E-Street Machine 20kW

Joined
Nov 16, 2016
Messages
60
Hello!
I am from the great state of Texas, and am building a high power E-Bike. I am going to detail the build in this post hopefully with lots of pictures.
Of course as true with everything else here, Everything is Bigger in Texas, this E-Bike is not an exception.

Coming in with an estimated weight of ~75kg with batteries, the specs for this bike are:
(570) Panasonic NCR18650PF Recieved
(1) [strike]Kelly KBL72501E, 72v 500A[/strike] Kelly KEB72121E 72v 550A Recieved
(1) Enduro Ebike Frame (Cheap Chinese one with 140mm width) Recieved
(1) QS 273 v3 Hub Motor Recieved
(1) [strike]DNM USD-8 Forks[/strike] Going to a Marzocchi 380 CR Recieved
(1) [strike]RockShox Kage RC Rear Shock[/strike]Ended up going with the DNM Burner Recieved
(1) Domino Throttle Recieved
(1) Cheap Thumb Throttle for Regen Recieved
(2) 22" OD Rims and Tires? Recieved
(1) 73Kg Me
In addition to this, I am implementing an arduino based cycle analyst replacement. The CA-V3 seems great but its pricey and I think I could do it better. I am going to be using adafruit neopixels to display power meter in much the same way you see a tesla do so, with a current meter and a orange color during regen. I will also use this with the temperature probes to monitor and display battery, motor, and controller temperature. I couldn't figure out a way to get around floating point calculations in this code (I'm not very good at it) so am using a separate arduino for[strike]throttle filtering[/strike] and mapping for separate power levels. Using a digital throttle inhibitor is far to much work for an amateur like myself, and would require a low-pass filter that either introduces inaccuracy or latency, neither are desirable for a throttle response. I will instead use an analog voltage divider circuit for different power levels.

The battery is going to be 560 Panasonic PFs in 20s 28p, to get a peak 20kw. It will be a tight fit, but In a render I was able to get 546 cells in there, and I didn't account for some of the curves on the frame, if I can't fit them all no biggie. Everything I read says these cells can do 10a continuous, but I don't think that would do well so I won't peak over that. I am going to weld 2 layers of 8mm*.1mm nickel strip onto these cells with an arduino built spot welder, as I can't seem to buy a JP Spot Welder anymore. According to posts here that is acceptable for ~10amps, right where I will peak per cell. I am going to use 2 AWG Wire for all the connections between battery/controller/motor up till the closest point on the swingarm.

The QS 273 V3 will require widened and reinforced dropouts which I am going to machine at a local shop. Fun physics tells me that an Approx 350nm (258ft/lbs) of torque with a 22 inch wheel, is theoretically 309lbs of force, or 1375N. Me +the bike will be about 150 kg. 1375/150 ~= 9m/s^2 of acceleration, provided I don't flip over. I am only half way through Physics so if this is wrong feel free to correct me.

[strike]As for charging this pack, I realized that even if I only used 4.5kwh per round trip it would still take me 15 hours with a quality 300w charger such as the ones from lunacycle. I have decided to build a meanwell knockoff charger with (7) 12v 30a chargers in series. This should be able to charge the (theoretical) pack in less than 3 hours, or 9 hours at 10a which is a more realistic rating for the cheapo chargers. I only have one on order right now to see if I can figure out how to current limit it as well as disconnected the negative terminal from ground to avoid having them short.[/strike] Going with the BMS battery S2500 2100w charger. Even with the $85 shipping (total rip off) it was about the same price as a meanwell abomination + the work required for it. I ordered it with the 240v option.

I will be waiting for batteries for a month at best, so I am going to find a auto battery for me to test my spot welder with.

Once I resize them I will upload them here, but full res images are in this imgur album: http://imgur.com/gallery/myski
 
I am looking forward to seeing how many of those 18650 cells you can actually cram into the Enduro frame. I think that's about twice as many as I have seen previously. :shock:
 
Darn, when I saw "Street Machine" I thought I'd be looking at this...

smgte_links_mitzieher_reise.jpg


... but it seems you're going in a different direction. Looks like fun anyway!
 
teslanv said:
I am looking forward to seeing how many of those 18650 cells you can actually cram into the Enduro frame. I think that's about twice as many as I have seen previously. :shock:

Me too, I took significant inspiration from this post:https://endless-sphere.com/forums/viewtopic.php?f=6&t=79612
Where he managed 480 cells and a charger. I don't know if I can fit 560- as the side covers taper near the top, something I did not account for, but I will try.

The 400a contactor is a tight fit into the upper part of the frame but it is about perfect.

The Kelly KEB72121E is a bit wider than the bottom of the frame which is actually less than 140mm, the width of the battery compartment. I will just buy a aluminum sheet to mount it to, because I can't bring myself to take the kelly apart, it looks like it it was waterproofed (where the aluminum cast fits to the plastic end pieces has some sort of hot glue or silicone sealant) and I didn't pay for the waterproof option, so I assume that comes standard on the KEB models?

Anyways the next order of business is finding a battery to test my welder with and then building a small vacuum former to create the plastic bits to cover the LED display.
 
I am very confident I will be able to fit all the batteries. Placing them in the side cover, (which is how much usable space I have) showed me this. Dont mind the ones that are negative terminal up.
Another thing to note about the Panesonic PFs that I got, is that the white ring you can see is actually a paper insulator to separate the can from the positive terminal. I didn't know if this would be included so I bought stick on paper insulators, and will still be applying them.
View attachment 1
Here is the QS 273 with a AA battery for scale... and the phase wires are even smaller than I expected, so at some point, that is going to have to change.
617094901198534758-account_id=1.jpg
Received everything except the front fork which is going to be a Marzocchi 380 CR and the actual bike bits like the front hub, spokes, cranks and the like.

A major thing I did not anticipate was the odd shaped pack and very high number of cells in parallel (28p). This has made designing the pack with current distribution in mind less ideal, as it was not possible to have all the cells adjacent to the next one in series, as seen in square packs.
batteryv2.jpg
I will just weld on more nickel on areas that have dis proportionally high current loads, adding more layers for the respective amount of cells.
If anyone can think of a better way to solve this problem, criticism is always welcome.
 
ninepointeight said:
Hello!
I am from the great state of Texas, and am building a high power E-Bike. ...

The QS 273 V3 will require widened and reinforced dropouts which I am going to machine at a local shop. Fun physics tells me that an Approx 350nm (258ft/lbs) of torque with a 22 inch wheel, is theoretically 309lbs of force, or 1375N. Me +the bike will be about 150 kg. 1375/150 ~= 9m/s^2 of acceleration, provided I don't flip over. I am only half way through Physics so if this is wrong feel free to correct me.


Looks like fun. What part of Texas?

Your math looks like it's in the ballpark, if you're sure about that 350nm from the motor at 20kw. What might be more important is to calculate how much torque you'll have at the axle shaft. Because your dropouts will have to survive that much force. The edge of the flat spot on the axle where it applies force to the frame is something like 6-7 mm from center.call it 0.25 inches. So if you're at 258foot pounds, that's a ball park of what, ~12,000 pounds of force trying to spread your dropouts? Whats your plan for the dropouts?
 
You've got the parts there to make a very useful range and performance vehicle! Excited to see it come together!
 
Now we are talking! Can't wait to see how this develops. It has all the makings and the right components.
 
Drunkskunk said:
ninepointeight said:
Hello!
I am from the great state of Texas, and am building a high power E-Bike. ...

The QS 273 V3 will require widened and reinforced dropouts which I am going to machine at a local shop. Fun physics tells me that an Approx 350nm (258ft/lbs) of torque with a 22 inch wheel, is theoretically 309lbs of force, or 1375N. Me +the bike will be about 150 kg. 1375/150 ~= 9m/s^2 of acceleration, provided I don't flip over. I am only half way through Physics so if this is wrong feel free to correct me.


Looks like fun. What part of Texas?

Your math looks like it's in the ballpark, if you're sure about that 350nm from the motor at 20kw. What might be more important is to calculate how much torque you'll have at the axle shaft. Because your dropouts will have to survive that much force. The edge of the flat spot on the axle where it applies force to the frame is something like 6-7 mm from center.call it 0.25 inches. So if you're at 258foot pounds, that's a ball park of what, ~12,000 pounds of force trying to spread your dropouts? Whats your plan for the dropouts?
I live just north-west of Dallas, in the Flower Mound area.

No plan for the dropouts atm, although the enduro ebike frame I am using should make it very easy to bolt on a nice thick piece of steel on the dropouts. Not sure if that is needed though, the dropouts seem quite thick on the frame.
I will probably only end up running 10kw until I can upgrade the axle and phase leads anyways.
Thankfully it is steel so hopefully I will notice any bending and widening before it becomes a real problem.

I am about to start assembling the pack, and after a test I found that Permatex 80855 RTV Silicone was almost as strong as hot glue when bonding 2 cells together. Additionally, it allowed significantly more play and movement between the cells. Not sure if this would be a pro or con to silicone, as it allowing chaffing between cells, but I am guessing the silicone will hold up to years of use and temperature changes better.
The Permatex RTV Silicone seems to release acetic acid fumes when curing, which I am worried could compromise the plastic shrink wrap on the cells and or the cell casing over time.

I am curious if anyone here has built a pack glued together with RTV silicone, and how it worked for them.
 
Looking forward to your build as well. That is some serious battery my friend.

I thought the QS 273 had 16 mm squared phase wires from spec sheets.....that is the equivalent to 5 gauge wire. That is freakishly large! What did yours ship with?

That is why I picked QS 205 for my motor....big phase wires you DO NOT have to change out compared to the competitions pee wee wires!

Looks great.

Tom
 
Update:
After much bumblefuckery and ruined mosfets, I have built my spot welder on the PCB courtesy of Kaeptnbalu. I did add the diode attachment which resulted in some fun(not).
IMG_20170308_184700969small.jpg
The diodes allow current to flow through the battery when otherwise not possible. So when my dumbass drops the leads onto the partially welded pack across series groups, the maximum current of 28 panesonic PFs over 1mm nickel flows through the diodes. I have ruined 2 sets this way. This also ruins the cell that it shorts on by promptly melting through the casing.
Swapping cells is not nearly as bad as anyone makes it to be. I glued this pack with RTV so it would be tough stuff to remove barring acetone and strong acids. (no thankyou). I simply cut the shrink wrap on the negative terminal on the cell and pushed it out. They are in there pretty tight so when I put a new one in there is no play at all.
View attachment 6
Also when welding, sometimes i left the current too high when welding 1mm strips directly to the cell casing and melted right through. This released a interesting smelling liquid. I replaced these cells as well.IMG_20170317_033137251small.jpg
I finally laced the motor into the 19x1.4 inch rim I got from treatland.tv which was a huge pain. Quite a bit of rim drilling and forcing the 12g spokes to bend before I could straighten them out with tension. I am as inexperienced as they come in terms of wheel builders, so I gave it my all, and it feels good and looks good, but I have not trued it yet (no spoke wrench).
IMG_20170401_141507243small.jpg

To connect the terminals to the battery, I am using 4 18 AWG wires per cell that will connect to a 2AWG bus wire on top and bottom of the battery for - and + terminals respectively. Yes that was more wires than I anticipated and should have used a higher AWG because they overlap, making my idea of saving the width useless. Oh well.
IMG_20170328_160214897small.jpg

In case anyone was wondering, this quantity of cells certainly fit inside this frame.
IMG_20170318_142124562small.jpg

Also, the BMSBattery S2500 I ordered in the 220v model for the full 2000w of charging came with a standard NEMA 5-15 plug. Horribly dangerous but blah blah something about life worth living blah blah.
So I put together this deathtrap and wired it directly on to the 2 phases of my circuit breaker. I wouldn't be surprised if this burns down my house before any of the super safe Panesonic PF's have a chance to. This probably belongs on r/osha.
IMG_20170124_162027645small.jpgIMG_20170204_141612740small.jpg
The worst part about all that is that it's crooked.


The next steps involve finishing the battery terminals, lacing the front wheel, putting together the headset/handlebars, mounting the tires, and then wiring the whole thing including the arduino that will track amps temperatures, and speed and log amphours and miles, displaying this on neopixel displays that will be waterproofed (somehow, ideas?) and a 16x2 LCD.
I also plan to have a small switchbox on the handlebars with some waterproof toggle switches to turn on and off the headlights, turn off ALL lights including displays (hehe :twisted: ), and activate a voltage divider for the throttle, which levels will be selected from the turn signal switch from one of those cheap motorcycle horn/headlight/turn signal combo things. If anyone is interested I can post the code for this somewhere.

Litespeed you are correct, the phase leads do appear to be 5-6 AWG, with thin insulation, which makes sense.
 
Holy shit. Lots of batteries. Could you upload bigger pics to a site like imgur.com so we can look.
 
flat tire said:
Holy shit. Lots of batteries. Could you upload bigger pics to a site like imgur.com so we can look.

Gladly! Here are the full res images of the photos in the post plus a few more of battery construction. Many are not taken well and were only ever for personal documentation, sorry about that.
http://imgur.com/a/yCBSf
 
The same project! Party a long time ago, but it has been slow due to lack of time ... :D

https://goo.gl/photos/5zt9D2fFkwdES4KN8
 
CATAWEB said:
The same project! Party a long time ago, but it has been slow due to lack of time ... :D

https://goo.gl/photos/5zt9D2fFkwdES4KN8
Good luck my friend, great minds think alike :D . Unless you have far more knowledge/experience/money than myself, you are in for hours of hurt, physically and mentally, depending on the specs of that bike. I assume those FedEx boxes have cells? What arrangement are you building the pack in? What controller is that? I assume it's the 120v or 144v KBL? What KV motor did you get? I have the 11.46kv version.
 
Sorry for my bad English ... :cry:
In practice this is the tenth e-bikes I build and I wanted to do something powerful, perhaps exaggerating a little. :lol:
1) Motor QS 273 4000W 4T
2) Controller KHB12601 Kelly 600A (I will plan to 400A)
3) Battery Pack 96V 24S (XXP?) (I do not need a high speed, but the point in having a great couple to take off road.
4) BMS Bleautooth with external Display :idea:
5) 5) I predict two wheels 21" inches; A) To avoid the hard work you have done you to let the huge engine is in a small circle. :mrgreen:
Before buying the motor thought to keep him in a circle 19" , but when it arrived the engine I quickly realized that a circle 21" goes much better. B) exploited the tires of moto enduro I remain. :idea:

It will be a hard and demanding job and unfortunately as I said I have very little time available during this period.
 
CATAWEB said:
Sorry for my bad English ... :cry:
In practice this is the tenth e-bikes I build and I wanted to do something powerful, perhaps exaggerating a little. :lol:
1) Motor QS 273 4000W 4T
2) Controller KHB12601 Kelly 600A (I will plan to 400A)
3) Battery Pack 96V 24S (XXP?) (I do not need a high speed, but the point in having a great couple to take off road.
4) BMS Bleautooth with external Display :idea:
5) 5) I predict two wheels 21" inches; A) To avoid the hard work you have done you to let the huge engine is in a small circle. :mrgreen:
Before buying the motor thought to keep him in a circle 19" , but when it arrived the engine I quickly realized that a circle 21" goes much better. B) exploited the tires of moto enduro I remain. :idea:

It will be a hard and demanding job and unfortunately as I said I have very little time available during this period.
10th ebike, first post, where have you been???
Excited to see this external display you have in mind.
I have one question, how did you modify the dropouts to accept the larger 12mm flat axle? I used a file and manually worked down the dropouts on mine, but it doesn't seat fully into the rounded section anymore, so it is less ideal.
 
In the past I only built e-bike motor Mid 750W / 1000W like this: https://goo.gl/photos/yqLeQaW5Ci2MyARW6 :lol:

This is the BMS and costs a lot of money:
https://goo.gl/photos/TBPm7mKXhDT6ixa36

For the drive axle housing I worked the metal taking it from 10mm to 12mm. To me it seems to be going well, but then again everything is still to test.
will put even # 2 reinforcements to the anti rotation because the engine seems very powerful and could tear the metal fork drive axle.
 
I need some help here.
I have finished the rear wheel and mounted the Shinko 241 19x2.75 onto my moto 19x1.4 rim, and after inflation, I have a question.
Some sections of the tire has this rubber bit sitting outside the rim.
IMG_20170404_205739875small.jpg
While other parts of the tire this sits inside the rim.
IMG_20170404_205750466small.jpg
The tire feels very secure at 25 PSI and shows no sign of being able to be removed, but which way is correct, and how do I get it into the correct state?
Full res images here http://imgur.com/a/HVr3z
Thanks for the help.
 
when installing tires, i personally inflate maximum pressure (for making the tires sitting correct), and than i deflate to the pressure i like to have.
it also may help if you spray soapy water on the rim and rubber (for lower friction).

what rim do you have?
 
It is called "no bead breaker" when the tire does not come out from the rim. Try to inflate much the tube to 3 bar or higher, you'll see that will come out.
In addition you should properly inflate the bladder because with that powerful engine you are mounting QS273 risks tearing the inner tube from the tire valve bar.
In fact I'm adopting a different method; I will make a further hole on the rim for a instalalre stops tire as motocross bikes / enduro, in doing so I'm sure that the tire from spinning freely. I hope I made it clear ...
 
Hi Ninepoint...

I ask how will you insert 2 adjacent cells files? Our frames are very similar,
I have a Vector, you have the Chinese and my difficulties I get to make two parcels of adjacent cells. See image

https://goo.gl/photos/8NvpBsdKyy14n9Qa9
 
From what I understand, the Electric Enduro Ebike frame comes with plastic molded side panels, as opposed to the vector frame with aluminum(?) side panels. The plastic molded side panels actually extend out past the frame, bringing the width to 140mm.
 
Back
Top