Project Triebflügel! Fifth movie uploaded.

Show off your E-bike creation here.
fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Project Triebflügel! Fifth movie uploaded.

Post by fellow » Apr 03 2015 3:14pm



This is the complete ebike with batteries inside the frame tubes, rear hub motor, controller and everything else needed. You have to look really hard to distinguish it from the normal bicycle! The controller is placed above the rear shock absorber with tollerance of less than 2 millimeters.
Image
According to wikipedia, "The Focke-Wulf Triebflügel, or Triebflügeljäger, literally meaning "thrust-wing hunter", was a German concept for an aircraft designed in 1944, during the final phase of World War II. It was a Vertical Take-Off and Landing tailsitter interceptor design. The Triebflügel had only reached wind-tunnel testing when the Allied forces reached the production facilities. No complete prototype was ever built."

Well, now it is :)


Last movie is screwed up by youtube, "stabilising" it all over the place. As I've already stabilised it in post processing, the end result was not pretty. Lesson learned. Most of the stuf was filmed with a mobile phone, highres pictures with Cannon DLSR, and the end scene with Sony RX100. Will order gopro and try to film in 4k next time.

Some pictures of the finished ebike:
Image
Image
Technical data
Tehnical data and the blog below is mostly for my own personal use so I can keep track of the air pressures, oil change dates, oil volumes etc. It is even intended for people that are building their first ebike and need quick and dirty info.

U_charged: 75.6V HOTC (18s2p). 151.2V HOTC possible because of 2 battery 75.6V packs inside. I'm waiting for the To-220 mosfets to evolve, 151.2V is currently not used but is "nice to have" for future use.
I_max: 7A cont, 14A peak. Clutch limited. Thanks to cwah for clutch information and research in this matter.
I_phase: Less than 30A. This is clutch limited too. At 45A, clutch is sometimes making noise (but it climbs like a goat, and do a wheeles with some help). Less range is another drawback. Choosed 30A, 45A is pushing it.
P_in_cont: 0.5kW.
P_in_peak: 1kW. Clutch limited.
Fuselinks: 18A@30seconds silicone insulated 0.75mm2 fuselinks per every 18s pack to protect the NCR18650PF cells in case the damage of the internal wiring. 36A@30 seconds in total in case main fuse fails. Every cell is internally CID/PTC fused.
Main fuse: Littlefuse 0326015.MXP 15A, 125VDC, slow blow. (15A=4h continuous, 20A=1h continuous, 30A=5 to 60 seconds) to protect internal wiring.
Wheels: Schwalbe Super Moto 26"x2.4" or Schwalbe Magic Mary 26"x2.5". Front and rear tyre air pressure is 3 bar. Tubes: Continental 26"x2.5", Dunlop valve.
Front suspension: Travel 150mm, air pressure 6 bars. Maximum theoretical frame travel is 200mm+. 50ml of 5W-30 motor oil in the stanchions, 100ml in total.
Rear suspension: Travel 170mm@215mm*54mm shock, air pressure 16 Bars. Rebound set to max anticlockwise. Dampening 0 (open, maximum softness). Maximum theoretical frame travel is 210mm@240mm shock (horizontal top tube).
Brakes: Front 203mm disc, Rear 180mm disc. Shimano Zee hydraulic disc brakes, F+R. 4 pistons per caliper, 8 pistons in total.
Crank: Shimano Zee 36t, Hollowtech II crankset assembly. (1x7 speed hybrid downhill solution)
Derailleur: Shimano Zee RD-M640, short cage, DH.
Freewheel: 7 speed DNP Epoch 11t-32t. This really pushes DH version of the short cage RD-M640 to its limits. Three diamonds 13t-28t freewheel shifted poor, and 36t front-13t rear was simply not enough over 30Km/h. FR version is probably a better choice on other frames. This frame has zero chain grow, so DH version fits.
Motor: ATF oil cooled Q100H, 201rpm (223rpm according to Russel/Motomech) at 36V nominal. n_max=(75.6V/36V)*223rpm*15.7=7350rpm at 75.6V, and 14800rpm at 151.2V (Reserved for future use). m=2224g without the freewheel. Internal gear reduction is 15.7:1 (Thanks to Motomech for his research). Maximum oil chamber volume is 80ml, filled at 50% (40ml). When filled at over 50%, motor will develop a heavy leak. Reinforced clutch, se the clutch springs chapter.
Controller: Frankenstein S06S, third generation (Thanks to Snickers for his research). Mosfets: TI csd19536kcs 100V. Rdsontyp=2_3 mOhm. Maximum recommended DC current is 30A (Q100H 201rpm melting point is at 20A!). Mosfet technobabble: Id_pulsed_drain_current=400A. Id_continuos_drain_current (package limited)=150A.
Alternative mosfet: Infineon IPP023N10N5 100V. Alternative mosfet at 150V voltage is: Infineon IPP110N20N3 G 200V. Rdsontyp=10.7mOhm.
Range first prototype: More than 23km without pedalling, without the external add-on pack at 20'C. Range is less than this at lower temperatures, and more at higher temperatures.
Range second and third prototype: More than 32km without pedalling, without the external add-on pack at 20'C. Range is less than this at lower temperatures, and more at higher temperatures.
Total mass, first prototype: Less than 20Kg. Second and third prototype: Less than 22kg(display added, heavier batteries, Scwhwalbe Magic Mary tires vs Super moto).
Top speed: Limited to 38km/h. More than 50km/h possible (but kills range drastically).

Battery
First prototype: 375Wh (36 Panasonic NCR18650PF cells x 10.4Wh)
Second prototype: 400Wh (36 Panasonic/Sanyo NCR18650GA cells x 11.2Wh)
Ebike charging time: about 3 hours at 2.5A~3A (rapid) bulk charge.
Capacity, voltage and peak current of the first prototype: 5.8 Ah, 75.6V, 36A under 6 seconds peak or 2.9Ah, 151.2V, 18A under 6 seconds peak.
Cells: Panasonic NCR18650PF, total number of cells is 36. Total battery mass 1620 grams + 180 grams wiring = less than 2 KG.
Capacity per cell: Nominal 2.9Ah, IRL 2.7Ah
Max discharge per cell (constant current): 10A
Max discharge pulse current per cell (5-6 sec.): 18A
Full charge per cell: 4.2V
Charging method: CV/CC
Minimum recommended charging current per cell: 0.6A
Rapid charging current per cell: 1.35A and beyond.
Nominal (storage) voltage per cell: 3.6V-3.7V
Minimum recommended discharge voltage level per cell: 2.5V
Target discharge voltage per cell: 2.8V(cell minimum voltage), 3.0V(resting voltage)
Dimensions per cell: 18.5mm x 66.5mm without tabs, 20mm x 70mm with shrink tape, solder and spring formed tabs.
Weight per cell: less than 50 grams

Frankenstein S06S hacking information for V1 2012-2014 and V2 2014-:
Picture number 29 Hacked S06S.JPG
X5=White=Speed Sensor
TS=Green=Throttle signal. Vcc=4.2V. When connected to 5.1V = cutoffs!
RED=+4.2V
TV=Yellow (thick)=Brake signal
Z=Black (my rig) or yellow(thin v2)=Pass
X6=Blue(my rig)=Speed sensor
X4=unknown.
X7=unknown.
XS=unknown. Acts wierd when connected to GND and 4.2V!


LVC=Voltage divider hack

Approx MCU voltages are: LVC if under 3.3V and HVC if over 4.0V. Locate "big" 15kOhm SMD resistor marked 153 and a smaller 1.2kOhm SMD resistor marked 122. Connect your resistor and zener diode of choice in parallel with 1.2kOhm SMD (to the ground, third empty pin hole close to those resistors). Best combination so far is: 3.9V zener diode only, without any resistor. LVC about 64V.

Other less than perfect combinations tested:
15kOhm // 3.9V zener diode. LVC=63.3V /18s=3.5V per cell resting voltage (3.1v/cell sagged). HVC test OK. Cutoffs from standstill at some voltages.
4.3kOhm without zener diode gives LVC 66.5V. Some HVC problems. LVC is tad wrong in my 18S confiq. Range is 20km without pedaling.
6.8kOhm // 4.3V zener diode + 10cm of 0.75mm^2 shunt across ACS712. LVC=64V, but can be drained down to 58.7V. Problems: Phase current to high, sometimes it causes Q100H clutch noise. It stalls at WOT from standstill. That simulation is here: http://www.falstad.com/circuit/circuitj ... 8+144+2%0A

30kOhm resistor // 15kOhm + 3.9V zener simulation is here: http://www.falstad.com/circuit/circuitj ... 0+false%0A

About the simulation: You can double click the voltage source to simulate different battery voltages, and change zener diode voltages, change resistors, disconnect components using switches and so on... I'm unsure if HVC exists, or ACS712 limits the current at very high voltages as it hits the phase current limit (see the phase current hack).

75V hack

To adjust LM317 voltage/current, use 2 x 270 Ohms 3W resistors for 18S NCR18650PF. This puts the resistors at the very edge of what they can take thermally. Tested OK. Better option is to change LM317 to 75V switched voltage regulator Recom R-78HB15-0.5 Another alternatives are:
WRN78U15-500B 3 pin, 15V 18mmx12mmx9mm.
WRN78U24-300B 3 pin, 24V. 18mmx12mmx9mm.
Traco TEN 8-4813 36V-75VDC, 525mA max, 100V 100msec.
Pay attention, pinout for the LM317 and 78xx series is not the same! (Untested, reserved for the future use). Change all high voltage capacitors to 100V. Change mosfets to 100V+ compatible. Add LVC resistor/resistors.

Phase current hack i.e, "snickers hack"

To increase the peak phase current protection limit, a shunt must be added between pins 1 and 4 on ACS712. Possible shunt choices:
10cm of 0.75mm^2 wire shunt increases the phase current from 25A~30A to 38A~45A. Those values are estimatied. Warning! This hack makes Q100H clutch noisy!
5 cm of 0.75mm^2 wire shunt doubles the phase current limit, from 25A~30A to 50A~60A. Those values are estimatied. Reserved for future use (another motor).

Image
ACS712 pdf is here: http://www.allegromicro.com/~/media/Fil ... sheet.ashx

150V hack

Traco TEN 8-7213WI 43V-160VDC voltage regulator instead of LM317. 170V max 100msec. Shutdown voltage 42V. I_max=533mA. Dimensions: 32mmx20mmx15mm. 30USD/piece. Another alternatives are:
WRB11024YMD-3W, 24V, 25mmx25mmx10mm(14mm), 125mA max. 144V(155V) max.
WRB11024YD-3W, 24V, 32mmx20mmx10mm(14mm), 125mAmax. 144V(155V) max.
WRB7215ZD-8W, 15V 32mmx20mmx10mm(14mm), 170V max. 10USD/piece.
LTC7138 max140V!
Pay attention, pinout for the LM317 and 78xx series is not the same! Change all capacitors and mosfets to 200V compatible. Recommended mosfet: Infineon IPP110N20N3 G 200V. Rdsontyp=10.7mOhm. Add LVC diode. Untested, reserved for future use.

Q100H 201rpm@36V clutch springs hack:
Spring holder hole diameter: d=4mm.
Outer diameter of the spring: 4mm<D<3.5mm.
Original spring diameter: 3.85mm
Original spring wire thickness: 0.35mm
Recommended spring wire thickness: As thick as possible. 0.75mm works well.
Spring tension: 1 turn per mm and as strong as possible. Ball pen springs do fit, but you need much stronger ones.
Spring length: 15mm. Cut it as tight (as long) as possible. If you do curse during the install, the length is perfect. If they do not fit, cut them and try again. If it's too easy to install, you probably need a longer spring.
Total number of springs needed: 3

How it's made, see the attachments below.
Attachments
Picture nr 1 - Proceed Epicon side shot.JPG
This is the frame chosen for this project, MSRP 2700USD/2500€ Proceed VRC 2. It's German made, giving the name to the whole project. Previously, I've had US, Russian and this one is German interceptor inspired. This frame is ridiculously light and rigid, as many parts are CNC machined.
Picture nr 2 - Increasing the travel to 150mm.JPG
Epicon suspension fork is choosed because it weights 1770 grams uncut steerer, and it's lower part is magnesium. Higher parts are aluminium. Here, I'm increasing its travel to 150mm. I could increase it more, but the visual appirience became unsatisfying.
Picture nr 3 - 52mm outside and 19mm inside.jpg
This is the plan for hiding the batteries inside the frame. Nothing new here, the same procedure invented for the Interceptor V1 one year ago. The only change is increased voltage and range, and lower mass. Cable insulation changed to silicon, and improoved cable weight to 180 grams only. Working temperature range increased: -50'C to +180'C. Fuselink(s) are dimensioned for the NCR18650PF cells, 18A for 6 seconds burst per 8s pack. No need to carry excess mass!
Picture nr 3 - 52mm outside and 19mm inside.jpg (97.5 KiB) Viewed 11965 times
Picture nr 4 - Battery inserting.JPG
Inserting the third 8s pack. Yes, I drilled a MSRP 2500€ frame. I'm silly sometimes:).
Picture nr 5 - What's inside.JPG
2 of 8s packs inserted.
Picture nr 5 - What's inside.JPG (188.63 KiB) Viewed 2273 times
Picture nr 6 - Batteries inside.JPG
100.8V inside!
Picture nr 7 - 24s 100.8V, 8s left to go.JPG
100.8V inside, another angle. Yes, no batteries visible and there is plenty of space in both top tube and downtube! :)
Picture nr 8 - Let's make some more batteries, there is sitll place! DSC_0071.JPG
Let's do some more batteries, while we are at it! :) Cells are easily replaceable and cheap those days. 36 NCR18650pf cells costs less than a full 80l tank of gasoline in Sweden. They passed 1 year torture test in Interceptor V1 at very high current, even depleated to 0V they worked as new. Those are "C grade".
Picture nr 9 - The white stuff is to absorb big hits as this is downhill e-bike. I'm using 4 of them per pack DSC_0072.JPG
The white stuff is to absorb big hits as this is a downhill e-bike. I'm using 4 of them per pack. Their other function is for a "press-fit" against BB - so the batteries are secured inside and cannot move at all. As most of the stuff, I proudly manifacture them myself ;).
Picture nr 10 Wiring inside head tube.JPG
Wiring inside head tube. A lot of wires! 4 or 5 pieces of 2X1.5mm2 would probably not fit inside. Opted for 2x0.75mm^2 silicone, 180'C approved and they act as a fuselink.
Picture nr 10 Wiring inside head tube.JPG (239.83 KiB) Viewed 2268 times
Picture nr 11 BLDC_IncreasingVoltage.jpg
The most usual way to increase the torque of a motor is inreasing the current. I'll try another approach, I'm increasing the voltage so I can keep the wiring and motor mass down. One "unwanted" effect of this is increasing the V_max. Another is higher core loss at high speeds, which are usually short bursts anyway. Third is decrasing the range. Fourth is hitting the "wall of air resistance", where the motor can't hold continuous power at high speed and starts overheating. I'll try to fight this by using the lowest KV motor avaible for the motor type, highest gearing avaible and by using the speed limiter.
Picture nr 11 BLDC_IncreasingVoltage.jpg (22.24 KiB) Viewed 11619 times
Picture nr 28.JPG
Finished product.
Picture nr 28.JPG (239.7 KiB) Viewed 2260 times
0-100 Triefluegel.jpg
0-100 Triefluegel.jpg (131.34 KiB) Viewed 2258 times
IMG_3401 ok resides.JPG
Last edited by fellow on Apr 11 2016 5:02pm, edited 50 times in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel!

Post by fellow » Apr 03 2015 4:14pm

Now, let's decide what motor to use. Any suggestions for a good climber hub under 3 kg that likes 100.8V? :mrgreen:

EDIT: Seeing that 0.3mm lamination is still in absence, I will probably wire it as 18s pack (75.6V), top tubes consisting of 2s packs. Rest will be filled with 1000uf capacitors close to the controller. It makes it 0.6V illegal according to new EU directive, which is often a good thing (from the engeneering point of view) :). Now I can print "illegal in all 28 EU states" on it and come close to "illegal in all 50 US states" as one ES user already did :).

A lot of new components ordered, more batteries, state of the art mosfets, capacitors and other stuff so now is the time for the long slow-boat wait. This blog will probably be updated when the mailman arrives.
Last edited by fellow on Oct 02 2015 7:06pm, edited 1 time in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

User avatar
fesanand   10 W

10 W
Posts: 98
Joined: Jun 01 2014 5:48am

Re: Project Focke-Wulf Triebflügel!

Post by fesanand » Apr 04 2015 5:23pm

Our ways goes in opposite direction: I finished my second build, a full suspension 58v, and thinking to start my third bike: A carbon lightweight hardtail, I need to pedal more, I'm gaining weight.
The point is that I got lots of information from your previous Interceptor project. The detailed review you made is a source of information for the exact motor + battery+ controller (performance) I want to achieve.

Good looking project and welcome to the full suspension world, I'm warning you: you'll ride faster due to the confidence given to a frame that absorbs everything.

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel!

Post by fellow » Apr 06 2015 10:41am

fesanand: I'm extremly happy with both NCR18650PF and Q100H motor used in that ebike. IRFB4110pbf mosfets were a big dissapointment. Those mosfets are nothing special (=weak link), S06S box needs FDP027N08B or csd19536kcs to really shine as the rest of the components. I used 1.5mm2 wires internally, which was overkill as one battery string cannot create more than 18A curent for more than 6 seconds (36A for the whole pack). I'm using silicone 0.75mm2 wires now. They act as a fuselink is something bad happends, silicone insulation is approved up to 180'C.

Second bad thing was using a lot of fuses in stead of one main DC 6.3mmx32mm and a fuselink calculated silicone wire. It is simply more elegant, lighter, easier to work with and more exotic:). When it shorts, there is only one main fuse to change. If fuselinks blow, then something went really wrong and I must take everything and check it out anyways. On this bike its a matter of hours. On the yellow one, its days as the internal wiring went thru the BB and took wery complicated path...

I'm just affraid that a increased weight will spoil the fun with the full suspension....
Last edited by fellow on Apr 23 2015 3:29pm, edited 2 times in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

User avatar
spinningmagnets   100 GW

100 GW
Posts: 10642
Joined: Dec 21 2007 10:27pm
Location: Ft Riley, NE Kansas

Re: Project Focke-Wulf Triebflügel!

Post by spinningmagnets » Apr 06 2015 10:48am

The design principle involved is now called a "tip jet" (even when rocket powered), and oddly enough...it works.

http://www.gizmag.com/dragonfly-df1-helicopter/14539/
http://en.wikipedia.org/wiki/Tip_jet

Triebflugel:

https://www.youtube.com/watch?v=bqqQ48KFXJE


User avatar
beast775   100 kW

100 kW
Posts: 1203
Joined: Jul 05 2007 11:55am
Location: victoria british columbia

Re: Project Focke-Wulf Triebflügel!

Post by beast775 » Apr 06 2015 2:36pm

Wicked hiding spot for the batteries! :D I bet you cringed when the drill bit broke through.
2 FINGER SALUTE TO ESSO.no buying gasoline for 7 years now.
I spent all my money on bicycles women and beer the rest i just wasted.

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel!

Post by fellow » Apr 07 2015 11:09am

spinningmagnets: Yes, that's the one! :) Impressive chopter on H202, and they say Lipo batteries are flammable :wink: :!:

beast775: Yes, a lot of cold swetting there. Inside headtube were originally 2 big holes already. Those are welding aids maybe, I've seen this inside a lot of frames. I just had to remove the badge and drill the outer one. Started with 3mm, then increased gradually to 18mm. Hole is now eliptical 20mm-22mm in diameter. Battery 8s packs are eliptical 19mm-21mm in diameter. Took one new picture (picture nr 10) to show that the big hole was not the only one drilled. There are two 10mm holes inside to accomodate 5 pieces of 2x0.75mm^2 wires emerging from the 8p packs, and additional PAS sensor that is not yet in place. USB cable and throttle cable will also be hidden inside the top tube off corse.

Dremel (my favorite tool besides my Makita) did fantastic work here! Fem analysis says chassis is as stiff as before. My tests on Interceptor show 1 year of use without failure on a much weaker frame, so this should be safe. There are massive welded plates outside that are taking the majority of the workload. This is the peak inside the headtube:
Image


Choosing the motor

Choosing the motor was not an easy task. Possible choices:
1. Central motor. Mass in the right place. Estetically unappealing.
2. Big DD motor like Muxus 3000 or the like: too heavy for the full suspension (rear suspension is air, no coil to keep the mass down). Too powerfull for the batteries, I don't want to ebike around with the aux pack conected all the time.
3. Mac 12T is too heavy, its 4.2Kg. It's peak power would match the batteries perfectly.
4. Q128H is heavy at 3kg.
5. Xiongda 2 speed is very interesting, but it is also 3kg.
6. Oil cooled Q100H revved at 7300rpm@76V thru the 15.7:1 reduction seems the only logical choice. Mass: 2224 grams without the freewheel. The only problem with this motor is that I already own 2 of them, both are oil cooled and revving 4k+ rpm@55V. I really hoped to try something new this time... At these revs, it will explode like a supernova, or make it's own gravity field. Or both :mrgreen: . I'm looking forward to the first test... EDIT: It didn't explode.

Meanwhile I'm waiting for the parts, take a look on how to increase the power 10-times, great youtube video done by subcooledheatpump@youtube:
Last edited by fellow on Jul 01 2015 2:33pm, edited 4 times in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel!

Post by fellow » Apr 23 2015 2:59pm

The mailman arrived... the bike is half-finished now. To do:
1. Paint the Frankenstein S06S in white.
2. Hide the cables and install the kill switch.
3. Install the fuse.
4. Lace the wheel.
5. Rain shower test
6. Call my frazy friends to test drive it, in case something is loose.
7. Test drive it myself:).

If it reaches steady 40km/h every time I ask for it I'll be more than happy. EDIT: It reaches 40km/h at 75V, 7A cont, 14A peak.
Image
Attachments
Picture nr 13 - Frankenstein in the making.JPG
Frankenstein in the making.
Picture nr 14 7 speed mass.JPG
13t-28t 7-speed freewheel.
Picture nr 15 36 cells inside.JPG
36 cells inside! What a mess!
Picture nr 16 hate that 7 speed sram.JPG
I dislike this 7 speed 1:1 SRAM shifter, but it should work well with Shimano Zee 1:1 pull ratio.
Picture nr 17 ergonomic study.JPG
Ergonomic study... The green button is "nitro" button doubles the voltage (not used at the moment). I just need a speaker and a Super-Hang-On music theme when i pass the 280Km/h... I mean 25km/h off course :). Ended up moving the bell to the left side.
Picture nr 18 have to hide the mess.JPG
Have to hide the mess...
Picture nr 19 Q100H mass.JPG
Q100H mass is 2224 grams.
Last edited by fellow on Jul 03 2015 2:53pm, edited 9 times in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! Now with video-teaser :)

Post by fellow » Apr 28 2015 5:54pm

Todays progress:
Picture 23 Almost finished.JPG
Almost finished!
And off course a video-teaser on the first page:).
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! Now with video-teaser :)

Post by fellow » May 12 2015 2:06am

I'm still waiting for the 75.6V charger... Here are 2 more pictures. The connector in the middle is 3-pin mini XLR, aux battery pack/charging port. The white box is S06S Frankenstein, not secured in place (fitting job). My worst paintjob ever, used wrong primer. Ended up scrapping the paint and doing it right (not pictured). The big hole above is where the on/off switch will be placed. This will be hidden by the saddle/seatpost support thingy, se one post above.
Image

Here is probably the most important picture on the ES:
Picture no 27 fairchild.jpg
Yes, we shold use 200V and not 100V mosfets! Shocking isn't it? I will change those as soon i find 200V that have RDS_on less than 4mOhms.
Picture no 27 fairchild.jpg (69.63 KiB) Viewed 11094 times
Attachments
Picture no 25 Frankenstein.JPG
Frankensten in the making
Picture nr 26 My worst paintjob ever.JPG
My worst painjob ever, used different primer and main color... Will repaint it again after the initial test.
Picture nr 26 My worst paintjob ever.JPG (240.78 KiB) Viewed 2263 times
Last edited by fellow on May 22 2015 9:19pm, edited 4 times in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

User avatar
crossbreak   100 MW

100 MW
Posts: 2800
Joined: Aug 02 2011 11:20am
Location: Germany
Contact:

Re: Project Focke-Wulf Triebflügel! Now with video-teaser :)

Post by crossbreak » May 13 2015 5:44am

Triebflügeljäger = Nazi BS :P
Yes, we shold use 200V and not 100V mosfets! Shocking isn't it? I will change those as soon i found 200V that have RDS_on less than 4mOhms.
You wanna tell us that we shall use our irfb4110 at <48V? What stuff did you smoke, my friend?

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! Now with video-teaser :)

Post by fellow » May 13 2015 6:42am

Fairchchild and Infineon both recommend extreme V_ds margin. Infineon recommends 200V to 250V for 110V systems. It is a good engeneering practice* to build things that last, we can't argue about that. At the same time, we have the problem with increasing Rds_on as the V_ds margin increases. One have to make a compromise, 11mOhm mosfets get very hot when used hard. Mosfets are getting better and better, Infineons 200V mosfets have about the same RdsOn as original 68V mosfets inside S06S (11mOhm).

I've smoked (pun intented) a lot of mosfets on my test rig since my wires are unusually long.

4110 are very old tech, I do not recommend them. I recommend those to-220 mosfets for 48V applications instead of 4110:
1. Infineon n-channel mosfet IPP075N15N3 G V_ds=150V, Rds_on=7.5mOhm High voltage margin, a bit high Rds_on for my taste.
2. Infineon n-channel mosfet IPP041N12N3 G V_ds=120V, Rds_on=4.1mOhm About the same Rds_on as 4110, but 20V higher V_ds.
3. TI csd19536kcs V_ds=100V, Rds_on=2.3 mOhm Same voltage margin as 4110, but much better Rds_on. Note: I blew 100V 4110 mosfets on 54.6V after one season, so 100V is actually too low if the battery cables are very long. The other 54.6V system is still going strong on 4110, but its battery cables are extremly short. That controller is running mostly cool, but does get hot when abused.
4. The lowest I recommend for 48V system with short cables is Fairchild FDP027N08B V_ds=80V, Rds_on=2.2mOhm. I have 10 of those in to-220 package that I do not use, PM me if you need them.

For this 75.6V system, I'll start with 100V TI csd19536kc together with extra capacitors close to the controller. I will upgrade to higher V_ds margin when those mosfets blow. Actually i have one S06S controller with 3 different mosfets inside for testing purposes, and one spare controller when the other one blows. The goal is robust system with lowest Rds_on possible, so I understand reluctance to high V_ds mosfets here on ES.

*http://en.wikipedia.org/wiki/Good_engineering_practice
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! Now with video-teaser :)

Post by fellow » May 22 2015 8:05pm

Ebike is now finished, but not tested yet. I'm waiting for the charger ordered one month ago. First picture/post updated.

Edit: Maiden pre-flight done today, with empty batteries (as a bicycle). Full suspension tuned to 6 bar in the front and 16 bar rear with huge help of my test driver friend. Gears crisp as a silk, no drag at all from the motor. Before ATF oil fill motor "cogged" badly. After ATF fill motor is smooth as a silk. Frame rigid and plush at the same time, 5W30 oil in the stanchions really helped. Vibration tests:pass. Water test: passed. No fire or smoke on power on :mrgreen: . Motor clicks with throttle, 5V at halls and throttle voltage ok. LVC seem to be active at about 3.3V-3.5V/cell, S06S behauviour at 18s still has to be checked as soon charger arrives. 36s mode not tested.

To do:
1. Charge batteries
2. Make a youtube movie:)
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

cwah   10 GW

10 GW
Posts: 4130
Joined: Jul 24 2011 5:42am
Location: Between paris and london

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by cwah » May 25 2015 3:01am

Your big enemy here is not heating but destroying the clutch or gears

I have been using the bpm at 72V35A on a small 20" wheel and clutch died after 3 months. I decided to change it to 48V 35A on 26" wheel and clutch also was destroyed after a month.

No way the smaller q100H clutch can survive at that power level
Help me find my stolen electric brompton: http://bit.ly/1a0vbBC and Bosch Sinus B3 http://bit.ly/1eV0WQz

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by fellow » May 25 2015 6:27am

Thank you, even if that is bad news for me. . . Then I will detune it to 14A*75.6V=529W cont, 1058W peak. That is the lowest the shunt accepts, and it is still plenty of power for my needs. Two of my Q100H motors survived one year at about 700W to 900W peak, ATF oil colled. Are you using ATF oil inside any of your motors? Small quantities less than 1/3 of the internal volume does not leak at all.

Modified S06S controller (V1) seems to dislike high voltage too, it starts spinning and stops in some kind of emergency stop at 65V. I've increased battery voltage to 69V and same thing occurred, so it is not LVC spooking. Will try another controller (S06S V2, the one with smaller capacitors) before I change voltage sensing resistors (1.2K and 15K close, but not connected to the 5V regulator). I had problems with this controller before.

My guess is that you are using some other controller than S06S?
Last edited by fellow on May 25 2015 9:54pm, edited 2 times in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

cwah   10 GW

10 GW
Posts: 4130
Joined: Jul 24 2011 5:42am
Location: Between paris and london

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by cwah » May 25 2015 7:20am

I've been using bmsbattery integrated sine wave controller at 48v20A:
https://bmsbattery.com/ebike-battery/70 ... ttery.html

It peaks around 900W with a front q100H. I've been using it for more than 4 months no problems, and no cooling required.

most of my motor overvolted remain cool... but peak power just destroy the clutch even if it's cool.

No sure what to do about that so I decided to get a dd to be worry free :lol:
Help me find my stolen electric brompton: http://bit.ly/1a0vbBC and Bosch Sinus B3 http://bit.ly/1eV0WQz

1KW   1 kW

1 kW
Posts: 341
Joined: Jul 24 2013 10:23am

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by 1KW » Jun 25 2015 8:12am

any updates?

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by fellow » Jun 30 2015 3:13pm

1KW wrote:any updates?
Yes, I've had a maiden flight (still waiting for a 18s charger = almost empty battery) and took some pictures of the finished ebike. New movie added to the first post.
Picture number 33 Side shot2.JPG
Side shot
Picture number 34 ISO shot.JPG
ISO shot
Picture number 35 Motor.JPG
Oil cooled Q100H
Picture number 36 ONOFF button.JPG
On/Off button
Picture number 39 Clean look.JPG
Clean look
It pulls as an animal, did my first mini burnout at 1Kw (concrete + loose sand + Schwalbe SuperMoto slicks). After hacking S06S's LVC, is seems that 4.3KOhm resistor over 1.2KOhm is not enough. LVC is active at 64V, so i have to fine tune it. It may be phase shunt too, I'll look further into this after my vacation. It was a lot of face wind, top speed was 34Km/h (1/2 to 3/4 throttle because of cutoffs) at 64V, with LVC kicking in every few seconds.

Unfortunatly I'm having really bad luck with aliexpress and chargers. My first charger (13s) took more than one month last year, my second charger (18s) never arrived (refund), and my third charger(18s) is not trackable. I guess I have to order again after vacation...
Last edited by fellow on Jul 01 2015 11:40am, edited 1 time in total.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

User avatar
chas58   10 kW

10 kW
Posts: 706
Joined: Mar 19 2013 4:05pm
Location: Detroit, Michigan, USA

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by chas58 » Jul 01 2015 11:05am

I love your builds, and am waiting for some videos!

Going for 18s2p this time eh? A little more speed wouldn't hurt. I'm using high speed cutes (300 rpm and 320 rpm) that get me 38 and 42 km/hr respectively, with me providing about 1/2 the power.

This guy took your idea and is doing a kickstarter project with it:
http://www.gizmag.com/maxwell-epo-e-bike/38248/

(although someone in Germany built a bike just like that about 2 years ago, I didn't ever see if he put it into production).
25^3 bike: 25 lbs, 25 mph, 25 mile range.
Road and Mountain Bike Cute Q100 builds:
http://endless-sphere.com/forums/viewto ... =3&t=49691

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by fellow » Jul 01 2015 11:39am

Thank you :)! I'll try to make some "real" videos soon :). 42 km/h is a great speed! If I can reach/sustain it I will be more than happy! :D

Edit: The charger arrived! Yesterdays top speed was 40km/h at 74V, measured by GPS (no wind, no pedalling, 6 km trip). Today, it accelerated up to 37.8Km/h, then down to 32 km/h, oscillating all the time at the same voltage of 74V. Second run today it reached only 32 km/h, cutting off at that speed. Last 7 km run reached incredible 54.9km/h at a slight downhill (must be a GPS error, it didn't felt THAT fast!).
54_9 kmph.jpg
54_9 kmph.jpg (148.15 KiB) Viewed 10038 times
I will look further into this, but this is probably due to some wierd controller limit (speed limit, erpm limit or 7A/14A limit is maybe to low at speeds over 32 km/h). I'm not using PAS and not using the display, which can be the reason too. The white speed sensor wire from the Q100H is connected to white speed wire on the S06S, maybe I should remove it?
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

User avatar
tumich   100 W

100 W
Posts: 188
Joined: May 28 2014 3:28am
Location: EU

Re: Project Focke-Wulf Triebflügel! 7k rpm of sofisticated p

Post by tumich » Aug 19 2015 4:42am

Beautiful build :D
My ebikes:
1. Sur ron
2. Treking bike with 250W PAS and 34Ah battery

Cheap, new, oryginal li-ion cells 18650 up to 3500mAh --> viewtopic.php?f=31&t=61608

User avatar
chas58   10 kW

10 kW
Posts: 706
Joined: Mar 19 2013 4:05pm
Location: Detroit, Michigan, USA

Re: Project Focke-Wulf Triebflügel! 7k rpm of sofisticated p

Post by chas58 » Aug 20 2015 2:13pm

I'm just thinking with 3500mAh 18650s comming out, that would be a great power source.

Some frame coupler would provide easy access, to batteries inside the frame, but it looks like it is expensive to retrofit something like that. to a frame.
Image
25^3 bike: 25 lbs, 25 mph, 25 mile range.
Road and Mountain Bike Cute Q100 builds:
http://endless-sphere.com/forums/viewto ... =3&t=49691

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Focke-Wulf Triebflügel! 7k rpm of sofisticated p

Post by fellow » Sep 04 2015 5:30am

tumich wrote:Beautiful build :D
Thank you:).
chas58 wrote:I'm just thinking with 3500mAh 18650s comming out, that would be a great power source. Some frame coupler would provide easy access, to batteries inside the frame, but it looks like it is expensive to retrofit something like that. to a frame.
I agree! Interesting times, new batteries are getting better and better!

Update: Kickstand added, new movie uploaded, new "artistic" picture taken showing the controller placement:
Artistic picture of the S06S controller.JPG
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

fellow   1 kW

1 kW
Posts: 326
Joined: Apr 01 2014 11:04am
Location: Sweden, EU.

Re: Project Triebflügel! Third movie uploaded.

Post by fellow » Sep 15 2015 5:55pm

The clutch didnt like 1000W+ peaks at all. I'm in the process of removing it and I will problably end up welding it. Motor is still cool at 75V and 14A~17A peak. When this upgrade is finished, I'll start my new 3kW 100V build so this thread will probably not get updated very often.
My 18S Q100H Frankenstein S06S project is here: http://endless-sphere.com/forums/viewto ... =6&t=68305

User avatar
r3volved   100 kW

100 kW
Posts: 1452
Joined: Jul 24 2013 4:01pm
Location: Waterloo, Ontario, Canada

Re: Project Focke-Wulf Triebflügel! 151.2V of sofisticated p

Post by r3volved » Sep 15 2015 6:15pm

Some kinda wizardry
cwah wrote:Your big enemy here is not heating but destroying the clutch or gears

I have been using the bpm at 72V35A on a small 20" wheel and clutch died after 3 months. I decided to change it to 48V 35A on 26" wheel and clutch also was destroyed after a month.

No way the smaller q100H clutch can survive at that power level

Post Reply