www.recumbents.com - Shell Design Software
www.recumbents.com
www.recumbents.com
Home | Profile | Register | Active Topics | Members | Search | FAQ
Username:
Password:
Save Password
 All Forums
 www.recumbents.com
 Technical / Bike Building
 Shell Design Software
 New Topic  Reply to Topic
 Printer Friendly
Previous Page | Next Page
Author Previous Topic Topic Next Topic
Page: of 23

shooky56
recumbent guru

USA
545 Posts

Posted - 12/31/2014 :  13:44:06  Show Profile  Reply with Quote
Those numbers have got to be too small but still relative to each other. When I figure out what my mistake is I'll correct them. 113 square inches is 10x11, that's impossible.
Go to Top of Page

Speedbiker
human power expert

USA
3760 Posts

Posted - 12/31/2014 :  13:54:08  Show Profile  Reply with Quote
Shooky is gaining speed fast....

I love that old "same drag as a truck mirror" gag. Of course they are comparing their vehicle drag to theoretical flat plate drag. Ha ha.
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 12/31/2014 :  13:59:48  Show Profile  Reply with Quote
Found my error, had remembered a number in my head. Looked up another number in the interim then typed the 2nd number in for the conversion from Pixels-squared to inches squared. The corrected numbers are:

Varna: 570 i2
Velox: 799 i2
Wedge: 638 i2

Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 12/31/2014 :  14:01:28  Show Profile  Reply with Quote
Thom: I suspect that is surely true, possibly one of those big tall things too. Believe they are now out of business as their website appears to be unpaid this year. Was a grand idea for an EV though, almost got talked into some investment money for that (whew).
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 12/31/2014 :  14:17:29  Show Profile  Reply with Quote
One of the things I got from Tony's build was the adjusted Q-factor, something that hadn't remotely crossed my mind. Of course, that was all hard-coded so I gotta fix that too. It is important because that definitely changes the foot box needed.

Had just started fixing the crank-foot hookup when I saw what Tony was doing and went "Doh!".

Go to Top of Page

alevand
human power expert

USA
2936 Posts

Posted - 12/31/2014 :  15:47:42  Show Profile  Visit alevand's Homepage  Reply with Quote
One reason for things sticking out is the flow near the body is much faster that the free stream, and could be 2 times fater, so putting something that sticks out into this flow is bad.



C:
Tony Levand
Go to Top of Page

Upright Mike
human power expert

USA
3803 Posts

Posted - 12/31/2014 :  18:46:42  Show Profile  Reply with Quote
Shooky, one more for your collection.... my Norus streamliner...
http://i47.photobucket.com/albums/f152/uprightmike/streamliner%20scaling/Norusdistancecompressed.jpg

Edited by - Upright Mike on 12/31/2014 19:00:51
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 12/31/2014 :  20:59:57  Show Profile  Reply with Quote
Happy New Year all. You know you are old when you're back home before midnight!

Tony absolutely!

Mike:

http://s1372.photobucket.com/user/shooky56/media/HumanPowerUprightMikesNoRush-Trimmed-Analyzed_zps246a32b1.png.html br / " border="0">
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 12/31/2014 :  21:05:25  Show Profile  Reply with Quote
Mike try this: For some reason the img tags aren't working for me.

http://s1372.photobucket.com/user/shooky56/media/HumanPowerUprightMikesNoRush-Trimmed-Analyzed_zps246a32b1.png.html
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 12/31/2014 :  21:08:00  Show Profile  Reply with Quote
There we go:

[URL=http://s1372.photobucket.com/user/shooky56/media/HumanPowerUprightMikesNoRush-Trimmed-Analyzed_zps246a32b1.png.html][/URL]
Go to Top of Page

Upright Mike
human power expert

USA
3803 Posts

Posted - 12/31/2014 :  22:46:15  Show Profile  Reply with Quote
Thanks Shooky! That frontal view must have involved some work, since my original photo had the bike leaning against the fence with the board all messed up too. I browsed your other picts and came across a funky one that appears to be me on an upright bike! Thanks for modeling me. If it helps I drag out my duplicate board and get some updated (better) photos, though its about 20 deg outside and I don't think I want to be posing in spandex right now!
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/01/2015 :  04:28:26  Show Profile  Reply with Quote
Oh you're welcome for sure Mike. That was just a model pic to show the upright torso example from something you'd ask for earlier. I've played around and can make a much prettier upright shell than in that gallery. It was just a "yes you can make the rider upright" experiment.

Lol no no no I don't want you going out there in this weather either!

You could help me get a warm fuzzy for the accuracy of those parallax corrections though. Didn't do the dual (width different from height) corrections again as I think my current estimates are so bad as to not matter much. If you could, or if you already know, would like to see the true maximum width and height of the fattest part of the main body. Kind of wonder how close I was. I can tell it to adjust so it sees the width and height properly and then the area report should be extremely close. You probably know how to get the measurements but if you are scratching your head, here's how I'd do it. Slide it against a wall with a flat floor and slide something square (a cardboard box for example) up against its fattest part and measure the gap. Height is the highest part minus the ground clearance of the belly

Work: Not nearly so much as you'd think. What I program in (Visual Studio) has libraries that easily import, rotate and display a picture so I could have done it there. But the biggest problem was getting rid of the unwanted pixels. It would be hard to do programmatically w/o a hard line on the model. Since I had to draw that line anyway, I used Paint.Net. Took 15-20 minutes to outline the body and remove all pixels except the body. Paint.Net uses the right click->rotate method (this is why I use the same interface) and rotation took a few seconds. Took about 10 minutes to figure out the pixel scale (your picture was slightly different than the Battle Mountain shots). Your resolution is about 62 pixels per sq inch, the Battle Mountain shots were around 35.
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/01/2015 :  07:08:36  Show Profile  Reply with Quote
Early on, I took a look at this:

http://www.makehuman.org/content/download.html

After messing around with a body designer, I've decided the above software is our best option. I can already import the

Believe the "Download Makehuman" button is safe (it links to their site not "www.LetMeStealYourIdentity.com").

I'll leave it to you guys to figure out their software. It isn't too bad to learn. Chuckled when I used it the first time at the "gender slider". My mind jumped back to an old SNL skit about "Pat" where we were never sure of their gender.

When you are happy that you have dimensionally modeled yourself correctly, go to the [Files] tab.

On the files page, click the [Export] tab and check the Ogre3D radio button. You can export your rider to an xml file. That format is the one that works with my reader.
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/01/2015 :  07:21:41  Show Profile  Reply with Quote
Oh, couple things. There is texture stuff too, basically the coloring of things like the eyes or different shading on the body. I can't yet import texture information and maybe never will (it's more eye candy than functional) but the textures don't hurt, I just ignore them. It's a long story and I won't bore you by repeating it (have posted some of those problems before).

What I'll do on my end is have a rider file that points to the mesh file you create with MakeHuman. The rider file will also import the starting position of your rider which you'll do on my designer. I'll give you some small adjustment sliders for seat slide and tilt.

So much to do and only 8 more lives left too...

Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/01/2015 :  10:27:48  Show Profile  Reply with Quote
The frame designer: The frame you see took me less than 5 minutes but it was probably the 50th time I'd built one so was getting pretty good at using my own software. The most tedious part will be sizing everything correctly for your frame, wheel spacing, crank position, etc. I actually use rectangular carbon fiber product for mine 4.5 x 1.5 but using three 1.5 tubes would fill the gap. Really all we need for this project is something to give us a visual of what might poke out of a shell.

Here's what the frame designer looks like, moving on to the rider position designer.

Top View:

[URL=http://s1372.photobucket.com/user/shooky56/media/Cycling/FrameDesigner4_zps844b50d3.png.html][/URL]

Side View:

[URL=http://s1372.photobucket.com/user/shooky56/media/Cycling/FrameDesigner3_zps8149db69.png.html][/URL]

The beginnings of the body, first order was to import into my software the export from MakeHuman.

[URL=http://s1372.photobucket.com/user/shooky56/media/Cproject13_zps42da96bd.png.html][/URL]
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/03/2015 :  09:11:01  Show Profile  Reply with Quote
Ready Tony's build, it became apparent how concerned everyone was about Q-factor and narrowing the pedal box (at least I suppose this is the aim of the discussion).

Had not planned for that in my software and was doing some work to allow a sizable chainring and spread to the center of the pedals. While it is possible the chainring could get so large as to be the "pokiest" thing, that would be a rare configuration methinks.

This morning as I began to write a bunch of mesh creating code to resize the BB-chainring-crank length-crank spread-pedal type, it cross my mind that really all we probably care about is the spread of the foot and the length of the crank arm. Basically it's the FOOT we're worrying about poking out of the shell, it's just that the position of the foot during a revolution is based off all that other stuff. The animation of the crank is purely symbolic anyway so what I now think I'm going to do is let you position the spread of your feet in the rider designer and set the crank length in the frame designer. That'll give us the correct displacement of the foot in space while pedaling.



Go to Top of Page

warren
human power expert

USA
6118 Posts

Posted - 01/03/2015 :  11:01:43  Show Profile  Visit warren's Homepage  Reply with Quote
That should do it Wayne!
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/03/2015 :  12:11:18  Show Profile  Reply with Quote
Thanks Warren, while reading all the tech posts, I'm amazed at the general knowledge of several of you guys. That's a good warm fuzzy if you think it'll work. Glad you replied it's a complete shift of where I was headed so was waiting to see what you guys thought. Now that I have the green light, "On it boss!".


Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/05/2015 :  17:24:13  Show Profile  Reply with Quote
In the immortal words of Roseanne Roseannadanna, "It just goes to show ya, it's always something.".

A bit of ugly background (for a non-programming forum at least), all the 3D stuff is done with triangles. The body in the first release was just a bunch of big pipes that I created by typing some information (math stuff) in a file, reading it in and using that to create some pipes for legs, feet and a torso. A much finer detailed human body has been created by a team of open source programmers with the MakeHuman project and I decided that it is just as important to accurately represent the rider as any other part of this before going through the massive work required to fabricate a shell from scratch.

That is the reason I'm putting a good effort into the rider designer.

A few days ago, I thought I'd let you guys use the MakeHuman open source project to create your own bodies and I'd import them. I could already import their meshes but in the last few days I made a "bone importer". What I have discovered is they have screwed up the export so the bones are not properly defined. That leaves me with a problem importing. If their export is errant, the import will be too.

A bone, in 3D modeling, is an unbendable piece, not really the bone itself. In our case a thigh "bone" is the entire thigh. I have to manually determine which of the 10,000 triangle definitions belong to the foot, ankle, thigh, etc. The problem is, all that work only pertains the the human mesh I created. Everyone else's would be different.

That's good news and bad news. The good news is you guys will have a one-stop shop to build your frame, rider and shell in one spot. The bad is I've got to give you a rider designer rather than use the MakeHuman application.

My plans right now are to ship my body mesh with the software and allow you to drag the individual "bones" (meaning one non-bending piece of the body) to the size and shape you need.

What every person needs to know to do this is precise measurements of the body. Basically the foot size will be the size of your foot in your cycling shoes. You'll need be able to precisely locate the distance from your heel to the center of the pedal bolt. You'll need the same for heel to ankle, knee joint and hip joint as well as quite a few other measurements. I'll get around to making some specific instructions to do all these but those measurements should plop your rider in spatial relationship to the rest of the recumbent frame and allows a close fit shell with a good confidence level.

I'll also make some approximations. The human back is basically a large curly flexible thing but I'll probably chop it up into only two or three pieces. The recumbent rider's back isn't particularly odd shaped but the upright guys will need the flexibility and head tilt to get a good picture.



Go to Top of Page

Speedy
recumbent guru

USA
883 Posts

Posted - 01/05/2015 :  19:48:44  Show Profile  Reply with Quote
Everyone should have a "personal challenge project"
Creating modeling software is a heck of a challenge.
"Good on ya mate" for giving it a try.
Seems like a long way around the mountain to arrive at the desired result ... a human powered vehicle.

Makehuman works fine in Blender.
Blender is a very good opensource modeling software. A complete streamliner could be modeled in Blender. Used properly, it can export a useful .stl file.
Perhaps, you don't know, there are several free, workable CAD model software packages such as TinkerCAD, 123D, SketchUp.

For me, the reason to learn CAD modeling is to make parts as quickly as possible.
Thing is ... every streamliner I ever built, as soon as it was ride-able I found things that needed to change. Going straight to a machining environment is quite useful.
To that end having a CNC router has been a great time saver. The one I built measures 4' X 4' with a 6" Z.
If ... I built a CNC today for the task of machine soft materials I would focus on the current growth of DIY 3D printers and opensource motion control software. Making a CNC could be low cost. TinyG or Smoothieboard for the motion driver with NEMA 23 motors would keep the electronics under $500.
Use square tube and bearings for the slide rails.
CAM software is also opensource. tgFX and RepRap are both fine with RepRap a bit more robust and user definable.
Build it upside down ... that is, mount the rails to the rafters in the shop so the floor space stays open when the machine is not in use.
In a practical sense 1mm tolerance for a complete streamliner is precision enough. Machining to .001" tolerance over a complete vehicle would be awesome.

Computer modeling is plenty of fun. Pedaling the machine is the most fun.

Steve Delaire

http://molten3d.blogspot.com
Go to Top of Page

Speedy
recumbent guru

USA
883 Posts

Posted - 01/06/2015 :  09:22:29  Show Profile  Reply with Quote
FreeCAD is another opensource modeling software that you may find useful. The user interface is closer to professional programs like SolidWorks or Inventor. It's possible your work could be added to their library as an add-on.

Steve Delaire

http://molten3d.blogspot.com
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/06/2015 :  22:19:00  Show Profile  Reply with Quote
Thanks Steve. Yes all that stuff is good and thank you for your support.

Happen to own a copy of Aspire which also models 3D and provides pathing for CNC bits. Forget now but seems like it was about $3000-ish.

The problem was the airfoil shape. Wanted something to make the shape and figured the NASA symmetrical airfoil formulas were a good start. Was essentially "done" at about the 10 day mark (except for output which may take a few days, probably a few hours for the first version).

The rest of the effort has been for the community. Although this forum seems to abound with intelligence, there's probably a lot of people coming in here that would find the multiple sources needed to model the frame into a mesh plus the exports/imports needed plus the kinematic tools (I know of no kinematic modeler other than those for programmers) to made the rider to see if he pokes out of the shell. So the rider thing kind of makes my modeler "a thing that doesn't exists anywhere else". You could hand craft the airfoil in a cad tool but mine makes it a perfect airfoil when viewed tangent to the surface at any side angle (up to the point of maximum thickness anyway). That would be a real trick in a cad program.

As far as the CNC machine, I have two 16 foot long aluminum I-beams to mount in my 11 foot garage ceiling... they are lying on the floor, oh actually no I put them on a high shelf made just for them. They were to be spanned crossways by two more that made a box for a dolly to traverse their length with a second dolly to traverse their span. A standard 3600rpm was to rotate a 4 foot x 1" shaft with a 1.5" ball router bit (about 2 feet of clear shaft + 2 feet of z). The rest of this was just a guess. Despite large I-beams and heavy duty everything that's a lot of space for flex and vibration. At those sizes, aluminums coefficient of thermal expansion would also be a factor. I just figured let it move very slowly when you finally near the finish cut and it'd be ok. We ARE talking about foam after all and the idea was to save the roughing in time.

Money was not really an object so long as it was $10K or less total (and it would be). Basically my idea was to carve the urethane foam block into a shape within about 1/4" precision and let a person finish smoothing it. There wouldn't be any absolute shape checks (like ribs are in a ribbed plug) but I thought (could be completely backwards about this) that most guys just wanted a decent shell, not a world record. In such cases "pretty close" is good enough.

BTW I have all the controllers, software and hardware, save the wiring and miscellaneous needed to build it just got sidetracked (I was working 6-7 days a week and it didn't take much) and never got back to it.

Also Steve, I plan to output CNC pathing. You pick the diameter of a straight flute bit and it'll dump the CNC codes.

As far as the effort goes, my posts are to keep you up to date if anyone is interested. Am enjoying the challenge as I happen to love coding anyway.
Go to Top of Page

Grant-53
recumbent guru

USA
544 Posts

Posted - 01/07/2015 :  09:00:08  Show Profile  Reply with Quote
The curves could be approximated as a series of arcs in a CNC program. There should be a number of prototype firms or job shops that could do routing jobs for a fee especially if the code is provided.

As for building your own rig I'd focus on cross bracing the frame and then the means of securing the work pieces. The machine I ran used a vacuum table to hold plywood sheets.
Go to Top of Page

shooky56
recumbent guru

USA
545 Posts

Posted - 01/07/2015 :  09:29:58  Show Profile  Reply with Quote
Steve:

Reports for updates: What I meant was the reports are progress reports, perhaps they sounded like complaining. Other than "sorry for more delay", they are not at complaints. I feed on the challenge.

BTW, I still have Blender installed. It was something I pulled down fairly early to look at on the advice of my nephew who is an outstanding coder and also a fairly rare coder type that likes to try new stuff. He'd used both Blender and MakeHuman and suggested the I might hook them up (same suggestion you made). After I played around a while, we discussed this on our Ventrillo server for an hour or so. He ended up at <may be paraphrasing a bit because I don't remember word for word>, "I don't know how you could possibly do that with any tools I've used or know about, maybe more like the graphics art studio software but that would be very expensive.".

If Blender could automatically output an animation based on crank arm, Q-factor, ankle joint to cleat distance, seat position, rider size, etc. it would be exactly the ticket for the automation piece of this as I could "sprite" the animation frames based on time.

For those who understand the term sprite in graphics programming I quoted it because it wouldn't be a sprite (which is like the cartoon animations we made as kids as we flipped sequential pages of slightly moved stick men). I'd import the meshes that corresponded to the new "posed human". Each pose would be a separate Blender export. To get 10 degrees of crank rotation precision would require 36 separate poses and exports and this would be required by every user of the shell designer. Doesn't seem like a good option.

We (my nephew and I) also took a look at kinematic modeling but those libraries are pretty complex (they will model all motion of a human). It's a lot of setup for the restricted motion needed for this software.

I have to do a small subset of what the kinematic libraries do (believe the proper term for what I do is called "reverse kinematics" as I move something and force the model to assume the shape). Fortunately the only modeling needed is the ankle, chin and thigh. It took me ..mmm. three days I think to get that math working right and will require a bit of a rework when I get the new rider designer finished. For those who noticed that sliding the seat back "blanked out the legs", the reason is I did not enforce "max reach rules" that would preclude the seat being too far to reach.

As far as a long way around the mountain. Yep I had planned to make a Georgi shell (just sculpt one) as I didn't particularly care about it being airfoil shaped but it got cold before I got to that part. Happen to despise cold so I take "Winters off". That left me with several months of "what do I do now". Coupling that with the fact that any shell is a major undertaking, I figured I might as well have a good performing one after all that work.

That got it started. My plans were to sit on my recumbent and mathematically describe the foot box and rider box for the software to ensure the shell was big enough for me to fit.

Knew about this site and had occasionally looked at this or that over the years and knew about Warren (probably live within 250 miles of each other). So many people on this site have given so much of their free time (that's what impressed me most). The community seemed quite close knit and that was about all the excuse I needed to launch myself into a more robust application. Admittedly it grew a bit at the requests of others but mostly at my "sheesh they're not gonna wanna do this or that" judgments about some phase of the design work.

Again I'd like to reiterate that the frame designer, in particular, isn't intended to be a designer for building. My offer to output a full sized printout was to make sure that what you put in my modeler fit the frame you'd actually built as a warm fuzzy before you started to build a shell described by this.

If anybody ever once posts or emails me that they used this software to build a shell and it fit and worked and they were happy with the result, every minute of this effort will have been worth it!


Go to Top of Page

Speedy
recumbent guru

USA
883 Posts

Posted - 01/08/2015 :  18:15:00  Show Profile  Reply with Quote


A number of CAD modeling software packages have the tools to create motion to the various parts of the model, human or otherwise. This can be used to check for modeling conflicts.
Blender can be used for this purpose. Personally, I find Blender hard to make engineering models with. It does art models, creature models, gaming and motion all quite well. It is above average useful for altering pretty much any .stl file.
FreeCAD I've only used as a modeler software a few times so have no idea of it's full potential.
It has a good .stl 3D model checker and repair tool for the 3D Printing process.
Being an opensource program it's possible your work here could be an asset there. Just a thought.
There are other work arounds to check clearances without motion.


Steve Delaire

http://molten3d.blogspot.com
Go to Top of Page
Page: of 23 Previous Topic Topic Next Topic  
Previous Page | Next Page
 New Topic  Reply to Topic
 Printer Friendly
Jump To:
www.recumbents.com © 2017 www.recumbents.com Go To Top Of Page
Snitz Forums 2000