Dane.Kouttron

[12.15.14]  Upgrades To The Flying Nimbus

Flying Nimbus Upgrades Ahoy!

Enjoy Nimbus MK1? The following details upgrades to the Nimbus project, to increase its load handling capability, range and night-time visibility. 

If you haven't checked out part 1, do-so now! (its a bit of a read)


What?
Battery Upgrade Rear Lights Wheel Upgrade Motor Swap Grease & Weld CAD Files Conclusion Image Directory

What?
The first build log of the Flying Nimbus project received quite a bit of interest. During the use of the contraption, I've observed a few areas of improvement and have documented them here. The prime updates are:
  • Wheel mechanical mount upgrades
  • Increased capacity battery
  • Rear Lighting

Battery Upgrade
The battery pack has an odd volumetric constraint, the 26650 cells will not stack vertically, and horizontal packing does not achieve the necessary pack voltage / capacity. So a very precarious spacing was solved with precision printed cell spacers.  I determined that using A123 26650 M1A cells, at 2.2-2.3AH/cell, i could squeeze in an extra 20% energy capacity in the same volume. It appeared that the stock 18650 cells didnt quite hold up as well over time and had fallen to an average of 940mah/cell. This could simply be that the cells were aged (manufactured circa 2008). Either way, more energy, more range ahoy!
And lo, a glued together 3d printed battery array!
To meet the volumetric requirements, i glued the cells to their respective holsters. The cell group is setup to create two 7S2P cell arrays. Its important to keep the modules 'straight' as any bowing in the setup would result in it not fitting correctly on the tray. Note the final cell holster is missing in this view.
Welding and Adding balancer leads
For the cell-cell connection, 3/8" wide nickel strip was welded to each cell using a DIY resistive welding setup. A 7S balance lead [link] was added and soldered between each group. The location of the balance connector is fairly short. An extension [link] can be used to bring this balance connector (JST, 7 position) out to an external balance charger.
Heat shrunk 1L bottle 'armor'  
Four 1L seltzer bottles served their country and were cut-up to form the heat-shrunk casings for the cell arrays. Unfortunately they were actually a bit too thick, the battery array barely fit. Back to the drawing board. Shown, the hole and mounting point for the 50A dc circuit breaker and xt-60 charge port.
Swankier battery heat shrink
The heat shrink used was purchased from [link] Its fantastic stuff. These battery modules were then adhered to the tray. Tape and extra packaging was used on the outside of the cell groups to aid in wear and tear. The balance connectors are also visible in the front of the module.
Machining the battery bay end-plate to allow for slightly more space
Well, it all mostly fit, but some secondary machining on the end plates was needed to capture the last cell group.
Balance charging
For charging, a Hyperion 1420i 500W balance charger was used. Shown in its swanky DIY harbor freight suitcase. More to come on the Hyperion setup!
Upgrade from 18650 [14S4P] to 26650 [14S2P]
Here's the reason for the upgrade, a move from 3.6 Ah to 4.75Ah, surprisingly 30% increase in energy density, and a lower impedance to boot! Data collected via Hyperion 1420i
<Note the plot titles need updating>
14S4P 18650

14S2P 26650

Shown are both packs back to back. Each includes the 50A dc circuit breaker [photo], while the newer pack includes a rear XT-60 charge connector.

Rear Lights
The red tail light [link]
I looked for a 'dense' led strip for the rear of the skateboard and i found the above. Its 12V, so feeding power from the control board back over was needed. For this first setup, a 12v dc/dc was used tied to the battery pack. This is unfortunately not terribly ideal, as the under voltage cutoff of that dc/dc is ~16v. If the led is left on, it would drain to a non-recoverable state. Ideally these would get routed power from the control part of the chassis. 
SWANKY
The upgraded nimbus sitting @ MITERS. The rear LED draws a boatload of power, and heats-up accordingly. Its blinding at night, and definatley keeps up the 'what the heck is that thing' effect. 

Motor Swap
Removing the tire assembly

The motor shaft finally succumbed to a bend due to the offset loading / lever action of the wheel topology. It took roughly 22miles (1 mile/day for a month or so) for this to occur. Its visible clearly in the far right photo. With support from the opposing side of the wheel, the offset-lever action could be reduced significantly, as shown above. Unfortunately, un-bending is not an option,  motor-swap it is!
Removing the previous planetary array and installing on new motor

a 3 pronged gear-puller was used to pull the welded planetary off the previous motor and re-attached on the new motor assembly. The Reeve'd gears are holding up well!
Alignment and re-capping
The planetary transfers torque from the motor via a key way. Alignment is fairly crucial. A socket was used to apply force to the center of the planet-carrier and wedge it back in place. Finally, the hub/ring-gear assembly is added on and screwed in place.
Removing the bead on the tire using a Bridgeport
To remove the rubber tire from the rim was a fairly non-trivial undertaking. The full weight of the Bridgeport was required! I used an aluminum tube along with a block to press the tire out of the rim. Thats a lot of force!

Wheel Upgrade
The wheel itself actually has a bit of a design flaw, the support for the furthest side of the wheel is lacking, creating a small offset in the loading on the rim. This causes the mating between the motor and the frame to flex over time. In order to compensate for this, a wheel-mating feature was devised. Attached directly to the rim, it should provide the required support and constrain offset loading.
Waterjet Time!
After some setup, and making use of the scraps of 1/4" 6061 aluminum I've been collecting. The purpose of the 'X' style of the rim adapter was twofold. The scrap i had wasnt big enough for a full circle, and the tire valve stem still needed a way to reach the outside world.
Bearing inset
Fits like a glove. The bearing holsters are cut so that the M6 bolt passes through, while the cuts for the top plate are sized for an M6 Tap. The tapped ring has more material available for the screw to grab into, whereas the spacer rings come fairly close to the bearing material. 
Assembled
The bearing [link] fits well in its new home. The end plate is tapped for M6. Shown, 2 bolts hold everything in place.
Welded bearing insert
After some quick tig-welding, the bearing insert sits on top of the wheel hub. Pre-heating was key to get some decent welds.
Shoving in the tire & Leak Test
It took a lot of screwdriver + force to fit the tire back in place. Both myself, krikby and peter were needed to get this beast in place. Many thanks for the help! The sink was used to test if the tube was punctured in the process of installing the tire. No bubbles (or at least no super serious leak)
Putting it all together
The hub motor and shaft adapter are attached to the hub motor and the new bearing plate clears the side mount.
Re-mounting, re-cabling and ready to rock.
After completing the rim modifications (listed below) the tire was re-seated, re-inflated and everything was quickly shoved back together. 

 Sealing and Adding A Grease Fitting
Waterproofing the vent.
The vent holes, documented earlier, were intended to allow airflow to the controller. Fortunatley the super-swanky AMC controller's dissapation is fairly low, and all the vent does now is let mositure in. To fix this, i used some housing-silicone goo and painters tape. The inside of the skateboard is covered to prevent the goo from leaking in. The silicone was applied to the outside in thick beads.
Goop on that silicone
After letting the silicon sit for a minute, a sharp utility-knife blade was used to scrape the surface flat.
ACCELERATING SCIENCE

Using an incandescent lamp, next to the silicone, the drying is accelerated slightly. Shown (right) is the view inside through the tape.
reference to: [link]
Time for a grease fitting!
A 0.20" diameter hole was drilled in the hub motor, and a 1/4-28 threaded grease fitting was added. This allows grease to be added to the wheel assembly planetary gear train.
GREASE and allenkeys
Shown (left) is the grease-gun and the re-assembled nimbus. After adding grease to the geartrain, it seemed to roll in a smoother, somewhat quiet-er fashion.

Backlash improvements.
Fighting backlash
After conversing with Adam, about backlash issues that appear over time, as the pile of set-screws wear into the hub-motor shaft, the idea of welding on the adapter came to light. After some quick milling, a flat was added to allow for a bond between the hub motor and the mating shaft.  Backlash in this mate (even tenths of a degree) cause shuddering during direction change and can result in an upset rider.


DLAB AHOY
Campus was fairly empty, and fortunatley a comrade appeared so I could weld (dont weld alone!). Some quick tig-welding later (on both sides) and they appear to be one! Note the weld was performed at ~120A (high power, short duration) to prevent thermal buildup in the motor assembly. Thanks again DLAB!
PACKAGE IT UP
With the motor, adapter and everyone else re-packaged, everything fit back together!

In The Field
Nimbus attempting to blend in at a holiday-get-together-shindig.

(There's other photos in the photo gallery)
Concluding Remarks:

Overall the upgrades went well, it took ~20+ miles of riding to cause the shaft-bending issue to arise. I was initially hesitant to modify the rim further, namely due to the tire seating/removal process, but thanks to some comrades, that went by better than expected. The rear LED is surprisingly powerful, and does run rather warm, I'm curious about its operating life, as it doesnt transfer heat to the frame well through the adhesive tape it's mounted with. I still need to water-seal the vent holes in the front part of the chassis in a clean fashion. Otherwise, I have received a lot of feedback on the hardware, people are really interested in riding it. Interestingly, riding it around during daytime has a binary response with nearby people, some notice right away ask questions and become incredibly overjoyed to see it, while others are in zombie mode and just see a hovering person in their periphery. Finally, ideas for upgrades are bouncing around. As of this writing the weather hasnt warranted a snow/ice tire just yet, but it does seem to loom in the distance. Riding a self balancing vehicle on a frozen lake would be fairly excellent. An improvement that has been sitting on the white board for a while was proposed by an incredibly ingenious coworker, why not replace the dead man's switch with a beam-break laser. Not only would this look fantastic (glowing green light) but it would (potentially) remove the 'failed dead mans switch' issue.  

If you have questions or comments, ask below or send over an email.
Comments:
HTML Comment Box is loading comments...

(be careful, im not responsible for your incredible acceleration)

Dane.Kouttron
Rensselaer Polytechnic Institute 
Electrical & Electrical Power
631.978.1650