A2B Metro restoration / modification - Fechter's first Ebike

Finally, after years of delays, I finally got the external controller mounted and working. I'm using a 12FET Powervelocity sine wave controller. Takeoff is totally silent now, without that nasty screechy sound of the stock controller. Also hopefully gone are the days of controller overheat shutdown and being stuck for 30 minutes.

I mounted the controller under the rear rack, as I don't have a secondary battery that fits there.

External Controller 1.jpg

The garden hose sized bundle of cables runs down the side of the frame:

External Controller cable route.jpg

I used a plastic Hammond box to contain the "hairball" that comes out of the controller. It takes up almost as much room as the controller itself. It is not sealed, but open on the bottom to let any water drain out. It's sealed on the top, so rain and road spray won't get in much. It also allows easy access to a terminal block there to test/troubleshoot voltages.

Hairball 2.jpg

I have the programming connector hanging out now, for setup, but it tucks away inside the box.

Also got the MMG moped tire mounted on the rear. It went on easier than the front one, maybe just due to experience.

Moped rear tire 2.jpg

Here's the before picture:

Rear tire before.jpg

After:

Moped rear tire.jpg

It's heavier for sure, but fits in the frame. This should totally cure any cornering issues.

The maiden voyage was not without problems. One of the phase wire connectors was not fully seated and went open. Of course it didn't run very well with only 2 phases. A little poking with the ohmmeter found the problem fast enough, and an easy fix.

The next problem was the current limit was too high and it tripped the BMS, which should be set at 50A. CA peak amp reading showed 58A, so needed to crank the amp limit down. The settings did not match actual, but were linearly related, so a little calculator work and reprogramming fixed that. After that, it was smooth as silk and very nice to ride.

I have the variable regen mod on the controller, but don't have the regen throttle hooked up yet. The brake switch gives a fixed level of regen that's adjustable in software. It engages slowly and has a nice feel to it. I was able to dial it in for a max regen current around 10A, which is about the max for my battery.
 
Thanks. The "hairball box" has worked out really well. As luck has it, I had a few problems on the initial ride that required metering some of the connections. Having the terminal strip in there made testing pretty easy. I spent a lot of time labeling wires and have a chart of the terminal block positions vs. function. If I need to go in there years from now, I'll have a chance to figure out what I did.

I was too lazy to run a wire to the handlebars for the 3-speed switch, so I mounted one on the bottom of the box where it is fairly hidden, but can still be switched while riding. I have it programmed to change the current limit since I run two different batteries with different current ratings. Sort of a California Class 2, Class3, Insane mode switch. I really like that feature.

I need to do a bit more testing, but so far I really like the sine wave controller. Having 2kW on tap makes acceleration from a dead stop much better. And super-quiet and smooth.

Here's a better shot of the box. The 3-speed switch is the little silver toggle. The programming connector is hanging out, but will get stuffed in the box once I get everything dialed in.

 

Attachments

  • Hairball  3-speed switch.jpg
    Hairball 3-speed switch.jpg
    85.2 KB · Views: 3,932
Nice A2B based build. The upgrades you installed make the A2B a real industrial vehicle. The variable regen willbe intereting to hear about. Does the regen throttle you mention turn in both directions and spring back to a centered position?
 
I haven't quite figured out the regen control yet. A bi-directional throttle like a Vectrix would be one option if I can figure out how to make one. Another option is a separate braking trigger mounted above (or below) the brake lever. I think I can make this. Meanwhile, I have the fixed regen set at a nice level and it works well.

I went out for a torture test ride today and I'm quite pleased with the climbing performance. I took it up a really steep dirt fire road that leads to the ridge above my house. While it slows way down, it keep going and I can assist with some pedal power. I didn't stop and measure the grade, but would estimate 20%. I have a measured 18% grade on a street nearby and it maintained around 15mph going up that. Both of these hills were fairly short, maybe 1/2 mile.

When I got to the top, I felt both controller and motor. Barely warm, but I bet the windings on the motor were much hotter than the outside. Test conditions were near ideal for biking, with a temp around 72F. The controller cools off almost instantly when I stop.

View from the top. My house is below somewhere. Mt. Tam is the one furthest away in the first pic.
View attachment 1
TL ridge 3.jpg

Right now I have it set at around 2kW and everything seems pretty happy. The BMS overcurrent setting is the limiting factor right now. I might have to try tweaking that up a bit, as the batteries didn't really seem to get that warm. It can take more. I think I can get it up to 2.5-3.0kW without too much work.

Startup from a dead stop is so much snappier than the stock setup. I can keep up with most cars up to 25mph now.

I played a bit with the field weakening feature too. Normally it would top out around 27mph on flat ground. With the overdrive engaged, it takes about 2kW to do 30mph, so that's what it does. On a slight downgrade, it will keep pushing up to 35mph.

On my 15 mile test run, it got 30W-hr/mile, which is quite good considering the terrain and punishment. The stock setup got around 20W-hr/mile. If I use the 3-speed switch to lower the current limit to stock level, the consumption should be about the same. I'll try this at some point, as I suspect the new controller is significantly more efficient than the old one.
 
Did a pretty grueling test ride yesterday. Lots of climbing on the fire roads that are extremely steep in spots. The weather was pretty warm, about 90F with little wind. I did manage to get the motor hot enough that I couldn't hold my hand on it, probably a little over 60C. The controller also got warm, but not nearly as hot as the motor. Going at slow speeds on the trails isn't good for cooling. I think if there was a little wind or if the ambient temp was a bit lower, I wouldn't be limited much. According to the Google topo map, the climbs had a vertical rise of around 600 feet over about 1/2 mile, but since the roads go up and down like a roller coaster, the total climbing was much more than the difference in elevation between start and end. On the really steep uphills, the bike slows down to about 10mph, which is a good speed for pedal assist, which I did. Not sure if it would really make it without pedaling, but I assume the motor power is waaay more than what my legs can do. I'm no Lance Armstrong.

On the climb section, the CA reported energy consumption of 106 W-hr/mile.

A few short stops got things cooled off and once I made it back to pavement and got some speed, things cooled off very quickly.

I really like the tires. The fire roads are mostly hard packed dirt and the tires really grip nicely. At no point was I losing traction or skidding. The regen feature really takes a lot of heat off the brakes even at relatively slow speeds.

Chicken Shack fire road: Overlooks SF bay.

Chicken Shack 3.jpg

Chicken Shack 4.jpg
 
Alan B said:
Not that you want to change it, but would a mid-drive fit on that machine? If you are going to use it for that kind of riding a BBSHD, Cyclone or LightningRods would handle it well.

It would probably be a lot less work :wink:

I've considered those for sure. The trade-off is in reliability and noise. I'm going to try pushing the limits of the direct-drive hub motor for now. With the 20" wheel and stock windings, the kV seems to be just right for what I'm doing. I also recently acquired two more of them for next to nothing, so I have an ample supply of spare parts.

I'm really enjoying the total silence of the motor. If I'm doing a really long climb on a hot day, taking a cooling rest half way up is something I don't really mind.
 
The BBSHD is fairly silent. If you want noise the Tangent is probably the most fun. :)

If you don't mind adding a few pounds to an already heavy bike, it would be fun to have both the rear hubmotor and the mid drive. Just use one or the other as conditions require. Reserve the mid drive for cranking up those steep trails, use the hub otherwise. I use a small thumb throttle on the HD.

PAS on the HD is also nice, then the throttle finger or wrist can get a rest.
 
Your A2B seems like a very functional machine as is. I have come to think the complexity of mid drive outweighs the benefit unless weight distribution is an important factor. The main benefit of mid drive to long term ebike builders is just to try somethings new.
 
mr.electric said:
Your A2B seems like a very functional machine as is. I have come to think the complexity of mid drive outweighs the benefit unless weight distribution is an important factor. The main benefit of mid drive to long term ebike builders is just to try somethings new.

Some mid drives are exceedingly simple, like the BBSHD. It is a simpler install (on a suitable frame) than a rear hubmotor kit. Bafang has done a good job with integration, cabling and plug and play connectoring.

The main advantage of the mid drive for me is climbing steep stuff without getting hot. The BBSHD on my RidgeRunner climbs to the traction limit without getting more than barely warm while the hubmotor will literally stall or burn up if the gradient is steep and long enough. It also makes a lighter bike with better weight distribution and is easier to take wheels off when fixing flat tires.
 
For sure if I was starting from scratch I'd go for a mid-drive, but since I got the A2B for nearly free, it's an exercise in optimizing what I have.
I'm pretty happy with it now, but it is a heavy sucker. I always have concerns about the gears wearing out in any kind of geared motor, but so far I haven't heard of anybody blowing the gears on a Bafang. As much as I get out riding, they would probably last a lifetime for me.
 
I did a hill climbing test on a local street with a pretty steep grade.

Here's the view from the bottom:
Esmeyer street.jpg

This gives you a general idea for the grade. Even cars struggle to go up this street:

Esmeyer street 2.jpg

Here is a measured grade profile. The average grade is 15%. There are sections near 20%.

Esmeyer Grade 2.jpg

I went up to the top with no pedal input. On the steep sections, I was slowing down to 10-12mph but it kept going. Anyplace the grade was less steep, the speed picked up right away. At the top I checked temps. There was some noticeable heating, but nothing was too hot to hold your hand on. Power consumption was 146 W-hr/mi.
 
I went on vacation to Oceanside, CA, which is near San Diego. I loaded up the bike in the back of the minivan. One problem right away was the stinky smell from the tires. Chinese rubber offgasses forever. I eventually had to find some large garbage bags to put around the wheels to keep the smell from permeating the van. Near where I stayed, there is a paved bike trail that goes about 10 miles along the San Luis Rey 'river', which is completely dry this time of year. I did 30 miles on one trip, which used a little over 1/2 the capacity of my ammo can battery. The trail is pretty flat, but there is a steady sea breeze you have to fight in one direction. Going about 20mph, the CA indicated 22 W-hr/mi average for a round trip. Not too bad considering the moped tires. I probably racked up about 100 mi during the week. I might need to find a better saddle for those long runs.

Bike trail:

View attachment 3

Large cactus along trail. Don't crash into that!

Img_0220.jpg

The mouth of the river where it goes into the ocean: You can barely see a bit of the bike trail on the right side.

Img_0218.jpg

There is a long section of beach and a road that goes along it. Lots of beach cruisers.
Oceanside beach at sunset:

Img_0228.jpg
 
Mid drives are neat and effective for steep offroad/trail riding, giving high torque at the wheel for little weight.
BUT.. Their big drawback is cost. At least double or more the cost of an equivalent hub motor.
Not many mid drives in the used adds yet at reasonable prices, either.....but plenty of A2b's ! :wink:
PS...Fechter.... Ever thought of dropping that battery weight down lower ?..maybe a panier pack.
Im sure it would improve your ride and handling.
 
Hillhater said:
PS...Fechter.... Ever thought of dropping that battery weight down lower ?..maybe a panier pack.
Im sure it would improve your ride and handling.

Yes, I have thought about adding a second one and placing both of them on the panier rack (keep it balanced). I could go over 100 miles with that much battery. Actually, my rear pack weighs in at 12.5lbs and I really don't notice any handling problems even on rough dirt roads.
 
Another epic ride today. This time up Scettrini Fire road, which I can get to from my house by bike easily. This climbs up San Pedro Mountain on the opposite side from China Camp State Park. Looks like I climbed about 800' vertical to the place I turned around. About half way up, I stopped to let the motor cool off and take some pictures. I think if I just stayed on the throttle all the way to the top, something would cook. There just isn't enough wind at 9mph to cool the motor very fast. It does cool off nicely if I park someplace with a good breeze.

Topo map of the area. The map contrast is sucky, but I thank Google for it anyway. Starting elevation was barely above sea level, maybe 30 feet.

San Pedro Mountian Topo.jpg

View along the way up. San Francisco bay and US 101 in the background.

spm1.jpg

Marin Civic Center in the middle. The building was designed by Frank Lloyd Wright

spm2.jpg

View from higher up:

spm4.jpg

Cooling off spot. My house is somewhere in the background slightly left of center.

spm5.jpg
 
Today, a road called Little Cat fire road. I've been lucky to get out a lot lately and the weather is perfect, but maybe a bit warm (around 80F).

Not so bumpy, but still really steep with some ruts and loose spots. The picture really doesn't show the grade well.
Lcat4.jpg

Looking to the east over SF bay.
Lcat3.jpg

Looking south, toward home
Lcat2.jpg

Good cooling off spot.
Lcat1.jpg
 
I finally got my temperature gauge connected and dialed in. The gauge unit runs off 5V, which I took from the throttle. Input range is 0-1v. I made a divider network and added a trimmer pot to calibrate the needle at full scale. I take a trimmer pot and dial it in to the resistance from the thermistor chart, then plug it in to the cable where it has a connector near the motor to simulate a given temp. Then I can dial in the calibration pot to get the right full scale. When the meter goes over range, all the segments start flashing, which is a nice indicator that it's time to let the motor cool off. Since the thermistor is jammed between two copper windings, it can get pretty hot before the stator heats up. I have it dialed in for a max of 150C and that seems to be about right

Temp gauge 1.jpg

I was able to do a nice long climbing ride without stopping to check temps. Motor shell never got too hot to hold my hand on it today due to nice breeze and cooler air temps. You can really see the thing cool off when you start going downhill but the regen generates a significant amount of heat. I might need to make a switch to disable regen at times when descending and wanting to cool off the motor.

I did the first part of the 680 trail, which is very nice. The numbers on the CA are for the entire trip and back to home.

680 Trail.jpg
 
In the quest for finding the limits of the motor, I replaced the shunts on the BMS so the trip point should be around 85A. Then I was able to increase the programmed current limit on the controller to 60A without tripping the BMS. On takeoff from a dead stop, there is a short (100ms) blip of 65A, then it settles down to about 56-57A if the wheel load is high enough. Never gets to 60A. I think this is really the limitation of the battery voltage vs. the resistance of the motor windings and phase wires.

Well, I'll just have to settle for 2900W as the maximum unless I want to increase the battery voltage. Not bad for a motor officially rated at 500W continuous, 800W peak. In the max. current mode, it climbs the steep hills very nicely but the temperature climbs nicely also. With the ferrofluid in the motor, I can stop when my temp gauge maxes out and it will cool down in a few minutes to where I can start climbing again. On a really long climb where I make several stops, the motor kind of gets heat soaked after a while and it does not cool down as fast. Time to look at adding fins to the shell.
 
Got Nailed!

Out for a nice Sunday cruise before the rainy weather hits, I was traveling along a busy street on my way to the dirt trails. Things were normal until I suddenly hit something in the road that felt like a big rock or piece of wood, but I never saw anything. Tire hopped a few times like running over multiple pieces of junk and it make a pretty nasty sound as something scraped against the rear fender a few tire turns and then disappeared.

Where this happened was not a place good for pulling over. I looked down at the wheel and it looked OK, so I continued.
After getting past the busy part of the road, I slowed down and looked at the tire again and it seemed fine, so I continued onto the dirt trail. About 1/4 mile later, I hear the flat tire sound, flop flop flop, so I stopped to check it out.

Img_0269.jpg

Nice nail head.

Taking the tube out of the tire while on the road is not a good option. It requires a tools that are too heavy to carry routinely, including a large wrench to loosen the rear axle nuts and a pair of tire levers made for motorcycles. Alternately, I carried a can of "Fix-A-Flat" for bicycles. I figured I would at least have a chance with this stuff and had success with it years ago. The moped tires are very stiff and thick compared to normal bike tires. Nearly puncture proof, but there's always Murphy's Law.

Img_0271.jpg

Unfortunately, I did not read the instructions before removing the nail. Big mistake. The nail was a full size framing nail which had gotten bent probably by hitting the swing arm as it went around, then was fully inserted by a few more turns of the wheel. It was hard to pull out.

View attachment 1

Well, as you might expect, with a 4mm hole in the tire, the tire sealant just all ran out onto the ground and the tire did not even begin to inflate. Bummer.

Luckily, the moped tires are so stiff the bike was easy to walk even with the weight of the secondary battery on the rear rack. I even tried riding it and it went, but I was worried about damaging the side walls of the tires so went back to walking it. I think I could ride on a flat front as there is less weight on it.

I walked about 1/4 mile back to the pavement and called my wife for a van pickup. How embarrassing. I could have walked the thing all the way home, but it was about 3 miles and that would have taken about an hour.

I got a flat on the very same section of road about a year ago. I guess it's an area that never gets the street sweeper and there's a lot of contractor trucks driving along there.
 
Next time I'll just leave the nail in!

I also found out during this that my little hand pump won't fit on the rear wheel due to the limited space between the valve stem and the hub motor. Guess I should have tested that first. Well I know now, and I found some metal 90 degree tire valve extensions that solve that problem nicely. The extension stays with the pump now. New inner tube just arrived so I can proceed with fixing the flat. Raining now anyway, so good timing.
 
Awesome looking ride photos. You have discovered some beautiful trails.
 
Back
Top