Fusin 1000W peak Geared Hubmotor kit w/ Cassette Splines/FW

[amberwolf]

Only got a little of the testing done today (too many non-bike things happened that I hadn't planned for).

Using the Sorenson DCS-55-55 lab PSU, I tested first the CA's calibration for current and voltage, and it appears at lower currents (just under 5A) to be close enough to my Fluke 77-III DMM that I would trust it. I got interrrupted so didn't get the chance to setup a higher 40-50A load like I'd planned, just a single car headlight. It also matches everything else I have close enough, though the Turnigy Watt Meters do read high (which is about what I remember).

CA vs Fluke vs Sorenson meters for current:

DSC06726.JPG (49.29 KiB) Viewed 1040 times

Simulataneous current readings, seriesing all the meters. In order from PSU to load: WU, Fluke, TWM1, TWM2, CA (note slight voltage drop on each succedding unit, with 16.0V at Sorenson source):

DSC06727.JPG (59.51 KiB) Viewed 1040 times

Voltage readings, no load, Sorenson, WU, Fluke, CA.

DSC06728.JPG (54.42 KiB) Viewed 1040 times

So I think the CA's readings can be trusted just fine, vs those of the Fusin controller and "analyst", whcih I don't have a way to check as directly as this right now. (I have a test I could do, running a known current from the Sorenson thru the shunt while the controller is on but not running the motor, simultaneously verifying the current with the Fluke, the Sorenson's meter, and the CA, but I have to wheel the bike in there again to do it, and I think open up the controller so I can access each side of the shunt).

Then I can see what wattage the Fusin "analyst" display shows, and see if that matches with the calculated wattage I should get based on the known current times the actual battery voltage on the Fusin at that moment.



Next up was some oscilloscope probing of the controller-to-display lines. It has 5 lines, the Pack Voltage In from keyswitch, Pack Voltage Out to the controller enable line, Ground, Rx from controller, and Tx to controller, for serial data transfer of displayable information and button presses for selections.

Scope is set to 1V/cm and 1ms/cm on all of the pics and vids below, with 0v at bottom of screen, except the dual-trace vid of both data lines, which is 2V/cm and 1ms/cm, with 0v for Rx on bottom of screen and 0v for Tx at center of screen.

Rx line is lower voltage (~3V p-p)

DSC06732.JPG (34.74 KiB) Viewed 1040 times

than the Tx line (about 5Vp-p),

DSC06751.JPG (37.46 KiB) Viewed 1040 times

even at the connector at the controller end. But it is also cleaner, with virtually no noise at the "resting voltage" of 3V, while the Tx line has a saw waveform (presumably ripple in the internal SMPS of the "analyst" display) overlaid on it's "resting voltage" of 5V.

Another noise is overlaid on that saw wave when the backlight is turned on. Neither of them is a very "deep" noise, meaning they aren't likely low enough even at minimum to trick the receiver in the controller into thinking any of the noise pulses are bits in teh data stream.

DSC06753.JPG (38.17 KiB) Viewed 1040 times

However, I suppose it is possible that the only-3Vp-p Rx line (where data is going from controller into display) isn't high enough voltage to reliably be read by the display unit. I still have to open the display unit and check the signal at the PCB inside it, to make sure it isn't degraded by the time it reaches there).

Strings of data are sent only once every second or so, maybe a bit quicker than that (you can see their timing in the video below), whcih explains the too-infrequent displayed-information updates. Means that the MCU in the controller is doing all its' usual work plus sending out these data pulses, so maybe it isn't fast enough to update more often than that (but I sure wish it did).



Since there's no way for the display unit to get any of the power or speed data directly, the controller must be sending it in the data stream. So when those two things fail to update, either the controller is not sending it, or the display is not receiving it.

I dont know if the display directly detects battery level and displays that (as it does have it on the keyswitch input), or if it is detected by the controller and displayed there. Simple enough to test, by disconnecting the analyst from the controller, and just inputting battery and ground to it's connector, and seeing if the meter responds. We'll see next time I get the chance to do testing.


The wheel sensor is definitely always working, even when the display fails to update or just plain reads wrong. Tested with teh scope on the controller-motor connector at the controller end. Two vids here, one continuation of the other. Haven't gotten to hooking the motor's internal speed sensor to the CA yet, as experiment to see if that would work. (it should).

[amberwolf]

Ride today at power level 7. No new issues, same problems with the display. No time yet to troubleshoot the problem more. Was hoping to today...


Motor and controller are still doing great, with only the typical sensorless startup problems that appear to be in the nature of sensorless hubs. As a side note, I have occasionally heard a whirr instead of a grindy sound, so I suspect that sometimes it spins up backwards just for an instant, before the controller figures that out and reverses it. Probably happens on any geared sensorless hub, but this is the first one I have had.


I apparently forgot to note down some info , so I left *** on those.
Ride to work:
***m s trip time
2.185miles
***mph max
***mph avg

21.6Wh/mile
0.834Ah
47.04Wh
24.14Amax

58.3Vstart
56.8Vrest
54.8Vmin


Ride home
***m s trip time
2.353miles
***mph max
14.6mph avg

20.7Wh/mile
0.845Ah
46.81Wh
23.98Amax

57Vstart
55.8Vrest
53.7Vmin

At least today only a few people tried to run me over. Considering the incredible amount of traffic (about 4-5x normal for a Saturday at this time), I guess that's pretty good. :/

[amberwolf]

Been too busy with other things to get data from the last few days' rides posted up till now.

First, some observations before I forget again:

--The PAS power level 1-9 on the Fusin display *only* seems to affect the PAS portion of hte system, not the throttle. If you use the throttle you have the full range of power regardless of what setting that PAS level is on, *except* if you set it to 0, where it turns off both PAS and throttle. I see no effective difference in max watts or max amps regardless of PAS setting when using the throttle, in any of the commutes so far.

However, if only using PAS to control the motor, that setting most definitely makes a difference. AFAICT, it is essentially telling the PAS what "throttle" level you want when pedalling. If you set it to 1, it's kinda like having a little tailwind, and you hardly notice it on level ground when it kicks in (assuming you have the bike in the right gear for the speed you're starting at--if you forgot to downshift when you stopped, you'll feel it for sure). Set to 9, it's like suddenly getting WOT once it kicks in, and is quite disconcerting if you're not expecting it (I keep forgetting this thing has PAS, as I never had taht before).

--PAS is great, for the most part, but it needs much finer control, and not just an on/off sensing ability, and it needs to have zero effective delay between stopping pedalling and stopping motor power--it takes almost a second to stop the motor now, and if you are not yet used to this delay, and planning for it, you could put yourself in harm's way.

The finer control needed is to basically sense actual torque on the pedal system, and apply power to the motor based on that. This would totally negate even having to have the power level setting in the Fusin display at all, as it could just follow your legs--pedal hard and it gives you full assist, spin your feet around like pushing a feather and you get minimum assist (regardless of speed at which you are pedalling).

But this will require a different kind of sensor than the standard PAS it comes with, and probably reprogramming the controller itself to work with such a thing, and that's gonna have to be done by Fusin (if they care to do so--it might not be cost-effective for them).



So here's the last three work commutes, including today. Each trip was at successifly higher PAS power level setting, though I am using the throttle and not PAS to ride with. (basically am not pedalling unless absolutely needed, which is very rare). Unfortunately I don't remember the wind conditions/etc. now, so I'm not sure why there is such a discrepancy between some legs of some trips regarding power levels vs otherwise identical conditions. I am also unsure why the trip distance is arounda tenth of a mile or more different on some of the days, when I am riding the exact same path each time. All data is from the LS-CA-SA, rather than the still-unreliable Fusin display.


06-04-12 To Work (PAS level
8m 47s trip time
2.332miles
21.1mph max
15.9mph avg

25.9Wh/mile
1.078Ah
60.276Wh
28.18Amax

58.3Vstart
56.7Vrest
54.7Vmin

06-04-12 Back Home (PAS level
9m 34s trip time
2.352miles
20.5mph max
14.7mph avg

20.3Wh/mile
0.863Ah
47.706Wh
23.19Amax

56.8Vstart
55.8Vrest
53.7Vmin

06-05-12 To Work (PAS level 9)
8m 45s trip time
2.197miles
21.6mph max
15mph avg

23.1Wh/mile
0.899Ah
50.714Wh
25.02Amax

58.3Vstart
56.8Vrest
54.9Vmin

06-05-12 Back Home (PAS level 9) (note that this trip I was carrying a 30lb bag of dog food on the front rack, as well as a 10lb water pump/motor, in addition to the usual stuff. Because of the weight on the front it was hard to steer so I rode slower by a couple MPH and I took a couple of short detours around higher traffic areas).
10m 59s trip time
2.348miles
22mph max
12.8mph avg

24.7Wh/mile
1.015Ah
57.957Wh
24.16Amax

56.9Vstart
55.7Vrest
53.5Vmin


06-07-12 To Work (PAS level 9):
8m 26s trip time
2.186miles
21.4mph max
15.3mph avg

24.6Wh/mile
0.957Ah
53.760Wh
24.27Amax

58.3Vstart
56.7Vrest
54.0Vmin

06-07-12 Back Home (PAS level 9) (note there were two major accidents blocking my normal route back home, so ride was longer, with more stupid people who were also trying to find ways around the blockages). (note: I forgot to reset the CA after noting down the to-work stuff above, so the back-home data is a continuation of that).
21m 36s trip time
3.216miles
23.5mph max
14.8mph avg

22.4Wh/mile
2.091Ah
116.16Wh
24.5Amax

56.9Vstart
55.5Vrest
53Vmin


As usual, the motor and controller are operating very well and reliably, and quietly.

Only the Fusin display "analyst" is not useful at this time, and I suspect it is actually a communication fault between it and the controller causing the problems. I have yet to have time to get back to the troubleshooting.

[amberwolf]

6-9-12, still operating ok, other than the display issues. In this hot weather (100F to 105F+), the motor is still only warm, checked at various times during and after the ride. Temperature seems to peak about 4-5 minutes after I stop at the end of the ride, based solely on feel. (want to install a thermal sensor once I have an excuse to open up the motor). Controller gets warm enough to notice with my knee resting against it, but that's all.

Using just PAS power level 1, still no pedalling.
work
8m 21s trip time
2.197miles
22.8mph max
15.7mph avg

24Wh/mile
0.937Ah
52.764Wh
24.58Amax

58.3Vstart
56.8Vrest
54.6Vmin

home
10m 0s trip time
2.352miles
20.9mph max
14.1mph avg

24.1Wh/mile
1.028Ah
56.07Wh
23.46Amax

57Vstart
55.7Vrest
53.3Vmin





6-10-12, still working ok.

Using just PAS power level 1, still no pedalling. Was a lot of wind on the way to work.

work
8m 49s trip time
2.2miles
20.6mph max
14.9mph avg

26.3Wh/mile
1.027Ah
57.343Wh
24.29Amax

58.3Vstart
56.7Vrest
54.4Vmin


home
10m 4s trip time
2.366miles
20.3mph max
14.1mph avg

21.3Wh/mile
0.915Ah
50.398Wh
23.8Amax

56.9Vstart
55.6Vrest
53.4Vmin


68 miles total so far on this motor and bike, so maybe 13% of the way to 500 miles (target for basic review).

[Herrsprocket]

I appreciate all the great feedback on the performance levels and the variety of testing you are putting this kit through, but,,, I'm so very anxious to see what is on the inside of the motor by you to having a look at the guts of this thing. Ya know, phase wire size, the quality of the inner rings and the planets (materials, width of the planet gears), stuff like that.

[amberwolf]

That will be coming soon-ish. I have to have time to make the tool to do so, and the time to fix anythign I screw up in opening it or taking it off the bike or putting it back on, before needing it for work again. Right now DayGlo Avenger and CrazyBike2 are in experimental testing "modes" right now, and I wouldn't yet trust either one for my work commute--they *should* work, but if they didn't, I'd be walking the bike and me the rest of the way to work in 100F+ heat. So at least one of the bikes has to be in "working mode" before I can experiment with another one.

Phase wire size is probably 14g at best, probably 16g or less, given the thinness of the whole power/sensor cable from the motor and the size of the pins it has.


Speaking of data collection, I didn't get any for yesterday's ride, as I forgot to reset the CA before I left, and since I charge thru it (monitored as regen) the Ah used started as negative, and other stuff would've required math to figure out, so I just discarded the whole day's ride data set. Was very windy on the wya to work, so I expect if I'd done it right instead, Wh/mile wouldve been up near 26 or more.

[amberwolf]

I again forgot to reset the CA, so no ride data from today.



Just a note to myself for another set of experiments to perform after the review phase, as described in another thread where I was thinking about it:

Typically, the gearmotor's freewheel makes sensorless operation a problem. They will run sensorless, but starting them is a problem unless the freewheel clutch is modified to locked, non freewheeling.

Fusin's new kit is sensorless, and it uses a geared freewheeling motor. Seems to work fine, with the caveat that just like any other sensorless motor, it sometimes gets a slightly "off" startup event and you ahve to let the throttle off and retry, which works 99% of the time. Maybe once I had to retry twice, in the time I've been testing the kit so far.
viewtopic.php?f=4&t=39877
Note that because it is a freewheeling motor, then this problem happens even if coasting at high speeds, nto just from a stop. If it were nonfreewheeling, it would only happen from a stop.

I haven't had the time to try yet, but my intent is to try their new sensorless controller on their old motors, too, which are also freewheeling geared motors. (well, two of the three I have here are; I'll also want to try it with the non-freewheeling geared one). I expect the startup issues will be identical with the two freewheeling ones, and that only with the nonfreewheeling one will the problem go away and only then if I am already moving.


(theoretically, one nice thing about a sensorless controller and a non-freewheeling motor is that you have instant full-power reverse cababilty, simply by pushing the bike backwards with your feet first and then engaging the throttle--I will test that theory with the non-fw fusin).

[amberwolf]

Well, I tested over a dozen tries of rolling the bike backwards and then engaging throttle, and it always starts the motor forwards, even when it misfires, so I guess that's taht for "reverse" feature--just wishful thinking.


Today I had to pickup some scrapped bits of a dsplay from work, so I took the Bell flatbed trailer using this bike. Handles fine with trailer on there, almost as if it wasnt' there, even when loaded with about 30lbs of stuff (about what the trailer weighs, I guess, so 60lbs-ish extra load on the bike).

I forgot again to reset the CA for the trip there, empty, but I did get the data fro the tirp back. Was very windy from various directions dpending on where I was in the trip, so that makes it really hard to know if the Wh/Ah usage is higher from that or from the trailer pulling.

Same riding as normal, only difference was the trailer. Which I forgot to take pics of before unloading and disconnecting it. I did ride a little slower in some parts due to unusual amounts of traffic, so overall speed average is lower.


10m 13s trip time
2.336miles
20.2mph max
13.7mph avg

27.2Wh/mile
1.160Ah
64.183Wh
24.93Amax

57.0Vstart
55.4Vrest
53.1Vmin

[amberwolf]

No work ride data today; has been about the same each ride and nothing different this time, except the wheel has been feeling a little odd for a while and I decided to see why when I got home. The spokes were even looser than when I got the wheel (since I did not re-true or retension the wheel, wanting to see if these 14G spokes or the typical rim would cause problems because of it--and yes, they did). When I started turning each nipple 1/4 turn at a time to retension the wheel, I found one of the nipples wouldn't tension at all and was stripped out, probably from the wheel not being tensioned enough, letting the nipple turn and eventually rattle/rub threads/etc., but could've been that way from the factory it was made at.

Fortunately I had some 14G nipples and spokes in my junk box, probably from the original 36V Fusin wheel I got from Dogman way back, which I bent the rim on in a skid/crash on DayGlo Avenger, and eventually unlaced that Fusin from the rim and put it in a 24" for CrazyBike2, leaving it's spokes and nipples available for future spares for situations like this one.

Anyhow, since the nipple had to be replaced, I'd have to take the tire at least loose, so I figured I'd also change over to the Infinity Armadillo-type tires from Ohzee here:
viewtopic.php?f=2&t=12500&p=595972#p595972
and test them out on this bike in place of the torn-sidewall tire I've been using on the rear,

DSC06863.JPG (38.07 KiB) Viewed 893 times

and the not-too-grippy semi-knobbyish old Landrider tire on the front (which is actually the whole front wheel from the Nishik-E).

So I pulled the whole rear wheel off to do that. While I was at it, I pulled the cassette off and looked around at my tubing/pipes to see if any of them were the right diameter to make that tool needed to remove the cover of this Fusin motor, and unfortunatley I don't have anything large enough that isn't far too large. I'll have to think of a way to do it some other way. I also noted that I will likely have to take the freewheel mechanism off first, as it may be in the way of the cover coming up (if it is attached to the axle and not the cover; I don't know; shoudl be the cover but I can't see in there without taking one or the other off). It, too, would require a special tool to take it out, so hopefully it's part of the cover. No more time to spend on that part today.

So no peeking inside the motor yet.


I started to just replace the stripped nipple...then I thought, if I'm going to true this up and replace the nipple to make a better wheel, why don't I just go ahead and relace the whole friggin' wheel, since it's laced offset by one (small space over valve instead of big) and it's laced wrong (no over-under interleaving of spokes). Then it might stand a chance of surviving these potholey roads without bending up as easily. I'm pretty sure that if I didn't have the minimal rear suspension, the rim would already be trashed because of the wheel being so loose and laced wrong when I got it. As it was, it started out being out of round, possibly with a flat spot already on the rim (hard to tell when I originally checked it, and nearly impossible to see in the video).

Proceeded to unlace the wheel, and mark it to relace correctly. Got it all off and then about 1/4 laced (first 9 on the disc side when Hachi decided I'd been at it too long (couple of hours at this point) and wanted attention. So I gestured her over and bent forward to scratch her chin, and almost shiskebobed my eye on a spoke.

That really really hurts.

I got lucky and missed the eye and only got the corner of the lower eyelid, but my left eye still keeps tearing up and sorta feels like I have a splinter in it, though there' snothing there I can see, with a mirror and a light. I tried to take a pic but I can't really seem to hold the camera, mirror, and light well enough to get a focus on it. Maybe later when I'm not so tired and shaky.


Eventually I went back to the wheel, and finished lacing and mostly truing it. Espeicaly happy that the valve hole is now in the open space.

DSC06861.JPG (48.67 KiB) Viewed 893 times

Can't fix the out-of round problem without screwing up side-to-side too much, and can't fix the side-to-side completely. Guess the rim's already bent too much, or I am just not able to concentrate well enough to figure it out.

Installed the tube and rim and then aired it up, and I got about 40-45PSI when suddenly it started hissing like the airhose came off the stem, but it was unfortunatley the actual stem that was leaking. Again.

DSC06860.JPG (22.83 KiB) Viewed 893 times

I'm realy really sick and tired of blown up ruptured herniated violated crappily made valve stems on tubes. I mean, I can patch most any place on a tube except there, and be sure it will hold at least for a while. But a stem....

Anyway, that's where i finally gave up, after 5 or 6 hours of trying to do all that plus holding my eye while being drowned in doggie slobber as Hachi and Nana tried to make it all better.

DSC06862.JPG (69.41 KiB) Viewed 893 times

I gotta take a nap soon cuz I am really really tired, and then I gotta go up to work and pick up two 35lb bags of dog food that I couldn't take home last night, not having the trailer or anyplace on the Fusin test bike to haul them on. No way they'll work on the front baskets; I alrady tried that with a single smaller bag a few days ago, and that was not much fun getting home. So this bike won't be taking them home; I'll have to use CrazyBIke2 to do it. Well, it's built for that, so at least I know it'll work. At least, as well as it always has. I hope.

More on this bike later, when I get time.

[amberwolf]

I changed the tube out (didnt' have any more good ones, so had to patch an old worn one, unfortunatley a thin one), and finished more micro-adjustments to the side-to-side truing after prying at the spokes to relieve tension, so now the wheel should be within 1mm total side-to-side, though the rim is still out of round by at least a couple of mm or more in spots.

I was going to try to patch the stem on the other tube and use heatshrink over the patch to hold it on, but I can't find the HS I know I have that's sized right for that (I just used some a month ro two ago on a tube from CB2!). So for now I just cut a slice out of a totally-split-open tube kept for "parts", and used rubber cement on it and on the valve stem to glue it on there good. HOpefully I will find the HS and then I'll put it on htere, and can reuse this otherwise-nice thick tube (though I will have to enlarge the valve hole on whatever rim I use it with).


So back to commuting with this tomorrow, as long as the tire still holds air by the time I need to leave.

[amberwolf]

Well, I guess my relacing is *more* out of round than it was before, or at least, more in a single spot. It feels like I'm riding on a widely-spaced washboard as the wheel rotates and goes up and down. I sat down to try to fix it earlier, and didn't even get the spoke tool out before I dozed off sitting there. Now my almost-shiskebab'd left eye (and right elbow, which I banged so hard on something last night that I couldn't even feel my hand for several minutes) is throbbing again, so I can't really concentrate on it (or much of anythign else...been a challenge working today).


I *think* the ride was better on the new tire, but hard to tell till I can fix the out-of-roundness. It did brake better, as side-to-side is much truer than before, so pads could be set closer and more correctly parallel, and the wheel felt less flexy especially in turns, as it is tensioned much better (though possibly still not "enough").

Overall it's more success than failure, so far, but I *have* to fix the out-of-roundness as soon as I have another couple of hours (at least) to fiddle with it, in case I screw it up worse somehow.

I will have to post the ride data later as I don't rmember where I set the paper I wrote it on.

[amberwolf]

I think this is the ride data (hard to read my writing for hte date):
to work
8m 46s trip time
2.193miles
21.5mph max
15mph avg

26.1Wh/mile
1.018Ah
57.078Wh
26.3Amax

58.0Vstart
56.5Vrest
54.3Vmin

to home
10m 6s trip time
2.355miles
21.6mph max
13.4mph avg

23.9Wh/mile
0.981Ah
53.811Wh
24.34Amax

56.7Vstart
55.5Vrest
53.2Vmin

SInce then I've done more work commutes, including one yesterday hauling the trailer again, full of 130lbs of dog food and groceries (plus the 30-ish lbs of trailer itself) on the way home. Typically Wh/mile seems to be around 25-26 for the commutes, with a little wind like we've been having. When there isnt' any wind, it's more like 22-23Wh/mile. With the trailer, even at about 3MPH less top speed most of the way, and mabye 2MPH less average speed, it was still almost 30Wh/mile. (I wrote the stats down for thesse trips on my work schedule but I dont' know where I put it--maybe left it at work).


Motor got more than a little warm by the time I got home with the food, running it at 700-900W pretty frequently for the start/stop/acceleration of traffic, but it wasn't really hot, despite the 95F+ night air.

Motor and controller still working very well, but the Fusin analyst display comm problem is still causing real issues. The main danger is that when I am riding along and suddenly the speed display goes up to some ridiculous number (45MPH is the fastest I have seen so far, when I am not even going 20MPH), the display disables the controller because it's faster than it's limiter is set to. If I'm in traffic, this means a sudden loss of power *and* I have to stop paying attention to the road long enough to power cycle the controller so it will be reset.

Also when I have to stop for some reason, usually a traffic light or stop sign, the same thing can happen.

I"m going to disconnect the speed sensor wire from the wheel, to see if that will allow it to still operate and use the analyst display/control for it's power-setting ability and whatnot, while completely disabling the speed limitation stuff so it can't leave me worse than powerless in traffic.

[amberwolf]

Today I was rebuilding the rear wheel again, because a nipple sheared off it's top.

I've never seen that happen before, even when I had an already very-high-tension wheel get pulled apart from a derailed chain on my old version of CrazyBike2--it almost ripped the nipples thru the steel rim, but it didn't break them.


But this wasn't just one of them. First, just the one, that broke I guess while it was sitting there overnight, as it wasn't broken when I charged the bike last night. I saw it today when I was putting the charger away, and rechecking the battery balance, taking the ammocan pack off the bike to see what the cells each read (it's balanced still very well, within 0.01V even though I only bulk charge and don't do any balancing). The spoke was just hanging there, and I wondered, what the heck?! and saw this:

DSC06867.JPG (25.56 KiB) Viewed 748 times

I examined all the nipples, and saw no others with problems. Then I squeezed pairs of spokes, and no problems at first, but the second round of that caused another to break! It was very loud, sort of a clank as the tension was released.

DSC06869.JPG (28.89 KiB) Viewed 748 times

So I took the wheel off, the tire and tube and rim strip off the wheel, and replaced those nipples with some off of 14g spokes that Ianmcnally had sent me, originaly bought for a Crystalyte motor but I guess never used.

Started retensioning everything, because now the wheel was really messed up for side-to-side and roundness. Got almost perfect roundness, within 1mm highest to lowest, and then CLANK! another nipple let go. I guess I had spent 2-3 hours on it at this point, at least, off and on between canus interruptus slobberus sessions.

DSC06871.JPG (47.78 KiB) Viewed 748 times

I was really not happy by now. I took a wrench and put it between spoke pairs and started twisting to both relieve tension and to break any more weak nipples before I have to ride the thing.

Nothing else broke. So I replaced the third one, and retensioned everything again, and spent at least another hour or two or more retruing it AGAIN. I had the thought several times that I shoudl probably just pull ALL the Fusin-supplied nipples off and replace them with the ones from Ianmcnally, but somehow I decided I'd be adventurous, and call it part of the experiment to see how long the rest of them last now that they're acutally part of a properly-tensioned wheel.

Note that the angles aren't even all that extreme:

DSC06870.JPG (25.02 KiB) Viewed 748 times

Now the wheel is almost normal, but it does have a slightly flatter area due to damage from not being laced or tensioned properly from the factory. I can live with it; it's nto that noticeable during a ride even at 20MPH--the roads themselves are much worse than the wheel is, at this point.


So there's one more area Fusin needs to improve on--QC checking the parts used to lace their wheels--I have at least 3 out of 36 nipples that had some sort of fracture fault in the brass, from whatever cause, that couldn't handle even the normal tension of a decently-laced wheel and the weight of a pretty normal ebike and rider. (I ahve placed no undue loads on the motor wheel itself; all cargo has been on the trailer or on my regular front bicycle wheel off the Nishik-E, which is performing perfectly fine as expected, an the motor wheel is on a suspension swingarm).

[amberwolf]

I have been pondering again.


I think Fusin needs to redesign the controller and display so that the display does not simply cut off the controller as soon as it's speed limit is reached, but instead feathers back the throttle, so that it keeps power applied but only enough to maintain the speed at or below the limit. Perhaps it could begin feathering back as it approaches that speed limit, but I myself wouldn't want that, as it means slower acceleration which could mean the same kind of traffic danger as simply having the power cut off completely can cause.

Right now, as best as I can determine, the cotnroller is reading the speed and the current, and sending that data serially to the display unit in bursts around a second or two apart.

When the communication is interrupted for whatever reason, the display fails to update. Sometimes, it seems to lock up because of this, and cease to acquire or respond to any new data. Sometimes, it locks up for a short time (several seconds to well over a minute) and then respond again, only instead of updating with the new data, it *adds* the new data to the old data--I'm not sure how this could happen, but it's the only explanation I have for the absurd speed readings it sometimes gets, such as 45MPH when I'm not even going 20, or sometimes I'll see 25MPH when I hadn't even reached 10mPH yet at a startup, but the system cut out because it thinks it exceeded the speed limit set in it.

I still need to do some more work on the communications problem itself, but have not had time when I was physically and mentally capable of it. I usually only get one of the three at a time these days, sometimes two, rarely three.

[amberwolf]

I unplugged the speed sensor at the input to the controller, which also disconnects it from the display analyst thingy. Everything still works as expected, with of course no speed display on there. Figured that'd take care of the problem of it disabling the controller due to false overspeed indications.

Nope. Well, yes it did stop that, but that isnt apparently the only thing that causes it--I just don't know what else could.


Unlike previous rides, where at least a couple of times on each leg of the trip it would fail to start or to cut out, due to false overspeed, this time it only cut out once, but naturally it was when I was about to make a right turn with a bunch of cars behind me (some of which were also turning. I didnt' have a place to pull off out of the way so I just stood on the pedals and cranked to get going a bit (I keep it in a lower gear, middle in front and largest in back, so I can actually help it start going a little when this happens). then keyswitch power cycled it, pressed the mode button to turn it on, and throttled up.

Unfortunately that means I don't know what it said on the display. I wish I had been able to just stay where I was and look at hte display to see if there was some other error condition, or if perhaps it is still getting a false speed signal somehow, causing the problem to continue. The only other thing I'm aware of that could shut off the unit is the ebrake signal, but it definitely wasn't that, or it would have not worked after power cycling it either--if the ebrake was stuck on, it would have errored out until it was released and re-power cycled (I've done that before, when I was holidng hte brake as I turned it on).


Anyway, so my next solution is that I'll re-hook up the speed wire, but I'm going to just take that loop of wire that passes the keyswitch enable line up to the analyst, and connect it to the enable pad of the controller in place of the wire from the analyst. Then I can keep it all on there and use it if I want, including hte power level changing buttons for PAS, but will never have to worry about the display analyst thingy shutting the controller off.


Thankfully I have tomorrow off, and so can probably do this tonight if I wake up in time from the involuntary nap I can feel coming. Otherwise, maybe tomorrow before I head out on a shopping trip to American Foods in Mesa, where they have lots of good but cheap food (mostly restaurant and other large-size packages; cans of various things the size of my head for $1.50-$3, for instance).

[amberwolf]

Just for giggles I tried to test an MXUS DD motor with this controller, since it's sensorless, but it didnt' work.
http://www.endless-sphere.com/forums/vi ... 84#p600184
explains why I think it didn't, but basically BEMF from the MXUS at it's lower speeds is probably insufficient for the hall-emulator PCB in the Fusin controller to create a usable hall signal from.

[amberwolf]

On my wya to work today, the display cut power to the controller multiple times, which was the last straw.

I was able to finally look at it a few of those times, with little traffic on the roads, and each time it was flashing the battery level meter on "empty", even though it is fully charged (or nearly so, not even 0.1Ah out of the 10Ah pack used up the first time it did this, and not even 1Ah used the last time just before pulling into work).


So at lunch instead of eating I cut the wire on the display connector that feeds the power back to the controller's enable line, and connected the controller side of that wire directly to the power line to the display on that connector, so that anytime the keyswitch on the headlight assembly is on, the controller is on. Display can now do whatever malfunction it likes and it won't shut the cotnroler off.


I reconnected th speedo sensor wire to the controlelr/analyst, so that I will get the speed display there, too, if i want it. Not much use isnce the display doesn't update it all the time, and it is such a slow update anyway.


I have scheduled a vacation for a week starting July 8th, so hopefully I will have time to catch up on a lot of projects, including troubleshooting the lockups that I think are likely comm error issues. Only thing I have to do before doing projects is move my bedroom from the room I'm currently in to one on the opposite end of the house (as it is easier to cool that room), which is at least a whole day's project, maybe two, shuffling all the stuff around including the internet cabling. Gonna confuse the dogs for a while, I bet.

[amberwolf]

I forgot to post last night (thought I did, maybe I typed it up and didn't save it?) that the motor and controller system worked perfectly after the above mod, no cutouts, however the BIKE sucked.

I suspect someone messed with it, because I had the following problems found when I was leaving work, none of which were a problem when I got there:
--front wheel QR lever was undone and unscrewed several turns.
--front left brake pad was UPSIDE DOWN and loose.
--brake switch on rear was loose, flopping around.
--rear derailer was severely misadjusted, would not stay on a gear.
--turn signal flasher wouldn't engage correctly because the little holder I have it in was crushed.
--headlight wire was pulled off the switch, and the switch itself broken, body pulled off the bracket I'd mounted it on.
--front tire extremely low
--front shock completely airless.
--something else but I can't remember right now.

All were relatively easy to fix or to work around, except the shock, as I can't hand pump it up--the pumphead won't fit down into the well to go over the valve stem. I have to use the electric pump at home for that. The shock does lose a little air each day but I had just aired it up the day before, and when i aired it up after getting home it did not lose any more than normal thru today. So someone probably let the air out deliberately.

Same with the tire, except I could hand pump taht up fine.

Headlight swithc easly bypassed by twisting the wires together. QR fork tightned, brake pad reinstalled correctly and aligned/tightened. Rear Brake switch tightened up. Uncrumpled the plastic holder for the blinker, and it works ok, though not as well as before. I think the bimetallic sheets of the blinker itself are slightly bent or misaligned.

Have to replace the headlight swithc, maybe tonight or tomorrow. I didn't have the camera with me to get pics before fixing everything so I could ride home, so I only have a pic taken tonight of the broken switch.

DSC06872.JPG (40.69 KiB) Viewed 852 times

What bothers me is that I don't know anyone I work with that dislikes me at all, much less enough to try to kill me (by loosening the front wheel and the brake pad, which could've jammed in the wheel and locked it, with the QR loose).

So that means it was probably a customer, and I don't know any of them that dislike me that much either. Or just some random person that likes to screw around with cyclists or ebikers. But it's in the store, in the warehouse in back. I think if it were out front where everyone can see it it might've not happened, but in the back room there's no one to see, most of the time.



Tonight I put the motro thru it's most severe test so far--the same one that overheated the old 36V version 1 Fusin: riding slowly, 8-12MPH, heavy load on the bike, frequent stops/starts. Had about 72lbs of dog food on the bike--6 on the rear rack on top of the ammocan pack, and the rest on and in the front baskets.

Was not really a smart thing to do, as I could not actually "ride" the bike normallly--any kind of turn was nearly disastrous, and I dropped the bike twice when I lost my grip on the bars. I just stepped off so didn't get hurt, but both times I had to take the stuff off the bike, put it upright against a pole and tie the brake handles down to keep it from rolling, then relaod everything and restrap it down. But I couldn't wait on the food--it was a "get it now or never" super deal, and is why i really like having the ability to haul such cargo on CrazyBike2. I just never know when this kind of thing will come up.

I had to do what I never lke doing, and rode the sidewalks almost all teh way home--only after I crossed Dunlap was I on roads, and that's almost no traffic (one car passed me). 1/8 mile of parking lots, and the rest sidewalk.


Made it home ok, no damage to cargo or bike, other than more scratches and little dings from the fall-overs. Oh, and the CA's button connector somehow got knocked loose in one of hte falls, so I had to open it up and reseat it (not off all the way, just enough to mae the buttons not work).

[GrayKard]

Damn, sorry to hear that someone vandalized your bike while at work. Need a security camera back there in the storeroom.

Great review, keep it coming!

Gary

[amberwolf]

Thanks--we do have one...just not at the end I have to park the bike at. When parked out front, it's right in the view of two cameras--but it's all the people going by since it's right next to the doors that do more to keep people from messing with it. But the higher-ups that manage multiple stores want it parked inside the store nowadays (used to be they didn't want it taking up space inside).

Anyway, it just teaches me to be vigilant about checking things before a ride.

[amberwolf]

I put the bike thru another heavy cargo load, though only about 60lbs of dog food this time. Small bags in the baskets, one small bag on the rear rack on top of the pack, and the largest bag on the front baskets over the top of everything.

This time I had it balanced well enough, with enough of the weight lower on the bike to make it rideable at normal speeds on the road, though corners I still slowed way down for, so I woulnd't have to lean into them much, just in case.


The one recurring problem with this kind of load is bottoming out the front shock, and not having enough pressure to rise back up unless I standover the bike and pull the bars upward to lift against the cargo weight--very hard for me to do. But it happens at just about any 1" or greater bump or hole. I don't really have a fix for it, but then again I hadn't intended this bike to do this kind of work...just that like always, when I find a good deal on stuff I need, I have to take it and haul it with whatever I'm riding at the time, somehow.

Still, the Fusin motor and controller are doing a great job even like this. Only 25Wh/mile, even with that load at normal 20MPH cruising speeds and 15-16MPH average speeds. Controller gets pretty warm at slower to middling speeds with these loads in this heat (>100F after dusk), and the motor starts cool but gets fairly warm (more than bodyheat warm) by the end of the ride, and continued to warm up after I got home, till it was hot, but not uncomfortably so, about 10 minutes or so later.


I still want to make a tool to open the motor up, and see what it's like in there. I have had several ponderings on it; I mostly just need time to actually cut pieces and file them, weld them together, and test out the idea. Dunno if that will be any time soon.

I have another idea but it'd require someone to CNC something out of aluminum block or something. It'd be like a freewheel removal tool, with a nut on one end, hollow to fit over the axle, and the castellations on the other end to fit into the Fusin cover.

[amberwolf]

Motor and controller still working fine, after disconnecting the Fusin analyst display's ability to shtdown the controller. Still get wierd stuff o n the display, but at least it doesnt' affect anything.

Nearing 200 miles on the kit now. No further spoke or nipple problems yet.

[amberwolf]

pulled a trailer full of dogfood home today (big clearance sale due to rearrangement of foods at work, discontinuing some and getting new ones). about 170lbs plus trailer weight, still performing fine for motor/controller. I put the taillight from the kit on the trailer, held down with one of the straps holding hte food to the trailer. Some of the bags were damaged so I put them in the big styrofoam containers we get fish in, and tied everything else around that.

Uneventful trip, just the kind I like.

[dnmun]

whatta load!

the reason the wheel was loose may be that someone was attempting to steal the wheel and never got it off. around here they steal everything off of a bike left out. wheels and saddles and racks. i see bike carcasses locked to poles all the time.

[amberwolf]

possible, but they'd have to have been in the store's back room doing it, while we're open and people were around. If I'd had it outside i might understand--theft is not uncommon around here.