Re: Do CoreXY belts need to be of equal tension?
May 03, 2016 08:33AM
ikcl,

When designing in CAD, I start a folder for a project. There's usually one file into which final or near final components get assembled into the overall design, and multiple files for individual component designs. When a part design is finalized, I stop messing with it. When it comes to big things like how to lay out linear guides for a CoreXY machine I am currently designing, I draw one layout, then make a copy of it in the same file and modify the copy for an alternative and compare the two. The winner stays and the loser gets deleted. The same thing happens in the individual component files, so each file evolves as new ideas are tried and rejected. The process is slow enough as it is, and documenting every rejected change/idea would turn designing the machine from slow fun into a genuine PITA. Unlike software projects, no one is ever going to need to "unwind" my design so all that documentation would be a waste of effort.

Supporting "the community" is nice and I try to contribute where I can, but I don't let "the community" determine my design choices or the CAD software I use. My designs are for my own education and use and I don't worry about others being able to duplicate them exactly. I prefer to make the best printing machine I can using whatever tools make the design and construction easiest for me. I don't like to use printed plastic parts and I don't mind including designs that require a milling machine. I use a lot of surplus machine parts that are unlikely to be available to someone else in the same combinations. Someone would have to be pretty determined and have pretty deep pockets to duplicate my designs. There's plenty that can be learned from looking at photos or CAD images (that's where I get some of my best ideas, thanks to "the community").

In the course of designing a machine I sometimes model "popular" components to check fit, clearances, etc. I post those models to Youmagine.com. I just posted a model of the E3D Titan extruder last week, for example, so others don't have to buy it first to determine if it's going to fit their machine.

I use DesignSpark Mechanical because it is very fast and very easy. I don't care if the CAD files aren't compatible with more "professional" or "open source" CAD programs that some people in "the community" prefer. If someone wants to look at my work they are welcome to install DSM and take a look (it's free). Otherwise, the images captured from CAD and the photos will have to serve as inspiration.

Edited 1 time(s). Last edit at 05/03/2016 10:25AM by the_digital_dentist.


Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]
Re: Do CoreXY belts need to be of equal tension?
May 03, 2016 12:49PM
Quote
realthor
I can't write python for my CAD smiling smiley .
I am using a proprietary (unfortunately) CAD and I have thousands of objects in the file ...can't easily port it over to OnShape or OpenSCad ... I have no time for that and I barely know this software (DesignSparkMechanical).

yyep... exactly - it's from years of experience working with software that i made a deliberate and conscious decision to track down "programmatic" CAD development tools.... you're now basically stuck with what you've got, and it's quite likely to be a good first design learning experience.

_my_ first major bit of 3D CAD design was about 5 years ago, with mm3d, where i basically did over 5,000 points *entirely* by hand with a mouse. at some point i woke up from a kind of.... fugue-like dream, weeks had passed and i went "i must be nuts!!! aaaaagh! i am *never* doing this again!" - and that's why i thought about how to do this in a much better way: it was when i tracked down openscad, then pyopenscad, then developed pyopenscadspline as a tool to drastically reduce the number of points.needed, and to be able to manage the entire CAD using a git repository.

it's a totally different technique that helps me keep my sanity.

Quote

I usually just delete whatever objects I notice are bad design or have some fault or need replacing, otherwise I modify that object. Not many options to revert back a change

well that's pretty s**t if you're used to revision control, isn't it? but if you've never _used_ fine-grained revision control then i would imagine that, not knowing how much easier that is, you'd never get frustrated at the necessity of chucking out entire swathes of work... because you'd consider it "normal"... me i'd just go f*****g nuts at the time and effort wasted compared to the application of software development techniques.

Quote

but on the other hand this software is stellar at fast prototyping, it's very easy to redo something, if not very complicated. I haven't had many issues with it so far but I'd like to switch to something open source in the near future ... I have my eyes on Antimony for some time but haven't had the chance to dive into it.

you're talking here to a very unusual engineer, who is a cross-over between the hardware and software worlds. the only proprietary programs that i use for any kind of CAD work is for PCB design, because software libre tools right now are basically not up to the job: no design rules, no checks, no verification.... gaah. pttoey.

whereas 3D naturally lends itself to using software to generate the 3D objects, and it's only when you get to "Boeing 747" style product development, where you *need* to document exactly what kind of nut and bolt fitted into Revision 374 of the part that goes in the toilet (i mean that literally), that you really really need some of the horrendous mind-bending excruciatingly-complex proprietary 3D CAD software available today.

... but for everything else: find something you like and can tolerate, and go with it. honestly if it needs a mouse, they're all the same as far as i'm concerned: they're all something i'd personally never use, so am not the best person to advise you. i have tried blender btw - it worked but god was it hard work.
Re: Do CoreXY belts need to be of equal tension?
May 03, 2016 01:01PM
I would only think of openscad or the programmatic type for the final job of releasing the design in the open. For everything else a visual CAD with rapid prototyping features, where I can iterate in 10 minutes 10 different designs is the best I can get. This is why i am calling DSM stellar. Otherwise I hate it for several other reasons (no TEXT tool?, no PART MIRROR -only sketchin interactive mirroring-, ...really!!!).

So far I feel I am about 90% into my 3D printer design ... I have to build it, print the parts, test them, then release everything. If/when it works. Not yet, so besides a wiki for the presentation not much else. 3/4 of the designs I have put in the forums don't exist anymore smiling smiley ... so no point in uploading them to a wiki, besides what you call a source for others to see where I went wrong etc ... but basically i would call that pointless information overload. People barely look to complete/proven designs...


RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff
Re: Do CoreXY belts need to be of equal tension?
May 03, 2016 05:07PM
Quote
the_digital_dentist
ikcl,

When designing in CAD, I start a folder for a project. There's usually one file into which final or near final components get assembled into the overall design, and multiple files for individual component designs. When a part design is finalized, I stop messing with it. When it comes to big things like how to lay out linear guides for a CoreXY machine I am currently designing, I draw one layout, then make a copy of it in the same file and modify the copy for an alternative and compare the two. The winner stays and the loser gets deleted.

i actually really like that idea - to have two versions side-by-side.

i guess here the key difference is in the style of the machine. the number of parts that i'm designing which are critically interdependent on others around them is actually really high, because i'm making a folding, stowable *compact* 3d printer. ordinarily you would make a single part (an idler for example), it would sit in a corner, on some metal or wood, and err.... that's it.

whereas what i'm doing is pushing the boundaries of available space. i'm down to 312mm width on a 214mm printbed. in the CAD, after designing the z-assembly, corexy and now putting in triple lead screws, i tried the "fold-down" position and f*** me i f*****g well got the f*****g front lead screw hinges wrong and they're banging into the bottom z mounts. but i couldn't have known that until i had actually designed 90% of the parts.

so this is where a constant series of iterations is *required*. i *can't* design a single piece and expect it to be perfect first time, because:

  • it is hinged so has to be designed in combination with the piece it connects to
  • it is a slot-fit into another piece which has to constrain x and y but be slotted in on z
  • it bangs accidentally into another piece which is to be designed in about 3-4 weeks time
  • when the hinge rotates it turns out that half-way along the rotation the piece catches on something totally unexpected

so basically what i am saying is: you're tackling simple projects where pieces are far enough apart so as not to have any interactions with other pieces, i'm tackling a stupidly-complex one where there are a huge number of inter-dependencies, such that i pretty much have to design everything iteratively.

Quote

Supporting "the community" is nice and I try to contribute where I can, but I don't let "the community" determine my design choices or the CAD software I use. My designs are for my own education and use and I don't worry about others being able to duplicate them exactly.

which is a really valid way to do it, and i think recycling of surplus machine parts is fantastic: my partner marie's father is a collector and recycler of surplus parts, and he is one of the few engineers that people go to in anchorage to get things like "aluminium engine block cracks" sorted out. most people would write off a cracked engine block.

Quote

I prefer to make the best printing machine I can using whatever tools make the design and construction easiest for me. I don't like to use printed plastic parts and I don't mind including designs that require a milling machine. I use a lot of surplus machine parts that are unlikely to be available to someone else in the same combinations. Someone would have to be pretty determined and have pretty deep pockets to duplicate my designs. There's plenty that can be learned from looking at photos or CAD images (that's where I get some of my best ideas, thanks to "the community").

so this is another area where our aims and mindsets differ: i've worked on high-profile software libre projects (i was the key reverse-engineer behind samba's NT Domains and before that the Network Neighbourhood (nmbd) for example), so it's natural for me to jump straight into "how can i make this for other people to replicate oh and i'd also like to sell some kits to people as well" mode.

so, realthor, that's what you need to decide, is, why are *you* designing your printer? make that a clear goal, and, if documenting your build is one of the goals, put that on the wiki page smiling smiley

Quote

In the course of designing a machine I sometimes model "popular" components to check fit, clearances, etc. I post those models to Youmagine.com. I just posted a model of the E3D Titan extruder last week, for example, so others don't have to buy it first to determine if it's going to fit their machine.

yeah that was really helpful and i've started linking to that forum post elsewhere, where i've seen people recommending the E3D Titan, so really appreciated there, dd.
Re: Do CoreXY belts need to be of equal tension?
May 04, 2016 05:18AM
http://reprap.org/wiki/SCOUTcorexy

something for a start ...

-how do I upload a picture for the "|image =" section in the upper right corner?
Edit: I managed somehow but didn't quite get the usual way to do it

- or how do I bulk upload pictures for the gallery ...

- why don't I have the green framing around the top-right description of the machine

is there a way to embed github folder for the gallery in the wiki? (so that it displays pics found in a github folder)

Edited 2 time(s). Last edit at 05/04/2016 06:34AM by realthor.


RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff
Re: Do CoreXY belts need to be of equal tension?
May 16, 2016 09:38AM
Quote
realthor
http://reprap.org/wiki/SCOUTcorexy

something for a start ...

-how do I upload a picture for the "|image =" section in the upper right corner?
Edit: I managed somehow but didn't quite get the usual way to do it

doesn't matter - it works smiling smiley

Quote

- or how do I bulk upload pictures for the gallery ...

it's wikimedia, there might be a script out there.

Quote

- why don't I have the green framing around the top-right description of the machine

no idea!

Quote

is there a way to embed github folder for the gallery in the wiki? (so that it displays pics found in a github folder)

no - just point people at the link, i set up a gallery on my server (photon, it's available as a debian package for example) then you can see i put "go here". basically wikimedia assumes that external links to images isn't safe, so you have to upload them. plus, external images could disappear, or contain a license that's *not* compatible with the wiki's terms and conditions and license agreement.

by uploading the images you *specifically* agreed to release them under a license that guarantees that they'll continue to be there. many people do not think about these things, and do not realise that they're important.
Sorry, only registered users may post in this forum.

Click here to login