Dr_Lulz's 8kw Berg Viper

[dr_lulz]

UPDATE:
this is how it looks now:

Stats:

• Berg Viper DH Frame
• 200mm suspension front + back
• Magura MT5
• Mxus 3k Turbo 4T
• 133x NCR18650GA in 19S7P = 79,8V * 23,5 AH = ~1500 Wh
• Kelly KLS7218 sine controller
• Topspeed 70 km/h
• 60 km range
• 40 kg
• ZERO noise while driving

OLD POST:

Hello guys,
just want to show a bit off

i made my first build in 2012 using magic pie motor running at 5kw now... but i have always been dreaming about building something "real"
a few month ago i found soaresdacosta's ebike build (https://endless-sphere.com/forums/viewtopic.php?f=6&t=38754&start=400) using a very much unknown downhill bike "berg viper". Its pretty much the best bike to use. It has 20cm front and rear suspension and a lot of space for the battery. Dropouts are 150mm as well. Frame is thick as f**k. Perfect!

i got it for 600€ second hand, which is .... damn cheap for such a bike
View attachment 5
View attachment 14

i did some testing how much batteries could fit



... about 140 18650's
at 11Wh per cell thats about 1.5kwh

battery box design:
View attachment 12

torque arm design history including a final 3d printed prototype


lasered out of aluminum by laserteileonline.de
1.5mm for battery box
10mm for torque arms
cost me 60€ .. damn cheap!


test mounting:


whole box is bolted together using thin right angle profiles


Batteries
batteries are a really difficult topic ... i want to go for 8kw peak but i want to double or triple my previous battery capacity (0.6kwh) as well.
I was thinking about using samsung 18650-30Q that would deliver 15A per cell which would be 140x 15A x 3.3V =7000W in sum. But i think this is too much, as these could deliver 7kw continuous. I need it for peak only. Also they are not very cheap.
Fortunately tumich had some "unwelded" 18650GA's rated at 10A on hand. From lots of tests on the internet it seems like those are unterrated, they can deliver 15A and peak even higher. I guess thats enough for me.

batteries arrived from tumich,
145x 18650GA ... 390€


battery fitting test

140 fit but its too tight

My controller will be a heavily modified Kuateng with 12x CSD19536KCS Fets (100V 2.7mOhm)
So i will be limited to about 100A which means i need at least 80V to reach 8kw
Best setup would be 20S 7P (84V peak) but since 140x dont fit i will probably have to go for 19S 7p = 133 Cells (79,8V peak). Other options would be 22S 6P but i think this is too close to 100V and kill the fets.

arduino battery welder using 3s 4000mah 60C Turnigy:


first test welds using 0.15 nickel x25mm :
View attachment 3
... so faar so good
Plan is to solder thick copper cable on top to be able to handle 100A


Torque arm mounting and motor
The torque arms will be glued to the frame. As life insurance i will put 2x M5 screws as well.


First time mounting mxus v3 4T that i got "used" (new) from ebay laced into 19" moped rim.
Cost me just 320€


damn .. this looks so good

Todo:

• Battery box mount, i will probably get some kind of perfectly fitting clamp for the downtube lasered ...
• Battery welding and configuration. Should i go 22S (=92,4V) or better stay at 19S? I dont want to kill the Fets
• Glue back on the derailleur
• Mount new Magura 5 double piston brakes (got them new for 110€ @ ebay)
• Mount controller below the downtube
• Mount KT Display
• BMS, Charger
• Cabling, Thumb throttle, Lights ...

 

[Tommm]

20s would be best for charger compatibility.

Battery layout configurator here: http://e4bike.ru/page/battery-shape-configurator

 

[Merlin]

from what i read here it looks like you know what you are doing but with this GA cells you will have a huge voltage drop if you want to squeeze out 100amps in 7p. 30q would be the better choice but now its to late =)

dont know who said they are underrated but this is simply not true. even the official tec sheet with internal resistance show that and every "vape" battery testing website too.

but if this is your first big bike with a 13kg Rear Wheel you will be happy =)

i had some years ago a scott gambler. but your frame has better space for the battery. dropouts look also good. will be a nice rocket =)

 

[Tommm]

from what i read here it looks like you know what you are doing but with this GA cells you will have a huge voltage drop if you want to squeeze out 100amps in 7p. 30q would be the better choice but now its to late =)

I just read that.

You do not want to pull more than 10A from them for longer than a short burst. Depending on your winding T value and top speed requirement, I would go either high voltage or high amperage, or a mix between. If the bike is purely for shredding, drop the top speed(volts or serial cells). You will still have plenty of power.
I have a 160 PF cell pack, so 20s8p and the voltage drop at ~72-75A (8-9A each cell on 10A cells) becomes brutal. You can't go full throttle on 35-40% battery or you will shutoff at 3.0v. Your cells are higher quality but I have 1 more in parallel, so it should be roughly the same. 6kw is doable, but I wouldn't go above. 100A for 7p would be very serious torture, with cells going out of balance with a lot of heat and a drastically shorter life span.

When designing your battery, take the rated amps as the peak amps, not the other way.

 

[dr_lulz]

do you really think i cant squeeze more than 10A from 18650GA?
If it takes 10A continuous then it should be able to handle 15A burst

right now i think about going 22S = 92.4V ... fets should be able to handle it

 

[Tommm]

do you really think i cant squeeze more than 10A from 18650GA?
If it takes 10A continuous then it should be able to handle 15A burst

right now i think about going 22S = 92.4V ... fets should be able to handle it

https://lygte-info.dk/review/batteries2012/Sanyo%20NCR18650GA%203500mAh%20%28Red%29%20UK.html
Define burst.
It will get to 75C in 5 minutes at 15A, as a single cell, not cramped into a pack. In a pack, it would be 2-3 minutes, and if you take into account you were riding before and after too, so we don't start from room temperature... Just no.

What top speed are you looking for? The obsession with super high volts doesn't make sense if your controller has phase weakening.

 

[soaresdacosta]

Well done.
I see this movie before.

Luís.

 

[dr_lulz]

do you really think i cant squeeze more than 10A from 18650GA?
If it takes 10A continuous then it should be able to handle 15A burst

right now i think about going 22S = 92.4V ... fets should be able to handle it

https://lygte-info.dk/review/batteries2012/Sanyo%20NCR18650GA%203500mAh%20%28Red%29%20UK.html
Define burst.
It will get to 75C in 5 minutes at 15A, as a single cell, not cramped into a pack. In a pack, it would be 2-3 minutes, and if you take into account you were riding before and after too, so we don't start from room temperature... Just no.

What top speed are you looking for? The obsession with super high volts doesn't make sense if your controller has phase weakening.

well, for me peak means maybe 10-15 sec MAX.
In general voltage should be as high as possible to reduce losses. I dont care about the topspeed

i am planning to go street legal here in germany so i will be limited to 4kw nominal power (not peak!) and 45km/h anyway.

unfortunately last week my heavily modified (12 Fet 2.7mOhm) KT motor controller blew after some testing. There was a big light arc and now its completely fried. I dont want to spend another 6 hours beefing up another one so i have decided to go for a Kelly KLS7218S which seems to be 88.2V absolute max. I would prefer to go to the max voltage as close as possible but since i can only fit 139 cells in my box and i want to put in as much as possible i will probably have to go for 19S7P = 133cells now. Another option would be 21S6P with just 126 cells to get a higher voltage. But i dont have enough power anyway so i will need every cell.

The kelly should be good for 200 Phase and motor amps peak (20 sec). Running 19S =79,8V off the charger and 15A per cell this would be theoretical 79,8V * 105A = 8.4kw. Maybe 7.5kw with sag (and even less when counting motor efficiency...). But i think this is enough :lol:

 

[dr_lulz]

anyway there was some progress the last month


i designed the battery box mount

View attachment 8
3d printed prototype


and also made a backplate for the torque arms including the derailleur mount (no CAD pictures, sorry)

both custom laser cut out of 4mm steel and 10mm aluminum

mounted:



(not yet final)

battery box mounts:

so far so good


Then i have also upgraded the brakes to magura mt5 double piston (110€ new) because i did not really like the avid elixirs that were on the bike.
I can only recommend them. They bite really bad. I can block the front wheel with medium pressure. I have also mounted a new and fat "floating" disc. Btw: i cannot make a stoppie because the rear motor is too heavy



View attachment 1
tight fit. There is a 10mm spacer between the mxus and the disc


Todo:

• Battery box mount, i will probably get some kind of perfectly fitting clamp for the downtube lasered ... <-- done
• New derailleur mount <-- done
• Mount new Magura 5 double piston brakes (got them new for 110€ @ ebay) <-- done
• Mount controller below the downtube <-- controller blew up
• Cabling, Thumb throttle, Lights ... <-- ordered
• Get new Kelly controller <-- ordered
• Weld 19S7P battery
• Sand and spray paint everything black
• Get a Display (Cycle Analyst)
• Order 19S BMS + Charger
• Get street legal

 

[RollerBoy]

Curious - Which size frame do you have?

 

[dr_lulz]

another update!

Bike is almost done, was running it on testbench a few times with limited current. I can already feel the power

Here a few more pictures that i made during the build

Battery:



View attachment 11
6 mm^2 copper, good for short 100A bursts (i hope)
View attachment 10
View attachment 9
View attachment 8
arduino spot welder

View attachment 6


Kelly KLS7218S
The controller has 18x 2.3mOhm fets placed onto a nice aluminum PCB
I am really impressed how good this controller is designed compared to cheap infineons. It has 100v -> 12v DCDC, 2000uF caps and most of the controller PCB are actually copper traces to transfer the current from top to bottom fets.
View attachment 5
30x40mm aluminium tube

View attachment 3
perfect fit!
View attachment 2
View attachment 1
you can actually see the external BMS next to the wheel


Todo:
- Mount Cycle Analyst, Lights
- Street legal

 

[dr_lulz]

Curious - Which size frame do you have?

The frame is 17". There is only one size

 

[dr_lulz]

Btw during testing i found that my mxus 3k seems to run very uneven. I was expecting a little more quality for its price.

See in the video: (The chain is shaking!)
https://streamable.com/s/2xz6s/qdpxnv

 

[liveforphysics]

Beware of trusting the shrink between cells to keep your series cell groups from shorting due to chaffing from vibration over time.

Looks like a fast fun bike!

 

[dr_lulz]

Beware of trusting the shrink between cells to keep your series cell groups from shorting due to chaffing from vibration over time.

Looks like a fast fun bike!

hehe, i have wrapped every parallel pack in 1 layer of kapton.
Between the packs are: shrinkwrap - kapton - kapton - shrinkwrap.
Also the batterys are glued together so nothing moves
Hopefully this is enough

 

[Smoke]

Nice looking bike.

Could you tell us about how you soldered the copper wire to your nickel strip?

How many watts on your iron, any surface prep or flux required, etc.

I want to solder 14 gauge stranded copper wire to some narrower nickel strip for making 1s2p subassemblies of which 3x are built up in to a 1s6p triangular module joined by the wires when they are soldered in to an XT90 connector. A 15s2p wiring harness will connect 30x of the 1s6p modules in to a pack with an effective layout of 15s12p.

Thanks.

 

[madin88]

Nice to see that there are still people doing a great job on a MTB conversion. With all those high power bike frames available today almost nobody does it anymore. The bike is looking really cool :thumb:

 

[minimum]

Could you tell us about how you soldered the copper wire to your nickel strip?

You need special flux for that.


Beware, it's acidic so handle it's contents and fumes accordingly!

 

[Smoke]

Thanks. That says AG, do I need to use silver solder with that?

Now I need to find that flux or an American equivalent.

 

[minimum]

I used regular 60/40 tin-lead solder, it's easier to use.

 

[dr_lulz]

Nice looking bike.

Could you tell us about how you soldered the copper wire to your nickel strip?

How many watts on your iron, any surface prep or flux required, etc.

I want to solder 14 gauge stranded copper wire to some narrower nickel strip for making 1s2p subassemblies of which 3x are built up in to a 1s6p triangular module joined by the wires when they are soldered in to an XT90 connector. A 15s2p wiring harness will connect 30x of the 1s6p modules in to a pack with an effective layout of 15s12p.

Thanks.

i have used an old cooking plate, set it to 400°, put the strip and copper wire on it and then simply soldered it

 

[Smoke]

So just a pre-heat and solder?

Can you tell me what type of solder you used?

I don't want to screw up due to a simple misunderstanding.

I wish it was like 20 years ago when you could get lead solder that just worked...

 

[dr_lulz]

So just a pre-heat and solder?

Can you tell me what type of solder you used?

I don't want to screw up due to a simple misunderstanding.

I wish it was like 20 years ago when you could get lead solder that just worked...

i have used lead free solder. But i would recommend to use leaded because i had to redo a few joins ....

 

[spinningmagnets]

Buy 63/37 if it's available, 60/40 is OK...I have used the most popular lead-free solder (SAC305), and all I can say is...stock up on 63/37, or 60/40...

 

[dr_lulz]

latest update:


Right now i am working on street legalization with german TÜV + electric braking + legal lights