Everyone,
Thanks for the input... I think perhaps I was not clear enough about my intended designs so I will elaborate on my requirements, intentions and plans (as of now):
For the initial build I had planned (if it will fit with the 8spd) to use Gary's Sprocket freewheel threaded adapter to acheive a dual freewheel in the rear at the drive side (right side). The pedal link would be normal #35 bike chain to a normal single speed 40-50t (stock on the 9FS) sprocket. That will provide the mechanics of the pedaling interface. The only possible problem I see here is with the fit of Gary's sprocket adapter to the 8sp SA.
The drive sprocket attached to Gary's adapter (or one similar) would be a 92t (largest extron sprocket I can find) #219. The motor would be plate mounted and partially enclosed in abs plastic for protection from the elements.
The drive sprocket would be roughly 9-13t which would combine with the front for a first stage reduction of between - 10.22:1 and 7.08:1.
[Current Motor]
Tower Pro 5330 9T has a normal kV of 235 in it's default Wye mode termination. When you bring out the Delta Phase wires (for a total of 6 phase wires) and power it in the Delta mode, the torque goes up.. the current consumption go down and the kv drops by (235 / 1.73) = 135.838 or for future calculations we will use 135.8. This is no load kv.
Idle Current: 1.3 A
Weight: 679g / 1.5lb
Dimensions: 63mmx105mm
Shaft Diameter: 8mm
Max Rated Power: 2900w or 3.8667 hp
Delta no-load kV: 135.8
Delta max efficient current: 33 A
Delta maximum loaded current (@60s) 46.25 A
Delta max no load rpm (@42v) 5,703.6 @ 100% efficiency
4,562.88 @ 80% expected efficiency
Wye no load kV: 235
Wye max efficiency current 57 A
Wye max loaded current (@60S) 80 A
Wye max no load rpm (@42v) 9,870 @ 100% efficiency
7,896 @ 80% expected efficiency
[Option #1]
Parallel RC Drive (dual right side) with 8 spd SA internal hub. Using one of Gary's adapters to run dual rear freewheels and drive sprockets.
FET based Delta/Wye Switching (Instant)
Front crank sprocket would be 40-50t (stock most likely one either the 9FS or the specialized), rear pedal sprocket would be 14t - standard #35 (maybe thinner like on the 9 spds).
Motor will be mounted either to the rear trailing arm on the 9FS (for constant chain length) or most likely below the dropout (where a kickstand should be) on the Specialized.
The output sprocket from the motor would be 8mm bore (or adapted) #219 and varying from 9t-11t depending on desired speeds and evaluation.
For this example we will use:
Parallel Drive configured -
14t rear #35 sprocket for pedal drive
42t stock front crank sprocket #35
11t #219 pitch 8MM drive sprocket on motor output shaft
92t Extron #219 sprocket mounted to gary's dual freewheel adapter
Primary Reduction: 8.36:1
Since the gear range on the 8spd internal hubs range from 1:1 down to 3.05:1 that would provide the following results (top effective speeds).
Secondary Reduction: 3.05:1 To 1:1
20" Bike (9FS)
SA8 Gear Mode Motor RPM 1st Stage Reduction SA8 Reduction Total Reduction Expected RPM and top speed Top Speed
8 Delta 4562.88 8.36:1 3.05:1 25.498:1 4562.88 / 25.498 = 178.9505 10.652 mph
178.9505 / 336 = 0.5326
0.5326 * 20 = 10.652 mph
1 Delta 4562.88 8.36:1 1:1 8.36:1 4562.88 / 8.36 = 545.799 32.488 mph
545.799 / 336 = 1.6244
1.6244 * 20 = 32.488 mph
8 Wye 7896 8.36:1 3.05:1 25.498:1 7896 / 25.498 = 309.6713 18.432 mph
309.6713 / 336 = 0.9216
0.9216 * 20 = 18.432 mph
1 Wye 7896 8.36:1 1:1 8.36:1 7896 / 8.36 = 944.4976 56.22 mph
944.4976 / 336 = 2.811
2.811 * 20 = 56.22 mph
As you can see in the above table, having a dual winding on the motor allows for efficient operation at full charge in the 10mph range (useable for pedestrian navigation)… this does concern me a bit as I would rather see the Top Delta speed and lowest Wye speeds overlap or even better have a 10mph top speed difference (8th in Wye should be 42mph) which would allow for a much wider range and smother use of conversion between wye and delta.
So this should explain better how and what I am trying to do, it is much like Gary's build and many others except I plan to use the Delta/Wye switching to reduce the KV and the requirement for a true initial reduction unit (hopefully) this means the parts required for this build are:
Part # Qty Description Source Price Weight
1 Gary's Dual Freewheel Large Sprocket Adapter for SA internal hub Gary $ 25.00 ?
1 Tower Pro 5330 9T Hobby King $ 49.99 679g / 1.49693876 lb.
1 92t Extron Composite Lightweight #219 rear drive sprocket (attached to Gary's adapter) Comet Kart Sales $ 14.50 ?
1 11t 8mm drive sprocket #219 $ 15.00 ?
1 Aluminum 2"x2"x24" (source material) to build motor mount Home Depot $ 7.99 .6lb
1 #219 chain $ 25.00 1 lb
1 SA 8spd internal $ 179.99 3 lb
2 Freewheels? Type, will normal type suffice for starters to be upgraded to ENO style ? $20.00 (for normal freewheels) .4 lb
$ 337.47 6.5 lbs (without parts 1,3,4 counted)
Just found this while looking for 8mm 11t #219 drive sprockets:
http://www.thescooterguy.org/whats_new.html
This is an 8mm chain, drive sprocket and freewheeling rear sprocket that looks thread on with an outter (pedal) sprocket (seems up to 76t) and an inner bike style sprocket. $125 from this place… never heard of it but will check it out for my (our) applications.
Wow, today is the day for discovery… again looking for the #219 drive sprocket (11t) I found this and at $24.00, I bought one just for the drive sprocket (the motor will be fun to play with I'm sure): http://www.allelectronics.com/make-a-store/item/DCM-636/36VDC-450W-MOTOR-11T-SPROCKET/1.html a 36v 450w motor with 8mm 11t drive gear intended for an iZip or something. Nice find, but I will likely want the drive sprocket on that one (for another project) but at least now I know the 450w iZip used an 8mm shaft and 11t sprocket which is priceless (and was free) since the parts (aftermarket bearing upgrade kits, sprockets, etc) should all be highly available and cheap!!!! Now I know what to search for.
A few bits I haven't discussed… Since the motor (in wye, the highest current consumption mode) can only handle 80A maximum and I plan to run it at 70A peak @ 60s maximum @ maximum of 42v, I decided to equip it with Hall sensors and for initial build and testing use an Infineon 9FET (shunt calibrated, reprogrammed but not upgraded) since it should handle the nominal 57A fine with the soldered down shunt at 42v max. I will work away from this as time permits but it's a quick, cheap and reproducible solution that should be under 50.00 for the common man as opposed to the HV110 for 200+.
One final note of research info… While attempting to replicate the bedini type motor (google it) I had an idea about handling the back emf (since I won't need it for detecting zero crossing due to halls) and tested it on the bench (in the crudest fashion)
Using a very tiny motor (max 100w, 2-3 lipo) and home made testing ESC I spun the motor (with a tiny prop attached for some load, indoors so no ambient winds) to max rpm at the voltage then cut the power. I had mounted the prop with a tiny neod… magnet and rigged a hall sensor to record RPM of the prop (headspeed). My measurements were from removal of power to stopping of the prop. I then changed the programming in the controller and added a few fets to the prototype board and basically created a phase inverter that feeds the current coming from back emf into a positive forward pulse into the next winding. This is only at 0 throttle and isn't at this time (I don't think) compatible with regen but… in my tests, when the controller was in this mode (and consuming no more than resting current) with 0 throttle the freaking motor spun a total of 2.3x as long.
This is not reproduced yet and is not scalable (with my current work) to a large scale eBike / HV controller (at least not for the near future)…. Additionally I want to test this against a traditional freewheel style rear hub of a low, mid and high end bike. If my initial findings are correct, this method consumes no more power than normal non-throttle operation would (and also requires no dissipation of the power via the controller), requires nothing more than a few additional fets and some logic in the controller and yet seems at this moment to allow NON-FREEWHEEL freewheeling.
For the skeptics… again I haven't reproduced this in useable scale and I haven't finished the comparisons under low power no load conditions however it does seem this will reduce the drag a non freewheel motor configuration will have upon pedaling. Before people start in about entropy and thermodynamics or conversion… this is not perpetual motion, rather it is the use of the already present back EMF generated by the stator in motion (the same back emf used for regen if you doubt it's existence) simply redirected back into the next phase coil of the motor in a semi "passive" manor (no power from batteries to return to motor).
Since no motor/drive train is 100% efficient this is subject to the inverse efficiency effect of the output drive line or more specifically if 80% of motor power is delivered to the actual wheel then only 80% of the power can be expected (actually more like 75% but… entropy and I haven't done the calculations plus inertia and gravity will help with this effect also).
This small charge pulse would slowly fade away after just a few electrical rpm (I think, haven't measured) due to the natural conversion losses but … it seems to me a much more efficient use of "freewheeling" than to use regen braking (though the two are not mutually exclusive).
While normally when we (the members) post design information and say this is to be considered proof of my thought and working process for the purpose of patent they are kidding… I am not.
While anyone is free to implement, test, use this for personal purposes… I have filed notice of intent to file patent (think pre-patent but guaranteed timestamp) and even commercial use and production rights will be free of charge but those will be subject to quality requirements to be compliant with the patent to ensure knock offs bring down the price (mass production) but the quality doesn't drop with them. My interest is far more in the EV community and to benefit the EV community than to cash in.
I will post further info about the status of this 9FS to RC build as determinations are made.
-Mike
