phut

After learning of the salt method I have been experimenting with it and PETG. PETG adheres well to plain glass (reportedly) and sometimes even breaks it during cooling (reportedly). I had been using the glue method: whilst can be reliable has been a hassle and eventually messy and I wanted to find something cleaner and quicker. Whilst I can't say that the dozen or so prints so far constitute extensive testing, the salt seems to do enough to allow the print to cool uncouple without breaking the glass, but still adhere very well when warm. It's also much more reliable than the glue that with PETG is sensitive to the tack level right before the first layer is laid, as well as usually being an uneven surface that disturbs the first layer. A risk though is that one day a print might be started forgetting that the salt has not been re-laid, and may result in damage to the glass.

3.3 B Since I have an original without an LCD interface, the extruder preheat and G-code sender is very useful.

Seems like the preheat function's (with USB printing) auto disable timer stays active even after you initiate a print. So if you preheat the hot end it ends up turning off the heat x minutes into the print.

Regarding the comments on cyanoacrylate, there is certainly a bit of technique to it and variations in the formulations. Loctite has a number of formulations that are designed to work in surfaces of different finish level as well as different types of substrates. If you've done it right then the bond is always stronger than any printed PLA or ABS substrate will be. That said, you can get really far with your run of the mill $2 for 7 tubes CA with a little thought to the preparation and mechanical design. Par down surface imperfections with a blade, take advantage of tapering to gain closer surface fits, take a page from the wood or metalworking design and use finger jointing techniques to maximise the characteristics of the glue etc. Of course, this is not to say it is an end all/be all, I regularly have used 2 part epoxies, shoe goop, yellow glue and other types as well.

If you look at the history behind it, not at all. In the early days of 3d printing, the community normed on '3mm' as the nominal reference to the larger filament size beside '1.75mm', and that stuck as the nominal size to the detriment to practical application in bowden based feed mechanisms. Whereas 1.75mm had more decimal places from the start as well as printer manufacturers that moved rapidly in the 'early days' to exert global influence to make it a de-facto standard the way it worked best for them. Manufacturers were told to make 3mm filament. And so they tooled up to do so - without knowing the caveats in practical application that the 0.15mm difference makes. There does seem to be a slow collapse of 3mm nominal market as the standing opinion is that '1.75mm machines perform better'. Certainly as UM owners we know that the contribution of the feedstock diameter to overall print quality is SFA compared to other areas the UM holds advantages in over other printers in the market segment the UM lives in. I'm just not sure the '2.85mm' market will come to rise to replace it, since 1.75 is now such a well established standard. With feedstock making a substantial proportion of TCO (especially on a relatively reliable printer, like the ones UM make), UM Brand has its work cut out for it in the coming years to maintain supply chain for their filament diameter of choice. Meanwhile my modified UMO is run on historically available in my region 3mm nominal feedstock achieving perhaps 95% of the total quality one might be able to if I ran 2.85mm feedstock with optimised hardware.

I lubricate the X-Y axes with molybdenum disulfide. I use a type that's carried in a light oil. Marked improvement in axes freeness and also much longer lubrication intervals. I haven't found the Z needed any re-lubrication from the little package of whatever it was called provided by UM (this is on a UMO) in three years. But I guess if I were to re-do that I'd use MoS2 there. As for the recirculating ball bearings, they're supposed to be sealed for life but their seals gave up the ghost after a couple of years. Haven't added lube but not sure how long they'll last - I suspect that to be a throwaway job when the time comes as I don't really have confidence in their quality anyway.

I guess 1/5th the value of the original machine is a bit of a steep price to pay. One of the issues surrounding eccentricity is that for these types of couplings normally the fit needs to be a interference fit, which on the ultimaker (as I remember it during assembly anyway) it's not. This is a bit difficult to achieve when the shaft is shared with the bushing, but in any case you would need to have the shafts modified. Probably the most straightforward way would be to have the ends of the ground shaft linear knurled and press the pulleys on that way. Obviously your bronze bushing is never going to come out ever again if you choose this route though. Probably not an issue unless you want to upgrade the bushing. ED- mind you this also means your linear blocks are never going to come out again, too.

I agree with this, herringbone gears don't have any intrinsic design feature that make them better for backlash. In fact you inherit an issue on reversing rotation of the gear where the main reason they exist (eliminate axial force caused by a helical gear) stops existing. I have gone over the 'what if' of replacing the gears before, and the most sound conclusion was to replace them with moulded MOD1 gears (example) which are reasonably obtainable and have excellent friction, backlash, noise etc. etc. properties in comparison to printed or otherwise DIY gears.

I was just looking through the Marlin sources in an attempt to understand how Marlin calculates the feedforward rate. The thing is that it doesn't appear to have the code in the calculation at all on the github sources. Does using build-me-marlin use a set of sources that include the Kc calculation? or is it just a feel-good parameter that doesn't do anything?

Tensioning the belts would be my first port of call as such behaviour is attributed to loose belts. The other port of call is try another slicer program. The overall dimensions - all parts will have some shrinkage. You will find the printing order (perimiter->infill-> loops etc.) will affect the way the overall dimensions vary quite signficantly. The best way is to use a path plan that give you your desired surface and strength properties (or as close to) then bump the dimensional tolerances on your model to compensate for the final bit of variation.

It could be a flow (filament feed) problem. Can you see any horizontal irregularities inside the ruts? As others suggested it could be a Cura problem too, did you notice the print changed speeds markedly on or around those layers?

One thing that I noticed on the laser cut items of the UM is that there are places where a degree of interference fit is used on purpose. The degree of interference that should be used varies by materials employed, so I would be a bit wary of the situation. Also where fits like these are involved (note I haven't done a UM myself) usually there is a degree of dimensional tweaking involved to suit the performance of the particular laser cutter that you actually use, to ensure that the fit is the correct dimensions. So if you are entertaining acrylic or a similar plastic material then I would definitely suggest trying to get some tweaked laser cut drawings as a starting point. And be prepared to get out the file/knife/CAD to do mods to make it work. Also Acrylic usually is not considered an engineering plastic and isn't sold as such. There are many details such as water ingress, creep etc. which aren't completely characterised in the commodity acrylic that places will be using. Also acrylic gives horrible creaking noises when moved together in contact. Though usually Acrylic is bonded using a solvent cement and not left as a tabbed attachment (in commecial settings usually acrylic fitments aren't tabbed at all though) (Just for laughs go to a place that does metal laser cutting and get an Aluminium (5083 or something like that) cut quoted too - you might be surprised by the price...)

I 100% agree. Ultimaker seriously needs to consider working with distributors in other regions. I don't think it's so much the machine as such. Whilst it is a hefty amount to pay (in the case of Australia there are also around 15-20% import duties payable by individuals vs. the usual 10% sales tax) I think the machine at this time is still worth it. When I bought it there was a pretty long leadtime though which wasn't great - I'm sure people would appreciate it if they didn't have to wait as long too - but I haven't met anyone who was considering an Ultimaker who was hung up on that. There should be no problems finding willing and competent distributors. I have spoken to a local company which distributes certain other machines like the UM and they're pretty much keen on taking on the channel if UM was willing to open communications with them. One major thing though is to do with the upgrades - I'm quite cautious about selecting the budget shipping as it is both unregistered and uninsured, and upgrades are easily in the vicinity of 70 euros plus shipping, which is not a small amount to place bets on. At the same time the guaranteed option is a very steep at 46 euros (that's what, 50% of what I'm shipping? if it was just the hot end that's almost 90%!). If UM could start a spare part and upgrade distribution network that would already be a very, very good start.

You know if you're going to use a dial gauge to level your bed to only 0.1mm even then you may as well adjust till you just touch the head and turn the M3 screws in by 1/5th of a turn (or whatever your need your 'lift' to be) instead.

For reference the MOSFET is listed as an STP55NF05L so should be able to do what I want to do with it. ED- scratch that! there's a diode (D3) which is connected to VCC/2 which would cause my 48V supply to feed 48V into the whole Ultimaker circuitry if I did this! This definitely can't be done without addressing this issue at least.

I suppose I could open the .sch with Eagle which shouldn't be overly hard to get.

Could someone tell me what the low side MOSFET for the Ultimaker PCB is please? I'm pretty sure I have a 1.5.x board I'd like to run my heated bed from 48VDC but don't want to blow the UM board up. Also an image export or pdf of the schematics of the sheild would be very useful.

A lot of PSU's have an operating range of loading ratios between the rails (3.3, 5V etc) where it will safely operate. When you're using it like for a heated bed you may exceed those limits (because the other rails are completely unloaded) causing the rail voltages to fluctuate and the PSU to shut itself down. You may find that putting a couple of watts of resistor load onto the 3.3V and 5V rails will allow the PSU to do what you want to do.

I have done a 0.02mm Z print once ever (they claimed it to be possible so I just had to try it!) - without too much tweaking from a normal profile in Cura, the results turned out reasonably well. In real life you can still see a slight texturisation if you're close up but to the casual observer it just appears as though it's matte.

One thing that I think is missing are macro keys to different functions especially within the print menu. So for example the ability to move the nozzle around on the keyboard would be useful, as well as possibly keys that can be made to run a certain set of g-codes on demand. I think this style of indication would be good. The precise input should probably be an absolute angle to avoid any ambiguity. I also do agree that Cura doesn't make good use of the screen when it comes to the settings and information. For the print screen for example, there should be only one panel that tells us about the status of the print and allows access to during-print tweaks, then another panel that has everything else to do with manual setup. The console is also very squished and doesn't appear to auto-scroll, which makes its presence somewhat redundant.

I think the ability to enter in your own custom temp sensor tables would still be very useful in the marlin builder though. As for thermal sensor reliability, even if we discount the chance of the temperature sensor failing (or use a 'fail safe' sensor) there is still the issue that the Arduino isn't a safety device and could be liable to control failure (and the issue with the boot-loader not allowing the watchdog to fire in the case of failure etc.). If you're really concerned about overheating or whatnot you ought to employ a last line of defence, which would probably best be a thermal fuse.

The base V2 hot end kit at 56 EUR isn't too bad but it's very much the shipping that kills the value for money in my case. I'm frankly more bitter that I bought the spare hot end set when I bought my printer (And that was not long ago) and then the V2 hot end comes out right round the corner with no reprise for those who might have bought into spare V1 parts :/ From the looks of it where the materials are pushing up the costs will be the TFE and PEEK parts - the PEEK part must have cost a pretty penny to setup being moulded (?) and TFE in general is a bit more expensive to obtain and machine than most metals.

Use G92 in the startup code to change the virtual position of the head. This will cause the print to be offset by the same amount the virtual zero is offset from the real zero.

I like this list! I don't have the V2 extruder but it seems like it's well done so by the looks of things I'll be getting one in the future... Notes: slic3r seems to be able to produce a bit more consistent quality than Cura does, but at the same time it's quite more aggressive doing retractions. Bertho's feeder upgrade or similar is mandatory to use slic3r to full benefit (in fact it's just mandatory for any machine imo!) The cooling shroud, I use a similar high-flow fan shroud that is like the one that came with the original machine except printed. The only thing different I guess I would do to my current one is have the outlet centered around the nozzle instead because sometimes features on the far side aren't adequately cooled. Not essential though, usually you can change the model orientation if you know the problem exists. I'm actually using the same belt tensioners too, one thing I had to watch out for is that if the tensioner protrudes from the block it can interfere with the head clearance and also the endstop switches.

I find this is the most extreme when your extrusion volume is excessive, normally due to filament and drive variations. When your extrusion volume is correct then there's usually no scrape, but what does happen is that the nozzle wipes a bit of material into the print that causes a line in the print (particularly in the cases where you might be printing at higher temperatures). I find the most problematic issue this causes is that the head pressure builds up if the overextrusion continues, because it means that the extruder drive has to slip to compensate. The actual scrape is usually not a problem though, normally the sprung bed takes up any movement that's necessary to let the nozzle pass. You can try Slic3r too, that has the option to move the head up on travels. But beware that this is also followed by the head pressure drop and results in extra ooze (which in slic3r you can configure to have a retraction before this). Also in comparison to Cura, I find slic3r very aggressive on retracting, which has caused me many a problem with filament grinding previously when I ran the stock filament pinch thing (versus my current sprung pinch roller).