Frame construction

[Jon Weir]

How do you guys make your frames? More specifically, what do you use to notch your tubes?

Since we use relatively small tube and very thin wall sections, I'm curious as to the method that people use to get a good notch.

Hole saw, abrasive grinding, end-mill setup, or otherwise?

[Jon Weir]

How do you guys make your frames? More specifically, what do you use to notch your tubes?

Since we use relatively small tube and very thin wall sections, I'm curious as to the method that people use to get a good notch.

Hole saw, abrasive grinding, end-mill setup, or otherwise?

[B Hise]

Jon,

We use several methods, depending on the diffculty of the notch. Some single angle notches and simple joints are easist to do on a setup( tube notcher or custom fixture, which I would recomend taking the time to make) with a hole saw/endmill or on the laythe. Ive always found using a vertical mill cumbersome. We use angle grinders for detailing/fixing notches as well as for tube ends that are going to complex joints. After a few notches on the fixture it can be tough to maintain precsion, so we just "cut and fit" with the angle grinder. Though it takes a few trips from the frame to the workbench its worth it. Hope this helps.

Bryan

[Patrick W. Crane]

this is definetly an art. Takes practice. If you are really super duper keaner, you could mate "up to suface" in you solidworks model (if that is what you are using) and bust out some end patterns.. but where is the fun in that?

[Big D]

Here's what we did:
We figure out the angles of intersection based on the solidworks model. The machine shop at our college helped us make a really sturdy clamping block, which holds any 1"O.D. tube, and mounts to the carriage on the lathe. The mill bit is held in the chuck, and the tube is fed into it. We have found this setup to work better than a hole saw, but the disadvantage is that if the mill bit is at all chipped or damaged, it likes to grab and fold under the top half of the cut. Unless you are the rammy, hurrying type, this isn't hard to avoid. It helps to feed in the direction of the bit's axis, as opposed to using the cross-slide to feed.

These tubes only need a quick touch up on the grinder, and/or hand file, but the angle will be bang-on.

Another concept that we have tossed around is taking a pile of tubing and some solidworks to a lasercutting business with 3D cutting capabilities, and cutting everything out so it fits together like a puzzle, with slots and tabs. This would reduce the amount of jig needed, but it might be prohibitively expensive, unless the work was donated.

[Agent4573]

Bench grinder.

[Cement Legs]

Ball nose end mill. Place your tube in the vise and the right angle... use the side of the end mill to get your first cut... bring the bottom of the ball nose section up to the level for the next cut and walk your piece through... Two cuts for the price of one. I've got it down to making one tube (ie a tube needing 2 different angles cut in each end) in about three minutes. The real beauty is that when you set them in place they all fit nearly perfectly! No messy grinding! Just a little hand filling to take down the rough edges and really thin material left from the low angle cuts.

[BStoney]

You'd be surprised how good you can get with detailing tubes with a 4" angle grinder and a 3/4" or 1" half round file....

Find a really anal guy/gal that is mechanically inclined on your team and hook them up with a pair of safety glasses, ear plugs, a vise and angle grinder and let them go to town!!

Take your time, as better fitting tubes are well worth it. However, we had a guy last year that could grind a tube in only a few minutes and it fit perfect just about everytime. Too bad he graduated....

B.J. Stoney

[rjwoods77]

One of my friends is working on a tube notcher for a senior design project that was my idea. I've seen it in full production equipment before and he is making a downsized unit. Not really my idea except to make it portable. It is a cnc tube notcher. Uses a turning chuck that welders use. A plasma cutter is mounted to a linear actuator. You enter in tubenotch data and the thing automatically cuts the profile perfect. He is working on the software end of things right now but it works real good for the simple nothces so far. He is designing it so you can use a cutter you already have. Should be nice when it is done. Put tube in. Hit go. Done.

[TA]

http://www.andkilde.com/tubemiter.zip

Is a small program that generates printable paper templates that you wrap around a tube to provide a cut line.

I think it was written by one of the Aussie FSAE guys a few years back.

Cheers, Ted

[Chris Clarke]

We model all the tube cuts in Pro/E, then unwrap the tube and create a flat state tube model. Create a scale drawing, and print off all the tubes wraps on a plotter. Cut out the wrap, and wrap around the tube. Paint the end of the tube so that the profile is on and use a grinder/files to remove all the painted material.

This is very time consuming but gives accurate results. We mainly due it this way due to lack of access to a machine shop, and the fact that our chassis has a lot of complex cuts. Needless to say, that after this chassis is done, I am going to build a tube notcher that should speed up the process dramatically.

We also model all our jigging, and print off full-size plots for those.

[J. Schmidt]

We've been using an old lathe to notch our tubes for years. Use an old lathe because the bearings will be shot afterwards. Put an end mill in the chuck and use the cross feed to feed the tube into the bit. If you jog the cradle back and fourth while feeding you get a smoother cut with less vibration and tool wear. Use the angle gage on compound rest to get your notch angle. Compound angles can be done with a scrap piece of tube and a magnetic angle gage on the free end of the tube. It's crude, but it works.

[Big D]

I agree with BStoney.... nothing beats a perfectionist and a hand file, but the automated ways are quicker, if you're into that sort of thing. http://fsae.com/groupee_common/emoticons/icon_wink.gif

[Cement Legs]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Big D:
I agree with BStoney.... nothing beats a perfectionist and a hand file, .... http://fsae.com/groupee_common/emoticons/icon_wink.gif <HR></BLOCKQUOTE>

I dunno... perfect fit (+/- .005"), 1/5 the time, even surfaces all around for welding. Ball-nose bit and mill http://fsae.com/groupee_common/emoticons/icon_wink.gif

[Kirk Feldkamp]

Hey Chris,

How do you unwrap the tube in ProE? Also, does anyone know if this can be done in Solidworks?

Kirk
UC Berkeley

[Denny Trimble]

Yes, it can be done in SolidWorks using the sheetmetal functionality.

However, for a one-off car, the effort required to split the frame members into individual parts, unwrap them, send each profile to production (paper cutouts or CNC Plasma/Laser cutter), index the lengths and orientations of each cut accurately, etc. seems excessive to me.

But, to each his own.

[BStoney]

Agreed. That would take awhile..but like Denny said, "...to each his own."

We used to use a fixture that could be used with a drill press and some hole saws, but it ended up being a major pain and it's just as easy to grind-to-fit.

[Chris Clarke]

Kirk,

I modelled all the tubes as surfaces, modelled the cuts in the chassis assembly, then copied the geometry to a sheetmetal part, and unwrapped it. Assuming that you already have the tube surfaces modelled, this is how you unwrap the tube in Pro/E.

Create a new sheetmetal part, (name tubename_sm.prt)

Open the part

To copy the geometry of the part from your chassis assembly

Feature

-create
-datasharing
-extcopygeom

open
-open your chassis assembly

default location

select surface refs
-pick the two sides of the tube you wish to copy

ok


To copy the datum planes for the tube

Feature

-create
-datasharing
-extcopygeom

open
-open your chassis assembly

default location

-MISC REFS
-DATUM PLANES
-SEL BY MENU
-CHASSIS ASM
-DATUMS
-SELECT tube reference planes (make sure you tube surfaces and reference planes are named for easy selection, I used numbers 005, 010, 015.., and for reference planes 005_a, 005_b, 005_d, 010_a, etc. This made selecting the planes by menu really quick throughout the modelling process)

Create the wall thickness

Feature
-create
-wall
-OFFSET...THEN CLICK 'DONE'
-CLICK THE TUBE
-OFFSET DISTANCE = 0
-MAKE SURE ARROW IS POINTING TOWARDS THE INSIDE OF THE TUBE
-ENTER WALL THICKNESS OF THE TUBE

(I was having trouble with thicker tubes not giving the proper circumference for the wrap, i.e. the circ difference in the flat state for a 1" tube would be 3.135 instead of 3.141..., so I made the tube as thin as possible to eliminate this)

CREATE A 'RIP'

Feature
-create
-rip
-REGULAR RIP
-DRAW THE RIP (single line), MAKING SURE IT GOES FROM END TO END OF THE TUBE

Create the flat state of the tube

FEATURE
-SET-UP
-FLAT STATE
-CREATE (SAVE AS THE DEFAULT FILENAME IT GIVES YOU...)
-FULLY FORMED
-PICK THE RIP THAT YOU JUST DREW
-CLICK 'OK'

Unbend the tube

Feature
-create
-unbend
-regular
-pick the rip on the tube
-unbend all

Denny is right, this is excessive, but it is the way that we have done it for a while. Once this one is done, I will be building a tube notcher, and creating a code to run in pro/e so it does the majority of the repetitive work for me.

Has anybody written a code for Pro/E (or solidworks) that output all point names, location, tube lengths, angles between intersecting tubes, and tube angles relative to the assembly co-ordinate system before???

Let me know if you want more info on modelling this in pro/e, as I have to document the modelling procedure for the next guy who get suckered into modelling the chassis anyways.


Chris

[Chris Clarke]

B.J.,

What method do you guys use for laying out the cuts if you dont use wraps? Do you just go off of the jigging, and file by eye till it fits?

[Matt Gignac]

Chris,

We use a similar method, but it's a heck of a lot quicker. We just make the tubes as normal in Pro/E as thin protrusions with the correct wall thickness and such, and usually protrude up to a surface (i.e. intersecting tube)

Once we get all the tubes designed in the assembly, we make a backup of the whole frame assembly, so that all the unwrapped tubes dont mess up our initial assembly.

Then we just do the following
1- Open the individual tube

2- Make a radial cut the length of the tube (i.e. a section of a circle) as small as possible. I think the smallest pro/e lets you go is .1 deg.

3- Now you can unwrap the tube. Click on applications, sheetmetal, then driving surface (on the toolbar on the right), choose the outer surface of the tube, specify the thickness.

For the thickness, you'll want to go as small as pro/e will let you (usually .002) so the thing you print out is as close to the OD of the tube as possible. Pro/E likes to mess up though, and if it wont let you take something small, odds are because there's something funny at one end of the tube (i.e. really thin section or some weird thing where you have a little piece sticking out). If this is the case, try trimming the offending piece off, and try again. For some reason, pro/e wont let you pick the minimum they specify.

4- Once thats done, just click unbend in the sheetmetal toolbar, select the edge to stay put, and voila.

Its really not that bad, I have it down to a minute a tube.

Matt Gignac
McGill Racing Team

[BStoney]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Chris Clarke:
B.J.,

What method do you guys use for laying out the cuts if you dont use wraps? Do you just go off of the jigging, and file by eye till it fits? <HR></BLOCKQUOTE>

Chris,
We fix all suspension points with our fixturing, then cut the tubes to fit inbetween. It's actually very simple and effective. We have guys that have gotten really good with an angle grinder and can make stuff fit with no more than 1/16" gap all around.

Hope that answers your question.

[Kirk Feldkamp]

Thanks guys. I'll give the methods a shot tomorrow.

-Kirk Feldkamp
UC Berkeley

[Chris Clarke]

Matt,

I had thought about copying the assembly and unfolding the parts that way, but I choose the other way because all the wraps remain fully referenced in case I want to make any changes.

One disadvantage of doing it using surfaces is that the surface remains after you unwrap it, so you have to create a layer, and hide it. Using surfaces did make the cuts really easy though.

I probably took me ~4 min per tube to model the flat state and create the drawing.

Chris

[Rob Davies]

anyone know how to unwrap tubes in solidworks.

Thanks, Rob

[GTmule]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by Rob Davies:
anyone know how to unwrap tubes in solidworks.

Thanks, Rob <HR></BLOCKQUOTE>

You can do each tube as a seperate part, where the tube, after being modelled is turned into a sheetmetal feature (you have to cut a "slit" down the side of the tube.

The EASY way to do this, however, is to do your model as an assembly (or better yet, a PART file) of SOLID tubes (as though it were made out of round stock, not tube) all cut off pretty at the ends and all. You then save as IGES, open it up the the trial version of Rhinoscerous, and use the "unroll developable surface" feature (you'll probabally have to do a split line along each surface, or, better yet, make your round profile out of two arcs, or a circle that doesn't quite come all the way around on itself while you're still in SW).

www.rhino3d.com (http://www.rhino3d.com)

It MAY work w/o the split, but I can't remember, it's been a while. This also works well when doing A-Arms out of Aero tube. http://fsae.com/groupee_common/emoticons/icon_wink.gif

[JConway-UofManitoba]

for pro/e...

advanced features, unroll surface....

from what I remember....
make the frame as an assembly of tubes using the piping function.
copy surfaces of intersecting pipes to the individual pipe drawings
solidify cut them so that you have crazy notched tubes
copy surface of the tubes
use the advance &gt; unroll surface

I think you'll have to hide the unrolled surfaces, but if you make simplified reps with the rolled/unrolled, I think you'll be in good shape for making drawings/dxfs of the cut profiles.

I really like the idea of printing off the patterns and cutting them on your own. What kind of cutters do you use?

You have any kind of trick to index the ends when you have a long component or bent component?

[JConway-UofManitoba]

Now from my new solidworks perspective, exactly as these guys have said, throw in a miniscule cut/rip down the axis of the tube and use the flatten command in sheetmetal.

you'll have to make your pipes with sweeps probably.

[johnnySV]

CementLegs,

can you help me understand why a Ball-End Mill is preferred to a Flat-End Mill for this application?

thanks,
john

[Didier Beaudoin]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by johnnySV:
CementLegs,

can you help me understand why a Ball-End Mill is preferred to a Flat-End Mill for this application?

thanks,
john <HR></BLOCKQUOTE>

With the ball end mill, you can use the tip of the mill to notch the tubes as well as the side of the mill. From my experience, the tube is a lot less difficult to notch with the tip as it creates a lot less vibrations. To notch with the side, you have to make sure your tube is really well held in the mill because you may end up with a tube in the wall/stomach/partner doing work on the lathe.

[johnnySV]

we plunge cut with a roughing flat-end mill and it works great. jigging, angle measurement, and patience seem to be key. i wonder if our current system could be improved upon by using a ball-end mill.... still dont quite see the advantages as we are not feeding in the horiz. direction. maybe i just havent done enough machining to understnad the functional differences between BEM and FEM ??

[Didier Beaudoin]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by johnnySV:
we plunge cut with a roughing flat-end mill and it works great. jigging, angle measurement, and patience seem to be key. i wonder if our current system could be improved upon by using a ball-end mill.... still dont quite see the advantages as we are not feeding in the horiz. direction. maybe i just havent done enough machining to understnad the functional differences between BEM and FEM ?? <HR></BLOCKQUOTE>

I've never thought of using a flat end mill that way to motch tubes... Maybe because I've always had a ball end mill at my disposal.

Anyway, I can still think of an advantage of a BEM, and it is that you can make notches easily at angles really close to 0 or 180 degrees by using the tip of it.

[rjwoods77]

I use a serrated end mill. I found that you almosty have to because a normal end mill will fold the side over when you get far into your cut. I mount the tube in a vise with v blocks. I turn the head to the correct angle and walk the edge of the serrated end mill into the tube in the x direction. Nice thing about doing it this way is you can get super accurate notching. If you modeled your chassis then you can just use your sweep line to determine the length of the tube. You cut the tube exactly to this length. If you have a computer operated chop saw, stick a tube in a lathe and cut it off exact, or cut a tube a tad long and put it in a mill and use the x axis of the mill to shave the ends to the exact length. I do it the last way and get tube that are within 5 to 10 thou or true length (due to tube vibration). I then angle the head to the correct angle and touch it off on the tube. Reset x and then move it in the correct amount that you can easily detremine with trig in 15 seconds. This has given me truely "no-gap" nothces. This is all dependant on the fact if your chassis is jigged up. This is the best way i have used out of the 10 spaceframes i have made in life. Grinding is bullshit. You only do that if thats all you got and/or the frame isnt jigged up( which means the tube lengths vary). Grinding takes forever. I tried doing y axis feeds ith ball end mills and found it to be really bad on vibration and the tool wasnt meant for pluge cutting. I tried drilling a just under O.D. hole first then I did it and it worked better. But it is just easier to x axis feed. The tightest notch i ever made on a mill was 14 degrees off horizontal. I had to hang alot of tube off the vise and i t vibrated a bunch but it was one of the best fitting notches on the car.
I cant wait til my buddy gets his plasma notcher done. We say it in a cannondale manufacturing video. They had one machine that fed tubing into a bender, bent the tube, and notched the tube on the all in one operation. Got us thinking. If we cut the ends accurate of straight pipes and bent pipes, we could just plasma cut each end with a programmable notcher. Cant wait.

[JadgPzIV]

In ProE you can modify the 'bend table' and ProE will consider only the OD while unbending. So you won't need extremely small thickness in sheetmetal. This modification is simple, you just have to open the 'bend table' in notepad and to change the equation to eliminate the ID.

[Gavin]

anyone have an easy way to draw in pro-e if i have an autocad drawing? i have our frame drawn in autocad as a 3d line drawing. all of our other parts are pro, so i want to draw it in proe so we can assemble the whole car. I have all the intersection points, so can i easly protrude from point to point without making a million datums? Thanks in advance

[Storbeck]

I am working on a matlab program that takes the length, diameter, and wall thickness of the tube, and the diamters and angles of the notches, all of which are easy to get off the solid model or the fe mesh of beam elements, and spits out a text file of g-code to notch the tube in the fourth axis of our cnc mill. Shows the cut-to length, diameter, and wall thickness of each tube as a comment that appears on the readout on the machine. There's a rapid out of the way and an optional stop between each tube so you can read the necessary info, go cut an apropriat tube, put it in the fourth axis and hit go. I think once it's working it should notch tubes about as fast as one person can cut them. Hopefully.

I'm also trying to come up with ways to import a list of coordinates of all the nodes(again fairly easy to get from the fe model),tell it which nodes have tubes between them and have it figure out all the angles on its own. THat's down the road though, I'm still working out glitches on the simple tube notching part right now.

[Joe Wilkie - SolidWorks]

SolidWorks tube frames - this is really easy.
1. Do NOT use sweeps (way too hard)
2. Do NOT export as IGES to another package (too much effort)
3. Instead, DO the following:

Use the weldment feature to create the tube structure. The trim/extend features in the weldments toolbar will do the hard part of creating your fishmouth cuts, etc. I find that using a combination of 2D and 3D sketches is easiest, using 2D sketches at various cross sections of your car, and using the 3D sketches to tie the 2D profiles together. Later, those 2D profiles will make excellent associative reference geometry for your body work.

Gussets can easily be added to rounded tubes as extruded features, since currently our "gusset" tool only works on square tubing where there is a pair of planar faces to reference.

A weldment creates a multi-body part file, with each body being a separate member. Your drawings will automatically provide you with a cut list and pipe stock lenghts.

To unwrap the pipes, you'll just use the SAVE BODIES command in your cut list, and save all bodies into an assembly. This will turn each body into its own part file and link build an assembly. Each pipe can then be slit and unrolled using the sheetmetal functionality as described in earlier posts.

I hope this helps,

- Joe
joewilkie@soliworks.com

[B.K.]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Rob Woods:
Grinding is bullshit. You only do that if thats all you got and/or the frame isnt jigged up( which means the tube lengths vary). Grinding takes forever. </div></BLOCKQUOTE>

Grinding is the way I roll. I helped make four Cornell frames, grinder and hand file all the way.

I can appreciate that for one round tube onto one round tube a notcher is surely quicker. But I can't remember more than two or three nodes per frame that were that simple.

Maybe I'm underestimating tube notchers, but I remember lots of nodes that had three, four, even five tubes going into them, all different diameters, some of them square, coming in at shallow angles. If you want to optimize your stiffness:weight, I assume you want the centerlines of all the tubes going directly into the node so the frame most accurately reflects your FEA model, so you've got to achieve some extremely complex geometry at the end of the tube you're mitering. That's when you gotta grab the grinder and do trial and error.

Using CAD to make paper patterns is intriguing, but it seems like it would take a really long time to get the frame exactly modelled, tube by tube, with end cuts and stuff, then print and cut out the patterns; time you could spend making the frame. And besides, on those complicated nodes, you're not going to get four or five tubes to end up within .025" of where the model expects them to be, so the pattern would only get you the general shape and you'd have to adjust it by hand with a "bullshit" grinder.

[Cement Legs]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by johnnySV:
CementLegs,

can you help me understand why a Ball-End Mill is preferred to a Flat-End Mill for this application?

thanks,
john </div></BLOCKQUOTE>

Yeah Didier pretty much has the jist of it. Like I mentioned in an earlier post if I had all of the measurements for angles and lengths I could cut almost all of the tubes for a frame along with a-arms in no more than 2 days. The best part about it is that the fits are nearly perfect (.005"). The onlly tubes that take longer than a couple of minutes are ones with angles in all three dimensions or in other words dont have at least one unit vector of zero.

I really believe the method would save a lot of people time so maybe I should do a quick video and post it .... ah mannnn http://fsae.com/groupee_common/emoticons/icon_razz.gif who doesnt have enough commitments all ready.... I'm transfering to Arts.