10kW, custom cut aluminum frame, QS205, full suspension

owhite

100 W
Joined
Aug 3, 2020
Messages
285
Greetings people.

A post discussing my latest ebike construction. My first bike was written up here:

https://endless-sphere.com/forums/viewtopic.php?f=6&t=108357&p=1629256

Another pic. This bike was great to basically test my motor/battery/controller configuration. This bike uses 24S1P, with four 16000mAh packs in series. Each pack is a liperior 6S 12C 22.2V used from rcbattery.com. It weighs in at 44kg. The motor is 3T QS205 (I believe) and it travels at 55mph and can suck down 10kw at peak. It may go faster but I simply dont want to try on my current frame.

I also learned a lot about ebike building on Gen1 in general, as well as discovering an open source bike controller:

https://endless-sphere.com/forums/viewtopic.php?f=2&t=109266

I have several design goals for my new bike. I want it to be waterproof, and want to take it on trails so the batteries and other other internals need to fit snugly. I also want wanted it to be thinner than the first version. There were too many times the large box on Gen 1 was really uncomfortable.

The other thing that I am going after is to simply be able to construct a bike from scratch. I may ulitimately go back to using a pre-existing bike frame, but this has been very enjoyable to make so far.

Also let me mention that I have no expectations that this bike will be lighter than before. It's probably going to be close to the weight of Gen1, or maybe 4 kilos more. I dont think it's a good idea to go with light on this design -- the box construction will probably fold up if I go with thinner aluminum. I have mad props for the 33kg illegalbike -- maybe I'll get to that level some day.

Another goal is that I want to see the overall CG to get lower. If you look at most downhill frames, there is all kinds of space between the front wheel and the downtube, and I started thinking about building the batteries directly into the downtube, and that led me to think it would interesting to mount the batteries and electronics into a frame that was specifically designed to house everything -- particicularly because I'm after a waterproof bike.

After looking around, I found this cavalerie mtb with a nice big downtube to serve as inspiration. Here was the concept I was after:
yr7drRN.png

Fortunately the cavalerie peeps also post very nice dimensional diagrams which helps with choosing overall lines in CAD. After a few weeks on design/layout, the resulting design is here:
spQjXXH.png


What youre looking at is profile view, which is probably a little confusing, but the idea is two batteries are sitting in the gastank area of a motorcycle, and the controller will sit on that box. Two other batteries will run along the downtube. Here's a cut away from the other side. The batteries are in yellow:
H2OvV7o.png


A friend also did a stress analysis to find areas that would fatigue over time. There are several reasons I dont think this is particularly accurate. Unless handled very carefully the application of forces in the software probably doesnt really represent real world conditions, but more importantly I dont think CAD can model (bad) welding particularly well. The analysis was useful for find areas that potentially needed to be reinforced, and I adjusted the design a bit as a result.
kNZgHJS.png


The fun part of this build is it's all designed to be cut out out of flat aluminum plate. Here are some of the parts involved. Different colors correspond to different thickness:
qhLQYT0.png


Another amazing thing about the world we live in is that you can send off these parts to sendcutsend.com, and get aluminum parts up to 1/2 inch thick cut with extreme precision on a laser. Sendcutsend also does bending. Dotted lines indicate where the bends will go. The longish parts with dotted lines are the primary downtube/battery box. You load up the files, send a payment, and the parts come back in very short order:
7ZkStpR.png


Now for the welding. I am not a good TIG welder, in fact I just started. This pic is about the best I can do, but I have a fairly shaky hand and I'm very inconsistent -- so I like to avoid putting welds on the most conspicuous outer edges of the bike. The reason those folds are in the downtube/battery box is to hide the welds on the underneath portion the downtube.

Another trick I learned along the way was to bend metal was to use a triangular router bit to score a line down the aluminum. Shown in this pic. Then you can warm up the aluminum a little, put it in a vise and bend -- you get a really nice sweet straight bend.

This picture shows shaky hand welded box versus a custom bend box. I much prefer hiding the welds.
JOeh7nD.png


These pics show the assembled parts, which is 2.7kg so far.
L0PRLX1.png

JRHDtHo.png


I'm hoping to keep the whole thing (with swing arm) under 5kg. I'm very happy with concealing most of my crappy welds, almost all of them are internal or on the underside of the bike. This pic shows the worst of it, and here is another example:
he3JySi.png
 
Flat Al build is pretty cool. I just watched a video about a similar concept, although it has nothing to do with e-bikes. https://www.youtube.com/watch?v=JkXDi0oETd4

In theory, once the design is solidified, it could be packaged up and sold as a kit (like Ikea furniture you assemble with a welder). The design and metal cutting is probably the most time consuming/prohibitive part of this for the average DIYer. And unlike a tubular construction, you don't need a fancy jig to hold everything together.
 
you don't need a fancy jig to hold everything together.
I agree that for the most part things fit so well you dont need a jig, but now that i've put this together there are a couple of places where I would pay more attention to alignment. And when the forks go on I'm definitely making some kind of alignment system.

it could be packaged up and sold as a kit (like Ikea furniture you assemble with a welder
actually I'm just going to post all my files as open source, and if anyone wanted to they could order the parts themselves from sendcutsend.
 
owhite said:
you don't need a fancy jig to hold everything together.
I agree that for the most part things fit so well you dont need a jig, but now that i've put this together there are a couple of places where I would pay more attention to alignment. And when the forks go on I'm definitely making some kind of alignment system.

it could be packaged up and sold as a kit (like Ikea furniture you assemble with a welder
actually I'm just going to post all my files as open source, and if anyone wanted to they could order the parts themselves from sendcutsend.

do you mind if I ask how much those parts cost to get cut out? I guess I skimmed your earlier post too quickly and didn't realize you ordered them from an online service.
 
owhite said:
The big order was around $350, and the swing arm is $150.
I've also ordered a bottom bracket shell, and headtube -- both around $15 each.
I'm tracking all the costs and will post when I get closer to assembling everything.

That seems fair. You'd probably have to pay at least half that just to buy the raw stock.
 
Awesome job owhite, can't wait to see how it turns out! :thumb:

I've been wanting to learn to TIG aluminum for years,.. what is the welder you wound up getting, and how is it working
out for you? Complaints,.. wishes?
 
A bolt together version would eliminate the welding. More people could put it together. Threaded holes would eliminate through bolting.
 
nicobie said:
I agree that would make it easier to assemble, but I wonder how much welding it together adds to the overall strength?

It's a worthy question. As-welded tensile strength values for structural aluminum are about 1/4 the heat treated strength. But a welded seam is often much stronger than a riveted or screw fastened joint.

Best is to weld and then heat treat.
 
I also agree with bolts/ rivets and structural adhesive in gaps/ overlaps. Aston martins are put together that way and they are meant to go 200mph.

dscn5751.jpg
 
I've been wanting to learn to TIG aluminum for years,.. what is the welder you wound up getting, and how is it working
out for you? Complaints,.. wishes?

After a lot of internet research I landed on the Primeweld TIG225X. I am just a newbie, but i think it is fantastic. It is very well supported by third party videos. I normally ignore instructional videos on youtube, but I have learned a lot about the basics of TIG welding with some of the vids. If you want suggestions I can send some.

I also am taking to heart what one person advised is that at a certain point your settings, your equipment, and your machine really dont matter -- in general what is really holding you back is your skill level. It's just like if you have oil paints, a brush and canvas that aint gonna make you a da Vinci. Once I bought the machine I exclusively focussed on my technique, and didnt worry about buying any new consumables or special equipment.
 
On the subject of overlapping versus welding. I will point out that I've done a little bit of both.

Note this picture:
ptGiX0D.png


I got obsessive about avoiding showing a lot of welds external to the bike. So there a few different tricks I used to make internal welds to hold the thing together. The other side of that area looks like:
jZzUZ99.png


There are circles cut in the aluminum, that I put red circles around in the picture. The holes are in thinner 2mm aluminum, and they allowed me to put welds in those holes to connect them to the thicker 6mm support on the outside.

Interior of bike: UGLY
8FggqqS.jpg


Exterior of bike: NOT UGLY
IYKApJe.jpg


It's my intuition those areas wont structurally fail, but you never know.

Then to make the outside watertight, I hit the external seam with aluminum brazing rod. The stuff goes on like butter, is not as strong as welding, but it's not like using an ugly epoxy seal either.

Notice also that in the UGLY pic, in the lower right hand side there is an example of overlapping weld of the 2mm to a 9mm thick holder that handles the bottom bracket, AND, it handles the bearing going to the swingarm. Now that area, I could see breaking at some point. There are some other structurally reinforcing materials in that area that you'll see when I put together a post on the swingarm.
 
The swingarm worked out well:
WlMu6FH.jpg


EDIT: I retired this swingarm. See below

Swingarm fit to the frame is challenging. These are all the things to take into consideration:
  • The width of your motor is probably non-standard, and if you're planning on using pedals finding the right bottom bracket and crank arm width.
  • The swingarm has to clear the crank and the chainring.
  • The swing arm needs to get attached to the front frame with bearings that can take the punishment.
  • The length of your shock, with sag, impacts the sitting profile of the bike.
  • Shockmount placement and how it directs forces into the swingarm and frame.


This drawing is probably meaningless but it was what I used to address all those issues.

YcTIkTe.png


One design flaw was probably my use of these bearings. I cant imagine those little castings taking a significant beating. Also, they were mounted using bolts, and even though I used shoulder bolts, I'm expecting them to start to get loose over time. If I make another version of this I would like to get a machine shop to mill the right hole sizes in the aluminum so I can just force fit bearings into place. But hey at least for now they look cool:
RMmyuZF.jpg


I splurged and got me a rockshox for the rear, that fit in very nicely:
yTr3Mlj.jpg
 
I recently got the bike to the point that I could peddle it around and within 5 minutes of riding the thing I could tell the swingarm had serious problem with deflecting back and forth.

There were two serious problems both from using flat 1/4 aluminum on the arms. One was the back end wagged back and forth like a fish, the other was a rotational deflection as shown in the picture. If someone had their hand on the top of the back tire and pulled sideways, and if someone was on the bottom pulling in the other directions it seemed like it would easily rotate 5 inches.

The other issue with the first version of the swingarm was that the bearings sucked badly, so it was time for a redesign, ordering new parts, and some nice tooling.

1dLV5pB.jpg


The large tool there is an adjustable reamer, which was used to create a relatively precise hole in the aluminum. (You also need a hole saw to make the pilot hole for the reamer.) The advantage of reamers over a drill is that the cutting edge of reamers is on the sides fo the tool, and they sheer off just the right amount of material for a good fit:
2G8Pn44.jpg


I made four in one afternoon without any problems. To assemble, I made a jig out of laser cut plywood, clamped that puppy up, and got ready for welding. I also threw in 1.5 x 0.75 rectangular tubing:
Xed06PS.jpg


Compare the final version to the first one -- it looks like it will be 500 times more rigid, and the bearings have no slop in them, so this seems very promising.

V6DArjU.jpg
 
craciunptr said:
I also agree with bolts/ rivets and structural adhesive in gaps/ overlaps. Aston martins are put together that way and they are meant to go 200mph.

The unibody for the original Honda Insight (circa 1998) was Aluminum with adhesive joints (no welding). Alas I never had mine up to 200mph :?

It is not a new concept for bicycles. My circa 1990 Mitaya CT-7000 frame uses adhesive joints as well (CF over Aluminum). There was also a French company built steel frames somewhat earlier than that. The bicycles were of course standard double diamond road frames with tubes in lugged joints.
 
if you prefer glued ot bolted frame, look at pole bike for exemple. but you can start the milling machine
 
rider63 said:
if you prefer glued ot bolted frame, look at pole bike for exemple. but you can start the milling machine

The "Voima" is interesting frame technology. Very little information on the Brose motor except that it is 36 volt and EU limited to 250 watts. Internally it uses a cogged belt gear reduction. The frame geometry is a bit too 'downhill' centric for my tastes: the BB is far to high and no provisions for cargo.
 
owhite said:
Getting some test rides in:
sEVd8f9.jpg


I quite like it compared to the old bike. There's still a lot of fish/wiggle on the back tire. The newer swingarm reduced it a lot but now it seems like it's either coming from the back tire spokes, or still from the swingarm.

I could have said you that a skinny non-triangulated stays are a big no-no from a handing perspective on heavy bike. You either want triangulation, or truly massive stays like on a motorbike basically.

Compare this:

FKbZnPuh.jpg


To this:

9kParFbh.jpg

r4k5ubch.jpg


Rough, but WORLDS better. My frame is still pretty 'wobbly', hopefully it will go away once I'll bond everything with carbon (replacing some tubing with carbon along the way).
 
I have been watching this from the beginning. 20kW should not be a problem with a good cell choice, controller setup.

owhite said:
The big order was around $350, and the swing arm is $150.
I've also ordered a bottom bracket shell, and headtube -- both around $15 each.
I'm tracking all the costs and will post when I get closer to assembling everything.

These pics show the assembled parts, which is 2.7kg so far.
Image
Image

I'm hoping to keep the whole thing (with swing arm) under 5kg.


Does that include BOTH of the swingers? All parts for this plan cut and shipped, ?

I failed to see if you mentioned the actual stock, name, and .. the gauge. How thick is that aluminum, in those multiple colors?

Are you gonna fix swing arm #1?

Wow. 6kg frame.. Wow. How thin is that stuff? Is it hard to weld?

I have a goal and it is to be able to cut metal like this. I have endless access to all sorts of aluminum plate and stock for 2.50$ a pound. All that aluminum would cost me.. about I have a machine that cuts this and am currently building a much more powerful one.. to hopefully cut this kind of stuff all day. That will then be welded together. Someday.

You are an inspiration. Ty.
 
Back
Top