Electric Dirtbike

jktobin1

1 mW
Joined
Jun 13, 2018
Messages
12
ebikeSKETCH.png


Hello everyone, I have been a member here for a few years now and have mostly just been researching and seeing what everyone is up to. After seeing this post https://endless-sphere.com/forums/viewtopic.php?f=28&t=88595 by Vertigo I was hooked and had to have something similar. After a year in design, I have finally started production and want to share my journey and where I'm at now. I am creating a series of blog posts on my website and will slowly be updating and adding new posts as progress is made.
My Website: blog.jktobin.info

I am an engineering student at Arizona State University and have been using what I have learned and the tools available to me in hopes to make a high performance, reliable, and sexy electric dirtbike. Using a combination of downhill mountain bike, electric, and custom components I want it to be lightweight and versatile. By making this post here I hope I can get some views on my blog posts and be able to answer any questions you may have. I have done extensive research and engineering on all aspects of the bike and want to share what I have learned. There is still a lot left to do and I will be keeping track of my current progress and replying to questions here on Endless Sphere.

Motor and Controller
The motor I am using is the ME-0907 brushless DC motor w/ thermistor by Motenergy. Originally the motor was designed to power a golf cart and produces peak 26hp with .0959 ft-lbs of torque per amp. With a maximum amperage of 220 amps for 1 minute, the motor itself will produce 21 ft-lbs of torque without any gearing. At 72v it is considered a high voltage motor and with proper gearing should be able to move the bike along at 50mph.

I know the importance of a good controller and I did not want to skip any corners here. I chose a kelly controller due to their high build quality and available support. The KLS7275D sinusoidal controller can output a consistent 200A and a boost amperage of 500A! This controller should be able to provide the motor with more power than it could ever need but will allow room for future upgrades while also keeping heat to a minimum. The controller, originally for a golf cart as well, has a lot of features I plan to take advantage of. ABS, regenerative braking, reverse, and customizable torque curves were some of the features that drew me to this controller.

Battery
battery-1024x635.jpg


Currently, the battery is my main focus and is taking a lot of my time and energy. Originally I planned to make a simple pack, spot welded together and then encased in heat shrink. After a lot of research, I realized that this was not the best way to go and was far from the best design. Tesla's 18650 pack engineering was very interesting to me and seemed bulletproof. Copying their design using a plastic spacer with aluminum bus bars and individually fused cells, I designed a battery that fit my frame very well. I am going to end the description of my battery design here as it is not finished and still has a lot to go. However, the wiring and electrical systems that are part of the battery will need some explaining.

Looking at the 18650 cells avialable, I chose the LG HG2 cells. This gives each battery 3000mah and a continuous 20A discharge rate. These numbers put the entire pack, which consists of 10 parallel groups and 20 series groups, at 30Ah and a continuous discharge rate of 200 amps. With a high amperage output pack like mine, few BMS units allow over 200A to flow through them and with my goal of regen braking I was unable to find any that would even come close to working. To solve this I decided to cut some corners. With large cell packs maintaining a consistent voltage during discharge and regen braking is much easier than on a small pack. Therefore, I decided that for discharge and regen I will not be monitoring individual cell voltage. Instead, I will be relying on my controller to cut power when the overall pack voltage is too low. With no protection there I needed protection for charging in which I am using a Supower 150A BMS. I will not be charging anywhere near 150A but the BMS was originally purchased to provide both discharge and charging. This solution is not ideal but will still allow me to have a reliable battery that can be charged on the bike.

Suspension
suspension-1024x502.jpg


A powerful, fast bike that cant takes a hit or hop a couple of curbs would not be very much fun. With the added weight of the electrical components over a real bike, I needed high-performance components that will keep up. Using full-blown motorcycle parts would be too expensive and extremely heavy. Therefore, I chose to go with the best downhill mountain bike components I could buy. The front forks are 2017 Fox Factory 40 forks for a 26" wheel. With a heavier spring, the forks should handle anything I will throw at it while still being extremely lightweight and durable. For the rear shock, I chose to use the fox DHX2 with a 650lb spring. Determining the spring rate for the rear took a lot of time and a couple of programs to make it happen. Starting with a program called Linkage X3 I designed my rear suspension geometry to produce a progressive spring rate that increased near the end of the stroke. This was to help prevent a bottoming out scenario which with the added weight of the bike would not be good for my rear shock. After getting the geometry correct I took a few measurements and plugged them into Fox's Mountain Bike Spring Rate Calculator. This made it possible to choose a spring rate that would still be rideable but allow for the correct amount of sag.

Wheels and Tires
The wheels and tires for the bike were more work than I thought they would be. I needed a strong, beefy wheel that would take a hit but still be able to mount up to the existing components I had already picked out. The solution to this was found in an article on electricbike.com. Using moped rim I could mount larger tires and have a stronger rim to ride on. Using the Yamaha Play Bike Front Rim from a TTR125 and the Shinko SR241 tires I was able to make a very nice wheel and tire combo that looks awesome. Sadly the moped rim did not quite line up with the mountain bike hubs currently available on the market. In the front, I needed a 110mm boost hub with an ISO disk brake mount. In the rear, since there will be no pedals, I needed a 150mm front hub. Both of these hubs by themselves are not very popular and were difficult to find.

Onyx racing, a company out of California makes specialty, high-performance hubs that were exactly what I needed. After contacting them I was able to purchase the hubs I needed. Now the hard part, finding spokes and nipples that will fit the motorcycle rim and the mountain bike hub. Thankfully someone had already attempted this! Holmes hobbies sells oversized nipples that fit into the moped rim and taper down to a standard bike spoke size which is exactly what I needed. Using Sapim's strong spokes for the added strength, I had the wheels built by a local wheel builder who was more than excited to take the project on. I think they turned out awesome and are the best-looking part of the bike so far. [will update with pics soon].

Drivetrain
Another pinch point for this project was the drive-train. I wanted to use a chain for its durability and strength, plus chains are much more readily available than belts and allow for more freedom in design. After reading an article on chain and sprocket design and doing some math, I decided to use a #40 chain with a 5:1 gear ratio. Sadly, the largest available sprocket at the time that would support a #40 chain was 60 teeth. That meant that my drive sprocket could only be 12 teeth. This, according to my research would be too small of a sprocket and can cause binding. Due to budget and time restraints, I chose to go with it anyway. Currently, there is a 72 tooth sprocket that is available which will allow me to have a 15 tooth drive sprocket and may be the way to go after testing.

For durability and availability, I chose a steel rear sprocket. This was an issue because from the supplier it weighed over five and a half pounds! With that weight the entire wheel, tire, spoke, and hub assembly weighted less. To mitigate this I decided to lightweight the sprocket using topology optimization. For a round disk, it was quite simple and I ended up having a local shop EDM the sprocket which removed almost two and a half pounds giving the final sprocket a weight of 3.12 pounds. I plan to possibly remove more material as the sprocket is still quite heavy and feels plenty strong.

Braking
As previously stated, I plan to give the bike the ability to regen brake. This is especially useful in hilly areas and will only give me more range. Mounting the rear drive sprocket to the same ISO 6 bolt location that a standard disk brake would mount to I did not have any options for a mechanical rear brake. The rear brake will exclusively be regenerative and will have a hand control on the handlebars similar to the front brake. For the front brake, which does most of the work, I went with the Shimano ZEE brakes and Ice-Tech 200mm rotor. This combination of a hydraulic brake and a well-cooled rotor will give me a lot of stopping power out of the front wheel. With the features built into the controller, I should be able to achieve braking performance similar to two mechanical brakes.

Electronic Accessories
With my engineering experience, I have begun the process of designing and manufacturing a PCB that will be the central power distribution for the bike. From my 72v battery, I plan to have a 12v rail that is used to power an Arduino, lights, and a horn. I have always been a fan of digital control and am surprised at the number of vehicles and parts that are still produced today that are analog. Following my interest, I purchased two IXFN360N10T-ND MOSFETs. These MOSFETs can handle 360A a piece and with two of them in parallel, heat should be kept to a minimum. Having an on-board Arduino nano will give me complete control of the ignition, lights, and any other electronics on the bike. This will allow me to do cool things like use a serial addressable LED strip as the tail light and directionals, switching between red and amber, or control secondary circuitry at the push of a button. Since I couldn't make it simple, I have embedded an RFID antenna into the handlebar to activate the bike's ignition with the presence of an RFID chip that I wear on my wrist. I hope this will be a cool addition to the bike and give it that extra little pizzazz and reduce the possibility of it being stolen. (I am aware that this will drain my battery if the battery is left plugged in but it's so cool I don't care)

The Frame
frame-1024x695.jpg


This by far has been the most complicated and difficult thing I have had to design in my engineering career. There are so many things to consider and without a team of people behind me, I was forced to take all of them into consideration. Starting with the wheelbase and head tube angle of the bike I slowly began building a frame to hold all of my components. There were many iterations of the frame as I learned more about frame design. Slowly I began to find out what is important and what is not. The frame will be manufactured from one inch 0.095 wall aluminum tubing with a machined head tube. The swingarm of the bike will be made from two-inch by one-inch rectangular tubing with CNC'd axle blocks and pivot points. These sizes were determined after extensive FEA simulations. I am having the frame manufactured by a local motorcycle shop that creates custom choppers and other tube frame street bikes.

To begin designing the frame, I determined some of the specifications I wanted the final bike to have:
Seat height: 33"
Wheelbase: 50"
Handlebar height: 40"
Ground clearance: 14" (half the overall wheel diameter)
Headtube angle: 67 degrees
Reach: 18"

Through a lot of iterations (not all documented), I arrived at my final frame design that meets most of these specifications:
V1
ebike1-1024x560.png

V2
ebike2-1024x554.png

Final Version
ebike4-1024x559.png


Some of the biggest challenges were making changes to one dimension and not causing catastrophic changes to others. With this final design, I feel that I have a good balance between seat height, ground clearance, and suspension performance. Here is a picture of the sketch overlayed on the bike. (mm are easier to design with, sorry)
sketch-1024x609.jpg


What's left
There is still a lot left to do before this is a finalized bike. The battery will need more design changes before it is ready for production and needs to be perfect as it is a very crucial part of the bike. Currently, I do not have a seat for the bike and don't plan to add one until the rest of it works as it should. This should be a simple addition and will make the bike look much more complete. The frame is currently being manufactured and will need a lot of refinement once that process is finished. Mounting holes will need to be added for things like the controller, battery, and motor.

Overall I am extremely happy with the progress I have made and am excited to ride this thing once it's done! I will do my best to answer any questions and throw as many recommendations as you can at me, I need them. I will hopefully be making a second post on the battery to gain more insight on my design.

Links:
Moped Rims for Ebike: https://www.electricbike.com/moped-rims-tires-hubmotors/
Chain and Sprocket Design: http://gearseds.com/documentation/deb holmes/2.5_Chain_drive_systems.pdf
Holmes Hobbies: https://holmeshobbies.com/
Shock Spring Calculator:https://www.ridefox.com/fox_tech_center/owners_manuals/08/WCeng/Content/mtbspringratecalculator.html

Components:
Motor:https://www.evdrives.com/product_p/mot-me0907.htm
Controller:https://kellycontroller.com/shop/kls-d/
Battery Cells:https://www.imrbatteries.com/lg-hg2-18650-3000mah-20a-battery/
BMS:http://www.batterysupports.com/72v-...lithium-ion-liion-lipo-battery-bms-p-296.html
Front Forks:https://www.ridefox.com/family.php?m=bike&family=40
Handlebars:https://amzn.to/2YSP1iz
Stem:https://amzn.to/2YVunu0
Headset:https://amzn.to/2YH2cUh
Rear Shock:https://www.ridefox.com/family.php?m=bike&family=dhx2
650lb Spring:https://www.jensonusa.com/Fox-Steel-Rear-Spring-30-Stroke
Rims:http://www.prowheelracing.com/yamaha-play-bike-front-rim-1-40-x-19-black/
Tires:https://www.revzilla.com/motorcycle/shinko-sr-241-series-tires
Tubes:https://www.revzilla.com/product/bridgestone-motorcycle-tubes
Front Hub:https://onyxrp.com/store/mtb-hubs/mtb-iso-110-20mm-thru/
Rear Hub:https://onyxrp.com/store/fat-bike-hubs/fat-iso-150-15mm-thru/
Drive Sprocket:https://www.usarollerchain.com/40BS...nished-bore-sprocket-p/40bs12-78-sprocket.htm
Rear Sprocket:https://www.usarollerchain.com/40A60-Sprocket-A-Plate-40A60-Sprocket-p/40a60-sprocket.htm
Chain:https://www.usarollerchain.com/40-Roller-Chain-p/premium-40-10ft.htm?CartID=1
Brakes:https://amzn.to/31y0iSX
Brake Rotor:https://amzn.to/31ts5DY
Mosfets:https://www.digikey.com/product-detail/en/ixys/IXFN360N10T/IXFN360N10T-ND/2116933
Arduino:https://amzn.to/2H3GGOL
RFID reciever:https://amzn.to/2Z1pPlJ
RFID bracelet:https://amzn.to/2KL28ZJ

Small Parts:
Kapton Tape:https://amzn.to/2H2vp1b
12awg Silicon Wire:https://amzn.to/2H3krs2
6awg Silicon Wire:https://amzn.to/2YXIaA3
Quick Connect:https://amzn.to/2yTcO31
Heat Shrink:https://amzn.to/31xsAwS
 
Grantmac said:
The only thing I'd point out is that there are a few 150mm hubs which have a 6 bolt brake mount on both sides.

Oh thats good to know! I saw a few out there but I couldn't seem to find one that was readily available. The wheels are already built but if things need to change that's good information. I'm not sure yet if just a regen brake will be enough in the rear but we'll find out.
 
If you really want a MX bike you also need Torque. This is only possible if you using a 2 stage drive. My recent build has a chain reduction 1:3 on the first reduction and the second gearing is 1:5. With this reduction you will benefit a lot more torque without the heat. The start from zero speed give wheelie power all the way up, with less amps. Just a Idee wich route I went.
 
jktobin1 said:
Grantmac said:
The only thing I'd point out is that there are a few 150mm hubs which have a 6 bolt brake mount on both sides.

Oh thats good to know! I saw a few out there but I couldn't seem to find one that was readily available. The wheels are already built but if things need to change that's good information. I'm not sure yet if just a regen brake will be enough in the rear but we'll find out.

Oset parts are useful for builders:
https://osetbikes.com/gb/parts/pre-2015-parts/hubs/rear-hub-for-all-20-0-models/
 
Xtr6 said:
If you really want a MX bike you also need Torque. This is only possible if you using a 2 stage drive. My recent build has a chain reduction 1:3 on the first reduction and the second gearing is 1:5. With this reduction you will benefit a lot more torque without the heat. The start from zero speed give wheelie power all the way up, with less amps. Just a Idee wich route I went.

If I'm correct, with my 5:1 reduction I should be able to generate roughly 95ft-lbs of torque to the rear wheel. With my large tire diameter I will be lucky to be putting 50ft-lbs to the ground but for a bike that weighs no more than 150lbs, that should be plenty for a good wheelie. By adjusting torque curves I can always make the torque curve much steeper and guarantee that the front tire comes off the ground. I'm looking for something that is reasonable to commute on so a goal of 50mph top speed seemed reasonable factoring the weight of the bike and rider plus some wind resistance. If things don't work out, swapping the sprockets out shouldn't be too difficult.
 
As a motorcycle racer, the weakness I see in your design is having the motor pivot in line with the swingarm pivot. The reason motorcycles have the front sprocket forward and lower than the swingarm pivot is to create anti-squat forces when the chain pulls tight on the top of the swingarm during acceleration. Large manufacturers like BMW and Husqvarna have played with designs running the countershaft sprocket at the swingarm pivot position in the past, but abandoned the concept because it didn't generate as much grip as the current (popular) design. From a geometry point of view, the V1 design, with geometry using a swingarm pivot point higher than the front sprocket would create much better rear grip on acceleration than your final design.

With the effort you have gone into to design this project, it would be worth calculating the front end rake and trail figures, as well as the rear end swing arm length, angle and anti squat calculations, to ensure the bike handles as well as its propulsion performance.

All in all, it's a pretty cool project you have going on. :thumb:
 
spinningmagnets said:
Have you assembled the battery pack yet?

I have not assembled the battery pack yet but plan to have that in my hands by the end of the year if things go as planned. Do you have any recommendations? I am planning on using dialectic grease between the negative contacts and the aluminum plates. Then, using kapton tape, tape down the plates to the holder and begin spot welding the fuse wire between the positive terminals and the plate. Finally, a thin sheet of non conductive material will cover all of the plates on each side of the battery and the whole battery will be sealed off in heat shrink. It seems like it will be enough pressure to hold the plates in place. I am working on a model of this to show as a better explantation.

nelso said:
As a motorcycle racer, the weakness I see in your design is having the motor pivot in line with the swingarm pivot. The reason motorcycles have the front sprocket forward and lower than the swingarm pivot is to create anti-squat forces when the chain pulls tight on the top of the swingarm during acceleration. Large manufacturers like BMW and Husqvarna have played with designs running the countershaft sprocket at the swingarm pivot position in the past, but abandoned the concept because it didn't generate as much grip as the current (popular) design. From a geometry point of view, the V1 design, with geometry using a swingarm pivot point higher than the front sprocket would create much better rear grip on acceleration than your final design.

With the effort you have gone into to design this project, it would be worth calculating the front end rake and trail figures, as well as the rear end swing arm length, angle and anti squat calculations, to ensure the bike handles as well as its propulsion performance.

All in all, it's a pretty cool project you have going on. :thumb:

Thank you very much! I was not aware that this was done for anti-squat! I'm going to take your advice and do some calculations with what I have to see what those values for anti-squat, trail, and rake will be. Do you know where I could find reference values for these numbers? A bike to compare it to would be awesome but I'm not sure if that information is readily available.
 
Capture.png


I am messing with a design to mount the battery. I would put poly bushings in these holes and run bolts through the frame and through holes in the battery spacers. This would keep the smaller vibrations down and help lessen the force on the batteries from impact. I am not even sure if this is reasonable to manufacture. I am waiting for my manufacturer to start building so I have a little more time for tweaks to the frame.

Capture1-1024x411.png


After Neslo's comment I really started digging into the rake and trail of the bike. I imported a model of the bike into the Linkage X3, marked all my points and let the program go to work. It calculated my trail at 3.8". With some research, I found a few articles that said the trail should be anywhere from 3"-6" with a lower trail meaning sharper steering. With these numbers, I think a trail of 3.8" is a good number but I am not too experienced in this area. I am going to see what other numbers this program will give me.
 
jktobin1 said:
Thank you very much! I was not aware that this was done for anti-squat! I'm going to take your advice and do some calculations with what I have to see what those values for anti-squat, trail, and rake will be. Do you know where I could find reference values for these numbers? A bike to compare it to would be awesome but I'm not sure if that information is readily available.

It will all depend on what you want the bike for. Is it to commute on the road, or is it a trail bike designed for offroad? I am more familiar with road bikes than dirt bikes and they are set up completely different to each other. Dirt bikes have way less trail and are set up to be less stable and more nimble, while most road bikes are designed for faster speeds and higher levels of tyre grip.

Most motorbikes these days have a rake between 23 and 25 degrees (the lower the number, the steeper the angle and the easier to initiate the steering) and road bikes have a fairly small offset to get a static trail of around 100mm (which is a good balance of agility and grip). I'm not sure what a good trail figure is for a dirt bike, but I know they are way lower than what road bikes are. A MX bike has the axle on the leading side of the fork to increase the offset and reduce the trail, which makes them very agile, but less stable. Pushbikes tend to run even less trail as they are designed for lower speeds, unless they are downhill mountain bikes. Triple clamp offset and tyre diameter have an effect on trail, so once you have your tyre diameter and fork length, you can then work out your rake and offset to get the ideal trail.

At the rear, once you know your tyre diameter, you need to work out the swingarm length and pivot height to get the swingarm angle you want as this will effect the grip levels of the tyre as the suspension cycles under acceleration. A university once calculated the ideal swingarm angle for acceleration grip to about 12.7 degrees, but the countershaft sprocket location, swingarm length and linkage, spring rate and preload, and the compression and rebound damping all have an effect on the level of grip the system develops. It was interesting to find out that American Flat Track racers lower their bikes to not only lower the centre of gravity, but also flatten the swingarm angle to reduce the rear end grip so the bike oversteers easier as they aren't allowed to run a front brake.

Other things that have a big effect on handling are wheelbase, wheel/tyre size, fork length, swingarm pivot point, steering head height, weight distribution etc. Race teams keep their ideal settings to themselves and won't share for competitive reasons and there are are companies like Computrac which charge money and use an algorithm to calculate the best geometry set up for the given frame dimensions of a bike; but, because of the complex nature of how all of the variables interplay, it's hard to find reference figures for the 'perfect' set up. There are books like 'Motorcycling Handling and Chassis Design' by Tony Foale that explain the physics behind it all, but it might be easier to just mimic and scale down a modern motorbike or e-bike design that suits the intended purpose of your design and then tweak the positions to get the ideals figures for your application.
 
nelso said:
jktobin1 said:
Thank you very much! I was not aware that this was done for anti-squat! I'm going to take your advice and do some calculations with what I have to see what those values for anti-squat, trail, and rake will be. Do you know where I could find reference values for these numbers? A bike to compare it to would be awesome but I'm not sure if that information is readily available.

It will all depend on what you want the bike for. Is it to commute on the road, or is it a trail bike designed for offroad? I am more familiar with road bikes than dirt bikes and they are set up completely different to each other. Dirt bikes have way less trail and are set up to be less stable and more nimble, while most road bikes are designed for faster speeds and higher levels of tyre grip.

Most motorbikes these days have a rake between 23 and 25 degrees (the lower the number, the steeper the angle and the easier to initiate the steering) and road bikes have a fairly small offset to get a static trail of around 100mm (which is a good balance of agility and grip). I'm not sure what a good trail figure is for a dirt bike, but I know they are way lower than what road bikes are. A MX bike has the axle on the leading side of the fork to increase the offset and reduce the trail, which makes them very agile, but less stable. Pushbikes tend to run even less trail as they are designed for lower speeds, unless they are downhill mountain bikes. Triple clamp offset and tyre diameter have an effect on trail, so once you have your tyre diameter and fork length, you can then work out your rake and offset to get the ideal trail.

At the rear, once you know your tyre diameter, you need to work out the swingarm length and pivot height to get the swingarm angle you want as this will effect the grip levels of the tyre as the suspension cycles under acceleration. A university once calculated the ideal swingarm angle for acceleration grip to about 12.7 degrees, but the countershaft sprocket location, swingarm length and linkage, spring rate and preload, and the compression and rebound damping all have an effect on the level of grip the system develops. It was interesting to find out that American Flat Track racers lower their bikes to not only lower the centre of gravity, but also flatten the swingarm angle to reduce the rear end grip so the bike oversteers easier as they aren't allowed to run a front brake.

Other things that have a big effect on handling are wheelbase, wheel/tyre size, fork length, swingarm pivot point, steering head height, weight distribution etc. Race teams keep their ideal settings to themselves and won't share for competitive reasons and there are are companies like Computrac which charge money and use an algorithm to calculate the best geometry set up for the given frame dimensions of a bike; but, because of the complex nature of how all of the variables interplay, it's hard to find reference figures for the 'perfect' set up. There are books like 'Motorcycling Handling and Chassis Design' by Tony Foale that explain the physics behind it all, but it might be easier to just mimic and scale down a modern motorbike or e-bike design that suits the intended purpose of your design and then tweak the positions to get the ideals figures for your application.

That was really helpful! Thank you for all of the advice. Currently I am happy with the rake and trail. I've found a few figures online that seem about right with what I want for the bike. I may be wrong but I also think my anti-squat values are adequate. Although not perfect, I feel they will suffice and give me a good balance while not requiring major modifications to my design. Here is a screenshot from a calculator I found online. After filling in my numbers I end up with an anti-squat percentage of 150 at the beginning of the stroke and drop to about 60% by the end.

antisquat-1024x595.png

https://www.datamc.org/downloads/motorcycle-anti-squat-calculator/

After the sag of the suspension that puts my anti-squat percent at just over 110%. I am not sure if that will be enough but at least it is over 100% not under. Let me know what you guys think and if I should make some modifications."Setup 2" is the ideal conditions listed on the left. "Setup 1" is my current design.
 
Wouldn't you want to have regen on the front, and have the rear lockable? After all, you don't want to lock the front tire, and you'll get more power back from the front tire under regen braking as when you brake, the load will be shifted forwards.
 
SquidBonez said:
Wouldn't you want to have regen on the front, and have the rear lockable? After all, you don't want to lock the front tire, and you'll get more power back from the front tire under regen braking as when you brake, the load will be shifted forwards.

front brake regen is a neat idea, but how would it be implemented?? you would need a front hub motor to slow the front wheel without a mechanical front brake..
 
Back
Top