So I have this idea, and I'm thinking its stupid, because if it wasn't, someone would have already done it. So I'm hoping someone can tell me *why* it's stupid, so I can abandon the idea..

I want to take a printed gear bearing, top it with a printed pulley, and put 8mm thread in the middle of the pulley to drive the lead screw, all in a single print. These are all in SCAD format, and I think I can probably put it together in an afternoon, minus the timing belt (side question -- anyone know how to join the ends of a timing belt?). At that point I'll have an N-lead screw solution that requires no extra parts and one less stepper motor.

I don't see any reason why it wouldn't work, but I am not a mechanical engineer and I'm assuming there's a good reason this hasn't been done. I just want to know what it is before I waste my time

 

Is it this one you mean?: X-Endstop, D-bot by Carlsen

I had some problems with the X endstop (from the D-bot) and my screws, they did not hit each other perfectly clean. So I made this one that has a large area where for the endstop to hit.

 

Can someone help me please. I cant find the STL for the part labelled "Z Bar End V2 20x60 (The outside plate).


Z_BarEndsV2 for 20x60.stl is the other side:

 

I have managed to answer my own question:
When I open file with the windows viewer I only see the one half (as above).
If I import the file into sketchup I see both parts:


I have not encountered this before.

 

Well I thought about that, but I couldn't find a way to anchor it to a stationary spot without a second bearing.. My plan was more to just use the planetary gear as a bearing. I would put a flange on the outside to hold it onto the frame, then put something like a 40 tooth pulley gear added to the middle and screw threads down the center of the pulley. The belt would drive the pulley directly. The only advantage of the planetary gear in the simplest design would be that it would allow for the entire Z pulley assembly to be printed in place.

 

Now in calibration and optimisation mode after the first few weeks of prints, looking good but by no means perfect. Biggest issue is the Z axis, which I believe is a common theme given the amount of discussion here. My 2 cents:

- front and back z screws dont seem to work for me , as there is significantly more weight on the back one, which causes it to turn slightly when the power is off. Then the deck is not flat when I power on later, and i have to level the bed again. Solution here seems either to mount them left and right with equal weight, or to link them with a timing belt. ( many seem to have worked this out already, just catching up!). I actually had the z-screws diagonally, which maybe made matters worse.
- Wish I had got a much lower pitch screw, mine is 4 start 8mm pitch and the platform drops on its own with only a little pressure or weight, so messing the bed leveling.

Am slowly learning the lessons. Great thing is that the c-bot is so modular, can change over without a complete rebuild.

 

Some show-and-tell time.
Just spent a week on Maui: When I came back I felt inspired to make a 3dprint of the island: Above pic is what I came up with. Printed on the diagonal on my 12x12 platform, it's 14" across.
I had to download a bunch on individual chunks of the island mesh from the terrain2stl web app, then seam them together in Maya, + add the base and text (quite a few hours of work...). Sliced in Simplify3d, 200 micron, .4mm Volcano nozzle at 90mm\sec, 20% infill (probably should have done 10% to help with some of the roof issues), 2 shells, took about 13 hours in Makergeeks Gray'matter Gray PLA.
I'm pretty happy with it! Next up will be paint.

 

Once I get the paint finished, I'll blog about the whole process from start to finish. I don't plan to release the stl's for free considering how many man hours it took me getting the whole thing assembled: I've printed & sold my (similar) Oahu project in the past.
But, I'd be happy to send them your way Carl as a thanks for your great C-bot project
However, it's not one stl file, it's like... 32: It was so much data for each map chunk, I couldn't get Maya to boolean them together successfully. So it meant a lot of arranging of separate models in Simplify3d, like a puzzle of Maui all made out of little squares

 

In case anyone is interested, I modded the direct drive back plate to add a mount for the type of endstops used in the rest of the design. It has worked out pretty well for me so far, though it's a bit ugly (I basically just tool one of the other endstop mounts, chopped off half of it, and attached it to the side of the direct drive backplate).

C-Bot direct drive backplate with cheap endstop mount

 

It's up and running!

I have some more calibration work to do. First print I had the steps per mm wrong for the extruder and it was only moving half the plastic it should have. Next up I think I am going to drop the Z driver from 32 microsteps to 16. The auto bed leveling is working very well, but it doesn't need the resolution and might be losing steps because of the reduced torque from the micro stepping. I can hear the Z steppers knocking like a missed step. Second print came out very well though, so I'm not convinced Z is losing steps yet. Firing off a taller print to verify.

The fan is blowing full speed in the video. I've got some work to do on the design for the fan shroud and tuning for Cura to do to get that under control and quiet things down a bit. Without the fan running, this printer is super quiet. Very happy with the way it turned out.

I see I've got some cooling issues still with the fan design and slicer settings. You can see the deformation in the christmas ornament top due to heat.

Next up, finish the heated bed controller, tidy up the wiring, cut down the rear lead screw, move the bed more towards the middle and get back to work on the generator. I have a lengthy list of improvements to make on the hot end carriage design.

Thanks to everyone for your help and a huge thanks to Carl for the design and AKEric for his tall print and front/back Z lead screw mods which I am using!

 

Just skimming through the past month's posts...
I can save you the work. I previously converted almost all the parts to .STP, and around half to .SLDPRT. This was back in November/October though...

I ended up just completely remodelling all of them with a few small changes in solidworks as converting files from STP or STL in solidworks ends up with solidworks' interpretation of how you'd solid model them, and in most cases it's absolutely horrendous for editing later.

Here's links to the .123dx files and STP equivalents Dropbox - C-Bot_123D.zip
And here's the link to the automatically converted solidworks files from STP Dropbox - Solidworks.zip

 

That change from 1Hz to 0.25Hz is most likely not to spec. FOTEK's from what I've just googled are 10mS switching delay for zero-cross detect.
My (mouser-supplied) OMRON SSR sepcifies 1/2 of the load cycle time + 1mS. And with Bang-Bang at 1Hz, My bed sits comfortably with +-0.5DegC.

Although it could well be some sort of software adjustment as if its 1Hz to 0.25Hz, that's a power of 2 somewhere (probably a timer scaler or something)

Are you using a genuine FOTEK? The counterfeit ones are renowned for being badly made/over rated and failing spectacularly.

 

I'm not sure what you're saying here about it being out of spec. Sure there is a possibility of the PWM on or off time being shorter than SSR switch time. However, if it is 10ms, a 0.25Hz 8 bit PWM, that means that each PWM "tick" corresponds to 15.6ms. That means that worst case scenario, 254 and 255 will provide the same power output, or perhaps every PWM output is actually 1/255th higher or lower power than it should be.

Did you think I meant 4Hz and not 0.25Hz? For sure that's too fast.

I don't have a genuine FOTEK. For switching this current (max 5.8A), I didn't think it was worth spending the extra money on it (especially considering I already had it). It is "rated" for 40A, which I would never trust it for, even properly derated to 32A max. I've never had a problem with it and it doesn't even show up as warm on the FLIR, which is where they tend to fall apart-- sticking on when they overheat.

Switching topics, I emailed Keenovo about my heater coming off they they emailed me back within a few hours saying they'd ship a replacement along with a few questions about my build to improve the future quality of their product. A++ would buy again.

 

You mentioned your SSR switches at a lower frequency than you set in your software config? I thought it might be because your SSR was delaying for too long looking for a zero-cross because it could well be non-genuine (as most are if not purchased from mouser/digikey etc)

In-case you wanted to smooth your heatbed temp at set-point 110degC, upping from 0.25Hz (4sec cycle time) to what you intend to get (1Hz), should help you achieve a slightly more stable heated bed temp.

Very interesting stuff about the heated bed. Did anything other than the IR camera image make you think it might be faulty? I think mine is keenovo also, but I don't have an IR camera

Austin

 

Me either and it makes me sad.

 

Oh no, Smoothieware is the thing that is generating the incorrect PWM period. When set to 1Hz it actually outputs 0.25Hz or perhaps 0.20Hz just by rough estimate watching the LED. The SSR isn't even in the equation. I think my original point was that the standard PWM frequency is 2000Hz, which won't work because of the whole mains frequency + zero crossing switching, so it must be changed if you're not going to use bang-bang.

I like sticking at 4Hz because each 8 bit PWM "tick" is 15.6ms which is just under 1 cycle of mains frequency (16.67ms). At 55C I'm on for ~266ms and off for 3.734s. At 100C I am on for ~600ms and off for 3.400s. Sure, I could get slightly less swing if I went twice as fast and probably still have the same resolution considering there are two zero crossings per period but it stays pretty tight this way.

I prefer the PWM over bang-bang because I can't tell when I am actually heated up with bang-bang. My bed will heat from room temp to 110C indicated in something like 3 minutes. However, the aluminum plate is still heating and the glass on top is still much cooler. By watching the PWM output you can tell when it is heated because the output drops off and levels, which is when everything is in thermal equilibrium. If it were bang-bang I'd have to go look at the LED and try to gauge how much it is on versus how much it is off.

Yeah the main indicator was the smoke coming out from under the build plate. The cork mat I had underneath was getting singed and blackened from the heat.

 

Nearly complete! The printer is essentially done at this point, but I am having a bit of a hiccup. I am able to move X, Y & Z independently using the manual controls in Repetier Host perfectly. I can home X and Z without issue. When I try to home Y, everything seems to go perfectly until it hits the endstop. The Y axis hits and stops like it should, but then the X axis goes crazy for about 5 second, erratically moving back and forth in short stutters and then stops. The printer thinks X is still at 0. I can use the controls to move X away from the 'new' 0 point and back, but if I hit home again for the X axis, it does not move and resets its supposed position to a new 0 point. Very strange behavior. Anyone have any ideas? I have had motors with wiring shorts on other printers I have built and those displayed erratic behavior, but it was consistently erratic. This seems to work fine until homing Y and then X takes a dump.

EDIT: I figured it out. I based my wiring on the excellent D-Bot variant, but set everything up physically to home front-left like all of my other printers. The endstop for Y was wired as max instead of min. All is well! On to calibration...

 

Glad you figured it out. While I was reading I was screaming in my head, endstop wiring!!!

When you get it up and running, here's a sample of what it can do after a little calibration work. I'm a happy camper!

 

That is a nice print! I cannot wait to get mine all dialed in. Time seems to be my scarcest commodity. That and every time I sit down I see something else I want to tweak before moving on to the next task. LOL.

The endstop issue didn't register at first due to the X axis jumping all around. In my head I was thinking some sort of short or cross-talk. Then it dawned on me that CoreXY kinematics are very different from what I am used to and it may have been a case of the Y axis trying to continue to move through the endstop causing the motors to alternately miss steps with the result being the X carriage doing a jig. Before even digging into the wiring box I put a clamp on the frame to prevent Y from reaching the stop and it exhibited the same behavior. Knew I had it nailed at that point. I love the process of problem solving almost as much as the rest of it!

 

First test "air-print" Woohoo!

E3D Volcano and VariPower heatbed FINALLY cleared customs and should arrive tomorrow. 24v/25A PSU is on it's way as well. I'll have to use the C/D-Bot to print off the spool holder and RAMPS enclosure as they're too big for my Kossel to handle.

 

Thanks!

Right on.

I hear ya on the constant tinkering, improvements and time... I finally had to say enough and just get it up and running, then do mods AFTER it was working. This was after 5 months of building...

 

That's the conclusion I'm finally coming to.. Hopefully my build plate gets in soon (supposedly 3 days shipping, but I ordered it Monday and still haven't gotten a shipping notice), and after that I can hook it up to the heat bed. I'm almost sad to finish it, because the fun will be over then..

 

My last few parts are finally rolling in. Today my E3d arrived and I started wiring it up before work. I need way more ptfe tube than the kit provides, so that'll get ordered with my pei surface. I too have been waiting for my aluminum to ship, 7 days now. The BOM did not list the shims necessary if you chose the aluminum spacers for the carriages, those are on the way and i can replace the printed spacers, misc fans once i finalize my smoothieboard enclosure. Wire management wasn't the big headache i thought it was going to be, it's mostly sorted. I've also been calibrating as i go along and testing each component as it arrived. The steps per mm on all my axis are set, all values are in my firmware ready for each accessory to get plugged in, etc. After minor tuning i've been able to run the gantry upwards of 150mm/s without the motors or steppers breaking a sweat. It was a fun demonstration and impressed the hell out of me but it's impractical (at least without a volcano upgrade in the future). Best of all i'm way ahead of schedule!

 

For anyone using a DC heat bed with a tall build, what did you use for your wires? The 22AWG wires I have for the rest of the build ard rated for 2.5A, and I need 40A or so with a distance of 3-4ft in order to reach all the way from the power supply to the top build plate position.

 

10 gauge - 25 amps at 12volts for 12x12 heater. What in the world are you using that needs 40amps?

Wire Gauge - DC amperage guide

http://assets.bluesea.com/files/resources/newsletter/images/DC_wire_selection_chartlg.jpg