as it should).

Having the pulleys be slightly eccentric doesn't necessarily lead to backlash - if the belts are tight they should still move correctly... but perhaps in a slightly non-linear way, such that the head doesn't move at a constant rate. (I imagine it you were printing a ruler, the tick marks might get slightly closer together in some positions on the bed, and further apart in others, depending on the position of the pulley).

It's not ideal, but you aren't the first person to see it. Some folks have actually ordered replacement pulleys that are made to tighter tolerances (me included, but I haven't fitted them yet). It's one of those things where, when everything else is really set up right, then having the highest quality pulleys is going to help you get the last bit of performance from the printer. But, in the ordinary course of things, other issues - like the ringing - are probably going to affect your print quality far more. Certainly you might want to discuss it with UM support, or consider buying different pulleys at some point in the future, but based on my experience, I don't think that it's going to cause you a huge problem by itself. As I mentioned before, I have a pulley that is that bad or worse, and I don't see any problems in my prints that I can definitely say are because of that pulley.

(Btw, just to be avoid any confusion, in spite of that 'Staff' tag below my name, I don't work for Ultimaker, receive any benefits in kind from them, nor even have any sort of inside info. I simply volunteer my time here advising new users, and sharing my own experiences, and (more importantly) the wealth of information i've absorbed from everyone else who has posted their experiences on here. )

 

Share this post


Link to post
Share on other sites

The "banding" is occurring right at the starts and ends of letters which makes me think it is a slicing issue. I would look at the gcode very carefully. If you aren't going to be able to tolerate minor bands like that then I don't think you are ever going to be happy with 3d printing.

However they might be fixable by changing slicing settings it's just that I don't know what's going on exactly. It could be underextrusion after having slowed down and then sped up (the extruder takes a bit to catch up). It could be infill too close to the edge and having a thicker skin setting would help (.8 instead of .4) it could be a bug in the slicer. It could be a solid infill layer. It could be you turned on "joris" which you definitely don't want to do on this model but has been known to cause similar bugs. If you email your gcode to illuminarti he's been known to look at those issues occasionally. It could be that your model is a bit messed up and somehow you have inner walls or an inner gap when you unioned the F with the cylinder. That might be visible in the "xray" mode and fixed with one of the "fix horrible" settings. It could really be a lot of things and I think all of them are slicer related. You could try kisslicer or cura 13.04 which had a very very different slicing engine.

Or just look at it yourself - go to the "layer" view mode and get to that layer at the top of the F and look at it very carefully. Illuminarti would actually check the extrusion amounts against the move distances.

 

Share this post


Link to post
Share on other sites

Thanks for the input. I have done all printing so far with the usb. so i dont save out the Gcode. This was printed with a 1mm wall and zero infill. Creating a hollow part.

The banding does actually occur in other places but it also occurs at letter ends. I cant see it in the layer view.

I already use Cura. which is great. I have no idea what Joris is, but i dont think its been switched on.

Re tolerating banding. I dont think Im being fussy. I just see prints, like those on Iluminartis page and or the Ultimaker promo pics and they look brilliant. I think its reasonable of me to want the my unit made and calibrated by Ultimaker for £1500 to be the best it can be with some tweaking. This is part of the issue isnt it, understanding what is normal and whats not. If it helps any the banding is worse on the run i did at 0.1 and better on the print at 0.15mm.

Looking at the part side on. The banding does look like that layer has shifted as for every negative band on one side I get a posetive band on the other. Max 0.8mm shift I would estimate.

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×

Important Information

Welcome to the Ultimaker Community of 3D printing experts. Visit the following links to read more about our Terms of Use or our Privacy Policy. Thank you!