Unusual ebike build-ALL condition, off road, sub arctic

oldswamm

100 W
Joined
Aug 21, 2010
Messages
169
Location
Bethel, Alaska
Hi, this first post has gotten long, as I added to it, waiting for the ES buyout crisis to resolve (then I got busy). (I logged on to post this the 4th, but…)
Please bear with me, and don’t let the length overwhelm you. Lots of ideas, they can’t all be bad (a room full of monkeys and typewriters comes to mind). Comment on the worst (or best). (And please don’t quote this whole 3000 word post.) And if you’re critical of an idea, say why, not just that ‘you don’t like it’, or that ‘it won’t work’.

I started this design about a year ago, shortly after I first electrified the bicycle I was riding last summer. IOW, I've only been riding ebikes for about a year, but I've been THINKING (and reading on ES).

As I make progress, it's my intention to append near the top of this post, or the next 2 ('electronics', or 'uses, conditions, other stuff'), and leave old postings to show in reverse order (dated). I will/might also change the title to indicate what part of the project I'm working on, keeping the 'unusual ebike build' part. When I have questions, I will try and post them near the top (right here).


The first question that comes to mind right now is whether the heim joints are a good idea (would probably seal them with felt after lubing, to slow the influx of sand.) Whether on the swing arm or the frame, they impose torque on the mount, making for extra weight. They’re cheap enough and easy to replace when they do wear. And yes, I rejected a multilink system as soon as it occurred to me.

The other question would be 'what questions should I have asked?'. There are dozens of unusual ideas here. If you think I’m actually in error somewhere, please don’t just say so, tell me why!
IOW, this whole post could be considered a list of questions.


Needed:
An aluminum 17" MC wheel, with or without a tire. PM me if you have one you can let me have cheap (I'm way to poor for this hobby). Also need a smooth tire (can be a worn out knobby) for winter studding.
Slightly lower on the list would be decent front shocks (air/dampened).
Then a narrow 21" aluminum MC wheel, w/wo tire, for the front.
(So far I haven't found any of the above locally.)

Guess I'll stick my first drawing here. (8/13/11)



I want an all condition bike for sub arctic conditions (to -40). Off road, mud (tundra bogs), snow (put on snow shoes and walk/carry the bike back to better conditions if you get into snow too deep and soft). Will be working on aggressive studding come winter.

I'm almost 60 and don't need speed, nor will I be doing much jumping (I hope). But, I have a hard time going by an obstacle without wondering 'if I can make it through there'.

I’ve wanted a 2WD motorcycle since I first saw the idea in Popular Science 40 years (or more) ago (300#, 10mph, ICE, many gears and shafts).

It's a big bike, built for me (6'2", 200#). 5' wheelbase. I will just be able to stand straddle of it with the suspension extended and the seat low and level.

At this point I haven't done any welding on the frame. I've plasma torched some bicycles apart, and made some test welds. I need to measure tube thicknesses, calculate some stresses, and decide if I have all the tubing I need (I think I'm set). All the pieces came from cheap bikes, mostly huffys, which I wouldn’t think would be chrome moly. All the grinding sparks look the same. I don’t have gas for the TIG, but I stick welded some pieces together cold, and was satisfied with all but the thinnest. I think with preheat, I will have no prob. Will experiment a bit more before I start.

Depending on the shock mount holes used (I was planning on 3 possible positions at each end), I will have about 4 to over 6" of travel in the rear. I have a nice, dual air, shock.

The front is a cheap motorcycle front end with about 5" of linear travel, for about 4-1/2" vertical. Someday I might replace it, but I'm probably more likely to just start over again when that day comes. Before the question comes up, I'm using about 22# of MC, including fenders. Heavy, but strong. Will clamp/bolt/epoxy an adapter to the MC forks for the bicycle wheel and brake caliper.
I had a head failure on a 10 speed 30 odd years ago, tire contacted down tube, stopped turning (relative to the frame). Bob did one of his 1/4 second Superman impressions. Landed chest first and cracked a couple ribs. I’m willing to spend some weight to see that that front tire stays where it’s supposed to!
I should point out that I have another bike with another MC frontend welded to it, that I intend to test ride to decide how I like the head angle, rake, and trail. (It only recently occurred to me that this could be done with pedal power.)

The axle clamps will be built up to at least 5/8" wide, and the clamp bolts will be 5/16, grade 8, fine, double nutted, and installed with purple lock tight. Of course I will also use the axle nuts. No slots, just set the axle in the V (if the bikes upside down), and assemble.

The frame got a bit complicated, but all the extra tubes don't add much weight, and, well, it's not like I'm paying anybody to do the welding. The main design criteria that caused all this was the desire to put almost any size battery (large box) low and to the front (and keep it protected). I basically decided to add a couple light tubes that aren't really part of the structure. (The rear controller could be mounted in the compartment above the battery box, if I decided I needed a bigger battery, but I honestly don’t believe I will.)

Deep breath, pause, I want an ebike that's immersible. If you ride through water, you will eventually hit an unseen obstacle and fall down. I, on the other hand, sometimes ride into water that's (much) deeper than I expected (when an ICE's intake goes under, it stops, did you know that!?). This means extreme care with the wiring, waterproofing all boards, etc. (See electronics) The battery box would be 'sealed' (foam rubber), but not waterproof. Probably good enough, but I would like the battery/BMS in a sealed bag....

Will use the MC fenders front and rear, maybe cut down some.

Besides the (short throw) pedals, I will have 'pegs' about 1/3 of the way back on the swing arm, and footpads to either side of the battery incorporated into the air/rain/mud deflector (lower fairing?)(not shown), that's part of the frame/battery box for different seating positions, or a passenger. Makes it look even more like a scooter (like I care), but a nice place for the feet when riding through water, or worse.

It WILL have an upper fairing of sorts and a small wind shield before winter.... (That assumes I get the bike ‘done’ before winter.)

Built with a pair of motors from Methods, a 6x10 in the rear with a 3" tire, for a max od of 22.5", and a 5x12 in front using a 25.5" od 26" bicycle tire (for now).

12 and 18 fet controllers are overkill, but if I end up bedding them, I will want to use them gently.

For handlebars, I plan to weld neck tubes from old bicycle front forks to a piece of tubing clamped in the triple tree handlebar mount, then insert bicycle handlebar squibs(?) in them for a strong, highly adjustable handlebar 'system'. Hope with the drawing this is understandable.
Of course standard handlebars will work.


About the drawing:

Just a side view, sorry. I hope most of you can guess the front and top views and can mentally extrapolate the 3D. The only places there are double tubes have light crosshatching (forks, rear suspension, handlebars).

The tubes are drawn as simple rectangles (in visio), so ignore the joints and square ends.

The rear hub and spokes are drawn as an X5 in a 16" MC wheel, I'm using a 9C in a 17" MC wheel but didn’t take the time to redraw it.

The front wheel is drawn as a small (2-1/2") 21" motorcycle wheel/tire.

There are 2 seats drawn to indicate the 2 seat tubes (not a bad place for the passenger if she's short enough though), and behind them is a sketch of the 'banana' seat I have in mind (at it's lowest).

I intend to have 3 mounting positions for each end of the shock, not 2 shocks like the drawing appears to have. :)

1/2" heim joints for the rear swing arm bearing?

Haven't gotten around to drawing the brake caliper mounts yet.

The ‘line hatched’ tubes attached to the neck are from the MC.

When you look at the frame design, realize that the battery box doesn't add much to the strength, but very little weight either.

The orange/tan colored area would be kevlar composite, with the battery easily accessible (with a key) on the left, for r&r. Probably with a 10ah 20s or 22s built from 5ah20c turnigy (allowing for 1" padding/insulation all around except 1/2" on the sides), but the box is long enough for 2 separate 6ah 30c (would probably have a cubby hole/tool box, accessed from the right side in the unused part of that area, instead).

The green lines at the bottom would be the ground with the suspension fully extended or collapsed (max travel in the rear), which, with the double air shock in back, should be almost never.


This is a sketch of how I plan to get air from the fans in the swing arm, into the motor. How round I get the Kevlar funnel in the brake disk will largely determine the efficiency. I don't figure I need much air flow, but when I need it is when I'm going slow or stopped. There will be screens over the fans, incorporated into the phase/Hall wire protection.

View attachment 1

I intend to ventilate both motors, as well as forcing air through the rear.

As to moisture ‘getting in’, if you have warm air enclosed in a chamber, like these motors, and you cool them (by riding through water?), they will form a vacuum and suck in air SOMEWHERE, if only through or around the bearings, and if there’s moisture, it will suck in with the air. WITHOUT ventilation, there’s no way for the moisture to get out. My 2 cents on that subject.

My worry is sand wearing the plating off the magnets. Guess I’ll see. Any opinions? I guess I could thicken the plating on the magnets. It’s nickel, right?

Thanks in advance for any criticism and advice. (I consider a critical post, if logical, of more value than praise.) :wink: :)

Bob
 
I suppose the first questions, or requests for comment here would be:

What do you think of the idea of bedding the controllers and/or making the bike immersible?

What do you think of my serial interface for the controllers idea?

To make it water proof, all boards will have to be sealed, including the hall filter boards in the new 9Cs.

Low voltage wiring connections will have silicone dielectric at the least. HV connectors will have the same with good physical separation.
Hi V, Hi current will be heavy soldered ring terminals, with round headed bolts into nylocks, with silicone dielectric, heat shrink (clean), 2 layers of good tape, properly applied, more heat shrink (I don't expect to have to disconnect to change a tire). :)

8ga batt and ground leads, separate wire and 50a andersons for each controller, with 30a andersons for precharge. 10ga phase wire (silicone, of course). (If something’s going to generate heat, let it be the windings where I have no control.)

Standard Xie-Chang controllers (I would like to get some of the new bare boards), except I'm thinking, seriously, of designing my own serial interface board for it. The only chassis wiring would then be high current, switched HV (the key switch), the serial wire, and either 12v (if available) or HV to the head.
(Winter project.)

Could use a variety of different 28 pin PICs.
Before you start a processor discussion, I have them, and am familiar with them. I also have PicBasicPro, which while ‘basic’, has commands for the USB, 1 wire, and hardware PWM (as examples), and will be more than adequate, and I can insert ASM if necessary...).

The board would have serial I/O to the head unit (at the handlebars).
Would have outputs to the controller for throttle (PWM-RC), x3 (rev), x1 (or2), EBS(brake), TB (antitheft?), other?. Has anyone gotten the lock (TB) to work on these boards. I could simulate it well enough to keep someone from rolling or pedaling it away by simply using reverse in moderation.
Inputs for the shunt (buffered (duh)), batt v, all 3 phases (I want max resolution, and minimum time delay here, as I want to use this signal for traction control). Could monitor the 12v (mosfet drive) line?
A small mosfet, PWM modulated to drive fans.
That probably leaves me with I/O for 10 single drop temp sensors, analogue or 1-wire, or with a tiny bit more SW complication, multidrop.

The same board (with a different prog and processor....) could be used at the head.
Would use a PIC 18f2550, so with a USB pigtail I could have, you guessed it, USB.
Outputs for an LCD (I want a serial or 4/8 bit interface LED matrix, but..).
Inputs for brakes, 2 3way, 2 throttles, 2 Variable brake sensor (VBS?) signals (would be used for variable regen), lights on/off, high/low beam, turn signals, ??? (a couple pins left, enough for a ‘virtual’ CA interface).
Would (hopefully, eventually) provide most of the functions of a CA, as well as serving as the programming interface (set 3way ‘speeds’, rev speed, etc, as well as what functions are enabled, and what switches control them (remapping).

Would be so nice to be able to control things, such as the 3 way. To go from lo to hi you would HAVE to go to med first, and if you change up without letting off the throttle, it could be gradual. (slightly?) (Programmable.)

And, of course, the same board could be used (the PWM through the fan control mosfet), to control voice coil operated brake calipers (enabling full ASB, and allowing smooth transition from regen to disk). :D
<edit>Also need tilt/tip, so we can incorporate AFB (AntiFlipBrakes). Would also enable automated wheelie control. 3/4 throttle just lifts the front tire, 7/8 rides back with no acceleration, and full decelerates by tipping way back and applying brakes (think how a unicycle controls 'it's' speed). You wouldn't want to tell anyone there was a computer involved :wink: :) <endedit>

The hardware and basic (as in simple/primitive) SW functions would be quite easy to implement. (If that doesn't start an argument, I give up.)

OK, , a smaller version could be embedded in the hub, to send hall and temp info to a dedicated input on the controller serial adapter (with a 4bit word for the halls, should be fast enough). The 3 hall inputs (on my board), connected to the X-C-controller, would then become outputs. The next step would be to give the adapter inputs to detect phase EMS zero crossing, for automatic sensorless conversion(?). More SW would incorporate timing adjustments, at least within the limitations of the poor old PIC.

All from wanting to eliminate the maze of wires.!?


Waterproofing?:

If I bed the controllers to make them water proof, they become disposable (pretty much not worth trying to repair), so I want them overrated/underused. Tested at 105v, then used at 80, (or 90 if I want to push it) peak. Try and use them at 1/2 the tested current.
Numerous temp sensors and thermal current limiting/shutdown.

For the FETs, I'm thinking of soldering them (the tabs), in groups of three, to pieces of .090" copper, 1.2"w x 1.5"t, then epoxying the copper heat spreaders to the aluminum heat sink (the epoxy would provide the electrical insulation). My intuition tells me it can take the thermal stress, but it's complicated, esp in the sub arctic. One or two vertical cuts in the HS, (under the copper heat spreader) would solve the problem (I think), if there is one.

I would then brace the board to the heat sink (kevlar braid/epoxy), and let it hang in a cavity in my 'frame cover’(?). (Would have to machine a groove in the heat sink for the board.) The controllers would then mount by holes in the (overlength) heatsinks. Eliminating the case allows me to stand the caps upright on top, and add a bunch of small diameter (.395) caps on the bottom.



The blue would be heat tolerant, cold flexible, ‘sealant’ (suggestions). Probably used on the CPU, caps, or anywhere else I don’t want epoxy getting under. The thing above the fet is a to92 temp sensor. Only the added caps on the bottom are shown (not necessarily in exactly the correct places).

I want to include one of these stereo/alarms:
http://cgi.ebay.com/300556645966
Would mount the speakers in a homemade fairing.
I could jump off and go in a store without worrying so much about someone walking or pedaling it away.
Not much of a stereo, I’m sure (not the 20Wrms they advertize for sure), and I don’t think it COMES with a simple audio input, but….
Bear.
If you’re walking, ‘they’ recommend you make noise, but even if you don’t try, a bear will hear you coming and hide. You won’t even know you walked a few meters from one. An ICE motorcycle or ATV makes plenty of noise, but I worry that I will pop over a hill on the ebike and SURPRISE her….

The power for 12v (stereo, headlight, LED lights, etc.), would(?) come from a small 4s2p battery pack made up from dead IBM notebook batteries. (Can anyone tell me for sure what ‘lithium ion’ technology they were using in the T20 family?) I’m thinking of building my own 4 cell BMS, ‘for the experience’ (providing charge from the main 50-90v battery). Even starting with old batteries, I would hope I could get nearly 3ah from 2p with 2ah cells, if I select the best from what I have. This battery would be left on the bike when the main battery (LiPo) is taken inside for charging, so if I can build it from free cells, I won’t feel so bad about leaving it out in the cold.
I think I would have to limit the charge v based on temp, because of possible extreme temp fluctuation (I’ve seen it go from -40 to +40F in less than 12hrs). How would I determine how much current I would have to pull off, to maintain a cells voltage, as it warms up. Extreme case, bringing it indoors from the cold with a near full charge.

<edit> After thinking of the above, I think a 3s 2ah 20c LiPo (for less than $10) would be lots more logical. If I can't find a commercial charger I can modify, I might still build my own BMS, processor controlled, switched, current and voltage limited charge regulator (from the main battery), and simple td431-transistor-resistor balancing, with cell v monitored by the processor. Possibly an independent analogue circuit to force charge shutdown on hi balance current (in case of main regulator failure).<end edit>

Fuses: It’s my plan to use an inline ANL fuse holder as part of the battery pack, and separate bulkhead mounted ANL fuse holders for each controller. I’m thinking 150a, 100a, and 60a, (just to save the wiring and motor windings when things short).

Enough for now. Will post more later, (and probably edit mistakes).
Thanks for reading my ramblings,
Bob
 
I’m in bush Alaska. I’ve never SEEN an ebike that I didn’t build!!! (At least that I know of.)

Where I live, 400mi west of Anchorage, it doesn’t get nearly as cold as some places farther north. I’ve only just seen it a little below -40, but the record is -52F. The bike, and almost certainly the battery, will (probably) never see -40, but where possible it should be designed for it.

I like to follow snogo trails, summer and winter. This involves crossing water, muck bogs, and other obstacles. Since the trails tend to form tunnels through the trees, you have to continue, (or turn around, and that would be admitting defeat!). In winter, trails are easy to follow because they’re always packed down, except of course when it gets warm, or there’s overflow from the river or springs, in which case you have to face slush, or find another way around. The things that slow me down in winter are tiny drifts on the ice or packed snow, that vibrate or bounce at high speed, and ‘whoopties’ made when snogos accelerate spinning (d’ KIDS). Also, when I refer to ‘trails’ you have to realize I don’t mean groomed trails like in civilized places, just places where people have driven repeatedly.

Thanks to permafrost buckling, even the roads are often much better suited to ‘off road’ vehicles than ‘road’ vehicles. Suburbans for taxi cabs, 4WD SUVs for police cars. Ex police cars from down below are popular cabs, but you see them along the road with wheels broken off all the time (ball joints and axles).

Bethel is isolated from the road system. The only way in, is boat in summer, snogo in winter, or airplane year round. Very few cars or trucks ever leave town, except on the river in winter.

There are only about 25mi of roads total, paved, gravel, and a few dirt (sand). The longest possible, on road, point to point ride would be about 8mi.

We’re on the Kuskoquim river delta, and have lots of VERY fine sand. ICEs tend to wear out twice as fast as elsewhere because of sand leaking past/through filters (usually the frame breaks and the doors fall off before the engine wears out though).

Back to my bike, I want it light enough so loading it in a boat is no problem (the battery will be easily removable). If you’ve been to AK, that probably says a lot about it’s intended use. I ALSO want to be able to load multiple hundreds of pounds over it saddlebag style and walk/run it up a hill, Ho Chi Min Trail like.
 
Back
Top