Sunday, 25 December 2011

Mendel90

I never understood why Mendel has a triangular prism frame. The way I see it, the frame only has two functions: - To hold the Y bars in a flat plane and to support the tops of the Z bars. It isn't good at doing either:

  • The main forces on the Z bars are in the direction of the X-axis and the frame has no strength in that direction. It wobbles when the X-carriage changes direction. 
  • It also doesn't ensure the Y bars are in a flat plane because there is nothing to ensure one end triangle is not rotated slightly relative to the other. 

After a trip down a cobbled street in Sheffield my Mendel behaves as if one corner of the bed is lower than the other three. This is impossible because it has a flat sheet of glass on it, but it isn't obvious what needs to be adjusted to fix it but it must be the ends of the Y -bars. The bed needs to be level to within about 0.05mm for good results printing 0.3mm layers without a raft. That is difficult to achieve when the Y axis is strung from bars at opposite sides of the machine.

Other problems are: -

  • It gets smaller at the top, so the maximum Z travel is limited by the extruder colliding with the bars. 
  • The sizes of the Z axis and the Y axis are tied together, so you can't change one without the other. 
  • It is difficult to adjust the axes so that they are orthogonal to each other and keep them that way if the machine is moved.

This machine is my attempt answer to these problems. I am calling it Mendel90 as I can't think of a better name at the moment. The 90 is to emphasise that the frame is based on right angles rather than 60 degree triangles.


Two flat sheets are mounted at right angles to form the XY and XZ planes. Two buttresses maintain them at right angles to each other. This relies on the sheets being cut at perfect right angles but in the UK you can buy sheet materials such as MDF or acrylic cut to size and they have good right angles. The only cutting I had to do was to cut the arch out with a jig saw. It doesn't need to be accurate and it could be done with a hand saw. The piece removed could be used to make the Y carriage, depending on the material.

The buttresses are bigger than they need to be. I took them all the way back to give me plenty of room  to mount my non-standard electronics, but it also has the advantage that the machine will sit on five of the six faces, making it easy to work on.


If the anti-backlash springs are fitted to the Z-axis it should print in all those orientations as well, which would be interesting to try. When printing directly on glass, parts come loose when the bed cools. If the machine was on its back they would fall out the bottom. Who needs an ABP? It might also solve the PLA ooze during warm up problem.


The gantry could be unscrewed and laid on its back over the top of the Y axis to make the machine more compact for travelling. In this case the buttresses could be slimmer to allow it to become even more compact.

I used B&Q style fixing blocks to fasten the sheets together.


I bought some of these and I printed some. They are a lot faster to print than Mendel frame vertexes! The economics are interesting: they are cheaper to print than buy, but while my machines are fully occupied making parts to sell, it is more economical for me to buy them. The printed ones are actually more accurate than the injection moulded ones! The holes are all over the place. I think they must be formed by removable cores and the tool must be worn allowing them to move.

I drilled pilot holes using a paper template. I did this by exporting DXF files of the sheets from OpenScad. I then hacked together a Python DXF reader and an SVG writer to make a program that generated drill centres. I printed them on a large plotter but it could be done with A4 sheets tiled together like the Darwin bed template.

The design is modelled in Openscad, down to the nut and bolt level, and is fully parametric so you can make any size machine and scale the rod diameters and motor sizes if necessary. The only limits are that eventually belts would need to be replaced by rack and pinion above a certain length. It also automatically generates a complete bill of materials for anything in the model.



See also: mendel90-extrudermendel90-axes and mendel90-finishing-touches.


Merry Christmas!



162 comments:

  1. Merry Christmas!
    Good to see you blogging again:)

    Nice work.

    Best,
    Laszlo

    ReplyDelete
  2. looks promising, could you describe the electronics you're using?

    ReplyDelete
  3. I once thought about frame similar to this. As usual my mind got distracted to something else :-) Good job.

    ReplyDelete
  4. Looks great, hoping we will see the printed parts on the web somewhere.

    ReplyDelete
  5. Also if you use the ABS juice for your heatbed, you dont have to care that much about the leveling down to 0.05mm (thats crazy!!). I regularly travel with my mendel in trams and subway and dont have problems ;-)

    ReplyDelete
  6. Nice, keep us updated on how the pseudo ABP printing works. I see you use what looks like T5 belts, any problems with slop? I've seen people use GT2 belts for that reason, but then all your parts look excellent so do you find them ok? Is it easy to get the smooth rods perfectly parallel and orthogonal? I've found it slightly tricky on my prusa to get friction down, do you just rely on drilling the holes perfectly?

    ReplyDelete
  7. @Pablo,
    I used the microcontroller described here: http://hydraraptor.blogspot.com/2007/04/brain-box.html with Pololu stepper drivers and an extruder controller described here: http://hydraraptor.blogspot.com/2009/08/time-for-new-extruder-controller.html. It should work fine with any of the normal Reprap electronics though.

    @Auzze,
    Yes I will publish the design and put instructions in the wiki. I will try to get it done over the Christmas holiday.

    @prusajr,
    I prefer to print directly on PET and simply wipe it once a day with Acetone.

    @James,
    I have used both T5 and T2.5 belts and not really noticed any difference between them. I will have to make some measurements to see if there is any advantage to GT2 as they are not easy to get hold of in the UK.

    ReplyDelete
  8. Another great design I want to build. :)

    But I'm also a bit worried about how to get the rods perfectly parallel when there is no way to fine tune them? I know I'm not that good at drilling precise holes, even with a template.

    ReplyDelete
  9. Don't worry about alignment. I have it covered. Details coming.

    ReplyDelete
  10. This looks great! I would call it the "Mendel-T" which is a play on words with the Classic car Model-T and also describes it's upside down shape. Hopefully that's not taken.

    If this were to be drilled by me, it would need some way to be trued. Some kind of adjustment on each axis for parallel and orthogonal tweaking.

    On the other hand, for a laser cut frame I would say leave as is.

    As a side note I would not recommend acrylic as it cracks under stress and vibration.

    Q: Why is one Y rail so much shorter than the other?

    ReplyDelete
  11. The plastic parts have slotted screw holes to allow for easy alignment with a set square.

    There are only 3 bearings on the Y carriage which allows one rod to be shorter and means that only one packet of ten bearings is required to build a machine.

    I don't think there is enough stress to fracture flat sheets of acrylic so long as it only has round holes and radius corners (It is notch sensitive).

    ReplyDelete
  12. Have you seen the Printrbot?

    http://www.kickstarter.com/projects/printrbot/printrbot-your-first-3d-printer?ref=card

    As you had said you modified the Mendel design to reach this point but the result is almost exactly the same as that 3D printer, albeit with MDF instead of printed parts.

    ReplyDelete
  13. Printrbot has nothing to support the top of the Z axis, so I expect it is even less rigid than the Mendel. It doesn't have printed parts instead of the MDF, it has a small studding frame underneath.

    Although it looks smaller it moves outside its own volume so the y-axis actually takes up slightly more space then mine for the same build area and the y-carriage is bigger.

    I expect this to perform better and have a bigger build area for about the same cost.

    ReplyDelete
  14. That is looking good and getting well refined, the plastic parts look less chunky than most machines which is really nice. Easy access is very a big bonus and I do like that the Z motors are at the base below the level of the build area.

    And it has the added bonus that you can paint or style the MDF / Acrylic sheets how you like, to make your machine unique! Nice.

    ReplyDelete
  15. I've been struggling trying to square up my latest frame... this feeds directly into my OCD-laden veins!

    Looking forward to the plans!

    ReplyDelete
  16. hi Nophead,

    I would like to ask, what is your protection against head crash? Ie. on the newest mendel, the top z motors lifts when the head goes against the bed. In the past the general solution was to have springs on the bed to align and also to have some springiness when the extruder goes for whatever reason against the bed.

    I dont see any protection here.

    also I dont know what is the sheet under the pcb heated bed. It looks like some rubber on top and glass textile on the bottom.

    Also the y carriage is aluminium or acrylic? Cant see from the picture.

    Best,
    Laszlo

    ReplyDelete
  17. If the anti-backlash springs and nuts aren't fitted then the Z nuts simply pop out of the bottom of the x-ends if you try to drive the nozzle into the bed. With them fitted it is possible to damage things. I imagine it would break the glass and then the PCB.

    The y-carriage is Dibond. It has a heat shield made from corrugated cardboard covered in aluminium tape.

    ReplyDelete
  18. Thank you for your answers.

    > If the anti-backlash springs and nuts aren't fitted

    Are there any springs on your pictures? I cant see any of it. Am I overlooking it?

    Best,
    Laszlo

    ReplyDelete
  19. "I never understood why Mendel has a triangular prism frame."

    I always assumed it was meant to reduce the amount of materials/components that couldn't be replicated. The goal of the project is to make 100% of the parts needed buildable on a reprap. By creating a triangular frame, they reduce the length of rod needed, and the number of corners requiring nuts/bolts/etc.

    While the concept of a robot that can completely build itself is grand, honestly I think a more practical approach to building the robot is warranted, and let the robots build bespoke things, rather than act as mass production factories - which they don't excel at.

    ReplyDelete
  20. @Laszlo,
    I don't have the anti-backlash springs fitted but there is provision for them. For some reason I needed them on my Prusa, as without I had 0.2mm backlash but I don't know why gravity was not sufficient.

    @steinman,
    Yes they reduced the rods from Darwin, but lost sight of the primary function of the frame IMHO. Four sheets instead of 12 rods and all those nuts actually reduces the amount and types of vitamins.

    ReplyDelete
  21. Nice work! This looks a lot more rigid and scalable than the default Mendel toblerone - is that the same reason you went with 10mm smooth rods (looks like 10mm at least)?
    Any reason you changed the standard Z-axis?

    I'll be swapping out my prusa frame as soon as your design files go online.

    ReplyDelete
  22. i love it. this is the same way i built my z axis originally except the buttresses where two right angle triangles cut from 1 rectangle. i saw the mendel and saw endless complexity when the reality of what you need is quite simple. you've nailed the simplicity of everything needed. did you manage to fit ribbon cable for the heated bed? this thing will be a workhorse of reliable quality. the printing on it's side could be a very interesting little extra, although in australia i think you would have to cool the bed maybe a fan would work because at room temperature after printing PLA parts are still stuck for me, requires little force to take them off but they are still fused. 3mm glass from a fax machine, room temp maybe averages 25-30. maybe GCODE to shake the bed slightly violently...

    ReplyDelete
  23. Hi nophead,

    I noticed one thing on your new machine. Is the hot end going to be to close to the belt for the x axis ?

    Geirhart

    ReplyDelete
  24. Hi again

    Forget my last comment when I looked at it again the belt is in front of the rods, and the hot end is in between.

    Geirhart

    ReplyDelete
  25. I had the privilege of seeing the Mandel 90 in action when Nophead brought it along to a build day at the Sheffield RepRap group. This was shortly before xmas and Nophead had been up all hours finishing it off so he could bring it along to show. I can confirm it works great and the build quality is absolutely everything we have come to expect from Nophead.

    Merry Xmas and a Happy New Year.

    ReplyDelete
  26. i tried a fan (also cleaned my glass with acetone) part slipped right off no trouble :) happy days. would mount my printer on the wall and put a bin underneath right now if the xy carriages had upward constraints. hanging printers on a wall also uses dead space clears the desk up again ! pumped.

    ReplyDelete
  27. Looks very nice and simple. Do you have any problems with the twisted belts?

    ReplyDelete
  28. wanting to upgrade my prusa mendel to something more stable in a few months, I was leaning towards the mendelmax...but now that I've seen this I'm really interested in seeing how this develops in the future. the simplicity and elegance in your design leads me to think it could be quite stable and robust.

    ReplyDelete
  29. Hi Rhys,
    The half twist seems to work well so far. If it proved to be a problem I could revert to cutting it and joining it back to back as here: http://hydraraptor.blogspot.com/2011/06/half-belt-hack.html

    @christian,
    It may even be more stable than the MendelMax because the only direction the dynamic forces act in are trying to bend about 100mm of MDF along its plane, where it is very stiff due to stiffness being a cubic law on thickness, IIRC.

    The MendelMax along with all other Mendel derivatives I have seen have the frame strong in the wrong direction. Even with something as stiff as aluminium extrusion it wont be very strong in the X direction. Shape matters more than material strength.

    ReplyDelete
  30. I've long been frustrated with those same constraint issues on the RepRap. Thank you for your work with this derivative.

    Your comment about shape mattering more than material strength reminds me of assembling cheap shelves and desks; the way the feel flimsy and unstable until that last step of tacking the cheap, weak particle board on the back to constrain the rest of the structure with its resistance to shear.

    I look forward to seeing where your design goes from here, and what derivatives others create.

    ReplyDelete
  31. I have to say of all the variants created so far this one Is exactly what I have been looking for.
    Love it and after just seeing the photos and diagrams It gives me a bunch more Ideas. I am looking at welding a frame like this together out of stainless. It would be much more stable on my boat than most other materials

    ReplyDelete
  32. Yes MDF is probably not the best thing on a boat!

    Acrylic sheet or even Aluminium plate screwed together might be easier than welding SS at perfect right angles.

    ReplyDelete
  33. well, I can't be certain about the perfect 90's on welding but Im pretty certain I can get the frame square to within a nominal amount.

    My main reason for also using ss is I work in food processing so have a LOT of access to ss.

    ReplyDelete
  34. btw, how soon will you be posting the printed parts files?

    ReplyDelete
  35. I hope to sort the files out this week. I am working through a snag list at the moment.

    ReplyDelete
  36. Very impressed with what you have done, I'm planning a build for my first printer and I have chosen the Prusa for it's simplicity, Your work comes highly rated, any chance I could pay you to print me your best version of the Prusa in Orange? I like the idea of the linear bearing system, but any other improvements would be much appreciated.

    ReplyDelete
  37. Sorry but I am concentrating on this machine so I am not printing Prusa's at the moment. Also orange ABS is expensive in the UK, so I would need to charge more.

    ReplyDelete
  38. This design reminds me of the Rockcliff MDF CNC machine. I built mine in 2007 and after about a year the MDF had warped in the damp conditions of my shed and became grossly inaccurate. I rebuilt using aluminium angle.
    My colleague has encouraged me to build a Perusa and I have been making the printed parts on my CNC rig in 20mm ply wood (mostly) and other thicknesses as appropriate.
    I got to say I don't like the 8mm stud bar as a rigid material. Using a single bar at the base to keep the vertical components vertical has to be flawed. (?)

    I am encouraged by your design to go out and buy some 20mm MDF and make one too.

    Thank you

    ReplyDelete
  39. Hi nophead,

    It's nice to see you posting again !

    I love this new design, and your analysis of the triangular shape of the Mendel encourages me in the way I have chosen for my first repstrap : right angles between XY and XZ.

    Instead of trying to print a Mendel with it, I think I should use my time to upgrade this one in that way !

    Best regards,
    Cyril

    ReplyDelete
  40. Are you planning on ultimately moving away from MDF to something a little bit more durable/sturdy/warp-resistant?

    Very enamoured with this design, probably going to build 2 or so, but thinking of going for ply myself.

    ReplyDelete
  41. I am going to try acrylic next.

    I haven't had any problems with MDF warping/flexing or not being durable in my home environment. It does expand and contract a little atmospheric moisture, which affects the z-limit switch. Mounting it on the bar instead of the bar clamp will probably fix that. Reprap firmware uses a bottom limit switch anyway so that would attach to the top of the motor and not be affected.

    I am told that soaking MDF in acrylic paint stabilises it.

    In my experience ply warps more than MDF.

    ReplyDelete
  42. For the MDF you can also seal the edges with drywall compound. Doesn't take long. Not needed for the face of the wood. Also allows the MDF to be easily painted.

    ReplyDelete
  43. I have read that MDF is porous enough to pull a vacuum through it on a vacuum table, so I think the faces will need sealing as well to keep atmospheric moisture out.

    ReplyDelete
  44. "The main forces on the Z bars are in the direction of the X-axis and the frame has no strength in that direction. It wobbles when the X-carriage changes direction."

    I made the same comment when eD and Vik started discussing the "wedge" reprap but no one seemed to give a damn about this xD

    ReplyDelete
  45. I was just looking at prices for acrylic sheets and it seems pretty inexpensive. I think a clear acrylic base would look pretty cool. What are the dimensions of the wood panels in this prototype?

    I'm assuming that acrylic rods would be terrible as far as reliability goes. Thoughts?

    ReplyDelete
    Replies
    1. The base is 483 x 450mm.
      The gantry is 483 x 323mm.
      The stays are 323 x 169mm.

      The version with 8mm rods and 200x200 build area comparable with Mendel is a bit smaller.
      The exact dimensions are likely to change as being fully parametric small tweaks can change everything.

      I don't think acrylic rods are accurate enough to use with linear bearings. They might work with felt, but you would have then built a pretty good Van de Graff generator!

      Delete
    2. Did you design it to be 19" wide on purpose?

      You could even use a bog standard rack mount shelf as a base to start from.

      Delete
    3. No just a coincidence. I do have a 19" rack cabinet about 5' tall though. I could put 3 or four in that, but I would have to evict the PDP11.

      Delete
  46. Would you recommend this as a first printer to build? I've been thinking about building one for a while and I'm just about to start ordering parts.

    ReplyDelete
    Replies
    1. I would (but I am biased) if you don't mind waiting a few weeks for me to sort out the files and the build instructions.

      Delete
    2. I think I'll wait and get started on the electronics now. I assume I can use any of electronics that have been used for the prusa?

      Delete
    3. Yes it will work with any electronics.

      Delete
    4. I'm thinking of finally building a printer (been looking in to it since ~ 2008) and your design seems quite interesting. As with you, I'm more interested in a design that works well and isn't too hard or expensive to build even if it needs more vitamins. (I understand why the core team consider it important, it's just not something that's so important to me.) So I can't wait until the files and build instructions are up, keep up the excellent work :-)

      Delete
  47. Very nice & well thought out. I'm looking forward to strapping this with plexiglass since we got a pile of thick plexi and no MDF.

    ReplyDelete
  48. Hi nophead,
    Thanks for your leadership. I've been trying to figure out which design to go with and I think the Mendel90 makes the most sense. I'll be trying to grok all this, sourcing parts, and waiting for your build instructions.
    I am interested in doing it in clear acrylic. And also in scaling up quite a bit. I'd curious about the warping problems that come with larger prints, but before building I'm going to try to understand which size printer can be made safely as far as it's components.
    Thank You,
    John (aka groundhog in IRC)

    ReplyDelete
  49. I am working towards making one in 6 and 10mm acrylic.

    I have made ABS parts up to about 200mm long but only when enclosed in a box at 45C. To make larger parts you would have to use PLA. For larger ABS parts you need the chamber at about 70C, which is too hot for a machine made from plastic parts. I.e. the chamber needs to be just below the glass transition for the plastic.

    ReplyDelete
  50. So it seems clear acrylic isn't going to be the way to go if I want to create a chamber, as acrylic doesn't seem to hold up to well at 70C. I've also thought about glass. But I guess everything in the build chamber would need to handle 70C+. Things begin to get complicated. Maybe that will be my second build.
    I've been told acrylic is a very bad choice. "It will crack" they say. I've worked with it before and it can seem brittle ... if I'm not careful with it when working with it. But it held up for it's purpose just fine. Shrug.
    As I look at what I'd like to print, I can imagine I'd be able to print everything I'd want if I can get to 30cm cubed. Even though I might not have much success with large builds, is it a bad idea to even make a Mendel90 with say a 35cm cubed build space? I figure I can still make small items then, maybe have some ability to make several at a time. You mentioned about belt problems after a certain size.
    Also, I'll use metric measurements throughout for the cultural experience.
    -- groundhog

    ReplyDelete
  51. Acrylic is brittle and notch sensitive but I think it will be OK in flat sheets with rounded corners and little stress. I have a 6mm acrylic door on my Mendel hung from hinges fastened by tapped M3 screws and that is no problem. I also made parts of my first extruder from acrylic. I will find out soon anyway.

    Long belts tend to sag but I think it will scale to 350mm OK. Any bigger than that and moving table is not ideal. In fact Mendel is about the size where moving head, moving table or moving head and table all yield about the same build volume.

    You should be able to make parts as big as you want with PLA and a chamber temperature around 50C. Making large ABS parts is a lot more difficult. I would use a Darwin / Ultimaker type layout because that minimises the volume that needs heating.

    ReplyDelete
  52. I'm going to go ahead and build a Mendel90 first, with a possibility of scaling up the bed. I'm still working on grokking all of this so I'm going to make changes only after I understand what I'm doing. I'm going to start with wood, although I really really want it to be awesomer (clear). But I need a stable foundation to build on.
    So, grokking and sourcing I go, awaiting your wise instructions :)
    -- groundhog

    ReplyDelete
  53. I've just become interested in RepRap machines and have been browsing various docs describing different builds.

    So far I've seen no mention of the material 'Tufnol'. I would have thought Whale or Carp (dearer) brand Tufnol far superior to MDF or acrylic for these machines. It's much used by toolmakers for jigs and fixtures in engineering. 10mm Tufnol would be far more rigid and stable than 20mm MDF.

    Jim

    ReplyDelete
  54. Yes I think it work well but is a lot more expensive.

    ReplyDelete
  55. What size are the MDF panels then? A piece of 10mm 'Whale Brand Tufnol' 400mm x 300mm is £21.54 from Davis Plastics. It seem to me to be 'spoiling the ship for a ha'porth of tar' to spend £100+ just for the set of stepper motors alone (for that's what they appear to cost) and skimp on the material for the 'chassis'.

    Jim

    ReplyDelete
    Replies
    1. The stepper motors are only £72.

      The base of the prototype is 483 x 450, the gantry 483 x 323 and the buttresses 323 x 169. B&Q cut them out of a single 1220 x 607 x 12 MDF sheet for £12.48. The Tufnel would be £214.59 for enough to make two machines, but not cut to size.

      The beauty of the fully parametric design is you can use whatever you want. You put the sheet thickness in the configuration file and whether you want to use wood screws, tapped holes or nuts and bolts and it generates the sheet plans and the bill of materials as well as the plastic parts.

      Delete
  56. [quote] The stepper motors are only £72. [/quote]

    Interesting! Where can you get them at the above price? One of the things putting me off so far is the cost of the motors.

    Jim

    ReplyDelete
    Replies
    1. I bought my motors first*, figuring that any cartesian robot/cnc that I might be interested in would be adaptable to some modern, beefy, standard steppers.

      Of course, that's a big first step down the slippery slope, but I think it constrains ones's future options the least.


      * After messing with some homebrew-junkbox electronics to see if it would fit some Teacup RepRap software.

      Delete
    2. I got a dozen steppers that had been taken out of medical microscopes off ebay for £1.50 each, pretty much the same spec as the ones makerbot use. Just need to get round to the rest of the machine...

      Delete
  57. http://www.zappautomation.co.uk/product_info.php?cPath=9_159_43&products_id=516

    ReplyDelete
  58. Just like several of the previous posters I have followed the 3D-printer movement for a few years and am also planning to build the Mendel 90 as my first printer. It looks really really promising, good work!

    While I am holding my breath waiting for the plans for your Mendel 90 I thought I'd ask you about your modelling of the plans. I have just begun discovering OpenSCAD, and while loving it already I find that as the models get alittle more complex that a few standard solids it becomes really slow to rotate the rendered view. It is of course possible that I am just doing something wrong :), but when taking a test object I've modelled in OpenSCAD to a local computer store and trying it out on several setups I found that NVidia GT5xx which seem to be in most current consumer laptops really suck at running OpenSCAD.

    What kind of computer you are using for modelling Mendel 90?
    Are you really able to rotate the view in OpenSCAD when viewing the whole machine?

    ReplyDelete
  59. I am using a Laptop that is nothing special. All the vitamins and printed parts have a render statement in them. That means it takes about 4 minutes to compile from scratch but it then pans and rotates reasonably quickly, even the whole machine.

    The renders are cached, so once loaded, I can change things and see the changes in a reasonable time. There are some issues with the caching and dependencies which Marius is working on. In the meantime I work round it by touching files.

    ReplyDelete
  60. How do you align the thing, if you don't build and drill the holes 100% accurate in the first place? Surely a Prusa is a lot more flexible in terms of alignment.

    ReplyDelete
  61. All the parts have slotted screw holes allowing a couple of mm adjustment. I have replaced the fixing blocks with my own design.

    It is very easy to align because the sheets force co-planarity and the right angle between Z and Y. It is then easy to use a set square to set the right angle between Z and X and between X and Y.

    In contrast you have to manually align everything on a Mendel as all four ends of the Z and Y axes have two degrees of freedom and there is no easy place to use a set square. The instructions suggest using a plumb line to set the Z bars vertical. I didn't find that practical for many reasons.

    ReplyDelete
    Replies
    1. Does that also apply to the fixing blocks you printed?( i.e. The replacement for the B&Q blocks) Otherwise do you not have to very accurately build the MDF frame. Okay you have adjustments on the printed parts, but if the frames out surely this is a potential problem.

      Delete
    2. Yes I redesigned my fixing blocks with slots.

      To drill accurate holes I printed paper templates with cross hairs, punched through with an optical centre punch and started each hole with a spot drill.

      I don't think it really needs that level of precision. As long as the blocks pull the sheets tight together you get the Y Z right angle. To get X Y you need to get the holes in the base correct to about 1mm.

      I will be sourcing a supply of CNCed sheets.

      Delete
  62. "All the parts have slotted screw holes allowing a couple of mm adjustment. I have replaced the fixing blocks with my own design."

    Okay thats good. Any news on when source will be available?

    Any chance that parts could be made available as STEP or any other filetype readable by Alibre or is it all done in OpenSCAD

    ReplyDelete
    Replies
    1. Maybe this weekend if I get my head round github.

      The design is all in Openscad and it doesn't have a STEP export, sorry.

      Delete
  63. Will you be making both the 10mm and 8mm bar versions available?

    I have to say I prefer the idea of the 10mm even if it means the build cost is more.

    ReplyDelete
    Replies
    1. It is all the same code. You set what size you want in the configuration and it generates STL files and DXF files to match.

      It is more like software than 3D modelling. It will come with makefile.

      Delete
  64. nophead, any chance of a bom/hardware list lying around? It'd be very nice to order as much as possible in advance while the plans are being finished. Or is there still big changes being made?

    ReplyDelete
  65. I am making lots of small changes everyday. The BOM is generated from the openscad script and depends entirely on what options you select. E.g, the diameter of the rods, thickness of the sheets, etc.

    I will make a couple of standard configurations, one to reuse Prusa bearings and rods, and another with 10mm rods like the first prototype.

    ReplyDelete
  66. Everytime I take a look at your site, something else amazing is published! Looking through this post got be thinking that this could be a better alternative for the metal parts you'vew replaced because as you've mentioned, perfectly cut parts are easy to source, affordably, and if you added a milling end to the x-carriage you could Reprap (subtractivly) new frame parts.

    Well done and keep up the good work.

    ReplyDelete
  67. Hi Nophead,

    Every day on the ICS reprap list people are asking if Mendel90
    has been released yet, we have the measurements for the MDF now. And it is now over a month since Christmas 2011 so when can we expect the files for the plastic parts so all of us can build a Mendel90?
    We have all waiting for this information for months so Nophead when will you be ready for release of the design?

    ReplyDelete
  68. I have completed the design today, so very soon. I have to set up make files to make the STLs and DXFs and tidy up the directory structure, etc.

    What is ICS?

    ReplyDelete
  69. Hey Nophead, way back at last Christmas you indicated that your model will be published, now in February there are still no models for us to build, there was a teaser on the size of the MDF, but we all want to see the plastic designs.

    When will these be on thingiverse? Wallace, RepRapPro Huxley and printrbot are all there why not Mendel90?

    There are usually a few comments a day on the reprap ICS feed wondering when Mendel90 will be forthcoming, so any day now?

    Nice design, we all want to make one.

    ReplyDelete
  70. Yes it was a big mistake blogging at Christmas, I should have kept it secret as I have been hassled ever since. I was off work then and had no printing orders so I had some time to blog. Now I am working full time during the day and working on this until the early hours of the morning every day.

    Wallace, etc are all finished designs that is why they are published. It has taken me this time to complete the design to my satisfaction. It is a full model with everything apart from the wires and electronics. It generates the bill of materials so that is always correct. I don't think anybody has done an Openscad model with that level of detail before. Marius has been great fixing various bugs and limitations for me. The Prusa 2 was published a while ago but I think people are still asking for a correct BOM.

    I have no idea what ICS is, so I frankly don't give a damn what is said on it. People can wait for this, go with one of the many other designs already published or design their own. It isn't difficult, just very time consuming, at least for me anyway.

    I just have the mechanics of publishing it to do now, which is what I am doing. It isn't as simple as putting a few STL files on Thingiverse. This is a parametric design which generates an infinite number of machines ranging in size and using different sheet materials, wood, acrylic, etc.

    ReplyDelete
    Replies
    1. Since we're not paying you, you should do it at whatever pace makes it interesting and fun for you.

      Count me as hassling you to slow down and do it right.

      Delete
  71. I'm very curious about your openscad experiences and how you are using it. As you said this is probably the most complex openscad model out there.
    A couple of questions if you can find a few minutes.
    Did the image at the bottom of this post come directly from openscad?
    Do you have any shortcuts for aligning parts when describing the entire assembly?
    Roughly how many lines of code are we talking about?
    What libraries do you use?
    Have you built up a library of your own? and if so what's in it?
    When can we see the files :P

    Also congrats on a sweet looking printer.

    ReplyDelete
  72. Yes the image is from openscad. It exports PNG files.

    The parts are aligned with expressions like: Y_belt_line = X_origin - ribbon_clamp_slot(bed_ways) / 2 - y_belt_anchor_width() / 2 - 5;

    About 6000 lines of code.

    No I don't use a library.

    Soon.

    ReplyDelete
    Replies
    1. Hi nophead
      How can I send you a private mail?

      Michael

      Delete
  73. I am on gmail as nop dot head.

    ReplyDelete
  74. nop dot head at gmail dot com

    ReplyDelete
  75. Hi nophead
    I just found the OpenMendel90 on Thigiverse. When I saw it I was little frightened and disapointed because I first thought it is your final version.
    Above all things I don't like the arrangenent of the x-rods put one below the other. The extruder placed on the side cause a torque between the two rods. That will put additional weight to the x- and z-bearings and probably weare out them sooner.
    Althoug it maybe have some advantages it looks not nice to me.
    So there is still no reason to hurry. Take your time. Your version is the best :-)

    Michael

    ReplyDelete
  76. Is that the final extruder configuration? How do you get the bolt and large gear out without taking off the extruder in case you need to clean it out?

    ReplyDelete
  77. I run the extruder backwards to remove the filament. I then undo the nut, which is only finger tight against a weak spring, and slide out the bolt with the large gear. See http://hydraraptor.blogspot.com/2011/12/mendel90-extruder.html for the tweaks that allow me to do that.

    I am designing a new hot end at the moment and that will be attached by a clamp, rather than the two bolt method at the moment.

    I will also make a smaller version for Huxley90 which may allow two to be fitted the Mendel90, with a tweak to the carriage.

    ReplyDelete
  78. Hi Nophead,
    I'm new to the RepRaping game I was just about to buy a Printrbot until I saw your design. It makes sense to me, it looks very solid and judging by all the posts you probably have revolutionised the Mendel . Will you be selling the printed parts for it in the future ?.

    Thanks for all your effort
    Buzz

    ReplyDelete
  79. Yes I am printing parts to sell and writing the instructions this week.

    ReplyDelete
  80. @Buzz, the other option of course would be to buy the printrbot, use it to print the Mendel90 parts and then transfer the vitamins from one to the other.

    ReplyDelete
  81. I doubt there are many vitamins in common apart from the bearings and motors.

    ReplyDelete
  82. I would expect with some tweaking of the mendel90 you could also make use of the electronics, hot end and heated bed and perhaps a few other bits and pieces like the belts and knurled bolt.

    ReplyDelete
  83. Wow loving some of the new concepts you've used in this new design, especially the use of ribbon cables! Though what is the red stuff your using to support the cables through their movement?
    Thanks

    ReplyDelete
    Replies
    1. Polypropylene about 0.5mm thick. I cut the first ones from a file folder but I now buy it from here: http://www.hobarts.com/store/polypropylene-solid-p-27.html. Postage is a killer for small amounts though so I am on the lookout for a better source.

      Delete
    2. Thanks for the fast reply! I think I'll try sourcing a folder first! lol

      Delete
  84. Thanks for your work on this! I started building one with an aluminum frame and your design is great!

    http://frankieflood.blogspot.com/2012/07/aluminum-mendel90.html

    ReplyDelete
  85. nophead, you are some kind of OpenSCAD wizard.

    I'm building a M90 as my experimental printer, and it was very easy to upsize the plastic parts for spare NEMA23's I had laying around and raise the Z axis up a few inches. Almost too easy, I'm worried something is going to be wrong and I won't notice :)

    ReplyDelete
  86. Hi Nophead,

    Thanks for a great design, I'm aiming at building a Mendel90 for a stable and easy to set up first build!
    One question (for a start anyway!) with the acrylic version on Thingiverse (10mm and 6mm thicknesses) have you used 3mm tapped holes or are they a screw and nut combo?
    Many thanks
    Colin
    Wellington, NZ.

    ReplyDelete
    Replies
    1. The acrylic version has 3.3mm holes tapped for M4.

      Delete
  87. Could one use cutting boards for the frame, such as these?

    http://www.amazon.com/Thick-White-NSF-Cutting-Board/dp/B001DI5VCA/

    ReplyDelete
    Replies
    1. Probably, if it is stiff enough.

      Delete
    2. I bought one, they are not flat enough.

      Delete
  88. Nop_head,

    you wrote:
    > I used B&Q style fixing blocks to fasten the sheets together.

    Would you happen to know any synonyms for those prismatic fixing blocks, especially what they might be called in the US? I've tried a fair number of search terms and so far cannot find them anywhere. Since I'm in a boot-strapping/RepStrapping situation, I can't yet print them myself. Any suggestions to find a US source of those blocks would be most appreciated.

    Thanks,

    Larry
    ursine @t Gm4il d0t c0m

    ReplyDelete
  89. Hi Larry,
    Sorry that is all I know them as and am not familiar with US hardware shops. I would be surprised if they are not available everywhere though.

    ReplyDelete
  90. Hi Nophead ..
    read all this with interest and want to plan on getting started .. finance plan is first ..
    if the Kit is still GBP 499 .. what's the cost of shipping out to Auckland New Zealand? Not in a hurry so slow economical shipping is the right option, please.
    Thanx
    Tony Burns
    p/s Would prefer to converse by email if that's acceptable .. how do we do that? .. t

    ReplyDelete
    Replies
    1. Yes the price is still £499. Shipping to NZ is £83 for standard 5 days+ with insurance. I use gmail and Google thinks I am called nop head so you should be able to work out my email address.

      Delete
  91. This is a thing of beauty! Very nice work indeed.
    Could any one help me to run the make_machine script please? I have tried just about everything to get the script to run but I always end up with an "SyntaxError: invalid syntax" message:

    C:\Python33\Mendel90-master>c:\python33\python.exe make_machine.py mendel
    File "make_machine.py", line 19
    print "usage: make_machine [mendel|sturdy|your_machine]"
    ^
    SyntaxError: invalid syntax

    What am I doing wrong? I don't know enough about Python to make any sense of the script itself.

    ReplyDelete
    Replies
    1. You need to use the latest Python 2.x, Python 3.x is a different language.

      Delete
    2. Ahh...... Yes. Many thanks Nop Head for your quick response! I think I could have been banging my head against the wall for a long time trying to work that one out!
      Denzil

      Delete
  92. Hi nop head, is there a chance of getting the absolute drill coordinates of the sheets? I'm planning on making a Mendel90 from sheet aluminium on a CNC router. I have looked at the SVG file but all coordinates of the drill holes seem to be off-centre. Thanks for the great work. Cheers and merry christmas, Florian.

    ReplyDelete
  93. The SVG and PDF files are for printing out paper templates with cross hairs. For CNC you should use the DXF files.

    ReplyDelete
    Replies
    1. OK, but when I open the DXF file with Inkscape, I get a nice outline of the sheet (eg. y-carriage) but all drill holes are made up of small line segments. How can I turn that into cicles and get the centre position of each circle? Can I export new DXF files from OpenSCAD and have circles instead of line segments in my DXF file?

      Delete
    2. OpenScad only works in segments, not true circles. I find them fine for CNC routing or laser cutting. What sort of CNC are you using?

      If you look at the python scripts I wrote to convert DXF to SVG drill templates you can see how I recognised circles and converted them to true circles and centre points, but they are not needed for CNC unless you have a CNC drilling machine.

      Delete
    3. I'll send the drawings to a workshop that makes aluminium front panels. They will need drill coordinates for their CNC drill. They can also tap the holes later to ship ready-to-assemble sheets. I'll experiment with your Python code to print a list of drill hole coordinates and corresponding diameters.

      Delete
  94. Where you able to use acrylic instead of MDF? Sorry if this has been been answered, but I searched the replies for "acrylic" and didn't find the answer to my question. I'm planning to build my first 3D printer and this design appears to be sturdy and inexpensive. To prevent the holes near the edge from cracking the acrylic, I plan to increase the area by 1 or 2 inches all the way around.

    ReplyDelete
    Replies
    1. Yes I made an acrylic machine and used tapped M4 holes. That is what the files in the mendel directory on github are. I had no problem with cracking but I much prefer Dibond aluminium composite as it it lighter, stiffer and cheaper. That is what the kits I sell use.

      Delete
    2. Where do you sell the Dibond kits?

      Delete
    3. http://forums.reprap.org/read.php?94,196585

      Delete
  95. just found your design... really like it

    but in which of the SCAD files do i define the work area? there are loads of tiles :-D

    ReplyDelete
  96. In the config files. For example the dimensions of the Dibond machine are defined here: https://github.com/nophead/Mendael90/blob/master/scad/conf/dibond_config.scad

    ReplyDelete
    Replies
    1. thanks :-D

      and i guess https://github.com/nophead/Mendel90/blob/master/scad/conf/config.scad is where i set sheet thickness if going for MDF sheets ? dibond is not common in denmark

      Delete
    2. No, there is a configuration for the MDF version here: https://github.com/nophead/Mendel90/blob/master/scad/conf/sturdy_config.scad. That uses 12mm MDF. If you need other thickness you can define them in https://github.com/nophead/Mendel90/blob/master/scad/vitamins/sheet.scad and reference them in the machine specific config file.

      Delete
    3. hehe, i think 12mm would be enough - if going ahead i plan to do as much DIY as i can, i have a normal mendel with threaded rod triangle construction.... plan so reuse as much from there i can

      But have an VERY crazy idea of a 40*40*40 cm build area...

      Delete
  97. waiting for main.scad to render... you have warned people in the read me.... takes a long time

    but then how do you run py scripts when you use the windows version of OpenSCAD ? your readme assumes the linux variant it seems... could be wrong thou... and google might have the answer but me and google are not good friends, always hides the answers for me

    ReplyDelete
    Replies
    1. I use Windows XP most of the time. To run python scripts you have to install Python 2.7.x from python.org.

      Delete
    2. Xp... What i would call an old timer :-D just joking... ok ao a python install it is...

      Delete
    3. Yes I am an old timer. I also have machines with Win7 but I prefer XP.

      Delete
  98. So i changed the sturdy config to:

    echo("Sturdy:");

    Z_bearings = LM8UU;
    Y_bearings = LM8UU;
    X_bearings = LM8UU;

    X_motor = NEMA17;
    Y_motor = NEMA17;
    Z_motor = NEMA17;

    hot_end = JHeadMk5;

    X_travel = 400;
    Y_travel = 400;
    Z_travel = 400;

    bed_depth = 428;
    bed_width = 428;
    bed_pillars = M3x20_pillar;
    bed_glass = glass2;
    bed_thickness = pcb_thickness + sheet_thickness(bed_glass); // PCB heater plus glass sheet
    bed_holes = [418, 418];

    i know the bearings are smaller but its part of the plan to reuse as much of the parts i have... are the other values total bonkers? i just doubled them up as i want twice the size.. i know its going to be hard to get a heated bed that size... i still have to thinker if i need to get a PCB made at the size or if i can use 4 normal ones and just wire them right

    machine.scad i changed on to be:

    include


    am i doing ok so far? still ned to figure how to run the py files

    ReplyDelete
    Replies
    1. Seems right apart from the fact you don't normally hand edit machine.scad because make_machine.py does that for you. Once Python is installed I just run make_machine sturdy but some people seem to have to do python make_machine.py sturdy if the file associations are not set correctly during the python install.

      You also need to install inkscape and blender and set your search path to include them.

      Delete
    2. if you have the time for it, what file would you need to look at to confirm that i did the edits right?

      and as a side Q... if i need smooth rods can you supply them in custom lenght if i have time to wait for it? i have a susspision that i cant source them local

      Delete
    3. looking at the config for sturdy again....

      i see this line: Y_carriage = DiBond;

      but since i cant get Dibond that easy could i get away with something else? and what should i then change the config line to?

      Delete
    4. tried to run the py script but got an error.. know what it is but why?

      C:\Users\-bo-\Desktop\Mendel90-master>python make_machine.py sturdy
      openscad -D $bom=2 -o dummy.csg scad/bom.scad
      Generating bom ... done
      Traceback (most recent call last):
      File "make_machine.py", line 17, in
      make_machine(sys.argv[1])
      File "make_machine.py", line 11, in make_machine
      sheets(machine)
      File "C:\Users\-bo-\Desktop\Mendel90-master\sheets.py", line 50, in sheets
      f = open(dxf_maker_name, "w")
      IOError: [Errno 2] No such file or directory: 'sturdy/sheets/frame_base_dxf.scad
      '

      C:\Users\-bo-\Desktop\Mendel90-master>

      Delete
    5. can i just use the sheets dir from the mendel dir? there was of course no sheets dir under sturdy

      Delete
    6. tried to copy the sheets dir from dibond and got a bit further... but failed again... its missing a file if you wonder why not in english at last lines

      C:\Users\-bo-\Desktop\Mendel90-master>python make_machine.py sturdy
      openscad -D $bom=2 -o dummy.csg scad/bom.scad
      Generating bom ... done
      openscad -o sturdy/sheets/frame_base.dxf sturdy/sheets/frame_base_dxf.scad
      -351.25 -429.5 351.25 429.5
      inkscape -f sturdy/sheets/frame_base.svg -A sturdy/sheets/frame_base.pdf
      Traceback (most recent call last):
      File "make_machine.py", line 17, in
      make_machine(sys.argv[1])
      File "make_machine.py", line 11, in make_machine
      sheets(machine)
      File "C:\Users\-bo-\Desktop\Mendel90-master\sheets.py", line 67, in sheets
      InkCL.run("-f", base_name + ".svg", "-A", base_name + ".pdf")
      File "C:\Users\-bo-\Desktop\Mendel90-master\InkCL.py", line 11, in run
      run = subprocess.Popen(["inkscape"] + list(args) + [" -z"], shell = False, s
      tdout = subprocess.PIPE, stderr = subprocess.PIPE)
      File "C:\python27\lib\subprocess.py", line 709, in __init__
      errread, errwrite)
      File "C:\python27\lib\subprocess.py", line 957, in _execute_child
      startupinfo)
      WindowsError: [Error 2] Den angivne fil blev ikke fundet

      C:\Users\-bo-\Desktop\Mendel90-master>

      Delete
    7. No we can't supply custom rods or any other custom parts.

      Dibond is a trade name for aluminium composite panel (ACP). I would be surprised it is not available in Denmark. It is typically used by sign makers.

      I have used 6mm Acrylic for the y-carriage successfully but it is heavier and not as stiff as 3mm ACP. I tried MDF but it was not stable with heat.

      The error means OpenScad failed to make the DXF file, most likely due to a syntax error. Have a look at openscad.log. No you can't use files from another machine, they will be different sizes and have different holes, etc.

      Delete
    8. Looks like it can't find InkScape, which it uses to convert svg files to PDF.

      Delete
    9. oki, at first i thought there was no sheets folder in there... but in the zip there is so the script must have deleted the folder before first error... i took a copy from dibond folder to get it going.. if the script finishes ok i will extract the whole sturdy folder from the zip run again

      Delete
    10. will search for dibond but never seen it in any shops... and many business's here wont deal with private individuals for some odd balled reason, they bark about VAT number etc etc....

      Delete
    11. how thick MDF did you try? i have pretty stable temps here... one side of flat is facing south and the other north... so i can open windows to regulate temp inside...

      Delete
    12. I think it was 6mm. It was a few years ago on my Prusa and one of the many things about that design that did not work for me. It did not stay level from one build to the next as the heat from the bed dried out one side faster than the other.

      In the UK we don't need to give a VAT number to buy something unless we want it VAT free from another country. In fact even a lot of businesses are not VAT registered so would not have one to give.

      Delete
    13. I don't see how copying files from another machine would get it working as the script always deletes all the files before it makes them.

      Delete
    14. dooh me stupid... i was thinking of the gantry part....

      the carriage yes... you are absolute right... and MDF coated in varnish would stink nasty when hot... plus the moisture cant get out

      but think i found one that deals in Dibond. now i just need to source some alu for the heat plate. might place some support posts along the edge and some near the center to support the extra size

      Delete
    15. the delete thing you are right... it deletes them, but then yes it could not find inkscape and also thought i had added openscad and python to my path... but i had not tested them out before running the script... and i looked at folders and found the sheets dir missing...

      but running of a fresh extracted zip now i figured the culprit... an error 40 cm from the screen... ie me

      Delete
  99. oki, just curious.... how would you do the heat bed if you where to build such an insane big printer?

    ReplyDelete
    Replies
    1. As it would need something like 400W I am not sure I would attempt it. Maybe a milled aluminium plate with ribs for strength but hollowed out for lightness and some resistors to heat it.

      Delete
    2. i guess its an dual PSU setup... and a solid state relay controlled by the electronics... its going to take some creative thinking for sure

      Delete
    3. Hmm.. the alu plate idea sounds the most reasonble way... a total of 0.36 ohm at 12V will give just about 33A drawn which is close to the 400W mark

      now my plan is to use 4 resistors so the heat is generated more even

      Delete
    4. I use 9 resistors on a 200mm bed so I would go for 25 or 36 on a 400mm bed.

      Delete
    5. So a thin alu sheet with 9 resistors... that is going to be fun...

      would you recommend insulation between the heater and plate below it? and if so what can you recommend? maybe just thin sheet of styrofoam and and thicker frame of styrofoam so air cant vent out below?

      Delete
    6. I fill the gap with several layers of corrugated cardboard.

      Delete
  100. and yes have looked near and far in the wiki, but what are the recommended settings for slic3r when printing parts?

    ReplyDelete
  101. upps found it in buyers guide

    ReplyDelete
  102. Have started to investigate the build process to see if i understand it, using the build instructions for the dibond version at the frame assembly (step 2) it mentions 2 alu square tubes under the base plate, but this alu square tube is not mentioned in the list of vitamins in the build instructions, nor does it show up in any of the BOM files created... is this a mistake from your side?

    ReplyDelete
  103. oh the alu square is in the vitamins list in the build instructions... but not in BOM created from the sturdy config...

    ReplyDelete
  104. ahh... see the light at end of tunnel... square tube is used on the dibond model to lift the base from the surface as the screw heads under the base... these would scratch a table surface... the sturdy model does not have this and does not need the square tubes

    ReplyDelete
    Replies
    1. Yes that is the case. The generated BOM should match the config files exactly. The only fully accurate build instructions are for the Dibond kit on Github. The RepRap wiki is somewhat out of date.

      Delete