I ordered two new 20Ah LiFePO4 cells from Sam at Yesa on Tuesday to replace the two that were damaged (see "Fire is Bad"). Cost was $70 per cell and shipping from Hong Kong was $45. The cells arrived last night along with an individual cell charger. Sam asked me to send him a picture of the cell arrangement in the pack so he could have the tabs soldered in the correct orientation.
The red arrows indicate the bad cells.
Yesa's website gives pretty good instructions for replacing cells in a battery pack. I will apparently need use the single-cell charger on the new cells to adjust their voltage to match the older cells before the BMS can take over. I've also got some other fire-related damage to deal with, like replacing one of the charger wires with melted insulation, and finding some HDPE sheets and building a case for the repaired battery pack. Unfortunately my work schedule will probably not allow me to get this done until next weekend or later.
I took my damaged pack over to Synkromotive today to have some experts take a look at it. We took voltage readings of each of the 12 cells. They each read about 3.3V, except for two cells. One of them read just about zero, and the one with the obvious damage read about 1.8V. Their advice was to just replace the bad cells and keep going. They also briefly studied the BMS that came with the batteries, and deemed it satisfactory.
So now I just need to email Sam at Yesa and get a couple new 20Ah cells, install them, then replace the charger wires with the melted insulation, and I should be good to go. After I build an insulating case for the battery pack....
Can anyone recommend anyplace that recycles lithium?
I had a harrowing experience yesterday while I was riding the Rezistor home from work. I smelled something funny (it always starts that way, doesn't it?), looked down, and saw smoke coming out from under the seat. I pulled over, cut the power, got off, and saw the paint on the exterior catch fire. I blew out the fire, lifted the seat, and saw some arcing between the corner of the battery pack and the inside of the scooter body. I disconnected the battery pack immediately. The pack fits pretty snugly, so it's not too easy to pull out, and I kept bumping the corner of the pack, causing more arcing and reigniting the body paint. I eventually managed to stick some insulating material between the battery and the body, pull out the battery pack, and ride home on the remaining battery pack. Fortunately, none of the smoke was of the "magic" variety, as everything else seems to still work.
What had happened was the corner of the pack had been rubbing against the scooter body and eventually wore through the outer casing. When the battery innards came in contact with the metal of the body, they shorted and arced, producing plenty of heat to vaporize the corner of the cell and ignite what paint was left on that area of the body.
Let me emphasize that the batteries themselves did not catch fire. The damage was caused by my own failure to properly secure and insulate the batteries within the vehicle. The LiFePO4 battery pack appears to have remained very stable, and the other pack was fine. I plan to contact Sam at YESA to see if I can buy a single 40Ah cell to replace the ruined one and try to get the pack back in working condition.
So here are the lessons to be learned: Secure your LiFePO4 batteries so that they can't move around inside your metal scooter body, and cover the battery pack with a tough insulating layer.
A few less dramatic issues to report: A couple of weekends ago I paid a visit to Synkromotive, a company not too far from my home that develops controllers. I met frodus (Travis), one of the guys who posts here, and he showed me the current state of his eVFR conversion (nice!). I also had some excellent conversation with the Synkromotive folks. It was nice to have my project looked at by some people who actually know what they're doing.
Sometime during my visit, the fuse to my 12V converter blew, so I went home and replaced it. I started riding again and almost immediately noticed an smell reminiscent of cigarette smoke, so I shut down and pushed the bike back home. I checked my 12V converter, and it no longer appeared to be converting - I had 36V running through all my auxiliary systems. This resulted in burning out about a third of the LEDs in my headlight (although the taillight LEDs were all fine). I bought two of these nonisolated converters from TS Motors and this is the second one that has failed. Frodus helped me find a new 12V (isolated) converter on eBay for about $30 and it seems to be working fine so far.
I knew that I had a short in my system somewhere, and I suspect that's why the converter failed. After hunting around, I finally found that the postive terminal of the battery pack under the left cowl was coming into contact with the battery rack. So I cut apart some cat-litter containers made of HDPE and made an enclosure for the pack to keep it from contacting any metal parts. That took care of the short. I suppose I should have done that for the other pack as well...
Also, I finally bought an LED-compatible flasher module and installed that last week. My turn signals still aren't working quite right, though, so I've still got some wiring to check. I've started working on license plate illumination and backup lights for reverse gear.
Rode the Rezistor to work yesterday and today. I haven't been able to find any electrical outlets on the outside of the (very old) building I work in, so I had to improvise.
Not that this is absolutely necessary: my commute is 6 miles round trip, which is well within the Rezistor's range. I just wanted to see what I would have to do if I really needed to charge up at work for some reason.
I just started taking voltage readings before and after riding and charging. Right after a charge, the batteries read at 42V. After sitting overnight in the garage it drops to around 37V - might I have a slow charge leak somewhere in the system? Then after I ride it 3 miles to work it reads 36V. I'll continue to take readings, and I should really start to check the individual cell voltages to see how well Yesa's BMS is keeping the cells balanced. The BMS on one of the packs is rather inaccessible, unfortunately, so I'll have to take readings from one of them and assume they're both behaving the same way.
I've been swamped on a deadline at work, so I haven't had time to get the turn signals working or to build license plate illumination and reverse lights. I ordered some LED replacement bulbs for the rear signals (about $6 for the pair on ebay) and some blue electroluminescent wire that I plan to wrap around the motor power cables for ornamentation.
I've also started some research on what to do about the bodywork. If I go with the bare-metal look, there's a product called Gibbs penetrant/protectant that will prevent and supposedly even reverse corrosion. I found some pictures of a roadster restoration that was left in the bare-metal state using this product. I suspect you have to reapply it at regular intervals. There are some local powder-coating companies that will apparently do the sandblasting for you as well as the powder-coating; I might be able to get this done for less than $300. I plan to ask about the possibility of clear powder-coating, but I might just go for the brightest color I can find for safety reasons.
This weekend I rode the Rezistor to Spring Scoot, a scooter rally held annually each year here in Portland. It basically served as the initial test run and range tests. I started with a 6.6 mile ride on Friday night and rode progressively longer distances Saturday and Sunday. I finished with a 14.5 mile ride Sunday evening, much of it at full throttle. There seemed to be juice left when I got home, although my headlight (and presumably the rest of my accessories) seems to have given out a few blocks earlier. The controller, motor, and power cables were only a little warm.
I've been recharging right after every ride back up to full capacity. There's a little LED on each charger that changes from red to green when each pack is fully charged. After an 11 mile ride on Sunday it took roughly 4 hours to fully recharge the two 36V 20Ah LiFEPO4 battery packs from YESA.
The motor and controller are working well. I noticed sometimes that when I twist the throttle from a dead stop, a brief rattling noise comes from the motor and acceleration is delayed momentarily. If I give the bike a tiny push forward with my feet, then ease on the accelerator very slowly, this seems not to happen. I'm theorizing that the controller gets "confused" when the motor is at a dead stop, not knowing exactly what the rotor position is, and the rattling sound is caused by the rotor flipping back and forth. When I give it a little push, the controller seems to realize that the motor is spinning forward and it engages smoothly.
I also noticed some differences in acceleration, probably due from the presence or absence of the 10-second temporary current boost to 250 amps from the controller. Without the boost, acceleration is pretty slow, even slightly less than I would expect from a 50cc ICE scooter, but it's pretty decent if I get the boost. I find that, when stopped at a signal, I can get the boost pretty reliably if I put the forward/reverse switch into neutral while I wait, switching to forward when the light turns green (and then pushing the bike forward - I almost never get the boost if I hear the rattling noise). I don't know how often the controller allows the 10-second current boost to happen, whether it's a function of the fwd/rev switch, some amount of time between boosts, or something else.
I can feel the motor drag when I coast; I presume this is the regen activating. I'm learning to use the brake levers as little as possible to maximize my efficiency. I'm also riding hunched over a lot when at speed, since the Vespa's aerodynamics are bad enough without the addition of my torso as a sail.
Sunday I clocked 35 mph with my GPS, which seems to be about my maximum on flat road. I'm assuming I'm getting 100 amps at full throttle. At some point I may send the controller back to Team Delta to increase the continuous voltage, maybe to 150 or 175 amps. The motor is rated for 100 amps continuously, but it's getting plenty of ventilation, and with the way I'm riding it's not really a continuous demand.
Saturday morning at Spring Scoot 14. Breakfast was at Mickey Finn's in southeast Portland. Behind all the scooters on the street, you can just make out the Rezistor parked way in the back on the sidewalk, next to a door through which I ran my extension cord inside.
Charging up at breakfast on Sunday.
Twist'n'Play Scooter Club awarded me the rally trophy for Best Rat Bike. It's sitting on the floorboard and yes, it has a big rubber rat on top. The other rallygoers had many questions, many of which I really couldn't answer completely because I wasn't done testing it yet. It was especially fun to demonstrate reverse gear.
In general, I consider this weekend to be a success. I built the thing mainly as a commuter vehicle, and Sunday I rode almost three times the distance of my round-trip commute on a single charge. I have yet to find out what my ultimate range is, but I suspect it's not much further than 16 miles.
The replacement controller from Team Delta arrived on Wednesday. Fantastic service from Dan, considering I just explained the problem to him on Sunday. I installed it and, at first, got the same "MOSFET short circuit" error message from the controller as before. I wondered if the contactor might be faulty, so I took that out, fiddled with it as much as I could (which wasn't much) then reinstalled it. But then it occured to me that the 12V converter might be faulty, so I left that disconnected. That must have been it, because the scooter worked after that. Changing the controller setup from "speed" control to "torque" control did the trick.
But without the converter, I had no lights or horn. So on Thursday night I pulled out the converter and tested it. It was bad...somehow I had blown it out and it was creating a short-circuit between the negative battery terminal and the B+ connection on the Sevcon controller - hence the MOSFET short error message. So Sevcon has been vindicated; the controller was not bad after all.
Good thing I had the foresight to buy two converters - I installed the second one, this time with a fuse on the + input wire. Lights and horn were restored.
Thursday night I also installed the turn signals I made (LEDs, acrylic, and 5-minute epoxy) and a couple of mirrors. Turn signals aren't working, though...I think the LED's aren't creating enough of a current draw to activate the flasher, or maybe I have something wired wrong.
I made a little "REZISTOR" sign to fit in the horncast strip, and got my wife to paint the logo on the legshield. I was hoping to have it ready to ride on Friday night, where I could debut the scooter at the annual Spring Scoot scooter rally in Portland, put on by Twist'n'Play scooter club. The rally would basically serve as the initial road tests.
Results to follow shortly...
On Sunday I contacted Dan at Team Delta, who sold me the motor and controller. I described what was happening with the Sevcon Millipak controller and he was able to clarify things for me.
Issue 1: I could accelerate only very, very slowly, or else the system would shut down and the controller LED gave me a 9-flash error message, or "rotor position sensor fault". Turns out that I chose one of the initial controller setup parameters incorrectly. The controller uses current sensing and Hall effect feedback to servo the motor in either speed or torque mode, and I chose the speed mode when I should have picked torque. I don't have the equipment to change the setting, but Dan said he would put another controller in the mail to me on Monday with the correct setting.
Issue 2: I turn the keyswitch and I hear no "click" from the contactor engaging, and only get a 3-flash error message from the controller LED. This indicates a MOSFET short-circuit, which is internal to the controller. Dan indicated that the failure of the contactor to close confirms that this is the case. So, it's a good thing he's sending me a new controller anyway. He also mentioned that this is only the first out of about 58 of these controllers he's sold that has failed in the field.
So while I'm waiting for the new controller to arrive, I've been working on my front turn signals, using my usual medium of LEDs, acrylic plate, and 5-minute epoxy.
Well, it's been an exciting week. I spent much of last Saturday getting the brakes working, installing the levers, cables, and switches. The brakelight switches install in-line in the brake cables within the headset, so I had to figure out where to cut the cable housings to get them in the right spot.
Then I finished wiring the headset, making the connections to the switches, horn, headlight, brakelight, turn signal flasher, and reverse light. I also finished the controller wiring and put new leads on the keyswitch, the wires of which kept breaking off. I installed the horncast and the taillight and completed the wiring connections for those.
Last Sunday I started off by putting in one of the battery packs under the cowl. The BMS is just an exposed circuit board, so I wrapped it with a plastic bag and some tape to give it some nominal protection and weatherproofing. It's wedged in a small space on top of the battery so it's pretty well protected anyway. Getting the battery in was a tight fit.
I ziptied the two battery chargers into the underside of the seat. I plugged each one into an orange extension cord and then stuffed all the cords in there as well.
I bolted in the controller and by Sunday afternoon I was ready for a power check, using only one of the battery packs. I put the motor next to the scooter, connected the power cables, turned it on. I hear a click as the contactor engages, and my LED headlamp lights up, so that's a good sign. But I get the flashing LED on the controller indicating some error...oh, look, the control wires aren't connected to the motor. Plug it in and it works!
So then I had to move the thing off the back deck of my condo and into the garage for final assembly and a test run. Pushing it down the hallway and down a flight of stairs was a chore.
On Sunday evening I put in the second battery pack, hooked up the motor, plugged in the battery leads, and was ready to test. I turned the key, engaged the forward switch, and slowly twisted the throttle. I could hear the motor and chain moving, but there was no forward motion. I looked down and could see the chain spinning around the sprocket, but the welds between the sprocket and main axle had broken.
So on Monday night I unbolted the swingarm assembly from the body, removed the rear wheel and brake plate, and took the swingarm and motor over to my friend Arden, who was able to fix it with his TIG welder on Tuesday. Tuesday night I picked it up and reinstalled it into the scooter. Thanks again, Arden!
Late Tuesday evening I tried another test run. This time as I turned the throttle I felt a slight tug backward and heard a "tink", then heard the motor spin some more while feeling no motion. This time, when I looked down, the chain wasn't moving with the motor, so something else had failed. I felt around the engine case for a bit and pulled this out.
The 11-tooth front sprocket had broken in half. The notch cut for the keyway had really weakened the sprocket hub, and it broke right along the keyway and one of the set screw holes. Also, it turns out I had wired my forward/reverse switch backward, although I swear that I had checked the switch polarity beforehand...oh well, at least it's an easy fix.
I went to McGuire Bearing on Thursday to see if they had a sprocket with a thicker hub, but all of their 11-tooth ones were basically identical, and they didn't have one in stock with 12 or 13 teeth. I called around and found a place that had both, and picked them up on Thursday. Both of them have much thicker hubs than the 11-tooth.
Friday night I put on the 12-tooth sprocket, only to find that my chain was no longer quite long enough. I made a chain that was two more links longer (you have to have an even number of links, of course), but then it was too long, and the chain rubbed on the engine case. I switched to the 13-tooth to see if that would take up the slack, but it still wasn't enough. This is frustrating. At least I got an opportunity to reverse the leads on the forward/reverse switch.
This morning (Saturday) I took the swingarm off and took it apart. I was able to elongate the motor mount holes slightly, but just enough to shorten the distance between the motor axle and shaft so I could fit on my original length of chain. This was a relief, since I had feared I might have to make extensive modifications to the swingarm. I reassembled everything and decided it was time for another test.
This time it worked. Sort of. I got forward motion, but if I accelerated too much the power cut out, including the headlight. If I reset the system by turning the key off and on again, I could again get forward motion, but it didn't last. I could get up to maybe 20 mph if I go really, really slow on the throttle, but then the power cuts out. Interestingly, reverse seemed to work a little better (and felt really weird) but again, the power cut out as I gained speed. I looked at the controller after the power losses, and saw the LED indicator blink 9 times. Sevcon's manual says this is a "rotor position sensor fault".
I suspect that Yesa's battery BMS is set to limit the current draw, and that this is causing the shutdown. Decibel1, another member of this forum, described a similar issue with his Yesa battery packs a few months ago, and I had kind of been expecting that this would also be the case with my battery packs. So I took the batteries out and disconnected and resoldered the leads on the the BMS so that it would be bypassed on discharge. Decibel1 seemed to get decent results with this approach.
I tried a few more test runs, but the behavior was about the same. I still get power loss after initial acceleration. Then after a few more resets, it stopped working altogether. I turn the keyswitch and no power gets to the headlamp, and the LED on the controller blinks 3 times. The manual says this means "MOSFET short circuit".
I poked around the wiring with my multimeter for a while, checking various connections and looking for a short, then noticed that I was getting a 40V potential between the positive battery lead and the scooter body. This seems wrong (and somehow unsafe), so I figure out that the exposed charger connector on the battery under the cowl is coming into contact with the body. I insulate that with some electrical tape, and the potential goes away. But I'm still getting the 3-blink message when I turn the key, and no power. I'm hoping I don't have to send the controller back to Team Delta for diagnostic and reprogramming.
I haven't blogged in a while, but it's not because I haven't been making progress. It's just that when I get free time, I'd rather spend it building than typing and uploading photos. Here's a summary of what I've been doing:
I had to reconfigure LiFeP04 battery pack 1 so it fit better in the space below the seat. This involved cutting the tape that held it together, cutting and resoldering some of the links between the cells, and extending some of the wires to the BMS.
I got some perforated angle, some hinges, and a bunch of nuts and bolts, and built a rack for battery pack 2 under the left cowl. Originally I was thinking this might have to be welded to the body, but I don't weld, so I ended up drilling some holes in the body. I don't think it looks too bad. It sticks out a bit underneath, but still looks better than an exhaust pipe.
I took apart the taillight and replaced the tiny bulb (it didn't have a brakelight) with some LEDs mounted in a piece of acrylic. It screws in very nicely to the existing taillight mount. I wired a couple of resistors in series with a bypass so it functions as a brakelight as well.
I figured out places to mount the main components. The main fuse holder, contactor, and 12V converter are screwed directly into holes I drilled in the body. I had a piece of angle left over from the B2 mounting, so I screwed that to the body where the fuel tank was, and I'll bolt the controller to that. B1 fits pretty well right in the space in front of that, and I don't think I'll need to build any special mount, maybe just a big rubber bungee to secure it to the angle.
The old horn runs on AC, so I bought a new DC horn from electricscooterparts.com (interestingly they are in Boulder Creek, CA, where my mom lives) as well as a keyswitch. The Chinese horn was much less expensive than a DC Vespa horn. I cut some notches around the edge of the horn so I could secure it into the horncast with screws. The previous owner gave me a new horncast along with the old one, so in the photo you can see the old on the left and new on the right.
The brake parts I ordered from Ptown Scooters finally came in this week, so I was really glad to get those. I'm reusing the original front and rear drum brakes, so I needed new cables and rear shoes. As I mentioned before, this scooter never had a brakelight (or turnsignals, or speedo...) so Steve at Ptown found me me some brakelight switches that install in-line with the cables, and I think these will work out nicely. Thanks, Steve!
I got the missing bracket for the center stand, so I finally installed that. This makes work a little easier. I also got bolts for the seat attachment.
The old switchpanel was a mess, so I built a new one out of three layers of 1/8" black acrylic that I superglued together. The rectangular rocker switch on the left is for forward/reverse, the round rocker switch on the right is the turn signal, and the red button is the horn. The LED blinks red along with the turn signal. I bought the turn signal blinker at an auto parts store.
Most of the rest of the work is wiring. Today, using an unbent coat hanger, I pulled the rear brake cable through the wiring channel in the body, along with several wires for throttle control, forward/reverse, and lighting (including white LEDs for reverse, the housing for which I have yet to build). Pulling the wires was a total PITA; I'm really glad that's over.
Yes, still ugly. And the Vespa doesn't look so great, either. I've realized at this point that this is really going to be two projects: the conversion and the restoration. I've never done a scooter restoration before, so I'm not really sure how far I'm going to carry it. I suppose it depends somewhat on how much fun it is to ride when I'm done.
Oh, and I got married. Huge thanks to my lovely wife for her continued patience and support as I finish the project!
HUGE thanks to my friend Arden for helping me assemble the drivetrain. A fellow member of the Oregon Scooter Club, Arden has an amazing shop with machining and welding capabilities. He probably saved me over $100 in machine shop costs, as well as providing some of the raw materials.
First I had to cut the fat end off the axle (lots of pretty red sparks!). The fat end is normally where the gears go, but I have no need for it and it's in my way. Arden used his lathe to cut a 1" hole in the center of the sprocket so it could fit over the wider ends of the axle. We then machined a steel ring with a 17mm I.D. and 1" O.D. and then cut it in half; this allowed us to center the sprocket on the axle.
I had to cut a big slot in the V9A engine case near the left bearing to provide clearance for the 33-tooth sprocket and chain. Things were still pretty tight inside the swingarm, so putting the sprocket on the axle was a fairly convoluted procedure. We first had to position the sprocket, then slide the axle through the sprocket and bearing, then attach a split collar that clamps to the 17mm axle shaft with set screws, then fit in the split steel ring between the sprocket and the axle. I adjusted the location of the sprocket to center it in the swingarm slot, tightened the set screws on the collar, then Arden welded the sprocket to the collar and the collar to the axle.
The advantage of using an old engine case for a swingarm is that the front connection to the scooter frame, the rear connection to the shock, and the axle bearing are all exactly where I need them and perfectly aligned. The main disadvantage is that I now cannot remove the axle from the swingarm without cutting the welds on the sprocket (or maybe by removing the left bearing; I haven't tried this yet). Hopefully I picked the right gearing and won't need to change the sprocket. Arden almost talked me into abandoning the engine case and going back to my idea of building a swingarm from scratch (see my previous posts with photos of the wood 2x4 swingarm mockup). I still think this would have been more work, with a lot more welding and potential alignment difficulties. If it turns out that I weakened the engine case too much with all the cutting, I may have to start from scratch anyway.
To mount the right bearing, Arden had some 3/8" aluminum plate from which he cut two 2"x3" pieces, then used his lathe to bore a hole in one of them with the same O.D. as the bearing. I needed two plates because the axle stops just over 3/8" short of the motor mount plate. I fastened the plates to the motor mount plate with a screw in each corner, after aligning them using the engine case (which now includes the captured axle). Although I was careful to position the bearing plates such that they were clear of all the various internal bits of the swingarm, I didn't pay attention to where the screws fell relative to the motor, so I ended up only being able to install three of the four screws. Redundancy can be helpful sometimes.
The next step is attaching the 3-phase Etek motor to the motor mount plate using four 3/8" screws into the outer four holes provided in the motor. Positioning is fairly critical, so I first used my plywood template to locate the holes and make sure they were clear of the internal swingarm bits. I fit the chain around the sprockets to check the tension, and discovered I needed to locate the motor shaft a little further from the axle; this required that I cut more out of the engine case (fortunately it's aluminum). I also learned that the 11-tooth motor sprocket will need to sit really far out on the shaft to align with the axle sprocket, so I'll probably have to extend the motor shaft with some of 3/4" rod and a screw into the hole in the center of the axle.