jonnybischof

Why would you want to use a thyristor? That's an element for AC-switching and dimming. I've never heard of any uses for a thyristor in DC applications... (Then again, I don't know everything, of course , especially with AC elements my education is very basic) The problem with a mechanical relay is, you don't want it to switch quickly because relays don't like fast switching. The best alternative would be a simple MosFET - no problems with fast switching (except maybe that you could generate some eletrical interference). Here's how I did it. It's very simple, I didn't even use actual connectors but just soldered the wires directly to the pins of the corresponding parts. It starts with a LED to show me when the thing is heating. Not necessary, but informative. Then I divide the voltage by 2 (voltage divider with equal resistor values) so that I don't have the full 19V on my MosFET Gate. Again, this is not really necessary, but 19V is a bit high (my MosFET has 20V max VGS rating). Also, these resistors provide some very basic ESD protection (not really, but they're better than nothing, and a MosFET Gate is pretty sensitive). Then it's the simple "MosFET as a switch" hookup. Make sure that the Mosfet sits in the GND path, and not in the V+ path! As you can see, the MosFET will connect the two GND lines together, not the two V+ lines!

Sketchup isn't very helpful with this problem... I'll just print that part with PLA, didn't have any problems last time. Great idea! Will try that instead of closing the top. Much simpler to just put a cork on the build plate

My print from yesterday was finished without problems. ABS prints quite well on a PEI build plate. Sticks well when heated, and pops off when cooled down... Well-invested money But it's difficult to get these PEI plates really flat. They need to be fixed on the four corners, but also on the sides. I noticed my part has shrunk significantly! With the heated chamber, it printed perfectly, no odd angle walls because of uneven shrinkage during printing. But - it must have shrunk while cooling down... I can't use the part because it doesn't fit the things it was designed to hold.. -.- (And I gave some clearance..) About bubble wrap: While it doesn't look too good, it certainly does a good job keeping the heat inside. So it's probably one of the best possible solutions (except maybe double or triple glazing... ) to close the machine. Closing off the top will probably decrease the heat-up time for the bed. I'm recording times atm for non-enclosed heat-up and will compare against the time it takes when fully enclosed. But once everything is on temperature, I think there isn't so much heat lost when only the top part is left open. - If it gets to temperature at all. On the UM1, if you don't enclose any of the sides, it takes ages to get it to 100°C, if it gets there at all. (Depends on the heatbed construction of course...)

I'm printing ABS right now. Part is getting rock solid... Have to say though, I'm using the pretty expensive "Smart-ABS" by Orbi-Tech. I only closed the sides (UM1) with bubble wrap (pretty good heat blocker), the top is open. But closing the top can help with temperature problems. Make sure there are no noteworthy holes in the sides or the bottom (especially in my case with the open top), because that would generate a nice ventilation... If the bottom and all the sides are tightly closed, then an open top is not much of a problem. You do need more energy because the heatbed will need to heat more, but printing-wise it shouldn't matter much.

...and burn their house down? I will probably make a webcam stream (don't care if strangers watch my printer work...), and implement an emergency stop function which is activated by SMS. There's a nice GSM-addon for the Pi which should be able to do that. But that project is still a few weeks or months away...

You have to set that album to "public", or we can't access it

Your settings seem quite ok... What material are you using exactly? (brand) I've had ABS which gave me the strongest printed parts I've ever seen, and I've had ABS which didn't hold together very well at all.. Could it be a humidity issue? I heard ABS is more susceptible than PLA. I've never kept my PLA particularly "dry". Usually it hangs on the UM for weeks... Maybe ABS doesn't like that?

Slicing with Cura only takes a few seconds on a non-stone-age computer. I think cloud slicing is going a little too far... Besides, gcode files can get pretty large, so the download will take some time, too. And you don't want to use a mobile phone for that. Maybe it's useful when you want to actually slice stuff on the Pi. But why would you want that? I doubt Cura will run smoothly on a Pi when you rotate and inspect a part. Or maybe it doesn't run at all (total Linux noob here...).

You definetly do have to solder in the 4.7k resistor on an Ultimaker 1. Maybe there are other electronics that already have the resistor in place, but the UM1 doesn't. It will not work without the resistor, because the resistor connects the power supply to the thermistor (making a voltage divider). Without that resistor, the electronics will always read zero.

I'm going for total rigidity ^^. The nut is simply clamped to the top part with 4 screws and a little plate that you can see in the last picture. The screws are positioned so that the nut fits exactly in between them, and the screws will need to be tightened pretty hard in order to hold the z-nut immovable, but I will see how well this works... On my first UM1 kit there is definetly something odd about the leadscrew. The further down I move the z-stage, the further the leadscrew leans to the front, eventually touching the slider blocks. I hope this will go away with my construction which holds the leadscrew exactly in parallel to the shafts. The 12mm shafts should be absolutely immovable - so keeping the leadscrew in line with them should keep it straight. There might be some tension building up, I hope that's not going to be a problem. Using a flexible coupling on the z-motor would probably be good, but I don't know if there's a kind of flexible coupling which doesn't allow play in axial direction (the usual flex couplings don't work, I know that). There are several different kinds of flexible couplings - and I don't know most of them. They are very expensive (30 to 50 EUR per piece) if you're buying from Misumi, so I don't want to just buy some of every kind and test them, as long as I don't know that it is necessary. I'll could try asking Misumi for free samples, or information on which coupling would be ideal for that application. Or maybe someone here in the forums knows which would be the ideal coupling? By the way, yesterday I found that the small linear bearings for the printhead are just as bad as the big ones. There is a lot of room for improvement here...

Ok, version 1 of my z-stage is ready for prototyping. I will make a prototype with 2mm acrylic. Won't be usable, but should look pretty cool There is an error I just noticed after putting all the stuff together for Pic 2: The front is a bit too far recessed. I will move it further to the front a bit (making the "nose" where the front center hole for the platform sits) shorter. I experimented with multiple hole positions. The hotend usually doesn't sit in the center between the x-y axes, but is somewhat offset. On the Original hotend, the nozzle is offset about 10mm to the right and 10mm to the front. Anyways, I figured it's not necessary to put 9 holes there so you can correct offsets in all possible directions, but instead I went with 2 holes (center + 1 correction). Now, you just have to turn your hotend the right way so that the offset fits. I made the arm as triangular as reasonably possible. As discussed, making it triangular downwards doesn't work. On the assembled picture, I added some quick dummies for the bearings (top only). As you can see, they protrude about 4mm, and that is their short side. They can be mounted upside down so that they protrude 16mm. That works for the top, as well as the bottom bearings. This is why I lowered my "cage" height again. If you go with the bearings upside down, then the arm can be made "thicker" by up to 16mm. This will make it stronger if necessary (if no steel is used). So, on wednesday I'll see if I can lasercut the parts @ the Zurich FabLab. If they fit together and the construction seems sturdy enough, I'll have them lasercut in steel within the next weeks. /edit: Just in case you wonder: My offset correction goes to the front and left, instead of the right. That's because I'll be using Nick Foley's E3D integrated printhead which goes to the front and left. The UM original printhead can easily be assembled in the same way. /edit2: Just had a minor epitaph on the toilet! (should spend more time there...) The bottom part is completely left-right symmetric. So I will just optimize the design for my "front left" offset and leave the centered hole away (this one is useless anyways). By just flipping the part, you can convert it to "front right" offset

Looks much better already! *thumbs up* Always glad to help

Looks interesting, that board.. I like the "Powerful Atmega2560 processor with 256 kB memory, running at 16Mhz" statement. LOL. My watch has a better CPU o.O (actually it doesn't) I see this board has an integrated heated bed driver (1.7mOhm MosFET, even from IR, my preferred manufacturer *like*). Very nice Seems like one more thing that hits this month's bill -> No power switch is actually GOOD! DC side power switches suck alltogether Can you still connect the Ulticontroller to this board?

Usually, silicon insulated wire is high-flex stranded wire. Otherwise, the wire wouldn't be so flexible... You can easily see that when there are lots of small diameter strands instead of just a few, thicker strands. Silicon insulated wire is in general well suited because even the low-quality chinese stuff is still pretty good. Never buy PVC insulated wire of non-industrial quality - complete waste of money. High flex indeed doesn't imply high lifetime under moving conditions. I've tested high-flex fire-wire cables up to 50 million cycles and even these would sometimes fail. But it's the best thing you can do. You never get 100% maintenance free stuff... I bought a large stock of this wire: http://www.distrelec.ch/Web/Downloads/at/_d/wxHF-3KV250_dat_d.pdf?mime=application%2Fpdf Corresponds to AWG 14 (or 13), built up of 672 strands at 0.07mm2. Suitable for temperatures up to 180°C, and has a max. resistance of 8 mOhms per meter. Even if you take 10 meters of this cable, it has no noticeable resistance (meaning energy loss). The heated beds are usually in the 1 - 10 Ohms range.

Imho there is a simple case that favors 3mm filament over 1.75mm: Assumption: Both 3mm (which is actually 2.85mm) and 1.75mm filaments are manufactured with +-0.05mm diameter tolerance. 1.75mm doesn't go lower than 0.05mm (I believe that is pretty much impossible anyways). Conclusion: 2.85mm +-0.05mm = 1.8% diameter error tolerance 1.75mm +-0.05mm = 2.9% diameter error tolerance Note that there is also +-0.1mm tolerance filament. This would make things even worse for 1.75mm filament (5.7% diameter error tolerance vs 3.5% for 3mm). So to me it's clear that 3mm filament should be favored over 1.75mm, because variations in filament diameter result in uneven extrusion which gets you bad prints.

You can use a much smaller diameter wire for the thermistor. Anything up to AWG26 (0.14mm2) will do just fine. I'd ONLY use high-flex cabling as this cable is constantly moved around, and using a shielded cable wouldn't hurt, too. The shield would just be connected to the GND wire (thermistor cables are GND and Signal). -> You could use a single wire, shielded cable and use the shield as the GND wire.

These connectors are made to be soldered. You can just take it and solder it back on the PCB. The plastic won't melt as long as you don't heat it up excessively. But especially if you use 12V, it would make a noticeable difference to just solder the wire directly to the PCB and leave the connector away. 12V, 120W means a lot of current, and it's better to have as little components on that line as possible.

Sounds like an ESD. But I've never heard of exploding thermistors either... That would require a lot of energy! If your thermistor exploded because it was overloaded with electrical energy, then this energy should theoretically also have gone through the wires and into the sensor circuit. The UM1 board does not have any ESD protection whatsoever on it's I/Os, so there could be some damage there... But you'll notice that if it doesn't work anymore. As long as it keeps going, just look the other way.. By the way, I touch the nozzle with tweezers all the time. Non-insulated tweezers. I'm usually wearing ESD-sandals, but we don't have an ESD floor... Note that insulated tweezers are still able to be statically charged!

This teflon part is the reason why I'm trying to switch to another hotend design... I've wrecked two of them up to this point. Disassembling the hotend when you have a clogged hotend (and mine kept clogging at the teflon part) always lead to damage on the teflon part for me. It's a major design weakpoint imho. Sadly I don't have any spare teflon parts either - trashed all of them except the one I'm printing with at the moment..

Are you sure you wanna do this to yourself? o.O Ok now, before the Chocolate Wars start, let's all settle down and eat some chocolate.. In peace... Swiss chocolate of course!

Erm *cough* That's exactly what I was talking about, of course!

Dang.. Got a problem with the triangular shape: If I want to raise the lever attack point, it means that I have to raise the bottom plate with the arms. But that also raises the bearings -> no effect. If I really want to do that, I need to separate the rear part of the bottom plate from the front part, meaning pretty much a complete reconstruction. But.. A good part of the stiffness of this construction comes from the fact that the same piece that holds the bearings also holds the platform. I don't want to lose that benefit.. So, I guess that I'll make my prototype with a more or less L-shaped arm. I can "fake" the T shape by mounting the bearings upside down and can then decide whether that is necessary or not. If it is, then I'll redesign the thing again and make a solid, triangular construction with a solid rear frame and a well attached arm. But the simplicity of the current design just calls to be tried out

Ok, suppose you're right.. Hmm.. Does it make a difference whether if I make a simple filled geometry, or an actual space frame (with the same contour)? Of course the space frame would be lighter, but it would also be weaker than filled geometry, right? I was thinking of using rounded shapes (similar to the state you've seen before), but maybe a bit smarter instead of just quarter circle segments... That was just the easiest thing to draw in Sketchup Fair enough It might even be strong enough with simple plywood, maybe just use two pieces per arm... I will try to construct the parts with Creo Parametric, so that I can easily make variants for different material thicknesses. I still hate Creo because it's impossible to handle, but Sketchup is just not suitable for precision constructions.. I'd very much prefer a design that you can just put together, tighten the screws, and it's fine. I hope that's not too much to ask from a total noob designer, but I guess I'll find that out soon Separating horizontal and vertical tongues seems to be the way to go if the parts are milled by CNC. The problem is, I have very little space on the top parts. And if I use only one vertical and one horizontal tongue, then the parts can be twisted. I think I'll try to get it done with square, lasercut holes. Maybe I'll have to slightly file away the corners on the tongues, but the laser's radius should be very small so that it only takes a little bit. Maybe the radii can be eliminated in the lasering process by drawing the lines a little bit too long, but I should leave that to the laser-guy.

The Swiss disagree! :shock: I demand blood for this unspeakable insolence!! GET ME SOME STONES :mad:

I think I understand what you mean, and I will definetly try to improve the arm geometry based on your instructions. Thanks! If I understand this correctly, then there are actually two different problems: 1. The arm itself must not bend 2. The z-stage as a whole must not be allowed to bend down, for example because of play in how parts are joined together, or play on the linear bearings. Now, if I change the arm's geometry from a "hard T" with 90° corners into more of a triangle, then I will make the arm much stronger and prevent it from bending. But, let's say the chrome steel arm is by far strong enough, even with a T shape, then I don't actually gain anything from this. Right? I don't know if it is "by far strong enough". I suppose it is, because after all it's just going to be a few kg which shouldn't be a problem for massive steel parts. But nonetheless - I will of course optimize geometry as far as possible. But this will not solve problem #2. So, I think on one hand it is necessary to have high quality bearings and shafts with little to no play. There are different tolerance classes, and the bearing must match the shaft in order for this to work. I don't know the specifications of the UM parts. To be honest, I couldn't make out any differences in play between the UM shafts / bearings and their Misumi counterparts. The only thing I noticed was, that the Misumi bearings slide much smoother on the Misumi shafts than the original UM stuff. Maybe, just maybe (it's almost impossible to feel this by hand), the Misumi combo has a tiny little bit more play. That's probably because the shafts are g6 tolerance which is always slightly smaller than nominal. The UM standard shafts might be h6 which can also be slightly above nominal. By the way, the UM bearings are still worse than the Misumi if both are tested on the Misumi shaft. So there definetly is a difference in quality. I also have to note that I bought the "Eco" Misumi bearings which are cheaper, and not made in Japan. Maybe, the really good stuff from Misumi is even better (should be...). And then, there's the matter of my construction. The different parts must be fit absolutely tight so that there is NO movement between the arms and the bearings. That's difficult, because the bearings are mounted to a part (possible play) which is mounted to the arms (more possible play). This is where things get difficult for me, because I lack the experience and education of a mechanical engineer. How do I reduce play between two steel parts mounted together? -> clamping them together with two screws makes it stronger than if I just use one screw, right? So there's a weak spot on the top plate (see picture below) which has two "hooks" that go into the arm, but are only mounted with one screw. -> The hooks will hardly fit 100% snug into their counter holes. But I have to make it as tight as possible. If I mill these parts, then I may have a big problem because I can't get the short sides of the "rectangular" holes straigt. They will instead have a 1mm radius. So, lasercutting these parts should give me a better result than CNC milling them. I hope the costs won't be too high on this... Ok, let's say I settle with lasercutting steel. If I specify the hooks to be 6x2 mm, how much tolerance should I give on the counter hole? -> I expect that my drawings will be adjusted so that the lasercutter draws the "exact" dimensions I specify. My approach would be to make the counter holes eaxctly 6x2mm, too, and fit the parts together with a hammer. Could that work? Maybe cooling the hook down with ice spray helps? /edit: re-added the picture from before so you don't have to go look for it: