jonnybischof

Nur noch eine kleine Anmerkung: Die Wave Vase muss (kann) nicht im spiralize Modus gedruckt werden, da sie eine fix programmierte Wandstärke hat. In dem Fall kannst du tatsächlich problemlos mit den Quick-Settings arbeiten. Die 38 Stunden kommen ganz einfach aus der dünnen Schichtdicke x maximale Höhe (ca. 3000 Schichten!) bei relativ langsamer Druckgeschwindigkeit (hohe Qualität...). Bei einem so hohen Objekt brauchst du sehr gute Haftung auf der Plattform. 20-30 "Brim" Linien dürften da helfen.

Die Vase muss vermutlich im "spiralize" modus, sowie ohne Top fill und mit 0% infill gedruckt werden. Ansonsten kriegst du einen Baseballschläger statt einer Vase Vermutlich brauchst du dafür die Expert settings - da ich immer damit arbeite weiss ich das jetzt gar nicht so genau. Sieh dir in der "Layer view" Ansicht die Vorschau an - sieht es nach Vase aus, oder wird sie ausgefüllt? Typischerweise werden Vasen mit eher grosser Schichtdicke gedruckt. Optimalerweise auch mit einer grösseren Düse (0.8 oder 0.6). Die Wandstärke musst du = Düsendurchmesser setzen (0.4 beim UM2). Mit 0.4mm Düse kommt das meiner Meinung nach nicht brauchbar raus - das Objekt sieht zwar toll aus, ist aber extrem fragil. Da ist es besser, mit 2-3 Shells (0.8 oder 1.2mm Wandstärke) und ohne "Spiralize" zu arbeiten. /edit: Nützliche Informationen zur Optik beim Drucken von Vasen von Taulman (Hat zwar nicht direkt was mit Vasen zu tun - aber mit Optimierungen bei transparentem Material).

I preferred the good old "Your avatar needs to be 128x128 px. Period". But many people nowadays wouldn't even know how to scale a picture to 128x128px. Usually, it's fine as long as your avatar is square. It'll just be scaled then. But that's not always the case...

They are all the same platform, and they all run Marlin. Differences are in the details, such as the power stages, analog input quality, and features (how many temperature sensors, how many motor driver sockets...). There are more advanced platforms (32bit Cortex M3 based, usually) which don't run Marlin but have their own custom firmware (such as the smoothieboard). And then there are the raspberry Pi and similar platforms. I have no experience with those, though...

Not a good idea...

Do you know what would be the highest selectable value for that command? If you need to wait for several hours, you could probably just queue up multiple G4 commands... Using pronterface you can easily test the command's behavior.

The only way to make a DRM system work is to actually build it into the printer's hardware. You'd need to replace the Arduino platform with a platform that connects to your cloud system, downloads an encrypted package containing the sliced model. Then decrypt it on the printer and print it. That platform would need to be very well built in order to not be hackable. Then, you just need to give people a reason to buy that platform. Maybe as Apple how they do that I'm not suggesting to actually do that - my point is that it's pointless to do that and if there was a printer with such a platform, I wouldn't buy it and tell all my friends not to buy it. /edit: I know it would be useful to have such a platform for professional designers to sell their superior products for 3D printing. But I'd rather have them come closer to the open source ideal than the open source 3D printing world to come closer to DRM.

Works for me now. (Firefox, Win7 64)

Why not just use a platform that is built for 24V like the Megatronics v3 or the RUMBA? You shouldn't run capacitors that are rated for 25V at a 24V DC supply. As a rule of thumb, capacitors should always have a 30% higher voltage rating than the supply voltage. The headroom is necessary for voltage spikes and irregularities. If you run an electrolytic capacitor too closely to it's maximum rating, it will deteriorate much quicker than it should. Ceramic capacitors won't go defective, but they lose their effective capacitance - so running a ceramic cap closely to it's maximum voltage rating renders it pretty much useless.

I read this on a datasheet for some electronic component. I liked it

I printed at 0.2mm layer height. Configurable would be perfect

I'm not sure if it's a bug or a feature, but I'm afraid this is giving me worse print results than before... The newest Cura messed up the printing order: Infill is printed first, shells second, and finally it does the "100%" infill parts (top, bottom, odd angle walls). My latest print shows lots of blobs on the vertical surfaces, and the parts where the infill connects to the shell sticks out. So I get vertical lines in my walls. I guess the reason is that if the infill is printed first (with 15% overlap), there already is some material when the printer does the shells. This leads to the material escaping outwards a little bit. This is not the case when the shells are printed first. At least I never noticed it before. I'm not sure if it always prints the infill first, or if it depends on the model. I can provide the model that gives me trouble on request (by mail only - it's property of my company).

"Metric will govern". Making anything new in imperial units is just wrong.

Diamond Age Plastics has a "tactical matt black" PLA. No carbon fiber or conductive properties afaik - but it's easier to print with because it's regular PLA... Haven't tested it myself, though. I wanted to try Colorfabbs XT CF20, but you should get a stainless steel nozzle if you want to print abrasive (CF) materials. E3D now has stainless steel nozzles (VERY expensive, but at least they should have a very long lifespan). I believe they fit an UMO hotend, but I'm actually not sure about that.

If you pay for (therefore own) the printer, you can charge your classmates for printing their models. With reasonable pricing - no shipping necessary, and making them deliver ready-to-print files to you - you can get back enough money over 5 years for sure. For example, my rate is 3 $ per printing hour - including the material. Friend price would be 2 $. That's reasonably cheap, yet keeps your friends from spamming you with print requests. I calculated that rate to cover material costs and get me back the printer's cost within it's lifetime. More or less. You can of course make your own, more elaborate calculations

Ich verwende gewöhnliche Plastikboxen vom Baumarkt mit Deckel. Die sind "staubdicht" (nicht IP staubdicht, aber da kommt effektiv kein Hausstaub rein). Habe eine Grösse gefunden, in welche so 5-7 Spulen rein passen. Bei PLA habe ich nie ein Problem mit Feuchtigkeit feststellen können, deshalb lasse ich den Spass mit den Silica-Säckchen komplett weg. Ich habe eher das Gefühl, dass "übermässig getrocknetes" Filament eher dazu neigt, zu zerbrechen. Mein Diamond Age und auch Faberdashery Filament ist immer noch in Top-Zustand, obwohl es nicht trocken gelegt ist.

Turbulences can cause the airflow to cease (at least partly) or become unstable. But without doing a simulation this is very difficult to quantify or prove. The crossflow solution works well mostly because it generates air circulation inside of the printer, with minimal "heat leakage" through the open top (!). If you interrupt that, you could get worse results even though there's more fans (especially if you only have a 100W heatbed and not a 200W silicone heater). More fans doesn't mean better cooling - it's a good concept that means good cooling. I'd rather try adding a shroud to the hotend that directs some of the airflow to the dead spots behind the nozzle (viewed from the crossflow fan). No need to add a fan, just some "blade" to redirect the air.

Makerbot should patent the deed of filing patents. They have patents for everything anyways...

I don't think using two crossflow fans is a very good idea.. The idea of the crossflow fan is to get a laminar airflow across the build area, then downwards at the wall, back across the base and up towards the crossflow fan. If you put in another fan, you could break this circle and generate a non-laminar ( = turbulent) airflow that might do more harm than good. Especially if you need to contain heated air within the printer. Just a thought - maybe it works just fine, I never tried it... Btw. my crossflow setup is not tested yet (I'm getting used to that sentence...). Still working out the final details of my custom build.

It's a replacement for the UMO shield, so yes it does contain all the other features (except they're better ). I've used good MosFETs and switching regulators instead of linear regs. And maybe a bit cleaner analog layout, but not that much. I bumped into some details this afternoon, so it's not yet completed. But maybe tomorrow evening I can show you the design. /update: Still work in progress - I'm having trouble with a certain TI DC/DC switching regulator that disappeared from their WeBench design software. Now I either have to wait until they fix that, or switch to another regulator (which would suck because 24V to 12V is already done).

Ready to join the fray with my own TMC2100 implementation! According to Trinamic sales, the ICs should be available in March. Why not Pololu compatible? Because the Pololu footprint is simply too small. I want a clean layout that can actually make use of the TMC2100's nice properties. Also, I want access to all config pins. My board should also work well without any additional cooling or heatsinks (that's what the extensions on the sides of the PCB are for). Current return paths are much better here than on the silent step stick version (simply because there's more space...), and I also bypassed the IC's 5V regulator using an external 5V source. This saves quite a bit of power loss in the IC, especially when driving the motors at 24V. I also implemented the minimal ESD protection measures for the driver (just 6 more capacitors) which should improve lifespan and ruggedness of the driver. It comes with an Arduino shield that accepts these drivers and has jumper fields for all the config options. I have to make some final changes to it (had to switch a few pins on the driver), so there's no preview just yet :(. /edit: By the way, the heatsinks on the sides are an experiment - I'll cut them off on one prototype and compare temperatures. We'll see if they're good for something... /edit2: Why not make an all-in-one platform? Because I already have a few Arduinos laying around, and this is much cheaper than making one large 4-layer PCB (the shield is just 2 layers). /edit3: Provided these work as intended, I'll be selling them within Switzerland. Sources will be available once it's tested. No use publishing an untested (and therefore not yet working) design...

That would definitely be the worst thing they could possibly do If people would just not buy cheap-ass filament and instead use high quality brand filament, there wouldn't be that much problems about material profiles and all that. I've been printing several kg of different PLA colors from Diamond Age Plastics. There was never any issue at all. The stuff prints perfectly, just the way it's supposed to be. The very moment a manufacturer limits filament choice by making proprietary cartridges, you will have to live with what they feed you. Either it will be cheap standard material sold for a lot of money, or it will be good material sold for even more money. And what if at some point they change their policy and get sloppier with the material? You'll be helplessly dependent on them... Seriously, just buy good filament. I'm paying less than 60 CHF per kg for the best PLA there is, (thanks to Dim3nsioneer's efforts ) and I know that I will never have a problem with it and get great print results throughout. Hint to Ultimaker: Don't get me wrong, but if I were you I'd stop selling the filament that you have now. Either don't sell filament, or only sell the good stuff. You could team up with Colorfabb who have a good name (not a fan of their PLA personally, but lots of people are), and you're both based in the netherlands which would facilitate logistics... I don't know if you're actually making a lot of income with the filament or not, but I'd consider that as a necessary way to further improve your recognition of a "manufaturer of high quality 3D printers". Because that's what you are! I've tried several rolls of UM filament (both ABS and PLA) and they were all low-grade. That's a pity because people will have a bad first impression of their new printer. About building a 3k $ high-quality printer - I've actually been working on that for a long time now. It will probably be more like 5k$, but that's still much less than even entry-level "pro" printers. Progress is slow, because I have a daytime job to do, but I believe I'll be getting there pretty soon. Right now I'm finishing a new interim electronics platform using Trinamic's new TMC2100 stepper drivers in combination with the standard Arduino & Marlin platform. The difference is promising (see here) and it's actually a very "cheap" upgrade. Cheap meaning it doesn't take much time to develop because it works almost the same as the previous drivers. The parts themselves are a bit more expensive, but still playing in the same league. My next step will be a Cortex M3 (eyeballing the NXP LPC1789 at the moment) paired with an FPGA motion controller and TMC260 stepper drivers. But that's a completely new design, meaning it will take a long time to even become a beta-testing prototype... Another thing about materials: I won't use ABS anymore. It's just not a good plastic. It's toxic, warpy, gives weak prints compared to PLA (as long as we're talking room temperature) or PET and it's a real pain to produce nice prints with. And damn it, IT'S TOXIC!! People are putting their open-frame printers on their desks and let them print for hours and hours, inhaling the fumes and nano-particles this sh%%t produces without thinking anything by it. A few years from now, this could turn into a serious shitstorm when people start getting sick... (or not, but I'm not taking my chances...) Whenever my parts don't need to be temperature-resistant, I use my trusty Diamond Age PLA. For higher temperature stuff, I use Colorfabb XT which is a bit more challenging to print with (although most problems with it are solved) but gives extremely durable results. I can send you a print sample if you want to compare it to ABS . Besides that, XT also doesn't cost more than 50 CHF per kg. (34 EUR for 700g iirc). It's limited in color choices (transparent is stronger than the new colored versions as I've heard), but still the best material for technical prints that I know of. Both PLA and PET have very little warping & shrinkage, so that there is usually no need to take these factors into account when designing models. That is another huge advantage, especially if you're not an industrial designer and don't have access to professional CAD software that might help you with these questions. /edit Thanks to the free choice of printing materials in RepRap 3D printers, there is a lot of development going on. Carbon-fiber and even metal enhanced materials are sprouting, as are things like PA6 and POM. Who knows what more is to come in the next few years. This takes place while the big companies would rather like to sit by idly and let the customers buy their overpriced filament cartridges because they don't have any other choice. /edit2: Forgot to add the link above...

Thanks for the detailed report MMartin! Very interesting insight for me - I've never seen a "pro" 3D printer before... I know a thing or two about state-of-the-art industrial machines (made my apprenticeship in electronics engineering at Besi Switzerland, formerly Esec), but not 3D printers in particular. Being an electronics guy, I'd mostly pick up on the differences here: The Ultimaker (including most RepRap printers) uses a very low-end electronics platform. Sure, it works (and not that bad either!). But it's really a primitive thing compared to professional electronics. An update here would be highly valuable, but takes considerable effort in firmware programming. The beauty of it is - once that development is done, the new platform won't cost much more than the old one. Imho this is just a matter of time. Shortly there will be further improvements in this field. Maybe already on the UM3? It doesn't take a mammoth-setup using a host-computer, a slave real-time computer, and high-end motor drivers for every driver. It just takes a bit more CPU power, better motion control planning & execution, and a clean stepper driver setup with good drivers. The stepper motors used in RepRap printers are the ideal choice imho - they're cheap, and they do their job very well when they are driven by a good driver. Closed-loop is not necessary because there is no demand for higher motor velocity than what's possible with simple stepping. I don't see any benefits from 5-phase steppers. There is no demand for more accuracy, and none for more power. Another necessary improvement is to get rid of NTC "thermistor" temperature sensors. PT1000 sensors are much more reliable and accurate, given you have a good measurement circuit. EMI, ESD immunity and machine grounding are also very important considerations that the RepRap printers fail to address. Afaik, the UM2 made a lot of progress here, but it's not quite there yet. Everything I've written about can be improved if enough effort and knowledge is put into the project. In the end, there won't be a significant increase in costs (there will be somewhat of an increase, of course). Just use different, more suitable parts. Then there's the mechanical aspect - high-precision shafts (not regarding diameter tolerance, but straightness which is usually not specified, or not kept up to spec!), good pulleys, couplings, gearing and so on. This is pretty simple - you get what you pay for. And finally, there are all the little design details that make the difference - and make the reliability of the machine. Good hotend design, good material feeders and so on. These things are continually improving thanks to the open-source development scheme. At some point, the RepRap printers will surpass professional grade printers due to one simple fact: They are open-source, transparent (value to cost), understandable, modifiable, user-serviceable, and much more affordable. Just take the printing materials - 500 CHF FOR A SINGLE ABS CARTRIDGE??? That is 10 times more than high-quality material for reprap machines. 10 times is a lot! Even if you can afford to buy one of these professional grade printers - you'll always have to calculate if it's worth printing a part or not, because the part will have a significant cost (not even mentioning that you need to replace build plates often). The real beauty of RepRap printers is that you can just print something and wonder if what you designed works or not when you hold it in your hands. Then make 10 more prototypes until it works. Who cares? The process cost you only a few bucks...

Shouldn't you actually lose a few millimeters? If you have two nozzles, you still need to be able to reach all places with every nozzle. What good is it if you have areas that are only reachable by one nozzle? You will lose "distance between nozzles" because the one on the right will not be able to reach as far left as the one on the left. And you can't say you gain a few millimeters on the right, because the nozzle on the left still can't reach that area..

http://colorfabb.com/bronzefill#tips-tricks-woodfill a lower heated bed temperature. You should try that. If your bed is too hot, the print will not stick well and will probably deform and melt while printing. /Didier beat me :( By the way: I noticed this with Colorfabb XT (which is a very different material, so it might not apply here!). I usually print at 30 - 35 mm/s (very slow). But sometimes it's actually better to print a little faster. Colorfabb recommends 50mm/s for Bronzefill. It may be that printing slower than that actually gives worse results (if the hotend moves too slowly, it can melt the printed plastic and make it expand, which is what happens with XT).