Saturday, June 11, 2016

The 2016 MarkKart - Power

So the next installment of Markkart 2016..... powerrrrrrrrr (c) Jeremy Clarkson. 

We're going backtrack a bit, months ago I had a friend hanging out in the China and I realized I could leverage something that most people in American can. Taobao! And specifically all the sketchy random motor vendors on Taobao! But if you clicked on the link you're probably going wait... whut....

(page translated via google) 
What the front page looks like
 But if we key in some strategic Chinese characters we get:

Much better.....
Taobao is probably the biggest internal Chinese vendor site, containing just about everything you can think about. From consumer goods, fashion, to sketch motors and electronics to some really legit goodies direct from name brand manufacturers. And yes I did say internal for the most part vendors will only ship within China, and don't speak much English. Aliexpress is the site that sends Chinese stuff to America but you often have higher prices, and longgg lead times to deal with (with a lot of broken English).

So anyways to what I was looking for: big brushless motors, Markkart has generally ran Hobbyking SK3 "63mm" (really 59mm) out-runners of various KVs. Problem is these were becoming an expensive diet at $80 a pop and often out of stock.... (Charlessssss) and  I burned up a few. So instead of being a good engineer and doing math I figured I might as well go find the biggest brushless outrunners you can get in China....

Meet the Melon (C) Charles Guan

This beast is an 80mm Brushless outrunner rated at 180kv with a factory spec of up to 48v. And in stock form is capable of dumping out 5-6KW. The term "Melon" coined by Charles in reference to it being about the size of a melon. They used to be sold on Hobbyking for about $100 known as a "C80" but long since discounted. The only place that carry them in the USA are e-bike shops and those go for over $180.

Browsing the Taobaos I figured if I could find the sketchy Chinese factory that made them a cheap source of them would be nice, and after much googling and searching I found a vendor that had a batch of "blemished" units for real cheap. (To cheap to show! Add to kart for price).

So of course I bought all 6 refubs he had and sent them to my friend in China, who received it shortly and motor-muled them back to the states for me. :D

Now these are designed for aircraft use, meaning they don't feature any real way to attach a sprocket or pulley onto the motor... which is a bit diffcult for gokarting uses. (I have a burning "dislike" of set screws to).

Probably something this sized
Now to the week before Makerfaire as I was freaking out..... The solution starts with milling a 3mm keyway into the motor shaft, it being steel this required a carbide 4 flute endmill and a bunch of cutting oil. Since I didn't wanna disassembly the motors, protection in the form of a plastic baggie and tape was used to keep chips/oil out of the motor. (Always use protection)

Indicating! Takes long then cutting
Now personally I don't like cantilevering a pulley/sprocket/gear off the end of the motor shaft..... espically the former two (gears deflect less). Taking a page out of Shane's Tinykart I wanted to extend the shaft and support it in a bearing to to make sure everything was supported.

Solidworks solves all things
Well that wasn't to hard.... Using Razor Scooter hubs had the advantage of already having a 72TH HTD 15mm wide pulley built in. And with the desire of efficiency and quietness timing belts were the way to go. Drive ratio ended being a 16T to 72T giving a reduction of 4.5:1. Which with 8in tires gave a top speed of around 35mph at 12S (way more than needed). There wasn't a lot of math done for this but from past experiences and observing what others have done this should work out...

Some interesting design notes from this, building FIRST Robots I have gotten quiet of a bit of experince messing around with the Gates HTD and GT2 timing belts. Both are modified curvilinear tooth profile vs the traditional trapezoidal shape of XL/L/MXL/AT/STD/etc type belts. The Gates belts also use a fiberglass cord to keep them from stretching and are built with a neoprene body with nylon tooths. And are incredibly strong. HTD is the older design, which patents has expired so you see it kind of everywhere, while GT2/GT3 are the newest Gates offering and offer nearly 30-40% higher ratings for the same size/pitch belt. The geometries are so similar that GT2 belts work on HTD pulleys, and visa verse, of course this down-rates the possible max loading abilities. It's interesting to me how such a little change in geometry (so little they are backwards compatible) leads to GT2/3 having pretty significant higher ratings.

Anyone curious how to design and build things with belts and not have a bad time with them I would implore you to read the Gates Design Manual (200pg PDF) which goes into great details on how to make these things work happily for infinite life. Engineering! Math! Science! Numbers!  Yeah I should do more of those things...

So back to reality, the cheapo casted Chineseminum wheels have a 72TH HTD 5mm pitch 15mm wide timing pulley in the back of them, now to improve on what HTD can do. I used a GT2 490T Belt with a GT2 16T pulley on the motor shaft end. Using GT2 belts on HTD pulleys improves the maximum loading abilities to a degree, the exact amount of that "degree" is a bit of a gray area.... But according to the design manual all HTD parts/belts should be able to handle around 95 in-lb of torque. While a purely GT2 setup will handle around 180 in-lb, so using a GT2 belt and pinon (weak point) this setup should be able to handle somewhere around 130 in-lb of toque for life. Yep, Gates design guide numbers are for life, so you can exceed these numbers by a fair bit just your belt life decreased dramatically. With the gokart bring a short duty cycleish (watch me eat those words) I felt fairly confident in these numbers. This is also assuming perfect tension...

With timing belts axial alignment is key to making sure they work, and also making sure your pulleys spin concentric. The top view shows how everything is supposed to line up. With a tensioner that is *gasp* non-adjustable. Using maths in CAD I estimated the size of the tensioner needed, and it's a 3D printed piece of nylon ridding on a needle bearing. If my CAD guess is wrong, I'll just scale the print according.

Now to my favorite part of this the output shaft that couples onto the motor shaft and the pinnon pulley. Thanks to a connection gathered from my battlebot-ing I had access to something called at wire EDM machine. If you don't know what wire EDM is, think of as hot wire cutting of foam but for metals. Video below (for some reason everything I could find was old).

For designing things that I don't have an unlimited budget to fabricate (which is always), I always try to a plan or at least a technique in mind before CADing up something that requires a 10 axis CNC machine or something. Knowing I could cut some parts on the EDM meant that I can do some things I normally couldn't like.... billet pulley/shaft with a integral keyway!

 This is basically un-machinable with conventional tools. The neat thing about the EDM is that the wire is only .01 in diameter letting you cut basically square internal corners. And the machine is very precise down to .0005in +- and can hold perfect squareness even cutting tall stock. The plan was to take a steel rod, and machine a blank on the lathe, getting the OD of the shaft part to the correct fit in the bearing. And the pulley side oversized so i can have the wire EDM cut the pulley. A pilot hole will be bored though the center for indicating the EDM. Also a jig will be needed to be made for holding the part....

Unforantly I don't have many photos of this process since it was a RUSH job as I was running out of time on the build. Something like T-minus 7 days to makerfaire. And everything to shit.......... yay.

Around midnight while on the EDM, trying to cut the pulley part of the shaft the wire kept breaking.... and between 3 of us we couldn't figure out what was going on. So a re-design on the spot and re-cutting some parts was required..... (was there till around 4am..)
Thankfully we we are running a modern EDM with all the features
not as billet-y 
So this turned into a two part assembly, the shaft got the internal key and bore EDM-ed into it and the end shoulder got turned down soo... the EDM pulley that had a bore broached would press fit onto the shaft. After which I had it TIG welded together so they should never come apart.

See that's how weld should look like..... 
And now the big question... does it fit.

Oh yes! What happens when you like machine things to correct tolerances and shit. The integral key is great, no tiny keys to loose or to retain. Also around this point I got the belts in from SDP-SI and they fit the profile perfectly so everything was working out..

Friday, June 10, 2016

The 2016 MarkKart - The Beginnings

Again after a long lapse of blogging I am back!

The months of March and April were taken up by the fact a group of misfits and I built a 250lbs battlebot and ended up on Season 2 of ABC's reboot of Battlebots! That's a long story and I'll post a more in depth article on that when more of the NDA's get lifted.

But we filmed Battlebots up till the end of Apr, and due to being crammed into a huge hot sweaty room of roboteers we all ended up sick including me.... so after a recovery phase, I realized I had like 2.5 weeks to build/upgrade/do something for the Power Wheels Race at the Bay Area Makerfaire..... 

For anyone that doesn't' know yet, Power Wheels Racing Series  is a event hosted at various Makerfaires where teams show up and either mod a Power Wheels (kiddie ride) or build something that ends up looking like a Power Wheels for under $500. And race it around a small-ish track in the parking lot of a Makerfaire, and perform a skirt for audience votes to collect "moxie points".

So with the state MarkKart (or whatever it's called now) was left it needed some serious work. Burnt motor(s), way to high CG, handling issues and a frame that was 3 years old. 

The 2015 kart... 

And a dead Sk3 outrunner
So after thinking about it... a complete re-design ended up happening.... oh hi solidworks: 

Missing some bits.. but close enough
The CAD for Markkart 2016 have been in the works for a bit, but during the first week in May while recovering from robo-pox I got to a point where I was pretty happy and could start fabrication. It is the most gokart-y looking thing, I've built so far. It started out more chibikart like and kinda evolved into more of a 3/4 sized tinykart thing.... THANKS SHANE.

I'll break up the different parts of the MarkKart build in to various post (making this up as I go) to hopefully cover the project in greater detail then I usually do. (Charles' blog is the inspiration there).

Having always done bolt together alum tube extrusion frames on the past karts (I have a burning "dislike" of 80/20). I decided to try something new, and that is having a frame welded together for me. That's right, OUTSOURCING. Over the years welding is a skill set that has eluded me, and TIG welding aluminum is up there on the welding difficultly level. So to be able to get this done on time and with the quality needed, a ringer needed to be brought in thankfully we have an amazing welder at the local Makerspace that got the bid.  

Frame design was probably the easiest part of the CAD, all 1x1x.125 6061 square tube and all done up with right angle joints. With some round bar welded in for the seat mount (razor ground force seat).

Boring I know right
Yep pretty boring.
Overall dims are 30x40. To ensure it could be welded quickly enough everything was designed so all the welds were with same thickness material and butt joints at 90 degree angles. Basically the drive boxes bolt on the back. Spindles go into the front gussets a seat in the middle.

But Markkk this isn't what a gokart frame is suppose to look like...... 
A real cadet kart bare frame

To get philosophical for a second here..... I am well aware a square thing made out of box tubing isn't exactly consider a "real gokart" depending on what circles you run in. Due to limitations of time and everything else something quick and simple took precedence over everything else. Also round tubing used in real racing things, is a challenge to work with and with little experince in that area trying to bend and jig up a real ish looking frame just wasn't gonna happen. Also on a personal level this is project wasn't meant to build the fastest or the most "real" gokart possible, but something fun and silly. So yeah square tubing frame for now. Maybe something different next year...

A hindsight thought is I have learned a lot about gokart dynamics and handling over the course of being around PRS for a few years. Some of the teams are legit and have built some really cool stuff (photos later) and talking to them has taught me quite a bit.... the limitations of the MarkKart has always been my design abilities for vehicle dynamics but that's improving over time... (FIRST and it's skid steer sigh).

Anyways onto the build! First trip to the local metal yard for a bunch o aluminum.

Then drilling the mounting holes needed for various gearbox and paneling bits on the Bridgeport.

So with welding, the biggest worry I had was making sure everything came out square. If you look at the design the front bar and the rear bar actually bolt on with gussets. This allows the drive system to come off and on easily and also provided a jig within the frame for welding. I was able to bolt it together, ensure square and then hand off the bars and part with a drawing to the welder. So he could weld in all the middle pieces. 

It's a rectangle
The black magic of welding
After about 3 hours of welding it was done! Honestly it went together faster then even I expected, and the final product was awesome! Everything within a 1/16 or better, and dead nuts square. 

Looks like the CAD model... weird how that works out..
Next was adding the paneling to hold the electronics and batteries, the 1/8 alum was going to be waterjet cut but they ran out of time on the machine so I ended up using the local CNC plasma machine. Not as clean of cuts but with some filing afterwards still works fine. 

CNC Plasma is stupidly fast
 Because I was worried about the holes lying up after welding, I didn't put any of panel holes in the frame. So had the plasma cut them in the panel, clamped it together and matched drill.

Only use for SLAs le
 Bind pop rivets are a personal favorite of mine, these are 3/16 aluminum ones. And with a $40 Harbor Freight pneumatic rivet gun they take seconds to install and make for a clean look since you can't see anything from the top.

 And it's done!

Notice the stockpile of spray paint in a photo above? Something I've learned is that just simply painting your projects makes them go 10x faster... well maybe not but it makes them look 10x legit. And makes other people think you're more legit.... since I was waiting on waterjet parts and various things to come in I figured ill go give it a try. Painting aluminum is always a bit tricky due to the fact nothing likes to stick to it. (Wanted to do powercoat but no time). 

From what I have learned, it's best to sand everything with a 300-400 grit disk sander to get the oxidation off the aluminum, this also buffs out most scratches and such. Then to lay down 2-3 coats of zinc based self-etching primer. Afterwards then 2 coats of your color. Paint of choice was rustoleum spray paint from the local Homedepot. 

Zinc primer is always greenish
That looks legit now eh. 

We'll see how this holds up, the real trick was letting it cure for 3 days untouched. As most paints don't fully harden for days so assembly will need to be carefully done....  At this time waterjet parts come in and back from coating! So fun times should be had teaser for next post: