Around the world on a solar ebike

[solarEbike]

UPDATE: This post is now obsolete. I've created a new version of this solarEbike Trip Planner which automatically imports weather data for any city in the world. The new version is available here: https://docs.google.com/spreadsheets/d/ ... edit#gid=0

I've pre-loaded the spreadsheet with weather data for Oakland, California so you will need to import data for your own city. I will add instructions for doing this in a follow-up post.

If you want Wh for a city other than Oakland, California, follow these instructions closely. It looks like a lot of steps but it's mostly just tedious. There's some repetition and it's relatively straightforward.

1. Make a copy of the spreadsheet: File > Make a copy...
2. Open to https://pvwatts.nrel.gov/pvwatts.php in a separate browser tab.
3. Enter a location in the search box. A city name is usually sufficient but be aware that the first Google Maps result will be automatically selected so you will want to specify "Vancouver, Washington" or "Vancouver, BC" and not just "Vancouver".
4. The SOLAR RESOURCE DATA screen usually has a map to help you confirm that the correct location was selected but the map is currently disabled. If the location looks right, select Go to system info to continue.
5. DC System Size (kW): 4 (do not change this, you will specify your panel size in the spreadsheet later)
6. Module Type: Premium (if using SunPower cells)
7. Array Type: Fixed (open rack)
8. System Losses (%): 0 (we will specify these later in our spreadsheet)
9. Tilt (deg): 0
10. Azimuth (deg): 180
11. Click Go to PVWatts results
12. On the RESULTS page, scroll down to Download Results: Hourly and download the hourly CSV file
13. Go to your spreadsheet and select the 0° fixed tilt tab at the bottom.
14. Select File > Import... > Upload and upload the CSV file you just downloaded from PVWatts
15. You will be asked for the Import location, choose Replace current sheet. This is the 4th option. Then, click Import data.
16. Go back to your PVWatts browser tab and click Go to system info to go back one screen. If you wait too long, the site may time out and ask you to select your location again. Don't panic. This is normal.
17. On the SYSTEM INFO screen, change Array Type: 1-Axis Tracking, all other values should be the same as for steps 5, 6, 8, 9 and 10 above.
18. Click Go to PVWatts results
19. On the RESULTS page, scroll down to Download Results: Hourly and download the hourly CSV file
20. Go to your spreadsheet and select the 1-axis NS tab at the bottom.
21. Import the second CSV file the same way as steps 14 & 15 above. Do not mix up the CSV files!
22. Go back to your PVWatts browser tab and click Go to system info to go back one screen.
23. On the SYSTEM INFO screen, change Azimuth (deg): 90, all other values should be the same as the last CSV file.
24. On the RESULTS page, scroll down to Download Results: Hourly and download the hourly CSV file
25. Go to your spreadsheet and select the 1-axis EW tab at the bottom.
26. Import the third CSV file the same way as steps 14 & 15 above.
27. Go back to your PVWatts browser tab and click Go to system info to go back one screen.
28. On the SYSTEM INFO screen, change Array Type: 2-Axis Tracking, all other values should be the same as the last CSV file. Tilt and azimuth will be disregarded for 2-axis tracking.
29. Go to your spreadsheet and select the 2-axis tab at the bottom.
30. Import the fourth (and final!) CSV file the same way as steps 14 & 15 above.
31. Go back to the Summary tab and check for any errors or broken formulas. You should see your new city name in row 17 repeated 4 times. If everything looks ok, your Outputs row should now be showing data for your new location.

You can save multiple scenarios by making copies of your spreadsheet. You don't need to repeat the import steps above unless you want to add a new city.

[Tommm]

If you had some test runs, what is your average speed and how far you go one day, and how many hours you spend moving?

[solarEbike]

If you had some test runs, what is your average speed and how far you go one day, and how many hours you spend moving?

I've logged about 2000 miles (3000 km) with various solar setups over the years, mostly short 2 day weekend camping trips in the San Francisco Bay Area between May and September. Average speed and distance varied a lot depending on the size of the solar panel I used, batteries I carried, availability of plug-in charging along the way, terrain, etc. My longest day ever was 144 miles (232 km), averaging 15 mph (24 kph) with 9 hours 33 minutes of moving time.

Michael Polak is a Sun Trip participant who tweets his daily numbers here: https://twitter.com/ArachneLabs

Does anyone else know about other sources with more of this kind of data? I know JLE uploaded log files recorded at 1Hz but are there any daily totals and averages out there for a long-ish solar ebike trip?

[solarEbike]

I've updated the solar modeling tool I posted earlier so that it now automatically imports weather data for any city in the world. Just enter your location, pick any date range and change the battery/panel specs to match your solar bike. Supports fixed and tilted arrays.

Use cases:

• Figure out the solar panel size you need to meet your daily travel goal.
• See if adding a tilt mechanism is worth the added weight, cost and complexity.
• Plan a solar ebike trip.

The new version can be found here: https://docs.google.com/spreadsheets/d/ ... sp=sharing

It can take up to 60 seconds to fully import and parse the data for a new city so please be patient while it loads. I'm looking for feedback on anything that needs clarification. I've tried to keep this as simple as I could without dumbing it down so much as to be useless.

PS: I've been playing around with it and it's telling me that the optimal design is to use dual-axis tracking whenever the sun is shining, store the energy in a 3kWh battery pack and only ride at night or when it is overcast. Some user discretion is advised.

[solarEbike]

I've had a few requests for details about the sun tracking tilt mechanism I built so I'll try to gather the key bits here in one place. This is not going to be a clear how-to guide with easy to follow steps. I've had lots of failures and a few successes and I'm still working on optimizing my design. If you're reading this because you're planning to build your own tracker be prepared to fill in some blanks. Post your questions here. I'll answer them if I can.

For starters, here's the trailer and tilt mechanism I'm talking about. In this video clip, I'm demonstrating the full range of motion which is about +/- 45° from horizontal. The tilt axis is parallel to the direction of travel. I'm controlling it with a DPDT toggle switch which simply reverses the polarity on the 12V DC motor of the linear actuator. I'll discuss automated tracking using an Arduino and light sensors later in this post.


I'm using a recumbent bike but there's nothing about any of this that is recumbent specific. The above trailer is my own design, optimized for low weight and a wide range of tilt travel. However, any bike trailer will work. Below is an early iteration of this design with no motorized tilt mechanism. This is a Burley Flatbed trailer with a couple of selfie sticks to prop up the panel. It worked just fine but switching the tilt from the left side to the right side was cumbersome.

The red support panel is 1" (2.5 cm) rigid building insulation foam covered with a layer of fiberglass and epoxy with aluminum inserts at mounting points for structural support. The construction is similar to how surfboards are made. It is reasonably lightweight and stiff but it was a lot of work to make. This might be the right approach for you if you're familiar with composites but it's a messy learning curve if this is your first such project. There are commercially available "sandwich panels" like this but they're generally very expensive as the lightest ones are driven by industries with big budgets: aerospace, racing, yacht building. Other solar bike builders have created supports with inexpensive aluminum extrusion products as ribs under the solar panel. I've avoided this approach because my ultimate choice of solar panel requires a fully solid support surface.

Assuming you have mounted your solar panel on a trailer or overhead as a canopy roof and it is free to pivot left and right, the next step would be to ask if motorizing the tilt is worth the added weight and complexity. Based on my experience, you should definitely be able to manually tilt your panel at least 45° for charging while stopped. Early and late in the day, the correct tilt angle can more than double your solar output. This makes a big difference for long distance solar ebike touring. You can accomplish this by leaning your bike against something while stopped.

Adding the ability to automatically change the tilt while riding can add 5-30% to your average daily solar watt-hours, although 10-20% is more realistic. The actual number depends on your location, time of year, travel direction, number of daily hours riding vs. stopped, number of cloudy days vs. sunny days (tilt plays a minor role when it's cloudy). Based on my experience, most solar ebike builders will be better served by getting a slightly larger solar panel and skipping the weight and hassle of a motorized tilt.

Me? I like a challenge and the cool factor of having a tracker was too tempting to pass up.

I used a linear actuator (this one). It's a 12V DC motor connected to a lead screw. They typically include limit switches which shut off the motor when the bar is fully extended or retracted. Some come with a built-in potentiometer which you can use to read the current position, although this isn't strictly necessary depending on how you plan to control the panel position. The one I used is not waterproof so that would be a problem on a long tour. It's also bigger and heavier than is strictly needed for a panel of my size. If I was doing it over again, I would consider using this one.

Finding the right anchor points for the two ends of the actuator was tricky. The way I did it in the video above works but the lower mount hits the ground during aggressive cornering. A shorter actuator would have allowed me to mount the upper end closer to the panel pivot axis. Also, the closer you can place your panel pivot axis to the center of mass of the moving panel the lower the forces on the actuator throughout its range of motion. I modeled this in CAD but ended up doing much trial and error with the physical hardware before finding the right mounting points.

Once you have your linear actuator mounted, you can simply control it manually with a DPDT toggle switch (double-pole, double-throw) like this. With the help of a handlebar mounted watt meter to measure solar output, you can make adjustments on the fly with the flick of a switch.

5322587f348316b35a1518cbcf0608ad.png (169.82 KiB) Viewed 3132 times

I took several test rides with this semi-manual setup but quickly got tired of making tilt adjustments by flipping this switch while trying to keep an eye on the road and the watt meter. Don't get me wrong. It can certainly be done. I'm just pointing out that this is a perfectly valid project end point for some. You have something that works and not having to stop and get off the bike to adjust the panel tilt angle is rewarding enough to justify the trouble you've gone to thus far.

The next upgrade step involves connecting sensors to an Arduino microcontroller and writing some code. If this sounds intimidating or you have no idea what I'm talking about, don't worry. I had never attempted something like this before and picked up everything I needed to know by watching YouTube videos and reading tutorials. However, I had quite a bit of experiencing writing code in JavaScript and Ruby before I attempted this so learning to write a "sketch" (programming instructions) in the Arduino language wasn't a big leap. If you've never written a line of code you're probably going to want to recruit some help for this part.

The automated tracking algorithm is very simple. Two light sensors are mounted at 45° angles to the solar panel. The Arduino reads the light input from the two sensors. If one sensor is getting more light than the other, the motor moves the panel toward the light until the sensor values are the same. Here's how I mounted the sensors. There are two pairs shown. The embedded ones in the panel are nicely protected from water and physical damage but I made the window a little too small so I'm trying a second set (with black heat shrink tubing).

The Arduino looks like this. The white breadboard is temporary. Once I get everything dialed it, I'll move it to a perf board and pot the whole thing.

I got most of the parts for my automated controller from Adafruit.com because they have great tutorials and I like the way they support the Arduino open source platform.

Parts list:

• Photo Transistor Light Sensors (2)
• Adafruit Metro Mini
• Motor Driver Breakout Board to turn the linear actuator on/off/reverse
• Adafruit Bluefruit LE SPI Friend (optional, enables connection to smartphone)

This is the part where I cross my fingers and hope that you got enough from this to finish on your own. I muddled my way through by reading the product descriptions above and then following all the tutorials on each page until I got light activated solar panel positioning. My code is an ugly mess so I'm reluctant to share it and the panel occasionally gets stuck facing away from the sun but it mostly works.

UPDATE:

I just remembered that you can add automatic sun tracking without an Arduino or any coding. There are several ready-made solar tracking controllers available. They're designed for stationary PV systems with single axis tracking via a linear actuator so they should just be plug and play. The light sensors are all a little too big and clunky for my liking so I build my own but I imagine these should work just fine on an ebike.

Here are some links I found. If these have expired by the time you're reading this just try a search for "solar tracker controller" on your favorite shopping site.

• https://www.ebay.com/itm/TinyTracker-HD ... 3962898243
• https://www.amazon.com/SZMWKJ-Tracker-T ... 4449&psc=1
• https://www.amazon.com/Tracker-Controll ... NAL+CO+LTD

There's even a complete kit with linear actuator on AliExpress right now for $110 US including shipping. Will it stand up to thousands of miles of dust, rain and vibrations on a solar ebike? Probably not. Will it get you up and running quickly so you can figure out if sun tracking is worth pursuing further? Possibly... assuming it doesn't arrive incomplete, broken or totally different from the specs and photos. (Sorry, AliExpress hasn't worked out for me in the past...)

PS: I just watched a YouTube video for one of these products and found a potential show-stopper: there's a 3 minute delay between position updates. That's reasonable for a stationary system but way too long for a mobile system. Check with the vendor before ordering. On my own solution, I use a 3 second delay. (Also, I need to add some additional checks to ensure I don't exceed the actuator's duty cycle.)

[wturber]

Cool. Thanks for detail all of this out.

[amberwolf]

FWIW, you can recycle old Lazy-boy chairs / etc for their actuators. I have a few of them from one here, though I don't know offhand if they're long enough throw to do what's necessary in this application.

[solarEbike]

A little progress update. After ten years of stubbornly refusing to explore overhead mounting options, I finally decided that having a small roof over my head would have enough value in keeping the sun and rain off my head to offset the perceived downsides: high center of gravity, aero drag, cross wind problems and obstructed view of the road and scenery. Here's what I came up with:

I made it just big enough to fit one of my four solar panels (4x6 SunPower cells). The other three are still going on the trailer. The tilting mechanism can do a full 90° tilt for charging while parked and about 25° while riding. Yes, that Ferrari totally photobombed my perfectly composed shot.

I can reach the indexing plunger and make adjustments single-handed while riding. The 8" aluminum ring is a turntable bearing (table-top lazy susan). The rest is custom machined aluminum, carbon fiber tubing and stainless steel fasteners.

Welcome to California. Weather today: Smoke. Blue masking tape is temporary while I dial in the optimal tube lengths. The aluminum end caps will be held in place with 3M Scotch-Weld 2216 Epoxy. Stainless steel hinges on top to allow flat packing for air transport.

It's a little more bouncy than I hoped but tolerable, I think. I plan to route the wires inside the carbon tubes and add some slim profile LED lights to the roof to increase my visibility on the road.

[thundercamel]

I like the Lazy Susan as a starting point. Nice work

[Tolt]

You've got a lot of nice, custom aluminum parts on your builds -- and a tiny shop (as do I). Are you doing the machining yourself? What machine tools do you have?

[solarEbike]

I like the Lazy Susan as a starting point. Nice work

That was a fortuitous discovery. I started out trying to design a pivot mechanism from scratch, went browsing on McMaster-Carr for bearing ideas, saw a bunch of very expensive and heavy slewing bearings and then stumbled across this $20 solution on Amazon. It makes these weirdly brittle chips when I drill it so it's not the 6061 alloy I'm used to. I'm rolling the dice here that it will hold up over time.

You've got a lot of nice, custom aluminum parts on your builds -- and a tiny shop (as do I). Are you doing the machining yourself? What machine tools do you have?

Good question. My workshop is a one car garage. I watch a lot of YouTube machining videos but I've never used a lathe, mill or anything CNC. I've looked into becoming a member at one of my local maker spaces to get access to some fun toys but never quite got around to it. I've been able to drill, cut, shape and polish aluminum mostly using a few low-tech low-budget woodworking tools and a lot of pig-headed perseverance.

Right to left:

• Drill press (donated) with a 6" cross-slide vise ($75 at Harbor Freight*). You can do a lot with a drill press and a wide assortment of drilling and cutting bits. Step bits are your friend.
• Band saw ($130 at Home Depot). This is a woodworking tool but it cuts aluminum with ease.
• 4" disc sander ($75 at Harbor Freight*)
• 1" belt sander ($53 at Harbor Freight*)
• 8" (200mm) digital calipers ($54 on Amazon)
• The purple rectangle in the center is a sheet of sand paper glued to a piece of MDF with spray adhesive. Very effective for polishing and shaping small parts.
• The perforated white tabletop under that is a home-made downdraft table. Attach the shop vac to the box under the table and sanding particles go down instead of covering the entire shop. Pretty sweet for nasty fine particles like carbon fiber.

* Harbor Freight has huge sales and crazy coupons so I actually paid significantly less for most of these. Also, I got these years ago.

I usually start by taking a few measurements, make a rough drawing in SketchUp, print out a 1:1 scale paper copy, stick that to some aluminum stock with spray adhesive and start removing material. A lot of my problem solving happens during the machining phase so it's nice to be able to make changes on the fly which might not be so easy in a CAD/CAM workflow without starting over.

Pro tip: You can import McMaster-Carr 3D components directly into Fusion 360 right within the interface. If you use Fusion 360 and haven't used this feature it will change your life.

[wturber]

Years ago I made a 72 tooth chainring out of 6061 T6 aluminum using a bandsaw, drill press, a wooden jig, and hand tools. You can do a lot with a basic woodworking setup and some aluminum. :^)

[solarEbike]

I’ve been watching 3D printing from the sidelines for a few years. Frankly, most of the output has left me underwhelmed: tacky keychains printed at super low resolution, desktop hobby printers which take too much effort to achieve acceptable results, cosplay enthusiasts who spend hours priming and sanding parts to hide printing artifacts, and low strength parts unsuitable for serious real world use. In short: meh.

However... I had a couple of design ideas recently which looked like they might be good candidates for SLS printing (what’s that?). Now that I’ve received my first parts from an online printing service, my initial impressions are that the parts look great and came out as well as I could have hoped. Time will tell if they are strong enough to withstand the abuse I plan to inflict upon them but I’m cautiously optimistic. The best part is the realization that if I break a custom part while I’m on the road for three years, I can update the design and order a replacement. I can at least pretend to be less apprehensive about losing access to my workshop.

The custom CA3 hood above was my second print order. I designed this in Fusion 360. Love that Loft tool. My goal was to hide the octopus of connectors in the back of the CA3 and provide some waterproofing as I anticipate the bike will see a lot of rain. I also needed a place to mount the Feniex Apollo unit I’ve been using as a daytime running light. I made the enclosure big enough to fit the Grin GPS Analogger as it prefers a short run for the data connection to the CA3. The other cables are routed through a 2” (50mm) polycarbonate tube attached to the front derailleur frame tube. The crank clearance is just a few millimeters so I was relieved that my measurements were spot on.

The small brim over the screen was supposed to reduce glare and improve daytime visibility but road testing cardboard prototypes revealed that it would need to be much longer to be effective and I just didn’t like the way that looked.

I have a newfound appreciation for people who design plastic parts for a living. The SLS printing process practically eliminates design constraints inherent in other manufacturing methods: no need to worry about undercuts, supports or minimum draft angles. It’s pretty idiot proof beginner-friendly. Draw the shape you want, upload to Shapeways and open your wallet. US$103 for this part… so not exactly cheap.

My first print order was a little simpler. This is the new motor assembly for the trailer tilting mechanism. It’s about half the weight of the previous linear actuator and fits inside the trailer boom tube to protect it from dust, water and crash damage.

The two gray pieces and the white end cap are 3D printed. I designed these in SketchUp because I’m much more proficient with it but it’s not really the right tool for the job. I spent as much time cleaning up unwanted geometry as I did actually drawing the part. Some fillets would have been nice but I made everything chunky enough and the walls of the tube help hold the part in place.

In the end, they printed just fine. The dimensional accuracy was spot on. I designed pockets for captive stainless steel nuts and they fit nice and snug. The inner diameter fits the motor perfectly and the outer diameter slides into the carbon tubes with just a little play. The gray pieces will be subjected to significant torque so I had those printed in their stronger “professional plastic” option at about twice the cost of the standard stuff. Total cost for these 3 pieces with tax, shipping and a “priority fee” to get it in 2 weeks instead of 4 weeks: US$97. The second order arrived in exactly 7 days without paying rush fees so the 4 week turnaround estimate on the first order may have been due to my bad timing right before Christmas.

Final thoughts? While I’m still not interested in acquiring a hobby 3D printer, some of the hype is justified. This will make a valuable addition to my bag of tricks.

PS: Love ebikes and 3D printing? Check out Tom Stanton on YouTube. Mad lad is printing drivetrain components with PLA.

[solarEbike]

I finally completed the solar roof panel and took it out for a test ride.

I put a world map with my route on the bottom of the panel

Some serious lighting, including running lights, brake light and turn signals (not shown here).

Progress shots of how the encapsulated PV module is attached to the carbon fiber honeycomb panel. GE Silicone II using mesh screen to apply a thin, uniform layer of adhesive "dots" over the entire surface. The sand is to apply even pressure while the silicone cures for several days.

Here's how I encapsulated my "junction box" for the PV and lights. Wires are routed inside the carbon fiber support struts.

[fechter]

Looking good. Did you have any wind when testing? How it behaves in a cross wind will be important. I suppose if it gets windy, you just flatten out the panel to minimize it's wind profile.

[solarEbike]

Looking good. Did you have any wind when testing? How it behaves in a cross wind will be important. I suppose if it gets windy, you just flatten out the panel to minimize it's wind profile.

A lot of my local riding is right next to open water along San Francisco Bay and on the new SF-Oakland Bay Bridge bike path where it's always windy. The wind has never been an issue under these conditions. The panel is 8 mm thick and the roof struts 12 mm so the wind profile is minimal when the panel is level. Also, the roof tilt can only go about +/-15° while riding because the roof struts hit my helmet. The full 90° tilt is something I only expect to use when parked and possibly on smooth roads in remote areas while riding without the helmet.

I haven't been able to test more extreme conditions like the winds I'll encounter on the Altiplano in South America or gusts from trucks passing close to me at high speeds. For these extreme cases, the backup plan is to detach the roof panel and strap it to the trailer which consists of 3 equal sized panels connected by hinges so they can fold for transport. I would only do this if couldn't ride any other way as I won't get any solar generation that way.

The kickstand shown above is a different matter. As light as it is, the roof panel changes the center of gravity so much that any slight gust can tip the bike while parked. Early on, I imagined I might need to use a couple of tent stakes and some guy lines to keep the bike from tipping over in the wind while parked overnight but now I see I'll need something that can be deployed every time I stop. I'm currently working on something similar to the Click-Stand except I'll put one on each side of the bike along with a parking brake for maximum stability.

[solarEbike]

Here's my mobile office setup. I have my 6 LiGo batteries held together in two groups of three with long M4 screws on opposite corners as indexing pins. This makes it easy to leave three on the bike to continue solar charging while I set up office in the shade nearby. A custom plug connects the three 36V batteries in series to give me 108V DC on a standard North American outlet.

The 29 watt MacBook power adaptor doesn't care that it's getting DC instead of AC as it would just rectify the voltage to DC anyway. Neither does the 110 volt, 500 watt immersion heater. It boils water for coffee, tea or instant noodles in 4-5 minutes using about 40-50 Wh of energy. That's not great when I think about how many miles I could travel on those same Wh but it's a lot more convenient than setting up the stove on a cold morning or in the middle of the day.

Yes. That is a ceramic burr grinder. I'm not an animal.

Justin explains this use of the LiGos here.

[bakaneko]

Nicely done my friend. I also experimented with a solar bike with my 500W gearless rear hub motor and 3x 30W solar panels logging about 2-3 Wh/mi at about 12-13 mph on average. It was very very fun and got many conversations started. I decided to retire it for now but I definitely want to redo it but with satellite grade (35%+) solar panels and at least 150W so I can get an average of 100W. A semi recumbent seems like a good fit for it. Here is the bike.

[solarEbike]

... I definitely want to redo it but with satellite grade (35%+) solar panels and at least 150W so I can get an average of 100W.

That's a nice looking setup. Yeah, you strap solar panels to a bicycle and you're going to get some reactions from people.

If you find someone willing to sell you 150W of 35% efficiency PV cells, please let me know.

[Cephalotus]

... Sometimes it's hard to tell if I'm going too deep and then I remember this is Endless Sphere and there's no such thing as too deep. Still, the view counts don't tell you much about how many people find your post relevant or interesting...

I do find your data, information and homepage SUPER interesting and I'm planning to build a solar trailer, too, but I'm way behind your capabilities, so I can only copy what I can make myself.

[rowbiker]

Add me to the list of people following this thread closely, and learning tons from the extremely well written and illustrated material. I'm not doing solar on a bike, but the technology your experiments illustrate are widely useful -- and highly appreciated! We who learn salute you.

[solarEbike]

Thank you, rowbiker and Cephalotus, for the words of encouragement... and the reminder that I haven't posted in a while! Here's an update on some odds and ends.

Following up on previous discussion about the legal implications of operating an ebike in different jurisdictions around the word, I took my hub motor to a local laser engraver and asked them to put "250W" on the side plate to make this whole contraption a little more EU and AU friendly. I think it turned out pretty well. This was done using a 120W laser and CerMark marking paste on bare aluminum.

My phone is rated IP67 so it should be fine in light rain but I want to be able to use it for navigation in all kinds of weather while keeping it plugged in. I couldn't find a waterproof case that allows charging while wet so I modified one of these US$10 pouches. The bottom edge is completely open but extends far enough to (theoretically) keep splashing water away from the charging port

rain cover.jpg (292.97 KiB) Viewed 2178 times

The traditional kickstand wasn't adequate for keeping the bike from tipping over during stationary charging, particularly when the roof panel is tilted to one side and it's windy so I needed something more robust. Using spare materials I had laying around from other projects I improvised these "landing struts".

landing struts.jpg (453.53 KiB) Viewed 2178 times

They're nice and light. There's a shock cord inside which helps them deploy quickly.

My original 3D printed CA3 enclosure took a hit and cracked when my new boot caught the front edge while pedaling. Something about pedaling through a tight radius turn and rotating my ankle... anyway, it happened a couple of times and could easily cause a crash so I revised the design to make it as narrow as possible, made a few tweaks and ordered a new print.

The updated design is more boxy and less rounded. The lip on the open end is reinforced to withstand more abuse.

This is the raw print before painting.

Sparkly paint this time because I secretly identify as a pre-adolescent girl.

CA3 enclosure 2.jpg (249.94 KiB) Viewed 2178 times

Two switches mounted on the bottom. One to turn the CA3 on and off independently of the motor controller so I can log stationary solar charging while the controller is powered off.

CA3 enclosure bottom button.jpg (284.91 KiB) Viewed 2178 times

And an on/off/on SPDT rocker switch to control the backlight: default, red and off. (Wiring info here and here.)

Here's the view from the cockpit.

• Left handlebar: CA3 remote buttons, left turn signal switch, 12V motorcycle horn switch, low/high beams, brake lever with TripWire switch to control regenerative braking, Rohloff (internally geared hub) shifter.
• Right handlebar: right turn signal switch, motor controller on/off switch, CA3 digital Aux button (to set level of motor assist), brake lever with TripWire switch to activate brake light, Domino throttle, phone charger.

The switches are these "Harley" after-market replacement switches. They're a tad heavy but seem to be well made and look like they're sealed against water. The switch housings are big enough that I was able to add a couple of momentary switches to remotely control the CA3. The wiring is kept tidy by running everything through two sunlight-resistant 12 conductor cables I sourced from McMaster Carr.

cockpit view.jpg (446.96 KiB) Viewed 2178 times

[rowbiker]

This is all super-ingenious, and I think easily qualifies as 'out-of-the-box' engineering. I just watched Justin's presentation (video) about ebike motors at the Vancouver bike fair, in which he mentioned mfgr's claims about their motors' rated wattage. I wonder what he'll make of your brazen "250" label! I'll be curious to hear about your experiences with this approach.

[solarEbike]

I just watched Justin's presentation (video) about ebike motors at the Vancouver bike fair, in which he mentioned mfgr's claims about their motors' rated wattage.

Good stuff. I just watched that video yesterday. For those who haven't seen it yet, the part where Justin discusses motor ratings starts at 19:47.

I wonder what he'll make of your brazen "250" label!

I have a feeling he'll be fine with it.

I'm neither a lawyer nor an electrical engineer but my understanding of the legal wattage limits is that they're intended to represent the continuous operation of the motor, not peak. But at what RPM? And at what ambient air temperature? Does the addition of Statorade to my motor change its rating?

In reality, my priority of distance over speed, my 36V battery pack and my 20" wheel size mean that I typically average around 250 watts on local test rides (see below). It's not the same as continuous rating but it's close enough for me.

The relatively steady 600W sections are my recent preference to use basic PAS for hill climbing, torque PAS for all other times and the throttle for extra acceleration around traffic as needed. I switch between basic and torque PAS using CA3 presets.


"Officer, are you asking why my motor is so much bigger than all the other 250 watt motors? Well... um... it's because this one has regenerative braking!"

[solarEbike]

Looking good. Did you have any wind when testing? How it behaves in a cross wind will be important. I suppose if it gets windy, you just flatten out the panel to minimize it's wind profile.

Here's some anecdotal wind data. This was a particularly blustery day on the San Francisco-Oakland Bay Bridge bike path, high above an open body of water. I don't know what the wind speeds were but the signage on the bridge warns drivers about the wind conditions.

This kind of wind represents less than 1% of my riding experience so far and it was completely manageable. The roof flopping around was annoying but the effect on bike handing was minimal. It would have been worse if I didn't have a steel barrier between me and the high speed vehicle traffic next to me.

Also, the roof moved much less as soon as I got off the bridge. I suspect that part of the problem here was turbulence caused by wind passing over the bridge structure whereas on land the flow is more laminar and parallel to the ground... except when vehicles pass next to you.




If the flexing of the roof struts gets unbearable I may replace them with stiffer, larger diameter tubing in the future.