Project Interceptor, oil cooled Q100H@55V, in-frame battery

[fellow]

*According to wikipedia, "Interceptor aircraft (or simply "interceptor"), a type of fighter aircraft designed specifically to intercept and destroy enemy aircraft." This picture shows the complete e-bike with controller, batteries and everything.

Technical data:
U_charged=54.6V (4.2 V/cell). 13s Lipo hybrid, 48V nominal.
U_cutoff=45V(battery pack votalge)-5V(voltage sag)=40V LVC (3.1V/cell - 3.5V/cell)
I_max=17A
Main fuse: Littlefuse 0326015.MXP 15A, 125VDC, slow blow. (15A=4h continuous, 20A=1h continuous, 30A=5 to 60 seconds) to protect internal wiring.
Other fuses=down/diagonal tube pack (8+5 cells) 10A, Seat+top tube pack (6+7 cells) 10A.
P_in_peak=0,9kW.
P_out_peak=0,7kW at 75% efficiency
P_in_wheel_in_the_air=54W (1A@54V) @ full throttle.
Motor: Bmsbattery Q100H 201rpm 36V@48V / Aikema ZYJZ-100, ATF-U LT71141oil cooled. m=2.2Kg. Total oil volume is 80ml. 15.7:1 internal gearing.
Wheels: Schwalbe Nobby Nick 26" 57-559. Front 1.6 bar, Rear 1.8 bar. Schwalbe recommends 1.8 bar min, but it felt too harsh offroad.
Controller: Frankenstein S06S, upgraded to 6x IRFB4110bpf mosfets (100V rated). The original 68V mosfets are not suitable for 54.6V!
Charger: 54.6V, 2A. Charging time 3 hours.
Batteries: Panasonic NCR18650PF in 13s2p configuration. Chemistry is LiNiCOMnO2/Li-Manganese/INR/ICR/IMR hybrid. m=1.2Kg (45grams per cell x 26 cells).
Efficiency: 17.6Wh/Km
Range: 18km without pedalling, without external add-on pack.
V_max=32 km/h (30Km/h usable).
Total mass: Less than 17Kg.

Images attached below.

 

[fellow]

I've tested the combo today. Both good and bad news.

The bad news are:
- Cruise control is terrible, 3 of my friends (test drivers) had a lot of problems with it. We almost burned the motor as it continued to give full throttle at standstill. After disabling the cruise control jumper and wiring the emergency shutoff brake switch everything went perfect.
- Disabling the external speed spensor and using LCD gives inaccurate readings (both speed and distance). If the throttle is in 0 position, and bike rolls it doesn't take it into account. Plain stupid bug. The speed sensor is a must for distance reading. Edit: This is probably caused by the Q100h freewheel.
-It takes ages from 30Km/h up to 36km/h. I'll say that the practical top speed is 30Km/h.
-Frame is too harsh for off road use. Next project will be changing to full suspension downhill frame.

The good news:
+The bike pulls as an animal from standstill up to 30km/h. Top speed is 36Km/h on flat road, very little wind. The last 6 Km/h are more of academic importance.
+It's very silent, a lot of smooth pull. No strange noises at all.
+Voltage sag of the 13s2p Panasonic NCR18650PF batteries is less than 5 Volts worse case. At 45V (3.5V/cell=empty) it sags to 40V and activates LVC. At 46V it doesn't activate LVC at all. Hot off the charger (54.6V) the voltage sag is only 2V.
+The range is as expected, about 18-20Km without pedalling. We've tested it full throttle all the way.
+Charging took 3h with 2A charger, everything was cool on touch.
+I've tested S06S at constant 54.6V for 3 hours, it was about 37'C at its hottest part. It is definitly 13s Li-Po (48V nominal) compatible. It was about the same temperature when driving a bit later.
+Q100h stays very cool at 48V+, almost too cool.
+I did "David Copperfield"* inspired showoff passing some people. The expressions on their faces were priceless.

*=backward pedaling and using the throttle at the same time

 

[eagle_eye]

Hi Fellow,
thanks a lot for the detailed report and pointing out some good and some bad news!

The point with the 22km/h speed limit if no display used would be really bad news. Can you explain a little bit more - with the display you can set the S06S controller to max. speed and max. assist level (power output), right. Now if you remove the display and set the jumpers at 1,2 and 3,5 for display-less usage - is the controller config to max. speed and max. assist level gone and only 22km/h possible ?

Here d8veh is pointing out that the S06S comes with default setting to max. speed (no speed limit) ...
http://endless-sphere.com/forums/viewtopic.php?p=835686#p835686
(from thread: http://endless-sphere.com/forums/viewtopic.php?f=3&t=56090 )

 

[eagle_eye]

And so where did you hide the LCD display, dont find it on the pics. It isnt possibly inside the frame also, is it ?! :wink:

[....] As I see it the LCD is a must if you want to ride 30km/h+ (hide it somewhere, It's ugly).
[....]

 

[fellow]

I don't like that display at all, speed info is unreliable when decelerating regardless of preferences in the hidden menu. I used my smaller "normal" bicycle LCD at the moment pictures were taken. Today I started using "My Tracks Android" GPS app, not very reliable eather as I had a lot of stops at red lights (0km/h points not showing up on the graph). Here is the result of todays 7Km long test run with LCD 2.5 that ended up blowing the 15A fuse and shorting the S06S controller:

 

[fellow]

Hi Fellow,
thanks a lot for the detailed report and pointing out some good and some bad news!

The point with the 22km/h speed limit if no display used would be really bad news. Can you explain a little bit more - with the display you can set the S06S controller to max. speed and max. assist level (power output), right. Now if you remove the display and set the jumpers at 1,2 and 3,5 for display-less usage - is the controller config to max. speed and max. assist level gone and only 22km/h possible ?

Here d8veh is pointing out that the S06S comes with default setting to max. speed (no speed limit) ...
http://endless-sphere.com/forums/viewtopic.php?p=835686#p835686
(from thread: http://endless-sphere.com/forums/viewtopic.php?f=3&t=56090 )

My second S06S does not seem to have speed limit without LCD as the first one, so I'm really clueless on Chinease quality control . Answered the rest using PM, long posts as always. UPDATE: After changing mosfets on both S06S units to IRF4110pbf I can confirm that there is no speed limit (LCD or not, it does not matter). So it must be a glitch caused by blown mosfets or something similar. LCD is not needed at all and is now removed.

 

[fellow]

New seat, sram gripshift and new main picture added. Old seat didn't match interceptors 18km range, as my rear managed only 7km on it. New seat is big, heavy and ugly, but matches 20km range of the batteries better. It is the only replaced component on the bike that actually increased the mass (about 100 grams). The size of the seat gives me the unique opportunity to hide the controller under the seat and finally get rid of that first aid defibrillator-sized kit.

The bike is still under 17kg (about 16.5kg, 0.5kg less than my original project/target weight)...

 

[BiGH]

Very nice indeed.

Are the batteries easily removable from the frame or is it welded shut after insertion?

 

[fellow]

Seat and top tube batteries are easy removable, only seat post need to be removed. Down tube battery packs (the lower diagonal tube pack) are much harder to remove. Bottom crank (because of cables are passing thru the bearing), glued badge and steering assembly need to be removed (to get the batteries out that way). Nothing is welded, you can see the top tube battery thru the 20mm hole behind the seat if you are trying hard enough. The very same hole is used to get all the cables out. The head tube 20mm hole is hidden behind the glued badge, but its non-permanent glue so its removable.

 

[BiGH]

Very impressed!!!!

 

[fellow]

Thank you. Here is todays test ride, after changing to IRFB4110pbf mosfets. At about half a test ride (0.5km mark), I had to do an emergency brake as the the petrol MC turned right into my path. Not expecting a MTB comming that fast I guess. The speed limit was 40 km/h. Note the consistent top speed way over 35 km/h first 0.5Km, that was not possible before. Acceleration up to 35 km/h+ (first few meters on the graph) is now taking epic proportions! Lots of motion blur at warp 10 speed. Top speed was 40 km/h few runs later, but again it took eternity to reach it (not recorded here). 30 km/h on the graph was uphill/climbing. Sorry the GPS program is in Swedish but it's the blue line (speed in km/h) that is interesting.

 

[ecruz]

Wow this build is amazing, you cant tell great job!!! Congrats!

 

[fellow]

ecruz: Thank you!

Today everything just worked with new mosfets, which is kind of a bizzare feel :lol:. Going to extreme 100V mosfet (almost overkill margin against 54.6V maximum battery voltage) prooved to be a wise decision. The lower rdsOn of 75V(IRFB3077, 2.8mOhm) and 80V (TI CSD19506, only 2mOhm) were tempting but Q100H is 12.6:1 geared and voltage spikes are always unpredictible. After all it's was the Fairchilds recommendation to use 100V mosfets for 48V system. IRFB4110 mofsets really do take hard abuse without complaining! The 68V rated chinease fets died very embarrasing/unheroic death, so I just wanted something bulletproof and 75V and even 80V felt to close to 68V of failed fets. The repaired/upgraded IRFB4110 controller was as cool as the motor, it's obvious that the potential of the combo is not fully utilized. Edit: The controller is cool on flat terrain only. The controller do get warm at climbing. For hilly areas, I recommend the 100V 2.3mOhm TI CSD19536 mosfets for frankenstein S06S mod.

100V 2.3mOhm TI CSD19536 were simply nowhere to found here in EU, so my choice was easy - The IRFB4110 were cheapest of the bunch and easy to obtain. I've manifacured new "amp rig", so soon i'll be able to increase the amps in the controlled enviroment, measure the temperature of the units and compare it with the GPS readings of acceleration and top speed. I even have 1 meter of double insulated 2G1.5mm2 wire @ 56,4V outside my bike, to test at different amps (as my internals are using this wire). Edit: After few members reporting Q100H clutch failures at 1KW+ power levels, I'll stay at current power level.

[youtube]I5cYgRnfFDA[/youtube]
http://www.youtube.com/watch?v=I5cYgRnfFDA

 

[Snickers]

I transformed the small S06S to a frankenstein version...
In complementary of modification for running at 52V (80 Ohms input resistor replaced by 390Ohms and one small capacitor for 50V to 63V and LCV by one SMD 15K to 22K )
I reinforced all the power line and add a shunt (~0,3 milli ohms) in // of the ACS712
I replace the cheap original mosfet by some Fairchild version FDP027N08B 2,2-2,7 mOhms(max), Gate Charge 137-178nC 223A alternative mosfet are possible but cost double and sometime not available in acceptable quantity for me.
Then I push the limit of the S06S controller by reducing the main shunt. At 53A, I obtained more than 2500watts passed though the small box...

I’m sure, the S06 could do more but this is enough for me. :lol:
Some picture and detailedhere(in French!)

 

[stuntmanmike]

Wow! Cycle Pro, I haven't seen one of those since the early 80's. When I was a kid a had a "Macho II", that was their super heavy & indestructible bmx.
-Killer!

 

[fellow]

Snickers: Wow, 2,5 kW are really impressive! You've inspired me to increase the amps further.
Stuntmanmike: Thanks.

Here are my findings with Q100H (oil cooled),frankenstein S06S (4110), and ambient temperature at 10'C:

14A, 54.6V, P_in=764W. Q100H is at ambient temperature (10'C). Very relaxed to drive, very silent.
16A, 54.6V, P_in=874W. Q100H is at ambient temperature (10'C). The motor changed it's tune to "honda high rpm-machine".
17A, 54.6V, P_in=928W. Q100H starts to follow the laws of physics and started to increase its temperature to about 30'C.
19A, 54.6V, P_in=1 kW. Q100H is a killowatt machine now! Motor is confortable to the touch, but its far from cool at 40'C. The motor starts to sound strained, more of a diesel engine than a high rpm petrol one.
20A, 54.6V P_in=1.1Kw. Q100H sounded as a mix of a diesel engine knocking, and a pulsejet. It accelerated furiosly, making the loud "bangs" and polluting the air around. Test aborted due to safety reasons because of massive traction problems on soft pavement. Yes, it tried to kill me. The ultralight alu chassis started to twist and oscillate. Simply put, the ebike loosed its relaxing feeling and got twisted RWD personality disorder. Q100H temperature 40'C-50'C. Dialed down to 17A. At the moment, its at 300Km+ mark.

The Interceptor MK2 project got finished during this sommer. Here is a teaser of it:
View attachment 1


By the way, the Interceptor MK3 project started today. Technical data:
2WD, 2 kW, full suspension downhill frame, carbon fiber.
Target mass: less than 15Kg.
Front motor: Oil cooled Q100H at 16A.
Front controller: Frankenstein S06S (csd19536kcs)
Rear motor: Oil cooled Q100H at 20A.
Rear Controller: Frankenstein S06S (csd19536kcs).
Charger: 54.6V or 72V. This will be decided at the end of the project.
Throttle: half-twist.
Single speed: 36/16

 

[spinningmagnets]

I keep hearing so many positive things about the affordable sine wave controllers concerning how much quieter motors seem to run. Do you think the oil-cooling helped with quieting down the drive, or are the sound benefits coming from the sine wave only? (did you try oil-cooling with another controller?)

The Faraday and the Grace both have cylindrical cells inside the tubes of the frame, very stealthy, but...they are very expensive. Nicely done.

 

[fellow]

Do you think the oil-cooling helped with quieting down the drive, or are the sound benefits coming from the sine wave only?

I've never used Q100H without oil, so I can't comment on that. I can compare Q100H 201 filled with 80ml of ATF (300km) with Q100H 201 filled with 40ml ATF(50km). The motor with half the ATF is louder, but it drips less oil(almost nothing). 80ml ATF motor is very messy. Both motors are not annoying, the sound is very pleasent and hightech, especially at some resonant frequencies in the middle of the register. My tires are knobby, making that noise dominant.

did you try oil-cooling with another controller?

I've never used anything else than Q100H and Frankenstein S06S, so I can't comment on that. Default S06S got very crappy mosfets. Once changed to better ones, it is a very pleasent controller. The other downside beside weak mosfets is a bad compatibility with a hall half throttle. The MK1 needed a 1N4003 diode and a 270 Ohm 0.6W metal film resistor, the MK2 ebike matches without. I'm not sure if this is a S06S issue or a bad tollerance half throttle issue, but something is fishy with quality control there .


Here is the shunt with 19A limit:

 

[cjh]

I love the stealth look. Excuse my noob question but this intrigues me so how exactly do you go about constructing the battery packs and wiring them all together?

 

[fellow]

I love the stealth look. Excuse my noob question but this intrigues me so how exactly do you go about constructing the battery packs and wiring them all together?

Its very easy to do, one pack consisting of 7 (or 8 cells) cells takes about 15 minutes to make. Buy the NCR18650PF cells with pre-soldered "ears", like those:

.

Bend the metal in the chape of an "<" on one side, and ">" on the other side. Solder them together so metal parts are flexible . The bound in between should flex a little, but must manage 25A+ without melting so use as big contact area as you can. Solder the wires on the ends of the cylinder. Proceed until you get something like this:

Use the shrink tape to form the long cylinder. You can see the shape of the metal on the lowest battery that is closest to the camera. I was inspired by this video, but wanted to have flexible connection instead:
[youtube]J7DsZ_Sspp4[/youtube]
If you choose to use Panasonic NCR18650PF, you do NOT need an complicated external cell menagement system. A simple LVC (Low Voltage Cutoff) that is already programmed into S06S will do for a start. Don't forget one 10A fuse per every 13s block! Batteries are PTC protected inside. Those are very different of normal "explode-into-your-face" cells. Prior to the project, I've tortured them in imaginable ways, even depleated them to 0V and overcharged under lab conditions. I dare to say that no other cell can withstand the torture as well as this one sucker can! They are compleatly short-safe, idiot proof cells to start with! There are many fakes, i bought mine from a known reseller in Germany and payed about 6€/piece.

 

[cjh]

Thanks, that's really helpful. Just another noob question as I am still learning about batteries. I get that that 6+7 cells and 8+5 cells both add up to the 13S but I am still a little lost as to the 6+7 and 8+5?

 

[cwah]

Great work. Do you have spare motor in case you burn it?

 

[fellow]

Thanks, that's really helpful. Just another noob question as I am still learning about batteries. I get that that 6+7 cells and 8+5 cells both add up to the 13S but I am still a little lost as to the 6+7 and 8+5?

I'm using (8+5)s//(7+6)s=13s//13s=13s2p configuration=26 cells in total (13s in paralell with 13s). One NCR18650PF cell is good for max 10A,4.2V, and 13s is good for max 10A, 54.6V. The controller was 14A so minimum 26 cells was needed(20A). There was a space for maybe 4 cells more, but the internal cabling needed an awfull amount of space during install and 14s and 15s chargers were impossible to found. 16s seemed to be next voltage, but there is no space for 32 cells inside. The other reason for 13s was that I found the silent 13s(54.6V) charger. 13s felt as a perfect voltage to start with.

cwah: No spare motor, but I have one more ebike (Interceptor MK2) and 2 cars. I lace my own wheels, Q100H is about 90 USD, so it's not a big deal if I need to order another. With that said others Q100 failed at 17A-22A without oil so it's very interesting to see if the oil helps.

 

[cwah]

How long did it take for your q100 to fail without oil?

 

[fellow]

Others=other persons here at endless sphere. Both my Q100H are still going strong!
One is currently at 17A, 55V, oil cooled and the other one is at 16A, 55V, oil cooled.