Prusa i3 problems

After being a MendelMax man for quite some time, I've decided to build my first Prusa i3 and could use a bit of guidance.

When I built my Prusa 2, the process went really smoothly. When I built all my MendelMaxes, the process went smoothly. But unfortunately I'm having quite a few issues getting my i3 assembled. The design... well... just seems kind of half-baked. A lot of the standard parts have been problematic for me, such as the Y-carriage belt holder, the X-ends, the X-carriage, the extruder... There are a ton of 3rd party parts that people seem to be designing and posting to Thingiverse, many of which differ very significantly from the originals. The extruder and X-carriage in particular have probably like 20 redesigns out there!

The lack of bracing on the lasercut frame seems really problematic and securing it with M10 rods and fender washers seems very inelegant. I like the braced frame styles but by now there are three of those and they're each ever so slightly different...

It doesn't seem like there's a ton of development activity going on in Josef Prusa's github repo and most of the recent work I see has been done to the box_frame style. Is the lasercut style dead in favor of the box_frame style or something?

It sort of seems like what's needed is a released design that's "good enough" that people can improve on. The unclear development, unfinished-seeming parts, and overall fragmentation are really confusing me. Once I'm done with my modifications and I have a machine whose parts all works really well together, I plan to release them but of course that just adds to the problem of fragmentation...

Has anyone else had difficulty getting it all to work together? What have others done? Purchased full kits that they know will work well? Resigned themselves to replacing half the printed parts with different ones when the default ones don't work out?

actual theres a new branch prusa i3 vanilla has the single aluminium frame complete they should be ok to use, i have 48 mm motors on mine and they was about 2 mm of the floor so i sunk the frame slightly and fixing it with m10 and washes looks sleek whole point of design really i have some default items on my prusa i3 build but once i start printing they will be updated as theres fault the the print not the parts check my blog the blue parts or originals but the rest or custom due to print error

---

Check my rubbish blog for my prusa i3

up and running
[3dimetech.blogspot.co.uk]

Hi,

I have built several of the i3 and the single frame without any bracing works very well.
The M10 rods on the Y-frame is a must compared to M8.
I use bowden setup and can only say I am very impressed with the print quality from this machine.
I have also changed from belts to braid and it workds superbly.

I sell frames in Adelaide and one of my mods to the frame is to be able to attach a brace, but I have not had the need for this at all.

Regards
Jan

I've built a few single plate i3 s and have sold several kits using the parts with the original design. The new vanilla branch is a consolidated branch that is the same basic printer with some things simplified. If you use an aluminum single plate you won't have to brace it. You may have to on a wood or acrylic but on 1/4" or 6 mm aluminum you won't have to.

If you have any questions about specific part of the build post them and there are several of us that can help.

Thanks guys. Looks like going with a braced MDF frame might not have been the right way to start. I'll get an aluminum frame. Anyone know where one can be acquired relatively inexpensively? My initial attraction to wood and acrylic was for cost reasons; $80 and up for an aluminum frame seems punitive. I mean, for that much, I could get enough aluminum extrusions to build a MendelMax instead!

As for the vanilla branch, it's cool to know it exists, but as of commit 1322723d6107255d6c2a7552aebfdafe5a24c7ba (17 days ago) I'm not seeing a ton of differences between that and the master branch and most of the issues I've encountered haven't been resolved. Maybe I'll offer patches for the issues.

Are most folks using T5 or T2.5 belts? What are you doing about tensioning with the standard parts?

The plates cost what they do because that's what it costs to get the material and have it cut in small batches. Depending on where you buy retail on an 18" sq 1/4" 6160 T6 can be between US$60-80. Of course it's less per piece in full sheets but it's not something that easily lends itself to small batches. If you can get water jet cut parts in one offs for under US$100 you're doing pretty good. There are shops out there that will do it but most have setup charges and minimum run times on the machine. And many won't do one off work. If you can find a one off shop you could spend between $100-150 to have a single frame/bed assembly cut. That's what is attractive about acrylic and wood, it's less expensive to work with and the material cost is low compared to AL. I had to commit to 100 to bring the pricing to where I could retail the raw blank for $75. You can have it cut using EDM but by the time you drill the holes you're up on labor and handling and really not saving much, if any in the long run.

Dibond is another good material but I can't find any small CO2 laser shops in the area to cut it (machines aren't powerful enough, even Pololu has passed). I have to go to an industrial laser shop and around here that's not much less than a water jet house. That's going to vary by area as well as the price of the Dibond. In small batch Dibond is more attractive but as the quantities reach wholesale AL becomes much more competitive and available.

I use T2.5 belts and pulleys for my builds. I see Prusa is now specing HDT3 belts and pulleys and I'll likely test those to see what the difference might be. A big difference in the vanilla branch vs the original branch is that the .stls are already generated. In the original branch you can still spec sizes on some of the components. For example in vanilla the belt guides are set to use 624zz bearings whereas in the original in the config file you can define them as 608zz, 623zz and 624zz. Not having to edit a config file and make the output files is easier for those without a programming/hacking background but it limits the flexibility in generating .stls and makes it more difficult to take source and modify it for your own use. For example taking the .scad files and editing them with your own mods.

That all makes sense. Has anyone looked into like 9mm thick acrylic? That stuff is super hard and oughtn't bend much. I might try using Ponoko to get a Prusa frame and carriage cut from that.

And heh, scads are the least of my problems. I'm pretty fluent in OpenSCAD by now. I just want all those nice luxury features that my MendelMaxes have spoiled me over!

Getting acrylic cut is pretty easy and the material is readily available. An initial concern would be distortion from flat on the material either from the stock or cutting process. You can't thread it so you'll need to bolt through but that's not a big deal. Some that buy the plates would rather bolt through and I've threaded and bolted through on the testing mule build I have. The holes need to be about 3.5 mm for a straight bolt through. Prusa's .dxf has 3 mm holes (I reduced them to 2.5 mm so it could get a proper M3-.5 tap) so you could have those cut and clean them up from that.

Since you're used to aluminum t-slots with MendelMaxes, you may be interested in this Prusa i3 mod using a minimalist t-slot frame by Stephanie Shaltes (beware the design is not finalized). [github.com]

If you're on G+ I suggest you look her up, she's detailed some of her work there. [plus.google.com]

I'm much more confident in the frame rigidity of this t-slot frame than on a single unbraced aluminum plate. It's also a lot easier and cheaper to source. There are only 9 t-slot pieces compared to the 14 for an old MendelMax 1.0-1.5 or 11 for a MM 2.0.

I've been very tempted to convert my AO-100 to this t-slot i3 design. I would like to have the front of my printer completely open and easily accessible like the Mendel90, i3 and MM 2.0.

If I may ask, what do you think is wrong with the X-ends and X-carriage? I like how the belt goes through the X-ends.

Using the i3 single plate @ 1/4" and a similar design using 25 mm sq extrusion the rigidity of the vertical frame is similar. Both under a load more than would be subjected during operation can have movement in the Y axis but not due to the material of the structure. It's due to the anchor points at the footprint of the base with the fastening point being the 10 mm rod and the pair of bolts on each rod securing the plate. When the structure is boxed or supported as on the AO-100 there is quite a bit more rigidity. That's where the benefit of the extruded parts frames using a box base have an advantage partcularly when the base is boxed using the outside points so the footprint is the same width as the plate/gantry frame.

I'd say that it's easier to source extruded parts as they are more readily available as they are used in other assemblies which creates an economy of scale. Were there the demand to be able to order 1000 or more at a time the price of the single plates would fall below the extrusion at retail. Once one starts getting into a higher price point unit with a fully boxed frame (for example the MendalMax printers) that's when the advantages of the extrusion really come into play. Though the plates are more difficult to source and at this point have a price premium in the overall scope of the build I think it's still easier for the average inexperienced user to get enough parts to get a completed build. The drawback of the Shaltes build for a new user is having to scale and print the parts. I do think that's a good design concept and with some refinement and a larger community effort behind the project it could potentially be a go to design for a first time build.

Quote
vegasloki
The drawback of the Shaltes build for a new user is having to scale and print the parts.

What do you mean by scaling the parts? As for printing you need to print the original i3 parts anyway so I don't see the drawback here. Or did you mean that one can easily get an i3 printed part kit?

The build is scaled when the conf file variables are defined and .stl files are generated. By using the parametric model the parts are scaled to the sizes and config needed for how that particular build is defined. For those without a *nix software background there is likely going to be a learning curve with installing make and forking the repo to be able to generate the .stl files. Obviously parts need to be printed and new users likely don't have access to a printer or many (if not most times) don't know anyone with a printer. It wasn't too many months ago that i3 parts weren't available and when more people started offering them i3 builds started to outpace i2 builds. Another issue is documentation as well as fragmentation of the existing builds. The docs are getting better though the variants are causing confusion in many new builders. That's one area where commercial kit dealers can add value and another would be the consolidation of the software toolchain. Those two parts are what I see causing the most issues with mechanical ability of the builder being the third most common issue. That delta will widen as the users move from maker/hacker types to general users. At that point I think many will find value in pre built printers even if there is a few hundred dollar premium for the service/product.

Prusa's new files are available in .stl but I don't know if that is to make things easier or to make it more difficult for people to modify the parts.

Honestly, providing design files in OpenSCAD format is already making it more difficult for people to modify the parts. I mean, WTF! was my reaction when I learned of that software a few years ago, and I had a second WTF! moment when I realized so many people use it in the RepRap community.

I am not a programmer but an experienced CAD user and to me designing in OpenSCAD is unfathomable.

Just for fun I just started remodeling the vanilla i3 in FreeCAD because I like to actually be able to see an assembly and take measurements from it if needed. In that regard OpenSCAD is useless.

Anyway I don't want to hijack this topic more than I have already...

If you want STLs, it's easy enough to just compile the .scad files, export your STLs, and forget all about OpenSCAD.

Is FreeCad usable at this point? I tried it about a year ago but found it extremely buggy. If things have improved since then, I and probably many others would be happy to make the switch.

NormandC Wrote:
-------------------------------------------------------

> I am not a programmer but an experienced CAD user
> and to me designing in OpenSCAD is unfathomable.


Thank you for proving my point...

I use Autocad and thought the same when I first saw openscad. Now that I can find my way around it I'd rather use it to tweak designs either editing them directly or using whatever config file is provided.

Quote
Pointedstick
If you want STLs, it's easy enough to just compile the .scad files, export your STLs, and forget all about OpenSCAD.

STL are not what I'm after, but files usable in CAD software (which STLs aren't). As I said I import .scad files in FreeCAD instead, at least when it works I get solid primitives I can get some data on.

Quote
Pointedstick
Is FreeCad usable at this point? I tried it about a year ago but found it extremely buggy.

It's still somewhat buggy but it hardly crashes on me anymore. Maybe I know enough to instinctively steer clear of what makes it crash?

I've been using it a lot. FYI seven months ago I converted the Lulzbot TK-0 (now TAZ) Sketchup files to FreeCAD. I started an assembly file but have yet to complete it. Since there is no relationship-based assembly module in FreeCAD yet (it is being worked on), the parts were positioned manually, like you would in an AutoCAD document.

My files are on Github, but they're long outdated now. Since then the Lulzbot people have changed things up but kept with FreeCAD as one of their design tools.

Here's the result of a few hours of work. The assembly is not complete. After adjusting the settings the conversion of the Prusa3-vanilla scad files into solid parts in FreeCAD went surprisingly well. It generates models constructed out of primitives (CSG) that can be edited in the GUI.

I think I'll try to put up a tutorial on how to convert scad files in FreeCAD, some people might be interested (I can't be the only one who can't work with OpenSCAD!  ) but I don't know where I'll post it...



I wasted a lot of time on that alu plate as I wanted it to be parametric! Cutting out the PRUSA text just for fun.

hi yes I have also enlarged my reprap prusa i3 ,but I am affrayed of changing just my (y) axis do you no how to do this without changing the hole program ,I can open the arduino and marlin firmware but looking at it you have to go though the hole lot and update everything !! is there a way of looking at the program on the printer and just changing that just the y axis .can anyone help please