alaris2

awesome work Bertho. I found a new spring for the one I built and it's running beautifully. It's great to finally be rid of extruder problems, many thanks. of course you realize the next thing will be to get the extruder running faster so we can print at >100mm/s!

uhh.. silly question but what do you usefully print with something that small?

it's cheaper and easier to buy 1.7mm filament here - so I wondered, if I change the bowden tube to a smaller ID.. is there any other reason why I can't print with 1.7mm filament?

the quality improvement with kisslicer is enormous compared with Cura unfortunately. do you have the model available? it would be interesting to see others print and compare the same model. also do you know that's the best a dimension 1200 can do? or was that just 'coarse' mode?

well it's free in that anyone can download a working version - you don't *have* to have the extra features and not being open source isn't the end of the world for most people. but I take it you're not a fan. well can others offer to help you finish your slicer? you've probably got enough on your plate already..

start and end code is definitely better yes. what's the plans for integrating better slicers and moving away from SF Daid? an interesting comparision of slicers recently from flo suggests kisslicer blows all the others away right now in terms of quality.

i've been trying to run some 'scientific' tests of fan designs on thingiverse. the way I'm testing them is with a a hollow cylinder, 1" diameter approx. build at 100mm/s and keep everything the same but the layer time initially and change the fan shroud to the design to be tested each time. with the layer time at default (10s) most fan designs seem to do a perfectly good job on the cylinder. but now change the layer time down until the print quality visibly suffers. the numbers you get will probably vary from machine to machine, but the relative numbers should be very similar. i typically do the experiment twice on different days just to be sure ambient conditions didn't influence it much. most designs have a handedness about them - ie. they cool well from one direction but if the head moves in the other direction they have less effect. a rotated L shape shows this quite well. this is quite off topic - I need to start a new thread on this at some point when I've finished with the tests.

it is important, but the title now opens it for all wishes - is that what you intended? cos I have a huge list of them here.. Kyle > I thought your comments were very interesting and especially agree with the making scrap faster one. But I think it would be very interesting to see the quality of some of the prints from people who've modded their UM's. I am printing at 100mm/s with as many software tweaks as possible and a Bertho extruder mod but still notice occasional failings on perimeters, alignment of layer problems (also seen on Ian's prints) and top surface imperfections. before deciding that the solution is a tlalexander extruder I'd love to see what it can do - I've seen far too many 'upgrades' for UM that turn out to be downgrades or sidegrades and are only 'better' because that person designed them or spent so long building it they convinced themselves it was better. one of destroyer's early comments regarded 'proof' by way of numbers, measurements or photos - there's a lot of rumor in the 3D printing world but we all (me included) need to demonstrate improvements are real and not perceived so we can concentrate efforts on further improving the best ideas.

that doesn't mean what we're printing is worth showing

what's the print speed? and do you have a photo?

I see you're abiding by new EU law and informing users about the cookies - good work Daid

it goes soft at 40 degrees? which kinda limits what you can do with it..

has anyone played with this setting? the 3mm seems conservative but what range of values makes sense? I'm assuming 500mm/s is out of the question I tried 5 already but didn't see any real difference..

I've been watching the estimated time on Cura which was fairly accurate previously, but is now woefully wrong. I just had an estimated 1hr 30 print take 2hr 45 mins. I think i've narrowed it down to the dwindle setting - switching this on seems to be causing the problem. the amount of material is also puzzling me - why if I leave all settings the same but change just the print speed, can I make an object in weight ? the length of material remains the same, the time decreases (as expected).

don't knock the default fan too much. my tests of different fan designs show it to be one of the best designs weird tho it may seem. OK the material and construction is a bit shoddy, but the one under the machine is just as bad (and noone seems to complain about or redesign that one!!??)

so I've built this. some observations on its performance so far would be: 1) makes a huge difference to build quality. first layer is going down much better than ever before 2) but there seems to be a lot of variability unless I completely compress the spring. my spring is not as long as yours Bertho, but should be as stiff. once it's wound all the way in, the spring becomes irrelevant of course and the marks on the filament come back too. however extrusion is then awesomely good. (there are still less marks than before - no longer a shoelace effect, but still some patterning. if not tight enough, I see smooth filament with occasional vertical scratch marks on it instead) 3) when wound fully in I can print at 120mm/s with no degradation of perimeter as is the case with the default mechanism. what speeds have you tried Bertho? 4) I have retraction off for now during testing, but I see threads are no longer 'blobby' and spider-like as they are with the default mechanism. they're almost solid. this suggests 150mm/s is almost possible???! I really like the analysis and theory behind this upgrade Bertho - well done. if you don't mind some constructive criticism however, 1) the long screw is a pain to find. I replaced it with an electrical screw instead and used a soldering iron to melt it into the plastic. 2) the tolerances are really close. the bushing was only 0.05mm wider than the bearing when printed. i spent a long time filing other bits to make everything fit and operate smoothly. 3) the bearing was a bit of a pain too. shame it wasn't something easier to obtain.. 4) note to others - make sure you copy Bertho's photos when constructing. you need to put the screws facing the right way and the blocks too, otherwise it won't fit. 7/10 for a really awesome upgrade - by far the best of the 20 or so I've tried so far and very recommended to anyone else - but be prepared to put some effort and workmanship into building it.

nothing says we can't steer it ourselves a little in their absence. the extruder and heated bed are both areas we could (and should) be looking at. Bertho has an interesting development on the extruder front in a different thread - help him perfect this design further. and heated beds - I've seen several ideas here but none seem to have been overly successful yet, printing in ABS is an important step forward. personally I'm looking at different fan designs and lowering the minimum layer time whilst measuring print speed vs quality to try to get good prints faster. others have been looking at retraction lately which is another important area. and ultifay> yes I could, but I don't think it's appropriate to do so in this forum. however, I will say one of them is American, it costs the equivalent of £410 (UM is about £1500), the spec. is only very slightly less than UM in terms of build size and precision (6" cube) and it prints ABS by default with a printed bed that comes with it.

your analysis is correct (or at least agrees with mine) - but consider also what happens if the 2 x axes are not perfectly parallel, movement of the head may exert a slight sideways motion on one axis causing it to slide back and forth

Daid's points #2 and #3 are absolutely correct from personal experience. however before adjusting bed height, consider whether you care about those bumps and how much you care.. the reason being, at the height you currently have (you probably used the tin foil levelling method?) you'll get really good first layer adhesion, actually it can be difficult to get ti off the bed when the print finishes, but change the first layer height to 0.1 or 0.2mm to reduce the 'bumps'. if you lower the bed as I have now experimented with, you'll find getting the first layer to stick is way more difficult. in fact I've not successfully found a way to get the first layer to work unattended - it always requires me to fiddle with it a bit.

even if it's dripping in lube you may still have this problem as I've noticed it occuring at higher print speeds. only the left axis is affected it appears as you've also noticed. at 150mm/s it's very obvious and I suspect the best solution is to redesign the end caps to have a point which pokes inward, pressing against the rod to remove play. someone also suggested packing washers or a ball bearing, both of which would be fine if there's room.

very nice work Bertho - really impressed with both pages! am building your extruder right now actually..

whoa! you just like totally changed the past now man! groovy..

this looks great Bertho - look forward to seeing the files to make it. thedudevt > I think your second thingi link was wrong (it's the same as the first..)

I doubt you're the only one who decided to put their hard earned money on UM so they could mess around with it, I can add a 'me too' straight away'. took me a while to read all that post, but the information is valuable. more important, it includes real numbers, real experiments and real engineering. so I agree, let's start developing something! For those that read all the way to the end of that thread, the reprap boys make an important point (actually a few) that the extruder is the first weak point that has to be addressed - the head is fully capable of travelling at high speeds, but the material being pushed through has to match that or the print fails. out of interest I tried printing the thin wall box they used - and it comes out perfect on UM at 100mm/s and at 200mm/s and at 400mm/s and at 2990348734mm/s took me a moment to notice it always printed in 4 minutes. the minimum layer time was overriding the print speed setting so they were all coming out at <100mm/s. just goes to show how selecting a suitable test print is uber important. I know you started this thread on nozzles destroyer, but can I suggest we discuss extruders (on another thread if you want) since improving the extruder has an immediate impact and benefit without having to change any other piece of hardware which makes it easier for everyone to get involved and try. the nozzle is going to be a bit bespoke and not everyone has access to the tools necessary to pitch in and help. I could start such a thread by summarizing a discussion I had yesterday with some engineers about extruder methods. Having explained how the UM system worked they immediately said 'put the extruder on top of the head' so then I outlined the reason for NOT doing that. after which they thought it was quite smart but 'why don't you use a pinch and push system?' they asked apparently this involves 2 barrel grippers that tighten (electromagnetic) when current is applied. as they tighten they are slightly cone shaped inside (a bit like Owen's clamp?) and so push the material running through the middle of the barrel forward. think climbing up a rope - you tighten one while the other is released, then use the other to tighten and push as the first releases. I'm not sure if I explained that well enough, maybe it needs a picture, but it seemed a good idea.

but it melts into a nice blob if you leave it in your car for a while.. just saying..