So over the last few weeks I went back to my roots a little and spawned an electric bike out of basically just the hot garbage in my robot/van/go-kart storage totes and a $20 yard sale frame. That post’s for (yet) another day, but what I also needed to do is charge it… and that is where things got funny.
I still have several standalone e-bike/scooter chargers from the 2.00gokart and Power Racing Series days, but they’re all for 10S (37V for LiCo chemistry or 33V for LiFe) lithium packs. I also have a 12 LiFe-only one that I used for Melonscooter, which is the one I will be hacking up.
This new e-bike runs 12S regular lithium cobalt, with a nominal voltage of 3.7V instead of 3.3. I just had to figure out how to crank the charging voltage up some, from the 43.2V of charging 12S LiFe cells at 3.6 volts apiece, to 49.2 volts for charging lithium cobalt at 4.15 volts per cell.
So let’s crack it open. Basically every switching power supply architecture has means to do fine tuning and calibration on the factory floor, you just had to find it. They usually can’t adjust far, which is why I started with my highest-voltage charger.
I did a quick scour of the Internets for whoever has done this before. Luckily, the usual suspects at Endless Sphere (still alive and buzzing in this day of easy social media share buttons) have done this before, and I found some useful information in this thread. I seem to have a visual match with one of the chargers posted in the thread.
These things haven’t moved much technologically for over a decade if not more, so it was easy to make the visual correlations with mine. Here’s the potentiometers on my model, which was sold by ELifeBike, still around today as PSW Power:
From the schematic posted in the thread, I deduced that the termination voltage adjust trimpot was right by the output status LEDs. They usually are some place obvious for the technician. At the time I didn’t know what the other two trimpots did, but figured they were charge current and charge termination current threshold (At some point, your charge current in CV mode falls so low you might as well call it good).
The way I was going to make these adjustments, obviously, was all live, all the time. So if you do this, just remember that even if the output sounds relatively harmless like 24 volts, switching power supplies still will pack a few hundred volts right next to that. So, avoid poking the wrong thing.
We begin the prodding by checking the charger as-is after scattering it on the bench. Hmm, 44.3 to 44.4 volts on the output, you say. That’s above the termination voltage you’d want for LiFe/A123 cells, but not enough to really hurt anything.
I began cranking the CV trimpot and (many many turns later) got the voltage up to about 49 volts. Counter-clockwise is increasing voltage, at the rate of what seems like 0.2 to 0.3 volts per full rotation. I was scared of running out of potentiometer, but it got there. These small vertical trimpots are usually 25-turn.
I generally bulk charge my EV batteries, so I don’t take them all the way to 4.20 volts. Historically I’ve popped them out once every few months and balanced the packs manually…. if at all. Beyond that, you make an assumption about how far apart you can stomach having the cells drift, and assign a little safety margin. For instance, by charging to 4.15V per cell, you are saying that the summation of all cell voltage deviations both high and low shall be no more than 50mV…which is quite a lot, by the way.
So that’s why I didn’t adjust it all the way up to 50.4V, which is 4.20V/cell for 12 cells. You only get 3 or 4 percent of charge going that high compared to 4.1-4.15V/cell and it just makes for a much more relaxing experience. I dunno why, but I expected some kind of instantaneous catastrophic failure as soon as the thing hit 50.0 volts.
Next, I wanted to mess with the charge current not for any hot-rodding reasons, though you know me, but to see where the adjustment is made. On the PCB, the two (what I think are the) current-adjustment trimpots are located right next to a dual op-amp chip, part number HA17358. One of them probably adjusts the CC stage current, and the other the cutoff current.
I just picked one and started messing with it, and hey, it’s the correct one.
Pursuant to the “You know me…” up there, I gave it a few whirls to bring the charge current up to 10 amps. The “rate of adjustment” seems to be about 0.3 amps per rotation, so I was turning this thing forever to get to 10 amps from 8.
To dial the current in, you have to actually be charging the battery. Luckily, this bike uses salvaged Overhaul and Sadbot batteries. They’re 6S and 6Ah each, and I run them in a “2S2P” arrangement to get 12S 12Ah. They’ve been sitting a while, so were discharged somewhat.
Finally, after keeping an eye on it for a nervous 30-something minutes, I decided to see if I could change the threshold current for ending the charge cycle. This is probably something that is utterly unnecessary, but curiosity!
As I watched the current drop below 1 amp, I decided to give this trimpot a few spins to see if I could induce the cutoff. This time, clockwise seems to raise the cutoff current. I spun it forever counterclockwise before I realized I should probably go back the other way, as I did not actually count any of the revolutions.
A few turns clockwise later and the green LED turned on, indicating the cycle is complete. Again, there’s probably no need to mess with this at all.
While I was inside, I decided to also go ahead and shore up the completely unprotected PCB with some conformal coating around the chips and sealing the connectors. This thing no longer lives in a climate-controlled building, so I figured it wouldn’t hurt.
Some day you’ll hear about the bike itself, I promise!