The day starts off in style with some roadside plums. I seem to be living the farm to bike lifestyle.
I think about going to the Anacortes Ferry Terminal, the official start of this route, with a vague idea about possibly running into other people who are starting the same tour. But it would be an extra 34 miles with the backtracking so I decide against it. I need to keep moving east to get ahead of the cold season that is going to be descending here soon. If there are other cyclists on this route now, I’ll probably catch up to them anyway.
I cook up the last of the farm-to-bike eggs around lunchtime. Doing it with the electrical immersion heater cost me almost 8 miles (13 km) of range so I’ll probably use this method only when I’m out of fuel for my stove.
I make it to North Cascades National Park tonight. As the sign shows, this is bear country. I was planning on getting a bear canister at some point. I think it’s going to be sooner rather than later.
This time-lapse video is fairly representative of today’s route. It’s very pretty but the tall trees near the road combined with the low sun elevation at 49° latitude in September means that a lot of my potential solar power is lost due to tree shading. Even at solar noon, the sun is only 44° above the horizon. I’m getting only ~1kWh of solar energy per day. I will need more than that to keep up this pace for the rest of the trip.
I ride out of Vancouver early on Sunday morning. The city belongs to groups of tai chi practitioners, surprisingly spry senior joggers, and one solar ebike putting along at 10 km/h (6 mph) because the battery is empty and the sky is overcast. I’m so slow that frail-looking octogenarians on cruiser bikes are passing me.
The locals seem startled when I smile and say hello. To be fair, some of them may be wearing headphones. I entertain myself by pointing forward and asking random strangers, “Is this the way to California?“
Five hours later, I’m at the border. The border agent looks me over. “Oh, that’s interesting. What is it?”
“It’s a solar powered ebike. I’m taking a little trip.”
“Did you build it?”
“Are you going to make it and sell it?”
“If I say no, will I be put on a list for suspicion of un-American activities and denied entry?”
“Have a nice day.”
I make it to the Bay View State Park hike/bike site half an hour after dark. There are no other bikers here. My ebike battery ran out of juice 7 miles from the finish line. Fortunately, I installed a small back up battery just to run the lights in situations like this. Thank you, past self.
Huge thanks to the whole team at Grin Tech for being such great hosts during my 36 hour layover in Vancouver. Your dedication and passion are an inspiration.
Special thanks to Justin, Aaron and Robert for coming in on a Saturday to shoot a video and build me a new wheel with an updated version of the Grin All Axle Hub Motor. My v1 hub was an earlier model and performed flawlessly for 3.5 years and 28,000 km (18,000 miles). The new version has upgraded seals and bearings which are a better fit for the long journey ahead.
Here’s the video update of my solar bike build.
Next up, cross the border into Washington to join up with the ACA Northern Tier trail starting from Anacortes. Maybe I will run into other cyclists who are starting the same route? I’ll be leaving Vancouver in the morning with 5% battery charge. The weather forecast calls for clouds and it’s 160 km (100 miles) with a rather heavy ebike so that should be interesting. I’m not bringing a plug-in charger on this trip because, um… that would no longer make this a solar/pedal hybrid? I haven’t decided how I feel about that. We will see in a couple of months. Rainy days and steep hill climbs may convert me yet.
The following entries are re-posts from Instagram. If you already follow me there then these will all be reruns for you. I’m finding that at least half the people I meet along the way are not on Instagram so this blog is a way to follow my journey without a big social media company getting in the way of the experience.
I set off at noon. The Sea to Sky Highway between Squamish and Horseshoe Bay has loud, heavy traffic requiring my full attention. The spectacular views of Beautiful British Columbia go mostly unappreciated.
Surface streets take me to Grin Technologies headquarters in Vancouver. My bike is practically a rolling show-room for Grin: regenerative braking–capable All Axle hub motor, airline-friendly LiGo+ batteries, buttery-smooth Phaserunner motor controller, and the brains to tie it all together – a Cycle Analyst running solar firmware.
My arrival “coincides” with Friday afternoon beer o’clock. Grin’s neighbors are having a barbecue and we are all invited. As a pescatarian, the word “barbecue” does not trigger an enthusiastic response. It turns out that DWAY are a Japanese fish and sake wholesaler for the Vancouver area and “barbecue” is an affair mostly involving grilled fish and expertly prepared sushi and sashimi. My kind of barbecue!
This is my proposed route. I’m mostly following the ACA Northern Tier to Chicagoland, then the ACA Bike Route 66 to Southern California. This is the shakedown cruise I originally planned for 2018 to test the bike, gear and rider before setting out on a world tour.
My previous post, How to Convert Your Ebike to a Solar Ebike, was an introduction to this topic. Please read that first, if you haven’t already. This post will cover a range of recurring questions I’ve been answering over the years via email and in comments on YouTube, InstaGram, and ebikeforums about the nitty gritty details of matching an ebike battery, solar charge controller and a solar panel to each other to create an efficient system which performs well over the full range of real-world conditions.
My goal is to help you design a system which will extract the maximum amount of energy from your solar panel whether it is cold or hot. Don’t worry if you don’t fully understand everything here on your first reading. Most first-time solar bike builders got things working without a full understanding of every last detail. I certainly did. You can always come back and use this as a troubleshooting guide if things aren’t working.
Some electrical engineers will feel compelled to point out that as a non-electrical engineer I have no business giving advice about engineering electrical systems. They are right. I’m just a fool with a blog who has been tinkering around with this stuff long enough that I’ve made most of the mistakes it’s possible to make. I’m here to share what I’ve learned so you can get your project up and running before you make your first expensive mistake and get hurt or get frustrated and give up.
I’m assuming you already have an ebike so you already have the three components on the right: battery, motor controller and motor.
For our solar upgrade project, we’ll be working with the three components on the left. You might be tempted to start with the solar panel but we’re going to start with the battery and work our way back to the solar panel because the battery’s chemistry, number of cells and cell ratings determine our system’s voltage range and maximum current when charging. These values will drive our choices for the other components.
The most common ebike batteries are 36V, 48V or 52V. These values are “nominal” voltages meaning they roughly correspond to the average voltage of the battery pack as it is discharged from full to empty. If your bike uses another voltage like 24V or 72V you’ll need to do your own calculations but the same basic principles still apply. For our purposes, we will need to know the actual maximum and minimum values. A “36 volt” battery typically consists of 10 lithium ion cells connected in series. A single cell is at 4.2 volts when full and the voltage gradually decreases as it is discharged.
The pack’s BMS (Battery Management System) will disconnect the battery to protect the cells from over-discharge when the voltage falls below the LVD (Low Voltage Disconnect) value which is typically 3.0V/cell but may be as low as 2.7V or as high as 3.3V. This cut-off point usually happens under load and as soon as the load is disconnected, the cell voltage may rebound a bit but if the battery is discharged slowly it may stay at the LVD value. Here’s a summary of the numbers we will need.
Nominal battery voltage
Cells in series
Lowest minimum voltage (2.7V/cell)
Recommended minimum voltage (3.0V/cell)
Highest minimum voltage (3.3V/cell)
Maximum voltage (fully charged, 4.2V/cell)
I used words like “typically” and “usually” above because there are many different flavors of lithium battery chemistry and some of them have different minimum and maximum voltage values. Make sure you understand your battery’s limits before connecting a solar charge controller. Over-charging or accidentally short-circuiting lithium batteries can cause them to emit toxic smoke or start a fire. The labels printed on your battery pack or the charger which came with your bike are good starting points.
You’ll also need to consider how much charge current your battery pack can handle. A small, inexpensive pack’s BMS may impose a maximum charge current as low as 2 amps. With a 36V battery, that means that if you connect a 100 watt solar panel to an empty battery (~30V), the charge controller will deliver more than 2A of charge current and the BMS may stop charging. Since neither one of our solar charge controllers has the option to limit charge current, you would need to select a smaller solar panel or get a different battery if you run into this limitation. Your battery cells are likely rated for higher charge current than the BMS allows so it may be necessary to bypass the current limit by charging the pack through the larger discharge wires but doing so may also bypass important safety feature like cell balancing and disconnecting the charge current if any one cell exceeds 4.2 volts or if the pack gets too hot. Bypassing the charge port falls under “advanced users only” so proceed with caution if you understand the risks.
Solar Charge Controller
Now that we have our battery’s critical numbers, we can move on to our boost solar charge controller. This device converts lower DC voltage from the solar panel (input) to higher DC voltage needed to charge our battery (output). If you’re getting the Genasun GVB-8-WP, you’ll need to order the model which has been factory programmed for your battery’s maximum voltage. Get the one that is as close to your battery’s maximum voltage as possible without going over. The currently offered models are 41.7V for a 36V battery, 54.2 for a 48V battery, and 58.4V for a 52V battery. If you’re getting the CTK-EV300 charge controller then you’ll need to program it yourself for the voltage you need. Remember to use the “SELF” voltage option as the built-in 24V/36V/48V presets are intended for lead batteries.
My background is in roof-mounted residential and commercial solar electric systems. In that world, we seek to optimize the design for maximum annual energy production from each solar panel. However, most systems end up being a compromise of production yield, cost and aesthetics.
On a bicycle, it is especially important to extract every last bit of potential energy from the solar panel at all times as this allows us to carry the smallest and lightest panel that meets our energy needs. On a long tour, we will occasionally end up having to pedal up a steep hill with no electric assist because the sun isn’t available and every extra bit of weight is going to be a curse. Selecting the best solar panel for our ebike project is a topic for a future post. For now, let’s focus on the electrical specifications and how they relate to the overall system.
This part gets a little tricky because a solar panel’s voltage and current is determined by the temperature of the solar cells, irradiance (amount of sunlight) and the electrical load on the panel. To ensure the charge controller will be able to charge our battery under all conditions, we need to account for temperature. The VMP and VOC values on the label are only valid at STC or “Standard Test Conditions”, defined as 25°C solar cell temperature and 1000 W/m2 irradiance. Most of the time, your solar panel’s cells will be significantly warmer than 25°C and less than 1000 W/m2 which is why your panel’s peak output is typically only about 85-90% of the rated Pmax value and the charge controller’s output is around 70-85% of the Pmax due to 5-15% DC-DC conversion losses. There are four things you need to consider.
The panel’s highest voltage should never exceed the battery’s lowest voltage
We’re using a boost charge controller which means that it takes the input voltage from the solar panel and outputs a higher voltage to the battery. If your solar voltage is ever higher than the battery voltage, your charge controller will detect it as an error condition and will refuse to charge the battery. Even though you typically don’t discharge your battery to the lowest voltage it can safely handle, if it ever happens while you’re touring you don’t want to get stuck in a situation where your solar panel voltage is too high to charge your battery. The solar panel’s highest voltage will occur on a cold morning when the temperature is around 0°C (32°F). In this scenario, charging has not started yet so there’s no load on the panel which means we need to consider the open circuit voltage (VOC). The solar panel voltage will be even higher at lower temperatures but lithium batteries should never be charged below 0°C.
For example, the VOC of this Sunpower SPR-E-Flex-100 panel is 21.0V at STC (25°C) but what will it be at 0°C? From the datasheet (PDF), the temperature coefficient for voltage is -0.28%/°C. Since 0°C is 25° below the STC temperature, we take (0-25) * -0.28% = 7% which gives us a 1.07 multiplier to covert from voltage at STC to voltage at 0°C for any similar Sunpower panel. In this case, 21.0V * 1.07 = 22.5V which is what our charge controller will detect on the input side. Since 22.5V is lower than the lowest battery voltage for our 36V, 48V and 52V batteries, this panel is suitable for charging a depleted battery under these conditions. If you want to connect two of these panels in series, you would have 22.5V * 2 = 45.0V on this cold morning which would be too high to charge the 36V and 48V batteries but might be ok for the 52V as long as it has a little bit of charge left (45/14 = 3.2V/cell).
Nominal battery voltage
Maximum recommended solar panel VOC at STC (label value)
Maximum recommended solar panel VOC at 0°C
The panel’s lowest voltage should never drop below the charge controller’s lowest MPPT tracking voltage
On the hot end of the temperature scale, our solar panel’s lowest voltage will occur on a very hot day when there’s no air flow and the panel is under load. To calculate this voltage, we’ll use the maximum power point voltage (VMP) and a cell temperature of 75°C. Using our Sunpower panel, this gives us a correction factor of 1 + (75-25) * -0.28% = 0.86. Using our VMP, we get 17.5V * 0.86 = 15.05V. This value is well above the 5V minimum input voltage for the GVB-8 and just barely above the 15V minimum MPPT voltage for the CTK-EV300 so these panel and charge controller combinations should perform well across the entire temperature range. You will find that you’ll get similar values with most “nominal 12V” solar panels but be careful with smaller panels which may have lower voltages which may drop below the minimum required input voltage on a hot day.
The panel’s maximum current (IMP) should not exceed the charge controller’s maximum input current
This is unlikely to be a problem with one solar panel but if you connect two or more solar panels in parallel to the same solar charge controller to increase current, you’ll want to make sure the total current is lower than the charge controller’s maximum input current. Exceeding this limit will cause a properly designed charge controller to “clip” or “self derate” by lowering the input voltage and reducing the output power of the solar panels to less than the maximum possible under the current temperature and irradiance conditions. You’ll be losing potential solar watt-hours and you’ll be reducing the lifespan of your charge controller by running it too hot. A poorly designed charge controller may overheat and shut off or fail. Fortunately, this is easily fixed by adding a second charge controller and connecting the outputs in parallel.
The Genasun controller’s manual says you can go up to 9A of input current. The manual that came with my CTK-EV300 doesn’t indicate a maximum input current but claims “≤ 300W” power. If we assume this was measured at the maximum input voltage of 50V, that would indicate 6A max input current. This is consistent with anecdotal reports I’ve read that this unit tends to overheat if connected to more than 200 watts of solar panels. Your input voltage is likely to be lower than 50V so it makes sense that realistic working limit is closer to 100-150 watts, depending on your input and output voltage.
The panel’s maximum power must not exceed the battery’s maximum charging current
Plug-in chargers come with simple, straight-forward charge current ratings like “2 amps” or “4 amps.” This is made possible by the fact that the AC wall outlet (mains/grid) is a constant and effectively unlimited power source. Your solar charge controller is designed to always output the maximum power possible under the current solar conditions. These vary significantly over time and depend on the size of your solar panel so it’s not as simple to know the maximum output current. It’s up to you to ensure that the maximum output current does not exceed your battery pack’s ability to handle the current.
The highest sustained charge current will occur when the battery is at it’s lowest voltage and solar conditions are absolutely ideal. Even if your 100 watt solar panel typically outputs 80 watts, on a cold, sunny day you may see it sustain 100 watts long enough to be a concern if your battery can’t handle the extra input current. You need to plan for this scenario by sizing your solar panel to match your battery.
For example, RadPower ebikes all ship with a 2A charger so we know the battery can handle at least 2A of charge. However, their 48V 14Ah battery uses Samsung 35E cells. According to the 35E cell datasheet, these cells are rated for up to 2A of charge current per cell. With 4 cells in parallel, that gives a theoretical maximum charge current of 8A for the pack. However, the charge port is a relatively tiny 5.5mm barrel plug which certainly cannot handle 8A and the BMS most likely wouldn’t accept an 8A charge on the charge port. In this case, you could try going to 3A or 4A on the charge port and seeing if the plug gets too hot and if the BMS will cut off or will allow it. If that doesn’t work, you could try charging the battery through the discharge wires (see warnings above). Samsung recommends 1A per cell charge current for maximum cycle life so I would suggest keeping the “typical” solar charge current to 4A for this pack. Occasional spikes to 6A or 8A during regenerative braking and super-optimal solar conditions will probably be fine.
If you’re reading this far, you’re probably pretty serious about building a solar ebike. Hopefully, this has been helpful in selecting the right components for your project. Maybe you figured out that you need two solar charge controllers instead of just one or that you need a second battery to handle all the power from your large solar array. If you’re starting small with just 50-200 watts of solar panels then you’ll probably be fine with a single solar charge controller and your current 36V, 48V or 52V battery. You’ll find that most solar panels fit within the current and voltage limits outlined here. Things can get trickier as you add more solar panels so remember to review all the specs when you upgrade.
Solar panels aren’t magic: a guide to understanding the ratings
Let’s start with the worst case scenario: you bought a 100 watt solar panel to charge your 500 watt-hour ebike battery with the expectation that it would charge your battery from empty to full in 5 hours. What could be simpler? 100 watts times 5 hours equals 500 watt-hours, right?
You unwrap your shiny, new panel, plug it into your boost solar charge controller, plug that into your watt-hour meter (because you’re a smart cookie) and plug that into your ebike battery. Trembling with anticipation, you focus on the meter’s display but instead of 100 watts, it shows that you’re only pushing 70 watts into the battery. You remember reading somewhere that tilting the solar panel to directly face the sun is a good thing so you try that and the power goes up to 82 watts. That’s better. But almost immediately, it starts to go down again, 81 watts, 80, 79… finally settling in at 75 watts as the panel warms up in the sun. You’re thinking “What’s going on here? Did I get ripped off? This solar crap is a total scam! Somebody owes me an explanation!!” Ok, settle down. Let’s break it down.
Did you get ripped off by a dishonest purveyor of solar merchandise? Probably not. Did the seller’s marketing pitch set realistic expectations about the product’s performance under real-world conditions? Probably not. And that’s the problem.
That 100 watt label is a nominal power rating, variously referred to as rated power, PMP, PMPP, Pmax, or Pnom. A solar panel’s voltage, current and power output varies depending on the temperature of the cells and the irradiance level (sun brightness) so the solar industry had to decide on a standard set of conditions under which solar panels would be tested and labeled. The conditions they picked are called “STC” or “Standard Test Conditions” and are defined as an irradiance of 1000 W/m2 and a cell temperature of 25°C (77°F). These values were selected not because they represent typical outdoor conditions under which the panels will be used but because they are cost-effective when flash-testing each panel as it comes off the assembly line in a factory operating at a comfortable room temperature.
For decades, this caused relatively few problems. Consumer solar devices were limited to things like solar calculators and tiny portable lights where the solar panel’s rating didn’t matter to the consumer. Serious, big-boy solar panels like the ones on used on rooftops were installed by professionals who understood the measurement conventions and no one was upset or confused by the whole thing.
More recently, prices have come down and efficiency has gone up until suddenly small but relatively powerful solar panels can be had for US$ 1-3 per watt. Supply and demand being what they are, we now have a very competitive marketplace full of inexpensive solar charging products and equally full of marketing claims. What a time to be alive! The average consumer doesn’t want to study for an electrical engineering degree before buying a solar panel and the solar industry hasn’t been able to get away from that STC rating scheme because it’s just too well established. That’s how we get 100 watt solar panels that only give us 75 watts under relatively ideal conditions. It’s not a scam. It’s just a B2B standard in a B2C world.
Warning: math ahead. If math gives you hives, shortness of breath and/or flatulence, take a deep breath and go to your happy place. I’ll try to be gentle.
This is the label from a typical 100 watt solar panel. Notice how in tiny print it says “Standard Test Conditions”? This panel was tested to have these voltage, current and power values at 1000 W/m2 and 25°C. What the label does not tell you is how to figure out the voltage, current and power values at any other conditions. For that, we will need to look at the datasheet (PDF). I picked this panel because apparently Sunpower is the only semi-flexible panel manufacturer who can be bothered to publish a datasheet. This would never fly in the world of big-boy glass-and-aluminum framed panels with bankable 25 year power warranties, but I digress.
If your irradiance level is 500 W/m2 because it’s slightly cloudy then you might expect to get 50 watts from the panel (500/1000 * 100) but only as long as the temperature of the individual cells inside the panel is 25°C (77°F). Because the cells are almost black, they heat up significantly in the sun. On a warm, summer day when the air temperature is 25°C and we’re getting full irradiance at noon, the cell temperature is going to be somewhere between 45°C and 55°C, depending on air flow over the top and bottom of the panel. On a really hot day, it can get up to 75°C (167°F). Yes, you could cook an egg on that. I wouldn’t.
Let’s assume a cell temperature of 50°C which is 25°C above the standard temperature (50-25). From the datasheet, we see that our panel has a Power Temperature Coefficient of -0.35%/°C meaning that for every °C of temperature rise above 25°C we lose 0.35% of power. That’s right. Photovoltaic (solar electric) panels lose power as they get hotter. I know, it’s kind of counterintuitive but you get used to it. With a 25°C rise, that’s a loss of 8.75% (25 * -0.35%) or we can call it a 91.25% derate (100-8.75).
So, why aren’t we getting 91 watts (100*0.9125)? Odds are, we’re not getting the full 1000 W/m2 of irradiance. Even with perfect panel positioning relative to the sun’s position in the sky, the atmosphere filters out some light, more so when the sun is lower in the sky or when there’s air pollution. Let’s assume we’re getting 900 W/m2 because we don’t see any clouds in the sky and it’s the middle of the day. That would mean we should be getting 82 watts (100*0.9125*0.9).
Since we’re measuring at the output of the solar controller, we have to take into account any losses there as well. No charge controller is 100% efficient. A high frequency DC-DC converter like our boost charge controller is likely to be 85-95% efficient under typical operating conditions. Don’t be fooled by the marketing copy promising “99% peak efficiency”. We could measure the charge controller’s efficiency with a second watt-hour meter connected inline on the input side but let’s just assume it’s operating at 93% efficiency which gives us 76 watts (100*0.9125*0.9*0.93). The remaining 1 watt loss is in the connectors and wiring.
So, how long would it take to charge our 500 Wh battery using a 100 watt solar panel on a clear, sunny day? Lithium batteries like to be charged using a CC/CV profile meaning that the first 80% of the charge is at constant current and the last 20% is at constant voltage. That last part is much slower which is not the solar panel’s fault so let’s ask instead “How long to get to 80% state of charge?” Using our 75 watt example, that would be 5.3 hours (500*0.8/75) assuming you start a little before solar noon and keep re-positioning the panel to maintain the optimal angle. If it’s cloudy or you’re charging very early/late in the day, it make take quite a bit longer.
I’m leaving out some details here like that the battery’s internal resistance means that you have to push in a little more than 500 Wh to fill a 500Wh battery but I’m running out of steam and I’m craving a cookie so let’s call it done for now.
BONUS: If you’re matching a solar panel to the input and output limits of your charge controller and to your battery, you’ll need to take temperature into account. The VOC and VMP values on the label are just a starting point. You’ll need the Voltage Temperature Coefficient from the datasheet for those calculations. If there’s interest, I’ll cover that in part 2. Let me know in the comments. Don’t forget to like, subscribe and smash that notification bell! Oh, wait, this isn’t YouTube. Yeah, I suck at social media.
This is a solar bike manifesto primer for anyone who has an ebike and is considering adding solar charging. It’s difficult to write for all levels from novice to expert so this advice is most likely to be helpful to someone who has installed a DIY ebike conversion kit. If your experience is limited to riding a factory, turn-key ebike and you don’t do your own maintenance and repairs then you will want to get help from a friend or family member before trying a solar upgrade. Hint: find someone who has their own multimeter.
I have logged 50,000 miles (80,000 km) on ebikes over the past 13 years. About a quarter of those were touring and road-testing miles with various solar panels for charging the bike. I’ve built several solar ebikes over the years and I’m pleased to report that it is indeed possible to combine these technologies. Whether or not it makes sense to do so will depend on your goals and your budget.
Maybe solar panels are not for you?
If you’re only doing a couple of weekend ebike trips each year, my advice would be to skip the solar panels, borrow a friend’s battery and plug-in charger in addition to your own and find electrical outlets along the way. In fact, anyone who has reliable access to electrical outlets at the end of each day will find that carrying solar panels on the bike is less convenient and more expensive.
But I WANT to put solar panels on my bike!
Ok, I get it. You’ve seen photos and videos of solar bikes and you want to get in on the fun. Maybe you’re curious about solar power and want to extend your range? Perhaps you want to use solar panels as decorative plumage to attract a like-minded mate? It’s certainly a conversation starter. If you don’t enjoy being the center of attention wherever you go, this might not be the path for you.
I’ve come to think of solar upgrades as falling into two broad categories:
Range extenders: 50-200 watt solar panels to supplement your plug-in charging. Expect to get about 5 miles (8 km) of added range for every hour of charging under ideal conditions with a 100 watt panel. This is a great beginner project because it keeps the cost and complexity low while you learn the basics. You can always upgrade later. Folding panels can be stowed in your panniers and deployed when stopped or strapped to your bike/trailer to collect energy all day. Most trailers will fit a 100-150 watt panel easily.
Off-grid touring: 200-400 watt solar panels for ultimate roaming freedom away from electrical outlets. Expect to get 50-100 miles (80-160 km) per day. You will need to mount them on the bike so they collect energy all day which presents some challenges due to the large surface area. Recumbents and cargo bikes are popular in this category but I’ve seen some awkward attempts to attach this much to a conventional upright bike.
These range estimates assume you can manage to pack light and always pedal at a moderate effort. Expect to consume about 15 watt-hours per mile (9 Wh/km) while averaging around 14 mph (23 kph). If you’re riding uphill all day, into a headwind, in the rain, without pedaling, then your mileage will vary. These are long-term average values. You’ll get more on a sunny day, less on a cloudy day. If you’re unwilling or physically unable to pedal, cut the daily range estimates in half.
You can use any solar panel you want but “semi flexible” panels made with Sunpower cells for boats and RVs (campers/motorhomes) are your best choice in terms of power per unit weight and ability to withstand rough treatment on a bike. You can find them in all kinds of sizes on your favorite shopping site for around US$2 per watt and up. Traditional glass/aluminum frame rooftop panels are too heavy and should be avoided.
The size of your battery has some bearing on the efficiency of your system but does not determine your range in an off-grid situation. Assuming you’re trying to maximize distance traveled in a day, a bigger battery means you can take longer breaks before your battery is full, at which point you have to choose between getting back on the road or wasting potential solar energy because it has nowhere to go. That’s right, longer charge times are actually a feature. Conversely, an undersized battery (300-400 Wh) coupled with a large solar panel may run into problems with too much charging current for the battery cells or the BMS to handle. In that case, you’ll need to explore getting a bigger battery or using multiple batteries with multiple charge controllers.
Speaking of which, how do you connect your solar panels to your battery? You’ll need a “boost solar charge controller.” Just copy and paste those words into your favorite shopping site. You should find a couple of inexpensive Chinese models with MPPT for around US$30-75. A charge controller is a DC to DC converter which takes the solar panel’s output and converts it to the voltage needed to charge your battery. The “MPPT” business means that it automatically adjusts to finds the “Maximum Power Point” at which the panel’s voltage times current produces the most power. This varies depending on solar irradiance and temperature so we need to have Maximum Power Point Tracking to always extract the most power from our solar panel.
The output can be programmed in 0.1V increments to match your battery voltage. If you have the budget, you should get a Genasun boost controller for US$205. These are not programmable but are available in fixed output voltages. The “WP” version is fully potted which means it is waterproof and vibration resistant. They use higher quality components and run cooler, which should (theoretically) make them last longer. Because they don’t need a big heat sink, they are lighter. I’ve done my own performance tests, measuring watt-hours while stationary (unshaded) and moving (mixed shading) and found that the Genasun controller produced about 2% more over the course of a typical day while touring. Here’s a comparison video created by a fellow solar bike enthusiast.
Most solar charge controllers on the market today are made for lead batteries and can only be set to increments of 14.4V, corresponding to nominal 12/24/36/48V lead batteries. You cannot use these safely with your lithium ebike battery. Some of these unsuitable chargers may even state that they support “lithium” but upon closer inspection you may find that they only support multiples of 4 LiFePO4 battery cells in series which happen to like being charged to exactly 14.4V (4*3.6=14.4). Most other lithium cell chemistries need 4.2V per cell so you’ll need 42.0V, 54.6V or 58.8V for your nominal 36/48/52V pack. Your charger will need to be configured to the exact voltage your battery pack needs. If you’re unsure, figure it out before you plug anything into your pack. The labels on the charger that came with your bike and your battery itself are good starting points.
These controllers have PV input ranges which will work with most solar panels — just make sure that the open circuit voltage of your panel (VOC) is less than your battery voltage when empty (around 3.2V per cell) or you may find that you will not be able to charge when the input voltage is higher than the output voltage under some conditions (low battery on a cold day). Understand the specifications of your battery, charge controller and solar panel, keeping in mind that the solar panel voltage is lower than the label value when it gets warm. More on that here.
You may be able to connect two small panels in parallel but with larger panels that will likely exceed the maximum input current so you may need multiple charge controllers. There are trade-offs to be made: for example, a higher input voltage will result in better controller efficiency than a lower input voltage but connecting panels in parallel will give you better partial shading mitigation which matters if any part of your bike or body casts even small shadows on the panel.
If you have a Bosch, Yamaha or Shimano battery… I offer you my condolences. These closed, proprietary systems make it much more difficult to modify and enjoy your bike as you see fit. They’re well-engineered systems, designed to maximize corporate shareholder value and minimize liability and warranty claims. These vendors have no interest in helping you with your wacky solar modification project or supporting inter-operability with equipment from other vendors. If you’re doing pre-purchase research and solar charging is important to you then brands which reject open standards do not deserve your business.
I’ve read that you can trick the Bosch batteries into accepting a charge from a non-Bosch source by applying +5V to the signal pin and keeping charge current at 4A or less. If anyone knows a similar trick that will work with Yamaha or Shimano, please let me know. I know several solar ebike enthusiasts who charge using AC inverters on the bike using the charger that came with their bike and an intermediate 12V battery but these workarounds are heavy and inefficient. They should only be considered as an option of last resort.
I have written extensively about my solar conversions. I mention this as proof of real world experience in this subject matter and not as an example of a low-cost beginner project. Hopefully, my build will provide some inspiration for what is possible. You can do it. Start small and keep it simple. Add more later after you’ve had your first success. It’s not rocket science. If budget is an issue, you may find used solar panels on eBay, craigslist or your local equivalent. Or reach out to local solar installers or RV/boat supply shops and ask if they have any returned, blemished or damaged panels they’re willing to donate to you. Most of all, stop “thinking about it” and get out there and start doing something about it.
My current build has a 315 watt solar array good for around 80 miles (130km) per day. Just for fun, I recently did a 207 mile (333km) single-day ride using 784 Wh from grid-charged batteries and generating 2266 solar Wh so roughly equivalent to carrying six 500 Wh ebike batteries. Here’s a video.
If you want to learn more, I recommend watching the following presentation by someone who is far more knowledgeable than I.
I started this site as a sort of repository for links to information about this project with the idea that I would have a place to occasionally post things that didn’t fit elsewhere. The FAQ is freshly updated but most of the content I’ve been creating is where the eyeballs are: deep technical dives on the design and build process are on the Endless Sphere forum, some photos are on Instagram (@solarEbike) and there’s a YouTube channel with a couple of videos. Here’s one that has gotten a good response.
Justin from Grin Technologies came down for Maker Faire back in May and shot a video for their YouTube channel. I think it turned out to be a better introduction for this project than anything I’ve produced. If this is your first visit here, I recommend starting with this video. Enjoy.