Nighthawk Motorbike EV Build

MattsAwesomeStuff

New member
Joined
Jun 7, 2018
Messages
28
I have converted a 1985 Honda Nighthawk 750 to electrical.


image_uavhsb.jpg

(Before butchering).

It is a rescue, it was being parted out as it's 33 years old and the owner (who bought it to restore) had botched his engine restoration. I paid $20 total, after he'd stripped it apart.

I used the motor from a small 1969 forklift, the speed controller (for now) from a golf cart, built my own matching splined coupler with a rusty hacksaw, and since this is a near-free garbage-build, I originally planned to use Lead-Acid starter batteries from the junkyard that I'd buy for the core charge and return for the core charge when I murdered them (starter batteries do not like being deep discharged).

But then I found a nearly bottomless source of batteries from recycled tool packs. Hallelujah Lithium. I began The Great Battery Hording last September. The rest of the bike has been basically done since then, I repaired and reassembled the bike (from a rubbermaid of unlabeled parts and a garbage bag of fasteners, knowing nothing about mechanics at all or how to even ride a bike). I'm just waiting on battery.

I ordered and have been running a pair of Zanflare capacity testers non-stop, 24/7, since February. I've been both pre- and post-charging on DIY 6x 5-bank of TP4056s.

And today, after... (Feb,Mar,Apr,May) ... 4 months, I've finally cleared my backlog and I have no more cells to charge.

I'm free.

Free at last.

... unless you count the few I'm saving to film tutorials and teardowns.

... until Monday, when I pick up more.


I honestly feel a bit lost, this has been such a present routine in my life for so long now it's a bit like standing naked in the grocery store (you know sometimes you just replace one habit with another). You know the routine. Wake up, mark values of cells completed through the night. Empty post-charging banks. Move night cells into post-charging banks (if needed). Take cells from the pre-charge-complete pile, load them in, start them up. Go brush your teeth. Pack up your gear, take it with you so you can cycle it all day. Every 4 hours get twitchy as it's about time to rotate cells again. Pack it home, cycle it in the evening. Load it up for the night. Brush your teeth, go to sleep. And now here I am writing this on my phone while I wander around the grocery store OH SHIT SECURITY'S HERE AGAIN BRB.


Where was I? Oh right, batteries. I have enough good, capacity-tested cells to about fill 3 milk crates to the brim. Average cells are around 2200mah. Smaller packs are the power-heavy 1300-1500mah packs. Larger packs and older packs are around the 2200mah mark. Newer medium-sized packs are 2500-2700mah. Somewhere in the ballpark of 2000x 18650s.


And 340x or so 21700s (3350mah average). 330 of these (11s30p) will be my practice pack-build for a moped.

Trivia: 1 milk crate vertically stacked with 4x5 square-packed cell holders, 4 tall, holds 30 packs or 600 cells.

Trivia2: 1 milk crate horizonally stacked with loose cells, hexagonally-packed without holders, (4 sections) is 256 per section or an even 1024 per crate.

Trivia3: I need only around 1320 cells for my bike. Because that's all the weight/space I have for them, probably.



Y'all are building powerwalls so ain't nobody got time for boring motorbike build stuff, so, lemme go through my cell setup.



Charge in-place:


First I tried to build a charger that could charge or at least test the various tool packs without disassembling.


Challenge: TP4056s are not isolated, if you charge up a string of cells in series (as in a tool pack) they're gonna assplode (I think? Never verified). So my solution was to connect individual, isolated power supplies to each. "Yo you cra-zay! Those are like, ten bucks each!" you say, nay nay, $1.25 from Dollar Tree. I was too cheap to spend another dollar on a USB--Micro cable for each, so I ripped apart the supplies, desoldered the 120v plug and the USB-A terminal, and wired them directly to the TP4056.


image_uafgwb.jpg

(spoiler, it didn't fuckin' work)


image_evmywa.jpg


I ganged them all up in series and used the only sensible, free wiring system available: Cat-5.

I made a 5-series charger that ended in the female network jack (half of a Cat-5 cable coupler, $1), and then started converting junk tools with spare network cables so the system was modular, you just plug in whatever brand and style you want into the base unit.

Oh, and voltmeters. I mounted 2 rows of them, intending to mount a second bank of chargers when this one worked out.


image_hikjtb.jpg


I didn't like how it worked, it gave really weird results for reasons I couldn't figure, various meters would flick or show nothing (and would randomize which meters if I unpluged and replugged). All cells would stop charging at the same time, rather than each blink out as they reached capacity. I think I was whispering nightmares into the ears of the onboard BMS with my charging system.

I couldn't even get good readings off the voltmeters with the power supplies switched off, so it was useless even as a testing unit.

Charge-in-place was abandoned. Oh Danny boy, the pipes the pipes are calllllling.


Charging Rack:

This is the 6 independent 5-parallel charging rack I built and have been using since February, using mostly recycled tool pack bits. Tool packs across all brands are 20v so they all have 5-series plastic holders for them (sometimes 2 or 3 stacked, but 5x). Hence, I had a lot of them.

Tool packs also have beefy-but-short springs for the battery retention clips. Perfect.

Brass thumbtacks from the dollar store. $1 for 200.

More Cat-5 cable, free.

6 more power supplies, 6 more TP4056s, 6 more panel meters.

Off-brand tupperware to the rescue and all the guts were nicely assembled.

Chopsticks and hot glue to keep the power supplies off the board.

BEHOLD!


image_hrpzja.jpg

(half racked so you can see the guts).

I was unaware at the time that TP4056s can share a common DC bus for the input (as long as the batteries themselves are independent), or I would've strung up an old ATX PSU and used 30 of the 4056s. Instead I cheaped out and only used 6 because I didn't want the ratsnest of 30 power supplies.

If you want to imitate this, a few notes:

1 - Place the tacks by pressing them into the wood strips before you mount them to the wooden base. Use needlenose pliers from the side (holding the tack like you'd hold a tophat).

2 - The tacks should stick over about 20% from the edge of the wood. Set them in place, then push them almost, but not all the way down.

3 - Wire the tacks by stripping wire and just making a circle or two around the tack's pin, then a few light taps with a hammer to lock the tack down.

4 - Set the wood height, (or the plastic holder height) so that the battery diameter protrudes 30% or so out from the wood. That gives your thumb somewhere to grab the battery. You're doing this hundreds of times and will rip your thumbnail off otherwise.

5 - If you purchased 18650 holders... you fool. However, cut the damned sides off that grip it so hard, these aren't going in an aircraft. Certainly don't have the holder wrapping the battery more then 50%. 30% is enough to stop it from going sideways. The spring tension is enough to keep them in place and you'll be able to easily grab them.

6 - The spring goes on the negative side. ALWAYS. It will gouge and short if you try sliding it on the positive side. This is industry standard even for AAs. The thumbtacks are great for the positive side because they are smooth and slide and only contact at the peak.

7 - Build it upside down from what I did. Positives upwards, springs at the bottom. That way you are compressing the spring into your tabletop. If you do it the way I did, when you load and unload batteries you are lifting upward on the springs and thus lifting the whole board off the table. Also, if you care about alignment, you care about positive, not the negative. Also, if you care about sorting, you can see the positive side this way, which identifies different cells by the color. If they're positive-down you can't see unless you put your face on the desk.

8 - Give yourself lots of spring room. You want to compress the spring 20%, not 50%. Size it big enough and you can fit 21700s in there too.

9 - Steel/Stainles Steel has 10/100x the resistance of copper. It makes for an okay resistor so you can slam multiple cells in parallel if they're within a half-volt of each other. Also if you screw up, put one in upside-down, or parallel a 3v cell to 4 other 4v cells, the zinc/chrome coating will hiss and release white smoke as it boils off and the spring softens from the heat. It's kind of like a audio-visual indicating fuse.



One thing I didn't expect is that the TP4056s to have such variable end-voltages. Some finish at 4.25v, some at 4.19v.

I don't know how the 4056s determine cut-off, I've heard it's based on a percentage of initial current? Or, some fixed mA flow rate? In any case, when you have 5x in parallel, you're going to have 5x the termination current, so it tends to overcharge them all a bit. I'd recomment using a separate 4056 for each cell, no reason to gang them up like I did if you have a common power supply (4056s are $0.10/each from China, spend your $5 and get 50).

Anyawy, I like to give back to communites that've helped me and, even though I just joined the forums, the Pack Builder has saved me from having to write code to do that myself, so, thought I'd share a bit.

I now have more cells than I can fit in the frame so it's time to build put them together and figure out how to mount them.


image_dprist.jpg
 
- Amazing crazy project ! (my last motorbike (sigh !) was the europen version of the nighthawk)
- Stuffed with sensible advices (most of them smelling like "don't ask how i know it")
- and inventive diy (cell holders seem to be an infinite field for cheap and durty research)
- Perfect story telling !

That's an obvious THUMB UP !!
 
The TP4056s have a 1% variance. So, of 4.2, that's a 4.158V - 4.242V swing. Which, some could potentially push to 4.25 as you've observed. Those are operating out of spec, obviously. Or, they are the fake ones and not legit.

The reason why the series connection of the TP boards to charge battery pack is because they all share a common ground. So basically, you created a dead short from end to the other through ground and to all connections. Poof!

Great work on the other stuff, tho :)
 
Those are operating out of spec, obviously. Or, they are the fake ones and not legit.

Are you telling me that the circuitboards I ordered 50 of for $5, shipping included, might not be genuine? Sir I am shocked that you would make such an accusation. How dare you. My projects don't walk around wearing knockoffs like yours may.

Clearly what's to blame here are the panel meters which I paid $1.37 each. Those must be out of spec. I won't accept otherwise. I can't.
 
LOL touche! :p Maybe the meters can be calibrated. Do they have a pot on the back of the PCB? A good way to test the TP results are to put a DMM on it while it's charging. Then you'll know where it's cutting out at.
 
Hi nighthawk. I am Vic in Toronto Canadah! eh?

regarding bulk charging, I have found that i can take 2 of the 4-cell holders, stick them together to create a 8 cell holder that parallels all 8 cells.

So all the cells over 2.8 volts can be paralleled 8 at a time. Then i leave them some hours, and they will all settle to the same voltage.

I then use a 8 cell holder with an attached 4 dollar "buck converter / li-on charger (cc-cv)" board to charge the whole bunch to 4.2 volts.

image_izvqrd.jpg

Then I transfer them back to the plain, 8 cell holder to let them rest for a couple hours, and then its off to the resting bin for a week.

note: I use a solar bus wire to join all the cells inside the holder by adding it behind the clips. The pins will connect the 2 halves with or without soldering.

sorry for the duplicate images (im new here)
 
So all the cells over 2.8 volts can be paralleled 8 at a time. Then i leave them some hours, and they will all settle to the same voltage.

I wouldn't be joining cells that far apart, you'll fry them by overcharing them in a short.

I then use a 8 cell holder with an attached 4 dollar "buck converter / li-on charger (cc-cv)" board to charge the whole bunch to 4.2 volts.

I also have one of those I clip onto a section if I want to charge it faster. Also, to charge a whole 4x5 set at a time:


image_nfxixj.jpg


Buck converter, panel meter, boost converter to maintain fan voltage, and then springs and tacks to rack the whole 20-cell assembly.

I clip jumper wires on if I want to use the buck converter directly elsewhere.
 
Back
Top