Evolve Carbon AT with 50km+ range

whitepony

10 kW
Joined
Feb 19, 2015
Messages
663
Hello everyone, I'm a frequent sphere reader and I needed some sort of webspace to explain my evolve brothers what they have to do in order to double their range with zero-invasive actions and quite little cost - thought this might actually be a good spot to drop this summary of my actions! :)

Some facts:
* Evolve Carbon has a 37V 10Ah LiPo-battery, cells with nominal voltage of 3.7, cutting off around 4.2V max and 3Vish min
* Balancer of the BMS has 11 pins due to 10S?P battery layout, the pin distance is 2mm
* evolve quick charger is a constant current charger @ 42V, stopping once the current becomes too low
* evolve silent charger is a cc/cv charger, thats ultimately loading the battery @ constant 41.4V into saturation
* the available space for a battery is around 62 x 14.2 x 1.5cm

Minmum battery infos (stuff I didnt know and had to find out):
* LiPo based batteries have usually a nominal voltage of 3.6-3.7V which is sort of an average, they max out at 4.2V and shouldnt be discharged lower than 2.5V. LiFe-based batteries run at lower voltage per cell (3.3ish) and they are NOT compatible with the stock BMS. Look out ONLY for cells that use 3.6-3.7V nominal!
* for the panasonic and other tests, check http://lygte-info.dk/review/batteries2012/Panasonic NCR18650B 3400mAh (Green) UK.html
* if you build a 10S6P setup like me, it means you use 10 cells in series and 6 parallel - basically you weld together 6 cells in parallel for a capacity of 6x 3400mAh = 20.4Ah and combined 3.7V nominal. Then you connect the 20.4Ah 3.7V packs in series which adds up the nominal voltage of 10 x 3.7V = 37V.
* the nominal voltage is not the max voltage - a cell can usually be charged up to 4.2V and discharged down to 2.5V (in case of the panasonics I used), so the voltage of the total battery will be 42V max and 25V min. The evolve BMS is cutting off way before 25V though to give the cells a longer lifetime.
* a BMS is needed to keep all parallel 3.7V packs at the same voltage. It would be really bad if you charged the total 10S pack to 42V and some packs went up to 4.2V+ while others remain at voltages < 4.2V (likewise with draining them below 2.5V). Thats why you have to split the 10S block into 10 sections of 3.7V nominal voltage, so that an electronic can level these sections out in order to prevent discharging below min or overcharging above max voltage.

Material list:
* 62 cells Panasonic NCR18650b with nominal voltage of 3.6 and 3400mah capacity (2 spare cells incase i got a faulty one or I screw something up) (paid around 300€ @ http://eu.nkon.nl/rechargeable/18650-size/panasonic-ncr18650b.html )
* 3m hilumin or nickel band (my hilumin band was 99.5% nickel) 10mm x 0.15mm for spot welding cells (10€)
* multicoloured 0.5mm^2 cables for balancer connection (5€)
* 2m black 6mm^2 cable for +- poles (2€)
* balancer port http://www.conrad.de/ce/de/product/...ennstrom-2-A-PHR-11-JST-1-St?ref=searchDetail with 11x http://www.conrad.de/ce/de/product/...astermass-20-mm-BPH-002T-P05S-JST-1-St?ref=nz I linked these things cause it took me forever to find the compatible ones for a plug & play solution! :p
* about 2m 10x3mm sealing tape (I used neoprene stuff)
* heat shrink tube 1m 195mm width

*Edit: after building my 2nd evolve battery for someone else, I agree with riba: the Sanyo GA, LG MJ1 and Samsung 35E are much better cells overall.

Additionally I bought everything I needed to make my own spot welder. Thought a while about how to connect the cells - was reluctant to solder, cause I didnt have a high power soldering iron and I was also afraid to overheat the cells in the process. Finally I also dont really trust soldering as much as welding, so I followed many nice guides to build my own spot welder. Ill give a quick recap on material list and how to build one in the next post! :!:

Dimension of the evolve carbon inside, its also where I measured balancer pins and voltages to check whether the evolve setup was compatible with regular LiPo batteries:
battery.png


I quickly drew the battery setup with keynote, using 18650 (18mm width, 650mm length) cells in a 10S6P setup. The "B??" tags in the picture are where the balancer cables are placed - basically a 10S battery will have 11 balancer wires in total, dividing up the battery into 10 3.7V sections!
Then I tried to figure out which cell would be best for the job and after some interesting discussions in this thread I decided to give a heads up on that process to make people aware of a few things: the factory capacity of cells is only met under very low discharge currents, if you suck stronger currents from a sell, the discharge capacity might get smaller and smaller.
I started with thinking about how much power the Evolve actually sucks from the battery under demanding conditions - the motor is set to use up to 350Watts and from my experience with the original 10Ah evolve battery, that is actually a very reasonable value for the power consumption in demanding scenarios (25km/h, lots of carving, hills).
So, 350Watts at 36V nominal battery voltage means about 10A constant discharge current. This divides up on each of the 6 individual parallel cells and comes down to 1.5A constant discharge current per cell. Now the only reasonable source I had for battery comparisons has been http://lygte-info.dk/review/batteries2012/Common18650Summary UK.html where I checked the charts. The correct picture to look at is the discharge capacity down to 3.0V (stock BMS cuts off there) with discharge current of up to 2A -> http://lygte-info.dk/pic/Batteries2012/common/CapacityTo3.0.png (cyan colored bars). There are many good cells as you can see, some less and some more expensive - the Panasonic 18650B with 3400mAh seemed to be a very good choice in terms of very high capacity for moderate money.

When I had enough courage and was convinced that the stock battery management would work just fine with my 18650 cells, I went for it and ordered everything that was needed.

Cells arrived:
cells.jpg


For spot welding, I used clamps and old pieces of wood, to keep them lined up well - in total 10 blocks a 6 cells (only use mild pressure with clamps, you dont want to kill your cells). I measured the voltage of every cell before connecting - in order to identify something thats broken and to avoid bleeding currents among cells of a pack:
setupcells.jpg


After spot welding with the hilumin band, it looked like this:
spotweldcells.jpg


I connected pack 2&3, 4&5, 6&7, 8&9 with a second layer of welded hiluminband, then added the balancer cables with seperate small Hiluminbands, so I could solder the cables before I connect them to the cells (didnt want soldering heat transfer into cell, make sure to use soldering flux!). At that point it helps to stabilize the whole pack with wide clear tape (the super cheap one)... it will take a lot of stress from hilumin connections between packs and will make the whole pack much easier to handle. From here onwards you have to be very careful to not shortcut anything on accident. LiPo cells are very strong and can vaporize connecting metal in seconds. I started to use a lot of cheap wide tape to cover anything I didnt want to touch ... better safe than sorry!
balancercells.jpg


In that picture you actually see the middle section of the battery - my plan was to do the middle connection with hiluminband and then fold the 2 long rows of batteries like a butterfly. Used 4 interconnections for physical stability! Additionally, I added 2 more layers of hiluminband to minus pole of pack 1 and plus pole of pack 10, to account for the larger currents there. The 10x0.15 band only works up to 10A!
butterflycells.jpg


The butterfly idea worked well - just folded the 2 rows:
finalcells.jpg


For shock protection at the side facing to the ground, I used a worn out mouse pad :p
mousepad.jpg


Connecting the 11 balancer cables to the connector, make sure to check every single one of them thoroughly with measuring the voltage (and keep the rest taped to avoid any shortcuts). it should always be 3.X V depending on factory charge. For the correct order, compare the pins with measuring your old battery balancer pins. If you remove the connector and look at it, cables facing away, it started with B1 on the left, then B2, B3, B4, ... you get the idea, could be different for your board though. Finally added some heat shrinking tube, had a very cold moist towel which I used to take out the heat the moment the stuff shrinked into place. sadly I screwed up that part, the heat shrink tube folded my battery into a v-shape, so I had to cut it off again and use the rest of the heat shrink tube to try it again - it worked better, but I didnt have enough anymore to do the full battery, so I used parts of the old tube and fixed it with tape. not proud of this :(
heatshrinkfail.jpg


I knew my battery pack would be around 3mm too high for the original case, especially with my mouse pad protection - so I had 10x3mm neoprene sealing tape prepared to lift up the cover a little bit. the holes for the screws were cut by some magic leather stamp thing I lent from my mum: :lol:
neoprenetape.jpg


Putting everything together - beautiful evolve carbon <3
evolvecarbon1.jpg

evolvecarbon2.jpg


Sadly I screwed up badly in that final step of a 3 day odyssey, canted the BMS slightly when fixing the carbon cover and tore off a mosfet from mechanical stress - it killed the recharging mechanism of the BMS. stupid me :|
I took the BMS out completely for now and put the board back together. cell drift wont be a problem right away - Ill buy a replacement BMS or find a suitable mosfet and put it in in a few weeks from now. if anyone knows what mosfet is meant from whats written on it (t470 GN4G4S), please poke me!
bmsfail.jpg



Did a quick test run last night with a semi-charged battery - it worked! Dont think I'll have to do a range check - its double capacity = double range. With how I ride my carbon AT, this will be around 30km. A maximum of 50-60km should be possible in the AT setup if you do the 18km/h non carving max range thing - I might give it a try, but its a little too boring for my taste. :wink:

Hope you enjoyed it, it was a lot of work - together with building my spot welder, it took me about 3 evenings. The battery making until the evolve was up and running again took me about 6-7hours total. Then again Im new to this, first battery I welded, had to think about things long and hard before I dared to do them. Overall a *thumbsup*-experience. Ill report back in a few weeks how the battery is doing and what my average range is.


p.s.: I keep updating and adding stuff, also taking in new input when things could have been done better ... incase you are wondering about the edits!
 
its me again! as a "p.s." for the double evolve range article, I wanted to describe quickly how to build your own spot welder for ridiculously low money for pretty much professional DIY battery constructions.

most of the things were actually common knowledge from various pedelec forums - best thread ever here http://www.pedelecforum.de/forum/in...u-punktschweißgerät-für-akku-verbinder.11421/ (but its sadly in a german forum). I think I built one of the most simple version though! :)

Material list:
* 2F power cap (40€)
* 2x 8mm thick and 10cm long copper rods (2€)
* 20x5mm copper bar, length 30ish cm (3€)
* very thick cables, I bought 25mm^2 which was probably overkill (10€)
* 6 ring connectors (2€)
* old battery charger, charge voltage 17V
* some screws and a wood plank

Basically you use the power caps stored energy 0.5*C*U^2 (C = capacity, U = voltage), to weld thin materials together within microseconds. So all you gotta do is:
* loading up your powercap
* putting the electrodes where you want to spot weld
* connect powercap with electrodes with a switch who can take 1000-2000 ampere and unload it in fractions of a second
* start over

Most use fairly complicated and expensive electronic things to switch - thyristors, and other stuff. I went for the archaic 20x5mm copper switch version (see pictures). made the electrodes from the 8mm copper rods, drilled a hole in their back, cut in a thread and connected the ring connectors of the huge cables with screws at the electrodes.

A tip for soldering 25mm^2 cables into ring connectors: https://www.youtube.com/watch?v=zssIlnwH8dY

here you can see the full setup sort of - powercap, copper electrodes, copper switch:
spotwelder1.jpg


and here the copper switch closeup - you can tune the gap on the left side with adding washers of different thickness on the right - if you make the gap on the left side too small, the force pulling the copper back will be too weak and it will weld together eventually.
spotwelder2.jpg



hope this helps - its very simple, very cheap and works like a charm (this is me in action on my very first battery pack - at that point i killed one cell cause my copper switch welded together - thats when i learnt that the gap needed to be large enough - it happened never again after that, but I was cautious anyway and put together the electrodes every time for a quick check if the switch actually disconnected the powercap): https://www.youtube.com/watch?v=cfbOmuBMLZc

have fun :D
 
Nice work! Just a few notes:

-Panasonic B cells are too weak, you will kill them soon. Better choice with same capacity are Sanyo/panasonic GA, LG MJ1, and Samsung 35E.
-hillium is shit, you should have bought pure nickel strips
-using that kind of switching for welder with uncontrolled timing is bad, you can get very inconsistent results, too weak or too strong. Also, caps are too expensive, it's better to use car starter battery.
-you can just replace that one mosfet, no need for new bms.

Otherwise, nice job :)
 
riba2233 said:
-Panasonic B cells are too weak, you will kill them soon. Better choice with same capacity are Sanyo/panasonic GA, LG MJ1, and Samsung 35E.
-hillium is shit, you should have bought pure nickel strips
-using that kind of switching for welder with uncontrolled timing is bad, you can get very inconsistent results, too weak or too strong. Also, caps are too expensive, it's better to use car starter battery.
-you can just replace that one mosfet, no need for new bms.

Otherwise, nice job :)

hey there, what makes you think the cells are too weak? the evolve carbon motor runs at 350watts and I rarely accelerate/brake, just idling at 25ish km/h and carving is about 90% of what I do. so lets assume that consumption of the motor, then we have a roughly 10A stream from the battery. divided by 6cells its less than 2A/cell. dont think this will be a problem? sure its peak will be more, but on average the stream is very constant and moderate. I also followed http://lygte-info.dk/review/batteries2012/Panasonic NCR18650B 3400mAh (Green) UK.html, which seems to indicate that its a very good cell.

with the switching: yea, well :p the 5€ copper solution isnt as good as the 40€+ parallel mosfet/giant thyristor solution, but I had very consistent looking spots like you can see on the picture!

about hilumin vs. nickel: I thought they were the same! My hiluminband says 99.5% pure nickel. :?:

do you know what kind of mosfet that is? I didnt find anything when using T470 GN4G4S in google. also I didnt manage to get it off the circuit base. any tips on that? :shock: I dont really know all that much about electronic components ...
 
riba2233 said:
-using that kind of switching for welder with uncontrolled timing is bad, you can get very inconsistent results, too weak or too strong. Also, caps are too expensive, it's better to use car starter battery.

one thing about the switch and powercaps vs. battery (I followed your signature and now I understand where this is coming from):

with a powercap (the 2F version is 40€ by the way, find me a car battery for that price), the idea is that not the pulse length of an infinite reservoir (battery, some coils) is coupled to the spot you want to weld, but instead the stored energy in the powercap is controlled. whatever I do with the crappiest copper switch in the world - in the end the energy 0.5*C*U^2 will flow through the spots i want to weld, and it will do that in a most controlled fashion (discharge curve of the powercap, usually a few milliseconds). if i want to change the amount of energy i send through a spot, i can either vary the powercap charging voltage or the capacity of the powercap!

in other words: spot welding with a powercap is in fact a very consistent thing, because you always send the same amount of energy 0.5*C*U^2 through the electrodes. :)
 
Yeah, you are right about the batteries, I didn't realize that the board is only 350 W, in that case they are fine :) But Samsung 35E would probably be a bit cheaper, and it's also a bit higher capacity :)


Now, I don't have a great opinion about those cheap supercaps. They are essentially fake, they do have rated capacity, but they are not true supercaps (to high ESR which limits the current). Real one cost a lot more. I did actually find new car starter battery for 40€, and not a shitty one, schneider, made in Germany :) And unlike cheap fake supercaps, it can provide enough current which allows for short weld times. You are dumping current into weld for too long. If you had timing circuit, and you adjusted it for let's say 8ms pulse which is what I use with my 40€ car battery, you probably couldn't weld.
I'm not saying it doesn't work, but it simply isn't proper way. Unless you were really lucky and found true 2F supercap for such low price. But it that was the case, your non timed pulses would be too strong, and you would be blowing your welds :) I also tried working with cheap supercap, cheapest I could find was 1F, it was around 45€, and I couldn't weld shit with it (with timed pulse of course), so I dropped that idea. I still have that cap somewhere around, and it's pretty much useless.
Also, most people already has car battery (in their car), or maybe they can borrow it etc. Supercaps are not that common so you have to buy one.


For the mosfet, it's probably N channel, so any N channel mosfet with rdson below 10mohm and Vds voltage over 75V would be fine. P75N75 would be fine and it's cheap. But wait for someone else to confirm that, I'm not sure with this particular BMS.
 
hey man, overall I agree with you, also about the cheap "fake" powercaps. its never true 2F, but in the end it actually works for sub 100€ total cost (electrodes, switch, powercap, superthick cable)! i tried to tore off the welded band for testing - basically got it off, but it had holes where the weld points were. if thats not sufficient, then I dont know what is! :mrgreen:
 
Well done sir! I did a much simpler hack on my Boosted board where the extra batteries are simply stuck on the bottom (the Boosted doesn't have a cavity like the evolve). They are attached beyond the bms and so have bms's of their own. I thought about building my own pack and hacking more cells into the existing bms but I thought it would be too difficult - you're an inspiration dude.
 
riba2233 said:
For the mosfet, it's probably N channel, so any N channel mosfet with rdson below 10mohm and Vds voltage over 75V would be fine. P75N75 would be fine and it's cheap. But wait for someone else to confirm that, I'm not sure with this particular BMS.

you were pretty spot on with the mosfet! finally found the manufacturer & type:

http://www.aosmd.com/products/mosfets/n-channel/AOT470
http://www.aosmd.com/res/markings/AOT470.pdf
 
Do you happen to have any more pics of the BMS?

I am inspired by the work you have done and I am hoping to do the same the next month or so.

I am trying to find a BMS that is capable of being on board and handle the pack. By any chance was the one on the Evolve an off the shelf unit?
 
not of the original battery no, but of the BMS. if you really want i can upload them, but the stock BMS isnt anything special really. basically you can use ANY 10S LiPo BMS for this, there are plenty around, most of them between 40 and 50€. i wanted to save that money, so I just went looking for a suiting female connector to the stock one (about 1€) and it works well. repaired my stock BMS by now too with a replacement mosfet! :D
 
eboostin said:
By any chance was the one on the Evolve an off the shelf unit?

and yea, it looks like all the others you can buy. uploaded 3 pictures, hope it helps.

bms1.jpg

bms2.jpg

bms3.jpg
 
whitepony said:
not of the original battery no, but of the BMS. if you really want i can upload them, but the stock BMS isnt anything special really. basically you can use ANY 10S LiPo BMS for this, there are plenty around, most of them between 40 and 50€. i wanted to save that money, so I just went looking for a suiting female connector to the stock one (about 1€) and it works well. repaired my stock BMS by now too with a replacement mosfet! :D

I actually meant, do you have any more pictures of the battery back you made (side angles, etc)? It looks really good, and then butterfly idea you had was fantastic! Huge inspiration and I will be following suit.

I am trying to find a suitable 10S BMS but I have got a bit hung up as I was reading that there are some that do not allow regen braking. Also, I am trying to find one that cuts off at the correct max and min voltage.
 
riba2233 said:
Also, one shunt resistor looks shorted, maybe you could clean that up :)

wow, youre right - maybe thats the reason why I got a fairly short range on my original battery. its untouched by me, the ugly soldering work is basically how I got it. :shock:

p.s.: actually when I first got my evolve, I just had like 5km battery range before cutouts started, it got repaired by adding 1 more of these resistors in parallel configuration (this piggy-backed resistor you noticed).
 
eboostin said:
whitepony said:
I actually meant, do you have any more pictures of the battery back you made (side angles, etc)? It looks really good, and then butterfly idea you had was fantastic! Huge inspiration and I will be following suit.

hm, I dont really have more pictures and the battery is encased in new thermal shrink tubes now after I screwed it up the first time. :( there isnt really anything else to see though anyway, the most important part is to use wide tape to relieve the stress among connections while handling the battery. Im not even sure if the butterfly idea is very smart, it was just one way to handle all these packs and the dimensions of the final battery.
once the battery is encased with a thermal shrink tube, it becomes quite a solid package, cause the shrink tubes usually harden after shrinking.
 
Back
Top