Radwagon 72v GMAC Phaserunner Build

Hello,

This is my first post on Endless Sphere. I've been reading and learning a lot from posts here and I thought it was time to share my first ebike build. To give a little background on myself, I am a former competitive cyclist who raced road, off road, and track (though never in danger of winning any of them) back in the '90's. I no longer race, am nearing 50, and live in a very hilly portion of Seattle. My riding now is taking my 8 yo daughter to and from school and some general commuting and running errands.

I own one of the very early Radwagons with the steel frame and have put a few thousand miles on it. Generally I've been quite happy with it. My daughter and I enjoy riding the bike together and I often find with Seattle's traffic I can get places faster on the bike than by driving. The only real issue I've had with the bike were loosening spokes on the rear wheel early on, but the Rad guys tuned and tightened them and they've stayed tight since then. I would also occasionally have to retighten the rear wheel bolts as they would loosen due to the power/regen cycles and no torque arm.

As much as I like the Radwagon, I am by nature a modifier and customizer, and thus decided to comprehensively upgrade the bike this winter. My goals are to have more power for getting up hills (there are some hills here in Seattle I can't pedal up with my daughter on the bike, even with the Radwagon on full power assist), more top speed for when I'm in a hurry or need to ride in traffic at speed for safety reasons, and to have a bike the feels good to ride. This last point means I want actual torque sensing and not just cadence sensing. One of my frustrations with the stock Radwagon is its lack of TS. To me, it makes the bike feel unresponsive and "flat".

I looked at mid drive setups, but after a lot of reading and research (a lot of it here of course) I decided I wanted to stick with the hub drive for simplicity, minimizing drivetrain stresses, and to keep regen braking. When I came across the Grin GMAC hub it seemed like a good fit for my needs. Enough power, regen braking, and integrated torque arm all fit the bill. So I specced out a 10T motor in a 26" wheel, Sempu TS bottom bracket, and Phaserunner with a 72v battery. With this combination I should have roughly 4x the torque of the stock Shengi motor on the RW, and a top speed of 35mph when I need it. I'll have throttle and the TS for true pedal assist.

I am a long term hot rodder and wanted this build to be clean and functional with minimal excess wiring or visual clutter. In researching batteries I found the Luna Cycles 72v 10AH fat triangle (https://lunacycle.com/72v-triangle-panasonic-GA-10-5ah/) which seemed like it would just fit in the rear triangle of the RW behind the seat tube. I knew if it didn't fit there, it would easily fit in the main triangle. So I ordered one up.

When the battery arrived I found it did just fit in the rear triangle. I could even still run the front derailleur if I wanted to, but I decided to simplify by going to a 1x10 drivetrain, so the derailleur was removed. The stock RW mounts its controller on two steel cross members above the rear triangle. I used those to mount the battery to the frame. I used a 0.25" thick aluminum plate that counter sunk allen head bolts into and adhered to the top of the battery using 3M 5200 marine sealant/adhesive (very strong stuff that doesn't let go EVER).

With the battery bracket made, I then wrapped it and the battery in a custom plastic case. I used Worbla Thermoplastic (https://www.eplastics.com/sheets/worbla?gclid=EAIaIQobChMIycnal_aj5gIViPhkCh2TuAqQEAAYASAAEgK0nvD_BwE) to wrap and seal the battery. Worbla is a great material for a fabrication like this, as with a heat gun it molds and (somewhat) adheres to the battery and (absolutely) adheres back to itself. Thus I was able to make a water tight case for the battery that barely increased the overall dimensions.

Here is the battery bolted into the frame.

In keeping with my goal to make this build as clean and clutter free as possible, I decided to make a small aluminum housing to mount above the battery in the place of the stock controller to house the new Phaserunner controller, the assorted wiring connections, and a DC to DC step down converter (to take the 72v battery power and give me 12v power for lights and to power the Sempu TS bottom bracket). Since the Phaserunner needs to shed the heat it generates to provide maximum power, I figured I could mount it upside down in the enclosure, so its bottom (which is its heat sink) would sit flush against the top cover of the enclosure and I could use the top cover to shed the heat. Thus I made the top cover out of an aluminum heat sink with fins for maximum heat transfer.

Here is the enclosure painted and test fit on the bike with all components and wires in place. Notice I incorporated an XT60 charging port on the left side of the enclosure for easy charging. The DC to DC converter is the black box in the very back of the enclosure. To ensure solid contact between the controller's heat sink and the heat sink cover of the enclosure, I have it spaced properly so when the top is bolted on, it is compressing the Phaserunner between the heat sink top and the battery (where the aluminum plate is). To further maximize heat dissipation from the controller, I'll use thermal paste between the Phaserunner and the top to make sure there are no air gaps. I don't know how well this will all work, but I think it'll be better than just strapping the Phaserunner to a frame tube.

Since the battery has no internal power switch, and I won't be plugging and unplugging the battery to depower the bike, I made use of the optional wiring on the Phaserunner that allows a remote switch to power the controller on and off. The switch I chose is a little push button that I mounted on top of the seat post between the seat rails. I made a little bracket that holds it there and then dipped the entire back of the unit (the front has a waterproof rubber cover) in plastidip to waterproof it. The wiring runs down through the downtube and exits the same hole in the bottom bracket shell that the wires for the torque sensing bottom bracket exit through. The wires are long enough to allow adjustment and removal of the seat post, and there is a plug to disconnect it (not shown in pictures) for full removal of the seat post when needed. From the bottom bracket the wires run up to the controller in the enclosure. It s a pretty stealth switch location that allows easy access for turning the whole shebang on and off (and emergency depowering in case of runaway condition).

For mounting the Cycle Analyst on the bars, I didn't like the stock mounting brackets. My current stem has a split shell clamp to hold the handlebars, so I drilled through the top shell and countersunk a bolt flush from the inside and mounted the smallest portion of the stock mounting bracket to that. Since the top of the stem shell is rounded and the bottom of the bracket is flat, I used a little JB weld to fill in the area between the bracket and the stem to further stabilize it. It allows the CA to mount relatively low on the bars and centered.

The stock Radwagon has a hard wired headlight, but its not very bright. I had already replaced mine with a Busch and Muller IQ-X, but I was previously still running the stock battery powered tail light. In an effort to increase safety and keep it simple (I'm a big fan of the KISS principal), I decided to have my lights always on whenever the bike is on. For additional visibility day and night, I acquired a hard wired 12v LED tail light and some white and red waterproof LED strip lights. I mounted the white LED strips on the front of the fork legs and the red LED strips on the vertical rear rack tubes. The strips come with 3M tape mounted to their backs, so I just cleaned the tubes with acetone and stuck the light strips to them. All are hardwired to the DC to DC converter 12v output and draw 16 watt total (7.5 watts for the headlight, 1.5 watts for each of four 12" long LED strips, and 2 watts for the tail light). The converter itself is tapped into the power feed for the Cycle Analyst from the Phaserunner. Per Grin, that power lead is fed through a 1.5 amp polyfuse in the controller, so at 72v there are 108 watts available. The lights will take 16 watts, and the cycle analyst and the Sempu bottom bracket will draw about half a watt each. So I should be well under the limit. Having the DC converter piggybacking on the feed to the Cycle Analyst means it also gets turned on and off with the controller, so there should be no power draw when everything is shut down.

Wow, just wow!

Absolutely love the craftsmanship and great ideas displayed on this build. Nice job so far!

Thank you. I'm pleased with how it is coming together. My GMAC 10T rear wheel should arrive this week (it was on backorder for a while) and I am still waiting on my new brakes (Shimano MT-520 hydraulic four pots), components (Deore M6000), and cranks (Prolite 50T with narrow/wide teeth) to arrive as well. Then I can get the project on the road and tested. Hopefully more to come soon.

I’m so thankful for you sharing with us. Look forward to more updates.

I have a 2018 Rad Wagon and have been dreaming up what electrical upgrades would work. I ride the cargo bike for similar reasons but I’ve got two kids to haul around. The Rad Wagon is a solid bike. I also take mine on some of our gravel roads and it performs really well.

Do you have any battery life/distance/time estimates with this setup?


Sent from my iPhone using Tapatalk

Teampeters, what kind of electrical improvements are you considering? More torque for hill climbing? More speed? Range? I’ll guess the latter is one goal since you asked about my anticipated range with my rebuild. My RW is an early bike with an 11ah 48v battery. Thus it only had about 528 kWh of total battery capacity. I never maxed out my range and ran out of juice altogether and the battery meters are a little fuzzy on the RW, but I could get over 20 miles running about 20-22mph from full battery down to one bar running on mostly flat (relative term for Seattle). My new battery has about 720 kWh of capacity, or about 36% more than my old battery. Thus running the same 20-22 mph on the same route, I should expect to get at least another 7 miles of range. If I run faster, it’ll be less.

The real benefit I expect to see is greater torque for climbing hills and greater efficiency when doing so. The stock Shengi direct drive motor is not efficient when lugging slowly up a hill. So I would burn more power to get up those hills on some of my regular routes than I expect to with the new geared motor.

I hope this answered your question.

So, this is a big post for a tiny project. Still, I really like these kinds of little details.

As previously noted, I installed a charging port in the side of the controller housing I made. The only panel mount XT60 plugs I could find were the ones with the male pins (see pics below for example). The BMS on the Luna battery has battery voltage available at the charging port at all times, so if something metal happens to touch those two pins, it'll short. I don't think the BMS allows a lot of amps to the charging port, but I'd prefer to prevent any inadvertent shorts there anyways and keep the plug clean and dry. I looked for a protective/weather cover for this type of panel mount XT60 connector, but I couldn't find any. So I had to make my own.

I used a spare female XT60 plug and removed the metal receiver cylinders from it using a little heat from a soldering iron and pulling them out with pliers. I did this to make the plug slide in and out of the panel plug port more easily. With the cylinders removed I used a little belt sander to shave off excess material so it'll sit about .25" proud of the panel when in place.

I wanted a "leash" for the cover so it can't get lost and is always at hand after I'm done charging. I ended up using a piece of big 9mm wide wire tie (zip tie). I dilled a hole in one end, used a heat gun to have it bend and hold the U shape I wanted that works to hold the weather cover into the plug socket, and then used Gorilla Super Glue to bond it to the top of the modified plug (super glues are great for bonding plastic to plastic!). I left a little of the wire tie extending beyond the end of the plug to make for a convenient lift point for removing the weather plug. I then plastidipped the assembly for a uniform look. I think it should work pretty well.

Great job; I got tired just reading the details.Will be interesting to see performance data.

I received the wheel from Grin on Monday (GMAC 10T, double wall rim, biggest spokes). Today I was able to take it out for a quick test ride. I am still waiting on my 1x10 components and don't have a 7s cassette to allow me to use my old derailleur and shifter, so I don't have a chain or gears on it right now. Thus this was a quick throttle only test ride. Running up the hill in front of my house (maybe a 4% grade) the bike had not trouble accelerating up to (and past) 25+ mph. I backed off the throttle before it was done accelerating. The Cycle Analyst was only showing about 1500-1600 watts being used even under full throttle going uphill. Max wattage should be closer to 2900w (40a at 72v). It was raining, the roads were wet and slippery and it was getting dark, so I didn't push it or do multiple runs. The return run down hill saw me above 30mph before I knew it. No question this setup has WAY more juice than the stock RW.

Issues to resolve:

1. Regen braking is not working. The brake lever switch is working and the settings in the cycle analyst seem to be correct, but I'm not getting any braking force from the motor. I've got an email into Grin to help trouble shooting this.

2. Why wasn't I seeing more wattage under full throttle? Not sure and will need to look into this more. Is there a separate setting that limits max acceleration? I put the question to Grin.

3. The GMAC has a built in temp sensor, but the CA was not configured to use it. I want that thermal protection and will activate this once I identify the correct settings. I think the correct settings are threshold for rollback at 110c and maximum allowed temp 130c.

Assorted notes:

The stock Shengi motor/rear wheel of the RW was 14.0 lbs without tire or tube. The new wheel with GMAC motor was 11.0 lbs without tire and tube, but including torque arm. So three pounds lighter.

My new battery is just under 9.0 lbs, including my aluminum mounting plate and plastic case. The old 48v 11ah Dolphin case and frame mounting plate weighed 7.7 lbs.

I'm not going to weigh all my accessory components as this is not a weight weenie build (a steel cargo bike would not be the best choice for that, but I'm confident I'm loosing weight on this build while getting about 4x the power and torque, and roughly 36% increase in range. I'll take that.

The GMAC wheel installed with no issues. The integral torque arm is pretty slick. I definitely had issues with the stock RW rear wheel loosening periodically as the throttle/regen cycles cause it to work lose over time. I'm guessing that won't be a problem with this torque arm.

I'm pleased with the general appearance of the bike. I'd even go far as to say that without the "caboose" (the kid containing railing around the load deck) and the running boards, the bike has good lines and looks fairly cool. I'm not going to compromise my daughter's safety and comfort for looks though, so those items stay as long as she still wants to ride/be seen on the bike with me (and maybe a while after when she doesn't). The wiring is pretty clean without much "kludge" anywhere. The handlebar wiring can still be streamlined more, but I'm waiting for my new brakes and shifters before worrying about optimizing that area.

More to come soon.

Justin at Grin got back to me really quickly with answers to my questions. Great service! The regen not working was due to the settings in the Phaserunner. It was set up for regen braking a 52v battery and not a 72v, so its overvolt regen protection was not allowing regen to work. With that change to the settings the regen is working beautifully, including the proportional function with the throttle controlling regen force.

My not seeing more than 1600 watts was due to the battery not being at full charge and the lower voltage limit setting was capping power. With a full charge I saw 2400w and I have now dialed it back to a cap of 2000w. I want to get a better feel for the whole setup before I push it harder.

The temp sensor is up and working too. I haven't seen over 50 degrees C yet, but then I have not been pushing it up long steep hills yet either.

I'm still waiting on my drivetrain components to arrive, so no feedback on the torque sensing and pedal assist yet.

The oem RAD PAS sensor wouldn't work with the Phaserunner?

I'm sure the stock Rad PAS could be made to work with the CA. The issue keeping me from having a working drivetrain right now is that the GMAC uses a cassette style gear cluster and I don't have one until my new one arrives. The stock Rad motor uses a freehub style gear setup and is not compatible with the GMAC. So I need my new gears/derailleur/shifter to arrive before I worry about my pedal sensing.

New toys arrived this week! First brakes!

From my time modifying high perfiormance cars, I believe it is critical that brakes are upgraded to match (or exceed) any increases in power and speed. The new brakes are Shimano MT-520 four piston calipers with matching MT-501 brake levers. These are the new Deore level brakes that use basically the same levers, caliper bodies, and brake pads as the four piston Saint/Zee/XTR/XT brakes. I went with the four piston setup for as much braking power as I could reasonably get. The stock Radwagon brakes are 180F and 180R with Avid mechanical calipers and levers. I had previously upgraded the front to 203mm rotors for more stopping torque, and I kept those for now. The MT-520 calipers are currently running the stock resin pads and have good initial bite and seem to be good in the wet and reasonably quiet. I can also upgrade to the finned Saint/XTR metal pads if I need more brakes later, but I don't think that is going to be needed. The calipers don't have ebrake cutoffs, so I used these simple magnetic switches (https://www.amazon.com/gp/product/B07C24RK65/ref=ppx_yo_dt_b_asin_title_o04_s00?ie=UTF8&psc=1). I mounted the main switch body to the brake levers using hot glue instead of the included double sided tape as I believe it allowed for more secure mounting. The magnets are mounted to the end of the reach adjustment screw. The magnets have holes through their center that are large enough for the allen wrench to pass through to the reach adjustment nut, so this mounting does not prevent future reach adjustment.

The brakes installed quickly without issues, including cutting down the brake lines and bleeding the front and back. I had to open up the rear wheel splash shield around the caliper a bit for clearance.

With the brake levers installed I was able to finalize my control placement. On the left side I have (from outermost to innermost) brake lever, thumb throttle, and then digital auxiliary switch (used primarily for selecting pedal assist level). I have the throttle on the left side as it is used for both accelerating and decelerating (proportional regen). At the same time I am accelerating or decelerating, I am almost always upshifting or downshifting, so my right thumb is busy. Hence the throttle placement on the left. To get it to clear the neighboring brake lever clamp and the digital aux switch body, I had to grind back the two inside corners of the thumb pad that were closest to the handle bar.

The right side (from outermost to innermost) is brake lever, shifter, and then tucked inside the shifter clamp and under the shift indicator is the button for the horn. The horn (https://www.aliexpress.com/item/32980840218.html?spm=a2g0s.9042311.0.0.6d6b4c4d8Zlq7d) is like the horn/alarm that Juiced cycles used to sell. It is reasonably loud and hopefully will be able to get cars' attention. It is very convenient and close to hand with this location for the button.

This control mounting is compact and intuitive for me and works and feels good. Visually it is pretty tight and clean as well. I could take it a step further by getting a different version of the shifter that is ispec II compatible and then it could share the same clamp that the brake lever uses. Even though that shifter is only about $30, I don't think it is really needed. I am pretty happy with this layout.

Nice work!

thundercamel said:

Nice work!

Click to expand...

Thanks!

I installed the Sempu bottom bracket yesterday, which required a little fiddling. It turns out my bottom bracket shell actually measures 66mm across and not 68mm as it should. Beacuase of that I was not able to get the stock 68mm BB to seat with proper 40nm of torque without it binding the bearings. So I had Grin send me the 73mm BB flanges for the Sempu, which solved the problem. The stock 68mm setup for the Sempu has fixed flanges on both sides of the BB and relies on a perfect 68mm shell width. While the 73mm setup has a separate lock ring instead of a fixed flange on the drive side. According to Grin, some of the 68mm Sempu ship with the separate lock ring setup, but it is hit and miss whether it is that or the flixed flange. Having the separate lock ring allowed me to use the 68mm non-driveside fixed flange and the 73mm threaded insert with lock ring on the driveside. I thought I might need to trim excess thread from the threaded insert, but was not necessary. I could have also found or made a 2mm shim/washer and used the stock 68mm flanges.

Also, I found out there was not room in the bottom bracket to run my power switch wiring through the bottom bracket shell and have it exit through the same hole in the bottom of the shell as the BB wiring. So I made a new exit hole for the switch wiring at the bottom of the back of the seat tube. It exits at the bottom to allow full up and down adjustment of the seat post height without worrying about it hitting and cutting the wire.

Anyway, the BB is installed with the custom wiring for it to receive 12v power from the step down converter instead of the Cycle Analyst (the CA can't supply enough power to run the BB when running over 52v system). Its all working very well and I've just been riding the bike and dialing the settings I like for torque sensing.

The new Shimano Deore M6000 1x10 drivetrain is a huge upgrade over the stock Altus 3x7 components that came on the Radwagon. They always worked and were reliable, but the shift quality, tactile feel, and precision of the M6000 stuff is SO much better. I got an 11-46 cassette (which is basically the same diameter as the GMAC) and for now I'm using just the 48t biggest chainring on the triple crankset that came on the RW. I was planning on just using it and removing the other two chainrings before I realized they are all riveted together. I'd read about how you want to use a front chainring with a narrow/wide tooth profile to hold the chain on in a 1x drivetrain. So far there has been zero issues with the chain wanting to walk off the stock chainring that does not have the narrow/wide tooth profile. I suspect this might be due to the longtail configuration of the bike makes for a longer chain and less extreme chain angles in the various gears and thus less side loading on the chain.

I opted for the GS version of the rear derailleur which is the medium cage version as opposed to the SGS long cage derailleur. I did this as I didn't need the greater tooth range of the long cage since I'm running a 1x drivetrain and the SG version has a different top pulley geometry that allows it to track the cassette closer for more accurate shifting. With either derailleur you need a hanger extension to clear a 46T cog like I am running. With that hanger extension I could run up to a 50T cog and possibly larger.

Just a few comments...hopefully they will help.

Awesome documentation and fabrication so far.

Yes, Justin and everybody at Grin Tech are extremely helpful.

Problem you are going to encounter...if you don't pedal, your motor is going to overheat. Been there done that so nothing against your project but as you go faster the wind resistance is going up with the velocity squared and my 12T MAC overheats now and I am running a 14s/52v battery and a rear tire that is 722 mm in diameter.

Things you can do to help...
A. Go to a smaller diameter wheel...the driven wheel that your motor is in.
B. Auxiliary cooling...lots of options and I know Justin plans to do some experimentation with the MAC motor and cooling at some point.
C. Lower the voltage...yea, I didn't think you would like that idea...I don't either :lol: .
D. Limit the amperage to your motor.
E. Pedal

Of course A. through E. above will help but everything is dependent on how much power you push through your motor and how much you contribute to reducing the motor power (pedaling). I don't want to discourage you in any way because I think you are on a roll . Just keep the overheating issue in mind as you make changes and develop your project.