Why I Changed My Mind and Built Myself an eBike

Readers of this blog will already know that eBikes are by far the most efficient method of travel, but that I still decided to keep pedaling with my own muscle. I figured that I was gonna eat those extra burritos anyways, so why not use them for miles on my bike instead of inches on my waist? Well, COVID changed my whole mindset on this. Read on to find out why, learn the pros and cons of various eBike tech, and how I converted my own cargo bike to electric for under $700.

Why Electrify?

The first thing to ask yourself is why you'd want to electrify your bike. My main reason is that it'd get me places faster and more easily. This is especially in an area with lots of hills, which can be painfully slow on a loaded pedal bike. I used to slog my kids up hill at walking pace. It kept me strong, but it took a long time.

Riding an eBike up hills is a whole lot more fun, and it will likely lead to you using your bike instead of your car more often. Just look at these two graphs of the route I ride to pick my daughter up from school before and after I electrified my cargo bike. The eBike trip on the right has a far higher average speed. The eBike went 5 MPH so little that the graphing software literally removed that mark from the Y Axis on its graph! It also had 1/3 less travel time even with a lower top speed downhill, probably due to riding the brakes a bit to stay with traffic, and no sprints up near 20 MPH at the end like I did with the pedal bike (I must have been feeling energetic that day).

eBikes really speed you up hills

The trip graphed above takes a bit over 5 minutes by car, but parking, carseat buckling and then walking from the car mean the whole car experience is just about as long as the eBike's 7:03 total time if not longer. This parity of travel time is really a game changer. It eliminates one of the main reasons I would ever jump into my car for a short trip (cause I'm LATE!!!). The higher average speed of eBikes and easier parking mean that for short trips they're basically as fast, if not faster than cars.

In the gridlocked area of Northern Virginia where I grew up traffic was SO bad. I would often beat cars with my bike even on 10+ mile trips. Traffic there has only gotten worse since I left. Combine this with an eBike's higher average speed and it seems like in NOVA it would get you to your destination before a car for most trips under 10 miles.

This points to a key feature of bikes that car supremacy has blinded most Americans to. A 2 way protected bikeway can move more people per hour than using that same space for a single lane of car traffic. This is because even though bikes move more slowly, they take up far less space on the road and their lighter weight and greater visibility requires less forced stopping for safety. A lane of car traffic in a city only moves 600-1,600 people per hour, while a 2 way bike lane can move 7,500!

Cars take up way more space than bikes and buses.

Car ads have confused us into thinking that the top speed of cars is the speed they continuously move at, but anyone who actually drives knows in their bones that this isn't true. Research shows that the average speed of car traffic on a given road network moves toward the speed of public transport/walking/bikes in that same area over the years. This is because more and more cars will clog the roads until they've made them so slow that other options are faster (AKA the Downs Thomsom Paradox). Using an eBike puts one less car on the road, reducing traffic and hopefully spurring our lawmakers to build more safe protected bikeways so there's even more cycling and less car traffic. This virtuous cycle ensures that the people who will still drive cars move along more smoothly.

There's also a financial aspect too. We know that eBikes get over 1,000 MPGe and require far cheaper maintenance than cars. Replacing car trips with eBikes will end up saving lots of money and reducing emissions. This is particularly true if you drive an inefficient truck. Maybe you need that truck for specific jobs, but saving it for those times and riding an eBike when the bed is empty will save tons of money!

The fun aspect just can't be ignored either. An eBike's assistance makes you feel super strong and speeds you up hills. You can still cut off the assist or go faster to get a workout, but that’s now a choice of riding and not a requirement. It makes riding feel as wonderful as it did back when you were a kid. When I bought my wife an eBike for her science work I took a few trips on it and just couldn't deny how fun it was. Seriously, find one to borrow or rent and you'll see.

Hauling nearly 300 lbs of batteries and cinder blocks down to the stream to power my wife's water sampling robots

The other big reason to electrify is to be able to haul seriously heavy loads with my bike. My usual cargo are my two kids, and they just keep growing every day. They're about 80 lbs together now, but they'll be over 100 in the blink of an eye and I wanted to make it easy to take them where they need to go quickly via bike. Beyond this my Madsen is rated for 600 lbs and I'd like to be able to truly haul that much up the hills of State College with it. The heaviest I've hauled so far was near 500 lbs of cargo + rider. While I could pedal that on the flats, steep hills were impossible and required me to walk it up with great difficulty. Adding electric assist should let me ride to the hardware store 5 miles away to pick up 200 lbs of rock, pipes, or lumber with my bike, a serious upgrade.

Why not Electrify?

Building a bike itself creates over 3x the emissions as an eBike's battery + motor

Some readers will currently be muttering to themselves that adding a motor and a battery adds embedded energy to your bike and thus wastes resources. This is half true because while it does add embedded energy, that's small compared to the total emissions required to build the rest of the bike. It's half false because eBikes actually reduce emissions per mile compared to pedal bikes. Even if you still ride a pedal bike and only use the eBike to replace car/truck trips you wouldn’t have biked it will pay back it’s embodied carbon incredibly quickly (see my thread on this here).

There are other studies that claim that pedal biking has lower emissions, but when you read through them you see that this is only achieved by not counting the emissions embodied in the extra food you eat to pedal. Their basic premise is that in rich countries people are gonna eat whatever they want anyways. In my original eBike post I agreed with this, concluding I was gonna eat that extra burrito no matter what, so the emissions embodied in it belonged to my gluttony, not my biking.

The COVID-19 pandemic greatly reduced the amount of miles I put on my bikes and helped change my thinking around calories. As the disease spread I became largely home bound and quickly gained my own personal COVID 15. Actively deciding not to eat as much as I can all the time has helped me drop most of this weight over the past year and honestly it hasn't made me any less happy. I guess I'm lucky to have a body that doesn't require starvation tactics to drop weight (people whose bodies can maintain weight with even less caloric intake have the option to lead even lower emissions lives, but it seems pretty rough for them). I'm going to roll with what my own body is capable of and keep my emissions lower moving forward by sticking with my current diet and propelling my bike the extra miles I need to go with solar electricity instead of burritos. If you're not growing emissions free potatoes in your yard to power your pedaling it's worth a shot to see if you can reduce the amount of food you eat and add a little e-assist to pick up the slack.

But the Battery!?!

The final objection to eBikes is largely about the batteries they're powered by. As I detailed in Saving the World With an EV section of my EV post there are some legitimate problems with batteries, but they pale in comparison to those from fossil fuels. Batteries for eBikes are a special case because they don't have the same battery management systems found in large EV packs. There's no liquid cooling, or warming of your eBike battery and few limitations on how deeply you can discharge it. These can all lead to an eBike battery degrading much more quickly than those in bigger EVs. The industry is also still young, and many batteries are sold by sketchy firms, who have been known to put inferior cells into packs or bundle faulty chargers. This can result in eBike batteries catching fire while charging and even exploding.

One way to avoid a bad battery is to shop with big, reputable companies, but another is to buy a different battery chemistry entirely. The standard chemistry used in most EVs and eBikes is nickle manganese cobalt (NMC). It's so common than many people simply call it Lithium Ion or Li-Ion. But, there's a new option gaining popularity now called lithium iron phosphate (LFP or LiFePO4).

NMC can store almost double the energy in the same amount of mass as LFP so it has been the king for a while now. LFP main competitive advantages are that it basically can't catch fire from charging/discharging and it lasts far more charge/discharge cycles. That means a single battery will last for far more miles and won't burst into flames on you. LFP also doesn't use cobalt which means that it avoids the child labor issues that some cobalt mining has. Finally, LFP batteries also perform much better in cold temps than NMC. An LFP eBike loses less range when it's very cold out than an NMC one, though its battery still shouldn't be charged outside when it's freezing (large LFP batteries used in offgrid solar arrays have their own little heaters to enable cold weather charging).

My surprisingly light LFP battery

For my eBike I selected a small 10ah 48V LFP battery (though in reality it outputs more like a 52V NMC battery because LFP cells have slightly higher cell voltage than NMC). It weighs just about the same as the 14ah NMC battery that my wife's Radrover came with, but that's without the hard plastic shell + locking hardware hers has. This little battery cost just $200 though, with charger! A Radpower pack costs $549, and you can easily spend $1k on a fancy high capacity NMC battery. If you need 100 miles of range from your eBike then that might be the right purchase to make, but in my case my town is only 5 miles across. It took me over a week to get my first full discharge of this battery using max assist on my heavy cargo bike and I still I got 29 miles of range. Even in winter I get like 20 miles per charge at max assist. LFP batteries are rated for 2500 charge cycles before they can only hold 80% of their initial charge. That means over this battery's lifetime it should be good for over 70,000 miles, more if I use lower assist levels. The electricity from my solar panels to charge this battery for 70,000 miles of range will cost under $100 total. Try finding a car that cheap to run!

How To Electrify a Bike

Okay, so now that I've sold you on eBikes, how do you get one? The simplest option by far is to just buy a new eBike. There are plenty of eBikes available for around $1k from retailers like Rad Power, Ride1Up, Propella, Lectric, Ariel, and more. The main issue with these bikes is that to hit this low price point they skimp out a bit on their non-electric components. My wife’s lab has a Radrover bike that we love, but it weighs 70+ lbs and literally has the worst bike shifter I've ever used. That doesn't mean it's unusable, it just means you often won't be able to shift into the lowest gear when you need to. The hub motor in its rear wheel does a great job of making up for this by being there to scoot you up hills even if you're in the wrong gear though, but the reality is my 1980s Cannondale has far higher quality components than this bike.

Upgrading to a Mid Drive Motor

You can buy an eBike from a major bike manufacturer that has truly quality components, but it's going to cost you over $2,000 (like the Giant/Momentum LaFree E+ or the Yamaha Cross Core) and you could easily spend over $4k (like the Specialized Turbo Como SL 4.0 or the Lemond Prolog). Moving up the price spectrum like this gives you bikes with better shifters and brakes, but also better motors. These high end bikes generally use mid drive motors (the Lemond is the exception, using an ultralight, low power hub motor). These motors use your bike's gearing so they can deliver more low end torque when you're in low gear, but also a higher top speed when you're in high gear. Check out this YouTube video if you want the full details (if you want a counterpoint on how hub motors can still be great check out GrinTech's writeup).

My main issue with the hub motors on fancy eBikes is that they don't seem very future proof. The bike frames are made with a very specific cut-out to fit the motor. This secures the motor better and lets them make it a bit smaller but if that motor dies in 10 years and there are no more motors of the same shape being produced then that frame becomes garbage. A good quality bike frame can last many decades, heck I still ride a 1989 Cannondale much of the time, so I don't want to spend big on a frame that might become worthless much sooner.

Bosch Mid Drive Motor. Note the large space in the frame designed to hold it

DIY hub motors completely solve the issue of a frame design becoming obsolete by attaching to any regular bike frame right where you pedals would have gone. You may have to do a little monkeying to reroute shifter/brake cables that run where the motor needs to go, but that's only true on a few bikes. My Madsen had this issue but as you can see getting around it wasn't too hard, I just moved them up over the bracket [UPDATE: with a bit of jiggling I managed to route the cables under the frame like they used to be, it brakes far better this way].

Note how my mid drive simply attached to the hole my regular pedals had connected through.

Adding a motor like this to a decent bike that you already own will always be the cheapest option. Even if you have to go out and buy a new bike it's still probably far cheaper to DIY that into an eBike than to buy the same quality eBike new. The question is which motor should you buy?

What Motor to use?

For my eBike conversion I selected a 52V Tongsheng TSDZ2 mid drive motor for just $425. I had to buy a couple extension cables to make it work with my cargo bike, but all that plus my battery was still under $700. That's a far cry from the official $1815 eBike kit from Madsen. This motor is a bit less powerful than the Bafang BBS02 motor that kit uses, but it's strong enough for my needs and it has 1 killer advantage over the Bafang, a torque sensor. Torque measure how much force your legs are applying to the the pedals and then the motor multiplies that force. This means you still have to actually work at pedaling while the motor makes you feel super strong. It also means that the motor responds very quickly whenever you increase or decrease your pedaling strength. They are usually only found on higher end eBikes with motors from Bosch or Shimano, but this tiny, cheap, Chinese motor has one.

The light little TSDZ2

Other eBike motors use cadence based sensors that track how fast you're spinning your pedals and use this to determine how much power your motor puts out. My wife's Radrover uses such a system. These might require 1 or two full rotations of the pedals before they change the power output of your motor, so from a stop on her bike I generally use the throttle to get rolling. Having a throttle guarantees that the eBike is either Class 2 or Class 3 and prevents it from riding on trails or roads that are limited to only Class 1 (like some mountain bike trails near my home).

Cadence based systems also really need brake lever sensors that cut motor power as soon as you brake. Otherwise you might stop pedaling and hit the brakes while the motor keeps trying push you forward for a second or two. Mid drive eBikes that use cadence sensors also often include a shift sensor that cuts motor power every time you shift to avoid damaging your gears. Torque sensor eBikes don't really need this as you can simply let up on pedaling a bit before you shift, just like on a normal bike, and the motor will almost instantly follow your lead. I like to leave a heavy foot on my left pedal when stopped at a light though, so I still appreciate the brake cutoffs that came with my TSDZ2 kit.

[UPDATE] After a year with my torque sensor I like it, but I don’t see it as necessary. The main use is when going slowly up hill following people walking. A cadence based bike is harder to do this with, but the torque sensor lets me do it easily. A throttle based cadence sensor eBike can slowly creep up hills too though and once you reach speed though there’s almost no difference between a torque sensor and a cadence sensor. Maybe just pick the cheaper option?

How Hard Was It?

My crazy solder setup

The most technical challenge in converting my eBike came from the fact that the battery I bought was just a bare battery with wires coming out of it. I had to use my soldering skills to attach connectors to this so it could plug into the bike and the charger. I hope that LFP batteries with actual hard plastic cases and plugs start being sold in the US soon. Even if they don't soldering really isn't that hard, and you can buy all the tools I used for under $100. [UPDATE: Another dad replicated my setup without soldering by using push in wire connectors, I wonder if they’ll heat up, but so far it’s working for him]

The next hardest part of the install was removing the cranks and the bottom bracket of the bike. This required 2 special bike tools, a bottom bracket tool and a crank puller. These were intimidating to me because I'd never used them before, but it turns you basically just hold each with a wrench and turn real hard. I was lucky enough to get to borrow these tools for free from Penn State's awesome new Bike Den, but even if you have to buy them both new they're under $40 total.

USB Battery, GPS Tracker, & Lights

Outside of those two pieces of complexity, I tacked on a bit of extra challenge by using my cargo bike instead of a regular bike. I did a little extra wiring to connect the 6V lighting output wires of the TSDZ2 kit to a buck converter that charges the USB battery my lights and DIY GPS tracker run off of. Then I had to figure out a way to mount the battery. Madsen's eBike kit includes a special custom built bracket for this, but it wouldn't work with my case-less battery. On a regular bike I could have just put this battery into a trunk bag on top of the rear rack, but obviously this wouldn't work for my cargo bike, which has a huge bucket on the back. In the end I drilled a couple holes in the bucket so I could tie the battery down under one of the seats in the bucket (see pics of this on my Twitter). Another hole let the power cable get to the motor and that was it. This mounting location keep its it out of sight and out of the rain (yay for Madsen's awesome weather cover). My son has complained that its "taking up space where his feet want to be" so maybe some day I'll have to suspend it from the seat itself so he can put his feet under it? [UPDATE: I did exactly this by strapping the battery to the underside of the seat, it fits there perfectly]. The most amazing dream would be rebuilding the whole kids bench seat to be the battery, but that's outside of my comfort zone right now.

Final handlebar setup

After this was done all that was left was zip tying cables to the bike and setting up the handlebars. The main challenge there was to to pull off my bike's grips (shoving a zip tie between the grip and the handlebar makes this a lot easier) so I could replace my brake levers. [UPDATE: The brake levers that came with my TSDZ2 were crap, one of them literally broke off in my hand while biking, luckily I was able to stop with my second brake and carefully ride home. I may have caused this by overtightening them, but it’s never happened to me with another brake lever in the past. If I was doing it again I’d just keep my existing brake levers.] If I used a kit without brake cutoffs I could have avoided this step, but since this is how I carry my kids around I felt like cutoffs were required. [UPDATE: with a torque sensor cutoffs aren’t needed] The display itself could have screwed onto my handlebars without needing to take off my grips, and the only trouble I had with it was that I had to push scarily hard to get the plugs for the eBrakes fully seated into the holes in its back.

Luckily, I didn't break anything and the bike is now performing flawlessly. It took me a lot of research to decide upon this, but now that it's done I can't believe I waited so long. Riding my kids around with e-assist is wonderful. I expect it'll increase the miles, even I, a hardcore transportation cyclist, put on my bike. Hopefully this guide helps you to do the same, so we can all profit greenly together. [UPDATE: I replaced enough car miles with this to pay back it’s embodied emissions in just 10 months, I’ll have over 1,000 miles on it before I’ve had it a year, it took me 2 years to put that on the bike without e-assist]