August 2008 Archives

My replacement servo-tester finally came in yesterday. I immediately tore it apart and replaced the potentiometer in it with a cable so that it could plug into my Magura throttle. The design of this replacement looks very simple. All discrete components: a 555 timer IC, rectifier, and some caps. I think I'll try to reverse engineer it so that I can just build it myself next time.

While I was waiting for the replacement to arrive (two weeks!), I took my Boxer almost entirely apart and placed the ESC and other electronics in the compartment where the old gas tank used to be. I also made a pannier for the batteries on the side. The result is a pretty clean ride:

The bike is very stock looking, which was one of my key goals. From the outside, the only indication that it perhaps isn't gasoline powered are the two SLA batteries in the panniers. The only real work that needs to be done is to clean up the remaining wiring for the servo-tester / throttle.

This is a neighborhood kid who helped me make the battery carrier. He insisted on being in the photo.

I'd like to revise the battery carrier to be a little more stealth. This first version was just a hack-up to get the things mounted.

Right now I only have batteries mounted on the right side. Two SLAs (2 x 12V = 24V @ 18 Ah) give me a top speed of 30 kph (20mph) which is just fine for the type of intracity commuting that I made this bike for. The bike has such good starting torque that it's low max speed isn't that noticeable.

I originally planned to mount a third battery or perhaps fourth on the left side. However, two batteries are giving me the speed that I want so I might put a regular pannier on the left so that I can carry groceries.

The following are some screen shots of the CycleAnalyst after my first test ride in this new configuration.

My range was 6.5 km on almost fully recharged batteries. This surprised me. I only pulled out just over 7.5 Ah before the batteries died. The batteries are rated at 18 Ah (20 hr) nominal and around 12Ah (1 hr). I was sucking juice out of them pretty fast, still I expected a little more capacity.

The Wh/km of this bike is excellent as it's very light and not all that fast.

The speed here is given in km/h. So that would be around 20 mph max and 15 mph average.

The batteries were seemingly dead at the end of the test run. As you can see, Vmin was 15 volts (this includes voltage sag/drop from power use) which was causing my CycleAnalyst to reboot sporadically and the ESC to cut in and out. Very bad.

Yet when I hooked the SLAs back up to my smart charger, it said that they were 42% full. Which is about right as I pulled 7.5 Ah out of an 18 Ah battery. Still, there was really no juice left in them. Maybe if I had let them rest a little bit, I could've gone a few more kilometers.

I'm trying to figure out what to do. The max speed is fine but the current range is a bit lower than what I had wanted. I wanted a range of around 12 km so that I could go to work and back without recharging. Some people mentioned that AGM SLAs take a while to "break in" so I guess I'll try to see if that's true and if performance gets better as time goes along.

p.s. Just a side note, Amax was 129 amps @ 24 volts = 3000 watts or 4 electric horsepower! This was just starting torque, when I was cruising it was sipping around 300 watts to keep me going.

I've been asked if I had to register my Piaggio Boxer in CT and get it plated and insured. When it was a gasoline scooter in Minnesota, I had it plated and insured (and have a MN title for it). When I got to Connecticut, I was told that because it was 49cc, it fell under the moped provision and didn't need to be titled, registered, or insured.

Here are the requirements for mopeds in CT:

a. goes no faster than 30 mph

b. displaces less than 50 cc

c. has an automatic transmission

d. less than 2 brake horse power

e. and after 7/1/97 the need for pedals is deleted from the requirements.

When I did the EV conversion, I think I remained within the moped parameters. The scooter is geared so it doesn't go more than 30 mph; it displaces 0 cc which is less than 50 cc; it has one gear and a centrifugal clutch, so it is automatic; brake horse power is a defunct and idiosyncratic measurement system with no easy equivalent in other systems; and it has pedals.

So I should be good. If I'm forced to register it, I have the Minnesota title which I can always transfer to CT.

Let's see if inserting this Youtube code will work.... my friend Dave from the NEEAA giving my little Boxer a spin.

Here are some pics of the low-voltage converter that I created using the National Semiconductor LM2576HVT-12-ND buck-converter. The chip pretty much did everything for me -- all I had to do was to add two power-smoothing caps, a blocking diode, and a RF choke (which I mistaking undersized; promptly blew up and which I jerry rigged a loop of wire).

In the top left of the photo above, you will also note my premade, off the shelf 12VDC to 5VDC converter. This powers my LED headlights.

The following photo is the back side. Please note that I am a really lousy solderer. The LM2576HVT-12-ND will take any input from 16-75VDC and output a steady 12VDC @ 3A continuous but they mention that with active cooling it could do 5A continuous. That's 36-60 watts, plenty for me.

This is the wiring schematic that I used:

Top side again. The setup is small enough that my ESC and UBEC still fit in the project case with plenty of space for airflow -- and for another DC-DC converter to fit in there in case 3~5A is not enough.

The resulting low-voltage converter and ESC "mounted" on my bike with duct tape. This is an older photo, I've reduced some of the wiring and since changed the SB50 connector to a PowerPole 45.

On Sunday, I finished making the bearing mount on the right side of the motor spindle. This will help take a lot of strain off the bearings on the motor and prevent it twisting out from the torque. Also, I shimmed the front pulley with the thin metal from an aluminium can -- it was just perfect to close the gap between 12.5 mm and 12.0 mm. As a result, the motor is working much quieter now -- I just love the sound it makes when it accelerates -- gotta post a video.

(And yes, it's made of wood, I'm a woodworker not a metalworker and don't have access to a mill. The 12mm flanged bearing itself is hidden underneath the wood, it has quality roller bearings to support the spindle.).

I got a new V-belt at the local autoparts store to try to make the gearing taller by forcing the rear CVT smaller, but the V-belt was too small. I'll have to split the difference and get a new V-belt later this week.

Instead of depending on my unreliable low-voltage system for fan power (not good that it cuts out at full throttle), I instead wired two of the fans in serial. That means each fan is being driven by 19 volts -- a bit fast but not something they can't handle. One fan is on the ESC controller, the other I mounted in front of the motor. Unfortunately I found out that the mount I used blocked some of the natural airflow, so the motor ended up running hotter with the fan than it did "naked." I'll have to make another mount that diverts more natural and blown air to the motor.

I've ordered more batteries for the scoot, so now I just have to wait.

I put the Phoenix HV-110 ESC in the same aluminium project box that houses my low-voltage electronics. I mounted a temperature controlled CPU fan on it and ran the fan off the 12V line on the low-voltage converter.

The fan was in a "blow" configuration but I think I will change it to a "suck" config as I think that is more efficient for this type of fan.

The servo tester is mounted on my handlebars. Once my Magura throttle comes in (backordered), I'll be able to use that as a regular motorcycle throttle. The switch on the dashboard is for the lights. I have two high-power LEDs mounted in the headlights. The new white LEDs are so efficient that I can use a 5V 1 watt bulb instead of the old 6 volt 10 watt incandescent.

Everything was mounted on with duct tape and the batteries put in a small canvas bag that I had and I was ready for test drive #2 with the Phoenix HV-110 and HXT outrunner motor.

Notes on the test run:

• I was very surprised that my lights and the CPU fan cut out on full throttle! Coming back home, I later realized that the DC-DC converter at the heart of my low-voltage system is rated 36-75V. I'm running the bike on a 38.4V NiMH system, but at full throttle I have enough voltage sag that the power module cuts out. I'm ordering a new 15-40V --> 12VDC power converter from Digikey and that should solve that problem. Redoing my low-voltage module is a pain.
  
• Argh. If I was able to use 48V with the Phoenix ESC than I could've gotten away with the current 36-75 volt DC-DC converter. Double phoeey.
  
• With 38.4 VDC 3.8 Ah, I was able to go approximately 3km (2 miles) at 20 kph (15 mph). I think I'll get 3 more of these packs, which would give me a total range of 12 km (8 miles).
  
• Dang my NiMHs were hot coming off the test ride. I really need more packs in parallel to take some of the stress off the single large pack that I'm using. Still no good charging solution, I'll most probably just have to use a bank of 19.2V chargers.
  
• The ESC was not warm at all, so it looks like the fan is working well. The aluminium enclosure also helps with heat dissipation.
  
• There was a rattle in my belt-drive system that I narrowed down to a slip caused by the difference between the 1/2" pulley and the 12 mm spindle. Half an inch is 12.5 mm, so there is a 0.5 mm gap that is rattling. I think I'll shim it with some copper shim material I have lying around.
  

I'd love comments!

Some miscellaneous notes that are bouncing around my head.

• The Phoenix HV-110 has a listed max volts of 50 volts, which I thought would mean I could use 40 NiMH cells, with a nominal voltage of 48 Volts. However, you are limited to 36 cells NiCd/NiMH which gives a max nominal voltage of 43.2V. This is a little bit of a bummer since I had planned on doing 5S packs of 9.6V which would have given me 48V. I think this is because although the nominal voltage is 48V in that case, the peak voltage (fresh off the charger) can be in excess of 60V and the FETs most probably aren't rated that high.

• It turns out that I could've bought the 180 kV motor instead of the 130 kV motor. As it stands, I'll either have to get a smaller V-belt or a larger front pulley if I want a max speed of 45 kph (30 mph). No worries.

• I'm wondering what the best way to charge my NiMHs is. They don't sell many 38.4 V chargers, so I may just stick with 2 x 19.2 volt chargers and charge 2 packs at a time. I may want to get even more chargers so I can do the entire megapack at the same time.
  

My replacement ESC from Castle Creations (the Phoenix HV-110A) finally arrived today, along with a Castle Creations USB programmer for it. The Phoenix HV-110A is considerably beefier than the ELF controller that I burned out. I reprogrammed the Phoenix using the very nice USB programmer (notably: BRAKING off) and installed it onto my Piaggio.

If you have the right QuickTime codec installed, you can watch a very boring video of my biking revving up to speed.

Yes, that is indeed duct tape holding the motor wires and everything else together. My anderson powerpoles will arrive on Monday....

On the road, the moped ran just great, with a maximum speed of 20 kph (15 mph) as predicted by the voltage * kV (24V * 130 kV) of the motor. I also very quickly tripped the 30 amp fuse that was protecting the batteries. Luckily I had bought a whole pack of 30A fuses and had them in my pocket. I decided to go to a higher voltage as that would increase my top speed as well as lower the max amps drawn at the same speed.

----

I wanted a bit more speed so I wired up the RC racing NiMH packs that I had gotten into a single pack: 9.6V 3800 mAH packs * 4 == 38.4V 3800 mAh megapack and plugged that into the system.

With 38.4V, my top speed increased to 30 kph (20 mph) and -- more importantly -- I could get to this top speed with popping the 30 amp battery fuse. However, if I tried to go up a steep hill, I would pop the fuse. I need to wire two 30 amps in parallel to get 60 amps -- or put in a 60 ~ 90 amp circuit breaker.

The ESC definitely got warm, in the high 30C range (90-120F). I think I will use active cooling for it. I have some CPU fans that I got at clearance that should do the trick. The motor barely broke a sweat, I do not think I will need to do much in the way of cooling it.

The moped was definitely a lot of fun zipping around the neighborhood with it.

---

It got dark so testing was over for today. I spent the night rewiring the ESC into the test-box for my low-voltage systems and put a big CPU fan on all of it.

All in all, very very exciting. All that remains is to get some more NiMH packs and wire them in a 3P 4S configuration to get a 38.4 V 11.4 Ah pack. This should be enough to get me to work with a comfortable margin to spare.