Oatnet's 'Kepler Friction Drive' build

[EVTodd]

Anyone like to offer any feed back on this "sync problem ? .. 63-54 , 250kv motor , 5 cell lipo.
I have tried re-programming the ESC (Turnigy KF 100) with different combo's of :-
... timing, 5 - 18 deg,
... start normal , soft etc,
...PWM frequency 12 - 24 khz

I still get what i assume is "sync failure" at med speed under load ..( screeching sound and motor stops with throttle still on)
Any suggestions for a optimum working set up.. Timing, PWM freq' etc ?

Make sure you don't have any kind of an issue with something binding or actually stopping the motor. I ran into a problem just like that today with my gf's bike. I didn't space her bearing cover out quite far enough and the roller shaft dug itself into the aluminum. I'm surprised she didn't pop a controller. She must have been pulling a huge amount of amps when it happened.

I thought maybe she threw a magnet in her motor (even though that would surprise me with that small amount of power). I tried turning the motor and it was very tight. Then I felt the motor mount and it was VERY hot. Easy fix though.

Anyway, if the motor is stopping for no other apparent reason look for mechanical issues too.

[Hillhater]

Thanks Todd, but it is definitely the "sync" problem ....... a suddend "screech" as it accelerates through low speeds, then stops until i throttle back and start again.
( FYI:-actually the rattle was a mechanical issue with the motor mount initiated by the sync problem ...but that is fixed)
I really have to baby this up through the revs to avoid the issue, and this is under little load ( bike is on a stand, free running)

On the bright side... at least i have been forced to figure out how to program the ESC using only a servo tester !!

So , come on guys, one of you must know a solution to the sync fault on these Turnigy components ?
(please dont say Hall sensors ,... i think that may be a step beyond me currently ! )

PS, Sorry Oatnet for hijacking your thread, but i am working on a very similar friction drive with similar motor etc, so i felt it was relevant.

[Kepler]

I am using that speed controller with no sync issues at all but I do remember haveing sync problems with that motor and an older Turnigy ESC. It worked best with the highest timing setting from memory. I would keep the PWM frequency as high as it goes and use medium softstart.

[Hillhater]

Thanks for the suggestion Kepler.
OK an interesting discovery .
Reprogrammed the ESC with 24khz PWM, 22 deg timing, and V soft start,.
Seems to work ok, but i can still "force a sync drop out occasionally if i slow and accelerate hard.
Now here is is an interesting factoid ... it seems that once the motor loses sync, it also causes the ESC to default back to factory settings.
At least i know the start mode defaults back ( its hard to tell what the other settings are actually doing)
It stays in "default" mode until i power down and when re-powered it comes back on my "personalised " settings ( Start mode at least since that is the only parameter i can easily see)
Is this typical for these programmable ESC's. ? or am i missing something ( i have rechecked afew times and the same thing happens)
If so, beware of what sync drop outs can do to your ESC settings !

[spinningmagnets]

MrBoots, thanks for the EV-Logix "Throttelizer" suggestion. I found the link, but the web-site doesn't seem to have purchase info. Who did you buy yours from?

http://www.evlogix.com/content.php?link=Rcthrottle

[Kepler]

Thanks for the suggestion Kepler.
OK an interesting discovery .
Reprogrammed the ESC with 24khz PWM, 22 deg timing, and V soft start,.
Seems to work ok, but i can still "force a sync drop out occasionally if i slow and accelerate hard.
Now here is is an interesting factoid ... it seems that once the motor loses sync, it also causes the ESC to default back to factory settings.
At least i know the start mode defaults back ( its hard to tell what the other settings are actually doing)
It stays in "default" mode until i power down and when re-powered it comes back on my "personalised " settings ( Start mode at least since that is the only parameter i can easily see)
Is this typical for these programmable ESC's. ? or am i missing something ( i have rechecked afew times and the same thing happens)
If so, beware of what sync drop outs can do to your ESC settings !

I have never heard of this issue of going back to factory defults after a sync loss event. Even stranger that cycling power returns it to the personalised settings again.

I hate programing via stick mode (or servo tester mode) Is so hit and miss. It als puts a big strain on the ESC in relation to heat build up. I always use programing cards so I know exactly what the settings are. For the sake of a few dollars, well worth the investment.

I use this one with the K-Force 100A. http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=7187

[Hillhater]

Yes it is difficult to "see" the actual settings without the card ( or USB connection?) but i saw the reviews on the card , and backed away !
but the "Start mode" is very visible, and that is obviously defaulting back so i assume other settings may be also.
It certainly explains why some settings seem to help for a while, but after a "sync event " they no longer appear effective !

Is there likely to be any significant "electrical difference between this 63-54 , 250 KV motor and the 63-74 200KV motor's that you and others are using , that may make it more prone to sync failure ?? ... or should i suspect the ESC to be "sync failure" sensitive and maybe try for a warranty exchange ?
Q ?:- ...Do HK really have an Aussie warehouse now ?

[Kepler]

I always thought it was more an ESC issue then a motor issue when came to sync problems. If you were in Melbourne, I would lend you a motor and ESC to isolate the problem.

I dont think HK have an AU warehouse yet but looking forward to them establishing one.

[Hillhater]

I dont think HK have an AU warehouse yet but looking forward to them establishing one.

Their site homepage has an "AUS Whouse" location , with a note saying "Opening this month "

[amberwolf]

Maybe that doesnt' stand for WareHouse.

[Hillhater]

Well, it actually does say .."Aus Warehouse".. (i am a lazy keyboard jockey) , but i have to admit they dont specify what "This month" actually is !

.....but there was an item in their News section a few weeks back that said the Australian Warehouse would start taking order from September 3 rd !

Edit:
From their home page..

HobbyKing.com® International Hobby Store.
HobbyKing has 4 warehouse locations. Hong Kong, China, Germany, Australia and USA.

[oatnet]

Please forgive me if I go over the basics of what I have figured out, for the RC bretheren I assume this newbie stuff, but it might be useful data for other experieinced hubbies trying to learn the nuances of RC setups.

Here is the servo tester specified in Kepler's sale thread.

1_0661.JPG (48.24 KiB) Viewed 943 times

The Cover is a strip of folded plastic. The dial pulls off and the cover comes right off.

The 5 spots that need to be desoldered

3_0667.JPG (48.41 KiB) Viewed 871 times

Desoldered. Each took a quick blast of heat and application of a solder sucker.

4_0669.JPG (45.39 KiB) Viewed 871 times

Other side of the board

5_0670.JPG (46.66 KiB) Viewed 945 times

Wired to a standard xlyte throttle connector - I wire my Pot throttles to the same connector. I am not sure yet whether I have + and - to the correct terminals.

6_0672.JPG (25.12 KiB) Viewed 945 times

xlyte hall throttle wiring - I use the same format for POT throttles

And a diagram of a POT just for reference.

pot.JPG (7.46 KiB) Viewed 864 times

Here is how I THINK the servo connector is supposed to be hooked up, with brown as ground, red as positive, and orange as signal. Note this is the unmolested servo tester, glad I got a spare!

7_0683.JPG (47.61 KiB) Viewed 943 times

Here is the ESC as it comes from the box. On the lower side, notice that the plastic cover sticks up. I originally planned to cut this down, but I discovered that by removing the labels, the plastic ends slide right off the wires.

a_0677.JPG (60.11 KiB) Viewed 945 times

Here it is with the covers off, loose in the middel of the photo. I secured the caps to the battery leads with a wire-tie so they wouldn't be bouuncing around on rough roads. I also put some electrical tape around the exposed end to avoid shorting. The ESC seems quite solidly glued in its housing.

b_0678.JPG (69.38 KiB) Viewed 943 times

ESC in the Kepler drive, with cables routed out the back slot

c_0679.JPG (52.98 KiB) Viewed 945 times

Hooked up for testing with the servo tester, note the yellow LED on the left side of the green ESC body. Glad I tested this way, or it would be running in reverse. After this test I swapped red and white phase wires. I am going to cut the motor cables short and solder the ESC right to the motor cables because it is subject to movement, vibration, and elements. It is easy enough to unssolder if I want to change the motor.

d_0682.JPG (68.26 KiB) Viewed 943 times

[oatnet]

I built up a little connector (the 4-pin CD audio connector for a PC works nicely despite the extra pin) and battery per Recumpence's excellent tutorial:
viewtopic.php?f=28&t=12196

However, The Phoenix ICE 100 comes with a 5a BEC, so I didn't need an external battery or BEC to run the servo tester. The 5v wire from the ESC is red, so I expected ground to be black, but the other wires were brown and orange. I guessed brown as ground and that seems to have worked.

I used the battery on the servo circuit when I first hooked everything up, I got some beeps, the motor jerked a few times and stopped. I couldn't get any more signs or sounds of life from it. When reading the instructions, I realized it came with a BEC, and I thought maybe I had fried it. I opened up the top of the kepler to observe the red and orange LEDs (It's still alive!) I later discovered that it only beeps when the motor is connected, and my taped test connects had come loose.

I went back to basics. I bought (2) servo connectors, so although I had modified the other one, this one I could use out of the box. I connected the ESC wire to the 'battery' position, nothing happened. Connected it to the 'servo 1' position, and voila, rotation! However, the motor was turning the wrong way, so I swapped the red/white motor leads, and then shot the demo below. I see I could have changed rotation in the software, but I don't want to have to ride home backwards if it loses its presets.


Next I need to look into the USB programmer, maybe the one with my methods controller will work, I have to lookup the pinouts. I think I'd like to set the timing to low (documentation says cooler motor/controller) and the start to very soft, does that make sense?

I was waiting for the Kepler interface, but I got a sense of the urgents and bought an evlogix throttlizer today. I got the 'low'current 120amp version, with dual sensors for the ESC and the motor.

-JD

[oatnet]

Wow, 100 views on this thread since yesterday!

I didn't get a chance to post these pics last night, dressed the phase wires and got a wheel spinning in one test. Didn't have much luck on other tests in any of the 4 seatpost clamp positions, but I only gave each one a few trys. I'll get more serious when the throttlizer arrives, and I get the recommended grip tape on the motor, and I should try cutting the fiber washers like Kepler recommended. BTW reportedly the throttlizer has a 'shrinkwrap' version that will fit in the 80mm x 40mm x 35mm accessory chamber of the Kepler drive. My PM here went unanswered but I got contact at http://www.evlogix.com/ evlogix@yahoo.com

-JD

finished_0686.JPG (58.27 KiB) Viewed 864 times

finished 0692.JPG (68.59 KiB) Viewed 792 times

[Kepler]

I wouldnt cut the damper yet. The pickup looks good. You will find once the grip tape is added, it will pickup every time. Good news on the Throttleiser being shrink wraped. I took my throttleizer out of its case to see if it would fit but no luck. Perhaps is a modified board. Looking forward to seing how the throttleizer matches the drive. Should work well.

My interface is coming along nicely but still a month away I would say.

[Nanoha]

However, The Phoenix ICE 100 comes with a 5a BEC, so I didn't need an external battery or BEC to run the servo tester. The 5v wire from the ESC is red, so I expected ground to be black, but the other wires were brown and orange. I guessed brown as ground and that seems to have worked.

-JD

I thought you were supposed to disable the BEC if you're running more that 3 cells lipo?
http://www.castlecreations.com/support/ ... _Guide.pdf

[Kepler]

Incorrect. The Phoenix range use a linear BEC which should not be used with more then 3S.. The Ice range use switching a BEC which works fine right uo to 8S so all is good

[oatnet]

Nanoha, on first read of the instructions, I came away with the same impression about the BEC. Then I started differentiating the models better, and as Kepler pointed out, the ICE model doesn't have the same restriction.

Kepler, sorry I couldn't hold out longer for your interface, but maybe on my next Order.

I discovered that Krylon makes a no-skid spraypaint, looked for it in my local Ace Hardware, and eventually found it next to the no-skid tape, instead of the paint. It was clear, instead of the black I wanted, and it required primer, and sanding... The paint was $5.99, wheras the no-slip tape was $1.29/foot for the heavy duty version. I went for a foot of the tape to save $$$ and hassle.

When I mounted the tape, I slightly overlapped it, such that in a forward rotation, the trailing edge lays over the leading edge, pressing them together. The tape was easy to mount, wrapping it all the way around the motor to keep it aligned, before pulling the backing to expose the adhesive, and cutting it to length. As Kepler said, it makes the difference (thanks for the heads up!).

I noticed that after carefully lining up a gap between the motor and tire, then bolting the seatpost down, the post and motor lift away from the tire, as the seatpost clamp contracts. Now I rest the motor on the tire, and when I tighten the clamp it lifts away to a good height.

I have a few drive tests below. The first two are with the seatpost clamp in the rearmost holes, and the last is with the top bolt in the frontmost hole to angle it down. The latter setting seems about right. If I get it just right, it seems to go on smoothly, otherwise it bounces around disturbingly. I assume this is from using a servo-tester as a test-throttle.

Later I mounted an ammo-can battery pack to the front (16ah24v a123 prisimatic, 7.45lbs, 20c discharge) and trimmed up the handlebars, cycleanalyst and throttelizer and it will be ready to ride. I guess I'll have to take the motor apart to mount the throttleizer temp sensor, can anyone point me to a guide on how to do this? Or recommend what high-temp epoxy I should use to mount it?

Make sure to click on the pics to see them full-size!

-JD

The tape is half on. This end of the tape will overlaps on top of the start just a little bit. When the motor rotates against the wheel, it will press the end into and over the start, keeing it tight. I hope.

NoSlip__0711.JPG (56.76 KiB) Viewed 725 times

Here is the drive in place with the tape fully mounted

Mounted_0717.JPG (111.21 KiB) Viewed 725 times

A shiny view

shiny_0725.JPG (64.39 KiB) Viewed 725 times

A matte view

matte_0727.JPG (58.84 KiB) Viewed 727 times

A lousy job of showing the hair of a gap between the wheel and the tire

gap_0718.JPG (88.22 KiB) Viewed 727 times

[Hillhater]

I notice in all those tests that the motor jumps and stops a few times, and you have to throttle off and restart to get it going again.
This seems to be the same "sync" problem that i have encountered under load.
Have you tried simulating a load on the drive yet ? (IE: brake ?) .. i would be interested to hear the result
Also have you calculated you roadspeed at max throttle? ... seems impressive

[SanFranRider]

. . . got a wheel spinning in one test.

I watched/listened to your YouTube videos testing the motor. Does it really sound like an angry dentist's drill on steroids that's being amplified by Metallica's sound system? What I'm saying, it sounded like this setup wouldn't be very stealthy nor conducive to a nice quiet ride -- or did your recording setup magnify a barely audible 30dB hum?

If the KFDS is really as noisy as it seemed on the YT videos, perhaps all the other benefits of the KFDS are moot and I'll be "forced" to go with a BMC hub motor.

[LI-ghtcycle]

I can't say for sure since I haven't had my 63-74 up and running just yet, but my E-fight 110 wasn't loud at all, so you can get very stealthy motors, and I imagine set-ups that are 5S with the smaller motor (43 MM?) wouldn't be as loud.

[Hillhater]

Well the Tgy 63-54 on 5S that i have rigged is also a screamer.
I am suspecting it is an "electronic" sound ( non mechanical) that may vary depending on the ESC settings such as timing etc.
PS:- I found it surprising that to program the ESC they use the motor as a "Speaker" for the programming tones

[Hillhater]

Hey , Oatnet, have you got your bike on the road yet ??
We need some feedback on the motor performance and if you are seeing the "sync loss" problem ?

[oatnet]

Sorry for the Delay hillhater, I have been hoping that the servo tester was the source of the sync problems, so I have been waiting for a throttleizer before street testing. I paid $95 for it on the 12th with a promise of a 3 day turnaround. It shipped on Wed the 16th but then it wasn't, then it shipped again on Saturday the 18th, and got here on Monday the 20th, so count on 8 days or more for delivery, longer if you are not as close as I am. I felt slightly raped by paying $25 for a proprietary build of $5 USB programming interface, but oh well.

These days Sunday is my only shot at free time, so I was all excited to wrap this project up, and get it on the road and off my mind.

That excitement quickly faded. I am really, really, really disappointed in the throttleizer's build quality. The power leads were different lengths, and there is no indication of which side is batt and which side is ESC. The temp sensor leads are short enough to be useless, so I have to clip them and extend them. The cable for the rarely-used usb interface is long enough to lasso cattle; after tuning that will haunt me endlessly unless I make this a rodeo bike.

The shrinkwrap covering was way too big to fit in the space, so it had to come off, and that is when the bad stuff really became apparent. I cut the pins for the LED, since one leg was poking into the casing of a white wire. The metal backs of the two fets had been pushed into contact, and the legs were not well completely seated before soldering, so a lot of current will go through the thin part of the leg. One of the black power leads had a split casing, like it was cut with wire strippers, then cut again longer. That power lead also had lots of stray strands floating about. While looking at the loose strands, I realized that it also had a cold solder joint that had cracked - if I put 100a through that lead I'd get a nice flash and wonder if it was something I had done.

I soldered on a standard throttle connector (I'd have like to see it come with one installed since 99% of us use the 4-pin xlyte style), but now I am frustrated by needing to do a lot of repair work to get this up to snuff (or at least what I have detected so far), so I am bagging on it for today. I probably won't get a response on which set of power leads to solder to the ESC today, so that means I'd have to wait until NEXT Sunday to have the time to attempt an install it again. Maybe I should just wait for the Kepler interface, as I was really pleased with the build quality on the Kepler Drive.

Hey Kepler have you considered building the swingarm with mounting points for hall sensors? A simple bolt-on with screw holes for the big and little motors, and voila. I think eliminating the sync issue with hall sensors would eliminate the slow-start issue (might even be able to start from stopped), and if I could pump watts in as volts instead of amps with a nice reliable sensored ebike controller, I could eliminate a lot of resistance heat. I might whip up a bolt-on on the veritical mill in my Saturday machining class if I can get time and material.

-JD

Here it is as delivered, in a lumpy, uneven shrinkwarp.

shrinkwrapped_0794.JPG (54.61 KiB) Viewed 758 times

With the Kepler - it looks like if you remove the shrinkwrap, it might fit the dimensions I supplied. Note the gunk trail on the Kepler drive where the tape was.

with kepler_0795.JPG (74.86 KiB) Viewed 762 times

Hmm....

bottom_0806.JPG (47.9 KiB) Viewed 762 times

I should have taken a picture before I trimmed the legs to the LED. The one in back was poking into this white wire. The tops of those two FETS were in contact with each other.

spikes_0808.JPG (48.94 KiB) Viewed 758 times

Measure once, cut twice. Notice the strands frayed about.

splitCasing_0804.JPG (52.09 KiB) Viewed 762 times

Here are more strands - I wonder if that one pointed at the leg of a capacitor is close enough to arc.

stray wires_0809.JPG (34.04 KiB) Viewed 758 times

Notice the crack in the solder epoxy on this cold-solder joint.

cracked_0806.JPG (51.14 KiB) Viewed 762 times

[Ypedal]

Holy Shit !~!!!!!!!!