DIY Electric Car Forums banner
1 - 20 of 28 Posts

·
Registered
Joined
·
236 Posts
Discussion Starter · #1 · (Edited)
Brand new, barely released into the US market, less than 2 weeks old and I'm already bored stiff with it! Nice platform, mostly good build quality with some missed attention to details, but a solid platform that rides really nice! I've already been taking it apart. I guess the warranty is voided...OOPS!

It has a 64" wheel base. It's just 2" shorter than the Zap. It rides like it's lots bigger than it is as a result. It doesn't feel or handle like a twitchy short wheel based scooter.



I took this March 8 2022. So yeah...brand new! It now has maybe 40 miles on it and I've charged it once. The battery came with a very light charge on it which was enough to ride maybe 5 miles and try it out as soon as it was assembled.



Anywhere there is a large "Dualtron" logo, there are RGB addressable LEDs under it. The included LED controller does some interesting effects via a little IR remote. The plastic parts (fenders) are nicely molded and not rough. Ignore the sticker glue that I had yet to clean off. Integrated into the back of the rear fender is 4 LEDS for brakes and directionals that are plenty bright.







Both wheels are this way. You can see on the left side, the valve stem. This was stupid! There is not enough room to get a 90 adapter in there and typical bike pumps lack a schrader fitting short enough to fit that little gap for filling your tires. The inner tubes REALLY need angled valve stems! On the motor cable, you can see where the outer insulation has torn away. Apparently Dualtron uses super glue to secure the cable into the channel in the swing arm. Pulling it back out tears off the outer insulation. Apparently you are not supposed to ever need to service your EV!



This was disappointing. I pulled it out of the packaging with a kinked rear brake line. The scooter has hydraulic brakes. It's just a matter of time before this starts leaking. A new brake line is on order and covered by warranty. I have a bike brake bleed kit. This won't be hard to get the brakes going again after I replace the brake line.




The accessory kit covers everything you need to assemble the scooter. The manual was written in clear English with a few typos here and there. It did a reasonably good job covering assembly. Some functionality that you can change settings on were lacking any explanation what they do. The charger is almost useless at 1.75 amps! 26Ah battery at 1.75 amps = 14 hours to charge.



It came with a decent set of tools that were perfect for assembling the scooter. I really didn't need anything else to assemble it.



A few days after owning it, I was taking things apart to take a look inside. There's a small cover that hides all the wiring and controllers in the very front of the deck. In it is a switch that is clearly labeled "Spare Switch". It literally does nothing and goes nowhere as can be seen by the lack of wires going to it. The dual charge ports plug into the battery wires via automotive grade bullet connections. I would have used XT30 connectors...and will.



The scooter has front and rear motors and when I pulled off the heat shrink around the rear motor bullet connectors, I found them like this. I wasn't pulling on them. Whoever put the phase wire connections together never checked if they were pushed together. The left most bullet is barely connected.



Crammed inside that space is a tiny 5a DC-DC on the right side, a mess of wiring, LED accent light controller, mosfet switch for the head lights, flasher module for the directionals and a dual 6 fet controller.

 

·
Registered
Joined
·
236 Posts
Discussion Starter · #2 · (Edited)
Power, acceleration and speed: (Meh!)
I guess if you don't know what an EV ought to be like, then you'll think the Dualtron City is pretty quick and reasonably powerful. I've been building my own EVs since 2015 and IMHO the City is under powered and slow. A friend suggested that it was designed for cruising around in a city and not intended to be quick. I can't speak to what Dualtron intended, just what I observe. Top speed I ever saw was 40 mph...while going down hill. Climbing slight hills, best I could do was 25 mph. From a dead stop at the bottom of an upward roadway, 22 mph was my top speed. I think I rode maybe 1 mile before I had all the power settings maxed out. Accelerating up hills is simply NOT going to happen! It supposedly has 3950 watts of power. To me it feels more like 2000 watts at most.

Braking:
Regenerative braking has 5 fixed settings. I changed it to 5 to max out regen and to assist the hydraulic brakes. The hydraulic brakes are OK, but not good enough to brake at the limits of tire grip.

Head lights:
For the rider, they are low and point where you need light despite being integrated into the sides of the deck. From the perspective of oncoming traffic, they are almost invisible since they are shrouded and very low to the ground where anyone whose eyes are just a few feet above the ground won't see them shining. There are no other forward facing lights.

Side lights and directionals:
The RGB LEDs are on all 4 swing arms, both sides of the battery box, up the steering tube and on the rear wheel fender. There's decently good visibility from the sides and rear. Anywhere there is a "Dualtron" logo it's probably got RGB LEDs under it. The rear brake and directional lights are bright, but clustered too closely together. I might have to do something about that and move the directional lights to the ends of the rear swing arms and use the center cluster for tail and brakes only. There are no running or directional lights of any kind facing forward.

Battery:
The battery compartment encloses a replaceable battery pack. There is no precharge or anti-spark feature. Plug it in with a ZAP if the filter caps are not already charged up. The lid does not seal and the bottom of the battery box has deliberately drilled holes in it so it can drain out. I rode 40 miles in semi wet conditions with a few puddles and some snow in places. I had water and dirt inside the battery bay after 40 miles! The design of the deck lid and the lack of seals is the main reason for water incursion. This was soooooo easy to NOT have happen and Dualtron "fixed it" with drain holes! Having a swapable battery seems nice and I guess if you deliberately keep a battery at your destination so you can swap it out, then this is good thing. Most people, they go where they go and they won't have a spare battery when they run out of charge. I found this to be more of a gimmick than anything useful to most people. Never mind the wasted space that wasn't holding more batteries! I'm fairly certain that the 16S, 26Ah battery minus that space wasting plastic box can hold 20S and 26Ah. I'm sure going to try to make it fit!
 

·
Registered
Joined
·
236 Posts
Discussion Starter · #3 · (Edited)
Upgrades in the works or already done:

1. Tubeless tires and 90 degree valve stems: The inner tubes make the wheels unbalanced. This is pretty common and something most EV makers never correct with wheel weights. The City was not different. First thing I did was take the wheels off the motors...which are conveniently bolted to the hub motors and removable. I then pulled off the tires, pulled the tubes, installed right angle low profile tubeless valve stems and then installed the tires again. The rims are split, but also have the lowered center for making tire removal easier when getting the tire bead over the rim. There was no need for split rims. I sealed up the joint with several wraps of PET tape. This will hold air for many years. Both wheels leaked slightly at the rubber gaskets on the valve stems, but I use Stan's tubeless sealant anyway and that soon sealed up. Filling the tires is no longer a giant hassle and they now weigh less, are tubeless and that means fewer flat tires.



2. Improved braking: The City comes with 160mm rotors. They are too small to provide more than gentle braking. I bought 203mm rotors and the bracket extensions for the calipers. Those parts arrived 2 days ago. I won't install them until I'm ready to mount the motors again.



3. Motor hall and temp mods: They lack halls and temp sensors in the motors. I notice that from a stop that sometimes the scooter kicks backwards for a second and then takes off once the correct direction of commutation is detected. Statistically, the controller will guess commutation correctly 50% of the time from a dead stop. Halls solve that 100% of the time and make motor control better under all conditions. The stators have the cut-outs for halls, but none were installed. I need to strip off all the outer insulation on the phase wires anyway since it is ripped off in several places. Adding a hall/temp cable to enclose with the heat shrunk cover each set of motor phase wires will get is a good idea.



4. Flipsky recently released a 100v, 100a max VESC controller in the typical form factor of most EV controllers. They even used TO-220 mosfets in it instead of surface mounted mosfets. I have purchased 2 of them to replace the under powered Dualtron controllers. This will get me FOC control, variable regen braking, field weakening, 100v/100a max and a nice color LCD dashboard in a CNC shell from DaveGA. They are ordered and will be here in a few weeks.




5. Dualtron makes an add-on front light bracket that fits all their scooters. I have purchased one of them for front mounted LED headlights and directional lights. I will need to add a higher amperage DC-DC since the head lights could easily draw more current than the DC-DC that Dualtron used can deliver.



6. Dual thumb throttles: The one on the right thumb makes it go and the one on the left thumb is variable regen. I do dislike the "Hook trigger throttle" Dualtron implements. The factory LCD in the hook throttle doesn't display the data I want anyway, so it's useless for more than speed.

7. Hub motor venting: In my XM3500 build thread, I have been posting about modifications around my mini lathe. The side covers for the City hub motors will probably be the first thing to use thosw mods in making vents in the side covers. Conveniently, both sides bolt onto the magnet ring. Venting is likely to be needed when I run the motors at closer to 3000 watts each.



8. 82v/20S non-removable battery: 16S is probably fine with 30% field weakening for a bit better top speed. I'll do all the other upgrades first and see how top speed is at 16S on the factory pack. I'm dead sure I'll never see that mythical 55 mile range that Dualtron claims and with the added motor power, I'll see probably more like 25 miles. I may reuse the 21700 cells from the included 16S pack and add more to them at 16S for more range. I'll be implementing a smart BMS for 16S or 20S even if I use the factory pack as is. Over BT, it will transmit status to the DaveGA LCD. If I end up building a 20S pack, I'll use some Panasonic 3400mah 18650's I have for that.

9. New deck: The factory deck implementation sucks. It leaks dirt and water around the battery door. The rubber grip surface is stuck down with a not very good double sided tape that you need to pull up to access several of the screws to take off covers. That tape won't stick down for long. All the deck screws are stainless steel and they don't look very good against the black of everything else. I have purchased black oxide screws to replace all of the exposed stainless screws. The new aluminum deck will cover the entire battery box and incorporate a seal to keep out dirt and water. It will get covered in grip tape that never needs to be peeled up. All the screws that need to be removed to access panels that were under the old rubber cover will now be under the 100% removable deck. It will lock down and mount the exact same way the existing lid does, just seal up the battery bay and cover the entire deck area.




10. Charger: 14 hours to charge on the 1.75 amp charger is a joke! I did it once and that was already way too many times. I first looked at what Dualtron had and their highest amperage charger delivers 6 amps and they want $150 + tax + shipping for it. 6 amps is still more than 4 hours. I soon found out the charging port is a GX16-3 connector and they are easy to find. I then looked for 16S ready made chargers and found this 8 amp charger that cost me $112 total on Amazon. It's set to 16S right now, but via internal trimmer pots can be set to 36v to 90v. 8 amps gets charging down to 3.25 hours and that's reasonable.


 

·
Registered
Joined
·
236 Posts
Discussion Starter · #4 ·
This will be a fun little project and won't take long to complete while I'm working on things for the XM3500 in the back ground. If once it is built, I still think it's lame, I already have someone that will buy it from me for whatever price I ask. I already have all the cells, BMS's and LED lights. The other things I need are ordered and the rest is just machining and wiring.
 

·
Registered
Joined
·
236 Posts
Discussion Starter · #5 ·
I will be over watting the hub motors considerably. As is typical for all hub motors, they self cool very poorly. Literally THE WORST cooling BLDC motor design there is! The stator is trapped inside dead air space with no direct path to cool air. Over heating and cooking the motors will be highly likely without improved cooling.



1. I'll be cutting slots in the side covers to let in outside air and to eliminate that dead air space. This will create turbulent air flow inside the motor and that's good.

2. Inside the stator is 3 voids. I'm going to machine aluminum blocks that fit in them. They will have fins on them to increase surface area and to use that turbulent air flow to increase cooling at the backs of the stator teeth where heat is the highest. The blocks will get glued in place with thermal glue.

NOTE: I won't gooping up my motors with nasty ferro fluid.

IF anyone reading this can give me feedback, I'd appreciate it. I don't know which will work better.
I'm thinking about placement of fins on the aluminum blocks and what will cool the best.
1. Thinner block against the OD of the voids with fins across the width of the motor and inside those 3 voids.
2. Fill the voids with a solid block of aluminum and cut fins in the ends that curve and are inline with the direction of rotation.
3. Fill the voids with a solid block of aluminum and cut fins in the ends that are perpendicular to the direction of rotation.
4. Fill the voids with a solid block of aluminum and cut many square posts in the ends so that airflow from any direction hits a surface.
5. Something else?
 

·
Registered
Joined
·
236 Posts
Discussion Starter · #7 ·
All that lathe add-on work got implemented for real last night to make the below part. I opened up one of the side covers in the Dualtron City motors. It's not quite perfect, but still pretty good. As I went along, I refined how I did things and so early on in cutting these vents, I made some small mistakes that I now need to go over again and clean up. I didn't notice them until after I had pulled the side cover off the lathe. However, getting it re-registered so it all lines up is pretty easy. I made a jig that sets alignment based on the perimeter screw holes. It uses a section of rod in the ER32 collet to get the side cover rotated to the exact location consistently before being secured in the 3 jaw chuck.

This is fairly exacting and tedious work. Both the inner and outer arcs are 20 degrees of rotation. The straight sides of each opening are 17 turns of the cross slide. Last night, I went over the arcs 3 times to clean them up. I did the straight sides once and they show it in places. I need to go over them again at least twice to get them as smooth as possible.




For a first attempt and having never used this setup ever before for something I cared about, this is acceptable. Still...I'll mount this side cover again on the lathe and clean it up a but more. The 2 holes in it will get filled with epoxy before I repaint it. They will "go away" and be invisible soon enough. The milling marks where I machined off the Minimotors logo will get sanded to clean up the surface better. This is just the "initial rough draft".






Before and after initial machining.




The result in place on a motor.


 

·
Registered
Joined
·
236 Posts
Discussion Starter · #8 · (Edited)
Future improvements to this process:

1. initial material removal will be slightly under sized of the final dimensions. Hogging out aluminum with a 4mm mill is not exactly fast and there is a small amount of chatter. Under sized on all 4 cuts will mean the chatter will not matter as much. The final light cutting to size at much lower tool pressures will almost completely remove any chatter.

2. Go over all 4 cuts at dimension twice to make them clean and smooth as possible.

3. My 4mm mill has 12mm of cutting height. I have some mills coming in the next day or 2 than have longer side cutting surfaces. This will mean not needing to adjust the mill in and out depending on where I am cutting. The inner arc is exactly 12mm thick material. This is the maximum limits of the mill and I leave a small shoulder if I'm off with the depth of cut even slightly.

4. I cut a radius on some HSS for removing the "Minimotors" logo. I got a decent amount of chatter with it. I bought a 6mm ball end mill. I think I'll try that to remove the logo on the next side cover.

I noticed 3 things to "tune" on the lathe setup:

1. The shaft that locks the spindle to the rotary table flexes a little. There is a long slit in it that I cut too long. It's almost 3" long and I really needed about 1". I can weld this closed and that will remove about 90% of the flexing. Maybe, make a new shaft? I have enough steel pipe to make another one. I can correct a small dimensional issue in it at the same time.

2. Ways and gibs need regular tuning. This is a fairly common thing needing to be done on any lathe. My compound is a smidge loose. I need to tighten up the gibs a little to remove this movement.

3. The carriage assembly has a lock on it, but the cross slide and compound do not. I am dependent on the lead screws to hold position and no matter how well I might adjust the anti-slack nuts, there's no such thing as perfection. Vibration from hogging out metal causes the lead screws to move a little as well. I need to add gib locks so the slides can't move once I set a position on them. This is something I've known about and put off for several months now.
 

·
Registered
Joined
·
236 Posts
Discussion Starter · #9 ·
I've done some refining of my toolpost grinder setup and tonight I cut another side cover with improved results.



This is cut in 2 passes. The first pass is hogging out lots of aluminum and slightly under dimension. Any chatter that happens is not important since there is a second clean-up pass. Each hole as a result gets 2 complete sets of cutting. The second cut has about 10-15 thousands of metal to clean out so the result is reasonably good.






 

·
Registered
Joined
·
236 Posts
Discussion Starter · #10 ·
I wrote a list of improvements I wanted to employ before cutting another motor side plate.

1. I used 2 passes. The first one is slightly under dimension and the second one is on dimension. The results were good enough that I didn't bother doing a third pass at dimension.

2. See number 1.

3. The new mills didn't have longer cutting surfaces like a I thought so I used the old 4mm mill instead and made sure I was always at a depth that made compete cuts through the material thickness.

4. I did get a ball end mill, but there is no logo on the other side of the motor to remove. I'll use that soon enough.

Tuning:

1. I made a new shaft for the lathe spindle and it is a tighter fit. However, it didn't solve 100% of the flexing issue, just reduced it a fair amount.

2. I took apart the compound and cross slide, cleaned them thoroughly and stoned the gibs so they are much flatter and smoother. I replaced all the lousy gib set screws with much better screws that are easier to access. Everything got reassembled and tuned until the ways were tight and slop free.

3. After retuning the cross slide, the little bit of movement in it was gone so I didn't bother with a locking screw.
 

·
Registered
Joined
·
236 Posts
Discussion Starter · #12 ·
The motors are setting up so the thermal glue can cure. I got the headlights yesterday and so I thought I'd mess with those. Basicly...decent LEDs controlled by a steaming pile of overly complex crap! The control module lasted an hour before it was dead for no reason. I wanted to greatly simplify it all anyway, so I didn't care and jsut needed to figure out how to power the LEDs without all the extra garbage.

The LED according to a couple of aliexpress sales pages is a ZES-3575. I googled for about an hour for this part and never found a data sheet for it. I would be directed to the LumiLEDs web site, so I assume they make the LED.



Mine are not wired like this, so who knows what this might do?



The LED has 3 wires going to it. 8v is applied across a black wire and a yellow wire to power the white LEDs. There is a red wire, but that has 2v applied to it to turn on the yellow LEDs. I guess there is a mosfet that gets turned on inside the LED package? Weird implementation. Inside each LED host shell is a small driver board with 4 modes . White, yellow, white/yellow and flash. You can access it by pulsing the +v wire. You can see one of the driver boards held to the LED host with a zip tie and thermal gap filler to help cool it in the background.



There is a lot of extra crap to make it complicated and poorly implemented. Most of that wire mess is now eliminated and nearly totally useless in its factory configuration.



The wired remote is a weird implementation too. One button could have done everything that these 2 buttons do and been much more logical. It has a red LED in it that basically indicates that the lights can be turned on. If the indicator is not lit, the head lights will not turn on.



There is a control box and this board was inside it buried under silicon. These are dual LED headlights with ZERO need for this level of complexity! The volt meter worked intermittently, the power LED flashed rapidly for no seeming apparent reason, there's a tiny momentary switch on the board that seems to do nothing. There's a 1 amp, 5v DC-DC converter for use for charging your phone. I doubt it makes more than .75 amps and either way is pretty useless for modern phones.




On the control board was 2 4407 30v P-channel mosfets. I used hot air to reflow them and then mounted them to SO-8 proto boards for use as a dedicated mosfet switch. The head light will be controlled by a .5a locking switch at the handle bars. It can't handle the amperage of the LEDs so I need the mosfets to do all the current carrying. How convenient that the shitty control board has all the parts on it I needed! 3 wires, 2 SO-8 proto boards, a pull up resistor and it's done.





It needs to be packaged up so it's all water proof and isolated from accidental shorts, but this totally works reliably for 3-4 hours of continuous use so far.



I posted a video on the lights.

https://youtu.be/c3cozbkwdYo
 

·
Registered
Joined
·
236 Posts
Discussion Starter · #14 · (Edited)

·
Registered
Joined
·
236 Posts
Discussion Starter · #16 ·
These scooters don't seem to sell very well as is. My buddy who is a Dualtron dealer has yet to move any others except the one I bought. People try them out after thinking they look pretty cool, but no one buys them it seems. My freind thinks they are simply under powered and too expensive for the level of performance you get from them from the factory. I totally agree and that's why this one is getting upgraded! I hope Dualtron is reading this thread and taking what I say into account for the second version of this scooter.
 

·
Registered
Joined
·
236 Posts
Discussion Starter · #17 · (Edited)
I've been gathering together parts, machining things and so on so far. Now that I have most everything important to really get this thing built, it's time to start taking things apart so that they can be implemented into the new and improved Dualtron City by Doc. That's a nickname I picked up many years ago...

I pulled the controller out. Getting it out of the tiny compartment and untangling the "million wire rats nest" in there could have been done in a MUCH MORE organized fashion! It's a total shit show of connectors going this way and that all mashed together into an incomprehensible rats nest! Clearly no one concerned themselves with servicing these scooters later or needing to remove the controller or DC-DC or other control modules if/when they die!



Getting the controller out meant documenting 20 connectors. They all had to be disconnected since they were entagled all over the top of the controller. It could have been a bit easier if I had noticed that most were all soldered together at +12v and GND and then I could have cut off some heat shrink and pulled just 2 larger conglomerations of wires that come together at a couple of 3.5mm bullets. NOT very good job Dualtron! I guess they have no concern for servicing them later!

Inside this nearly cube shaped box is 2 6 fet controllers. They have a single cable crossing from one to the other for the passage of throttle, brake, eco mode and possibly very basic programming settings. That little rats nest of small connectors has nothing to with controller wiring, but is rather "how" Dualtron decided to intregrate power to LEDs and other things. Lots of things are integrated around the controller wiring instead of being independent.



I thought this was amusing. 30 amps and 60v. The scooter at full charge sits at 66v and most of the controller parts are 80v or so. 30a x 66v = 1980 watts x 2 = 3960 total watts. I think this is where Dualtron gets there wattage rating from for the EV. 30 amp controller...lol...haven't used a controller that LOW powered since I stopped messing with RC stuff!



This is reasonably well done...albeit with very low grade lead free solder that takes TONS of heat to get it to melt. I rarely solder anything at 750C, but that's what it took to get this stuff to transition from metal paste to liquid. Copper reinforcing goes right up to the battery wires. Good current path to the mosfets. For a very inexpensive controller, this is OK work!



NEC D79F9211 MCU. It's a pseudo sinusoidal MCU and gets used quite a lot in low end sinusoidal ESC applications. The blury through holes in front of the MCU is the real programming port or "DATA" port for the MCU. In there will be +5v, gnd, RX, TX and reset.



I was pleased to see that they used real mosfet drivers instead of a discrete low side/high side configuration done with many small parts on the board. The IR2103 is a reasonably good mosfet driver that gets used in far better controllers than this one!



There is nothing wrong with Magnachips mosfets either! I use them extensively for BMS and controller upgrades. This is the MDP10N027 which is good for 100v with a pretty good Rds of 2.8 mOhms and 120 amps in the dye with 416 watts of heat tolerance. 30 amps from this controller could easily be 75 amps with a max of 100 amps. Good little mosfet!



Both controllers mount inside the controller shell with thier heat spreader facing down. The bottom side of the controller shell is fairly flat and had thermal paste between it and the aluminum battery box floor which is 3mm thick. For this wattage, this is adequate.



I was pleased to see that both of the large caps were properly secured to the board so they can't vibrate and move and break off their legs. This 1000uF cap was glued down quite securly and took a fair bit of careful picking and cutting and prying to free it from the board to see the MCU under it. "Good work!" to whoever was in charge of this detail to make sure fatigue was not a problem. To the left of the cap is 2 orange wires. I'm pretty sure they are what applies BATT+ to the board. I haven't looked yet, but I bet BATT+ goes up to the LCD and then a mosfet inside it turns on and applies BATT+ to these wires/ The second orange wire goes to the other controller to "enable" it. Effectively, this is applying BATT+ to the onboard DC-DC and powering up the MCU. On the bottom of the board, they are labeled VCC. The 3 white wires are labeled SC. I don't have any idea what they are doing.



The JST 2.0 connector behind the 3 wires is the interconnect of control signals between the 2 controllers. I haven't checked and probably never will, but based on other implementations like this, there is +5v, gnd, throttle, brake, eco and maybe one other signal. The 3 wire cable in front of it is programming for the various modes? The wires in it are SD, RX and TX. That could easily be CANBUS or some other simple serial comms interface.



It took 8 tries to finally get this picture of the shunt! It's on the negative side of the battery despite that "B+" next to it. You can see the blurry red wire that is really the BATT+ or B+ this silk screening refers to. Anyway...really small shunt! I'm guessing that Dualtron has whoever makes these controllers, swap out the shunt for a larger one for the "higher wattage" controllers. They aren't taxing anything else on here...so why make more wattage complicated?



On the bottom of the battery box and right under the controllers is this "bash plate". I had not looked at it closely until I took it off to get to the controller mounting screws under it. I was astonished to find it was plastic instead of aluminum. I was sure, based on it's location, that it was a heat sink for the controllers and that's not the case. It sits right where THE BEST place to put cooling for the factory controllers should go! Go figure!

 

·
Registered
Joined
·
236 Posts
Discussion Starter · #18 ·
After I messed with the controllers over the weekend, I extracted the LCD and the switch cluster. I'll go over the LCD later and lots of people hate the "hook" throttle and would prefer a more traditional thumb throttle. I'll post about modding the LCD for a thumb throttle later.

The switch cluster is essentially an aluminum block with a small board inside and several low amperage switches. Frankly, none of the switches are adaquate for their intended application except ECO. I'll have to build a high side mosfet board like I did for my added LED head lights for each one.

Upper left is the headlight switch. The small purple wires go to +12v which can be seen in the solder pads to the right of the switch. The single red wire soldered to the switch goes to the factory LED head lights. This really needs a mosfet switch between the head lights and the contacts. I imagine that over time the contacts get scorched and unreliable. K2 is the directional rocker switch. It's fine as is since it sees control signals only. Lower right is the horn switch. Left is ECO. Center is that weird flashy tail lights thing. All of these very low amperage switches turn on the full current load to whatever they power. They are failures waiting to happen! I'm in the process of unsoldering all of them from the board so I can possibly add 2 LEDs that flash for when the indicators are on. Rewiring this thing ought to be interesting!



 

·
Registered
Joined
·
236 Posts
Discussion Starter · #19 ·
I took the pack out of the box to check out the BMS, cells and capabilities to see if it could handle the new current requirements...





This was inside the plastic box and is a snug fit in the space.



Inside the outer wrapping is the actual pack and it fills the outer wrapping except for the space next to the BMS.



The capacity isn't correct. LG INR21700M5OLT cells are 4300mah. In 5P that's 21Ah, not 25Ah.





The pack minus all the space wasting shell fills the space poorly. Even if I kept the exact same cells and cell count, I could probably reduce its size by 10-15% from how it is built now. There is 2" of width and 5" of length that is unused.

 

·
Registered
Joined
·
236 Posts
Discussion Starter · #20 · (Edited)
Both hub motors are totally reworked. They are both vented, have halls and temp sensors now. They still need connectors added, but that will happen after I decide how I want to mount the controllers and how long I need the motor wires to be.






I'm now working on the dual Flipsky 75100 controllers to fix their various weak spots.

Last night I sealed up the closed ends to the shell. I'll never need to take these off so I used thermal glue to make them semi-permanent. The silicon grommet for the internal LED got sealed down too.






The heat spreaders are smallish. They are nicely machined, but under sized for the application. I'll machine new ones that fill as much of the side wall of the shell as I can get to fit inside. This will give me a best case heat transfer out of the shell and to better heat sinking since the shell is pretty dismal in this regard. They had a single layer of kapton tape on them and that's likely to have failures. They should have used kapton insulators! There was no thermal paste applied to the mosfets so OF COURSE heat transfer was lousy!



Last night I took apart this charger to adjust it to 82v. While I was in there, I went to both voltage extremes...50v to 133v...impressive!

 
1 - 20 of 28 Posts
Top