mounting hall sensors to R/C brushless motors

[liveforphysics]

Tried it out. The position of the sensors is critical. 1/4" retarded, and it will try to start in either direction, then buck around and not run. 1/4" too advanced and even in no-load it trips the breaker on the 26amp 84v power supply. The motor also gets very warm very fast, even with no load.

On battery power, even at 84v rather than 42v, it feels weaker than the RC controllers, and the top speed is LESS ??? Feel like it hits a limiter or something.

The current shunt is not modified, so it should only be pulling whatever it's set to pull.

The throttle response is the deal breaker. It's a HAZARD. I don't even have a chain on my pedals, so I rely 100% on the motor, and this is not safe in traffic or trying any sort of spirited/technical riding. 1-2 second might not seem like a lot on paper, but when you are trying to negotiate traffic, it feels like an hour, and it's not safe.

Overall, I'm pretty pissed. If this "soft start" can't be solved, I'm going to throw out my 12fet and 18fet boards. What a bunch of work to make a soggy feeling terrible to ride safety hazard. The last RC controllers at half the voltage felt 100000times better for the 10mins before they exploded.

 

[PaulM]

Would this work with your Kelly controller Luke? I seem to remember you mentioning one of those early in the 18 hp thread.

 

[liveforphysics]

I have a giant Kelly, and I would gladly buy another of the same, except they can't reach the full RPM potential of the motors. The RC controllers can reach the full RPM, and maybe this infinion can too with some settings changes and/or bridgeing the shunt. However, unless the lag can be fixed, it's just not safe to use in traffic, or off road. Anytime you lift off the throttle, and go to re-apply, you feel this light gentle push that does almost nothing for about 1-2seconds (just long enough to get you hit by a car), then it accelelerates with full insto-wheelie torque.

Since the throttle response is delayed, you naturally move your wrist further than you need to, then it hits and wants to stand on it's tail, you let off, then it's back to waiting to go again. Freaking retarded and dangerous. The throttle of a machine MUST, absolutely MUST respond with the humans input, and directly with that input. A throttle should NEVER increase the torque output on it's own, well after you made the throttle input. That sort of crap causes you to launch out into the street when you only wanted to creep forward, or be sitting helpless when you need to get out of the way of something quickly.

Completely recklessly unacceptable. I'm going to mail them all back to China in a box with fresh dog-doo if the giant gaping critical flaw can't be corrected on them. I don't care about the money, but I have a day of my time into modifiying every part of them to be as durable as possible, and a full weekend of getting sensors mounted and whatever to make them "run". I'm pissed.

 

[methods]

Lol....

Yea... I am just going to dump the rest of the boards I have at cost.
I am willing to bet that he did not burn the JTAG interface and we could -COULD- re-flash the chips but the chance of Keywin emailing me the code is about as likely as him transferring $1k into my paypal account for no reason other than my good looks.

He reads his email and I have sent him 3 without any response which means he does not want to talk about it.

Luke - The original controllers had no "soft start". A request was made by a member here and that is how it came to be... I am sure Keywin can reverse that request but it wont help us. Maybe we can pick up a couple 18 fet boards in a few weeks when they get it sorted out.

Your work is not a waste as I am sure there are plenty of people around here looking to buy a 4110 modded controller.
Just dont blow it up!

And btw: That thing is current limited and your RC is not. Dont even try to compare apples and oranges :? That thing has a 40A current limit

If anyone is looking for fresh 18 fet boards with no fets and no wires PM me. Still in the original packaging, not even opened yet. They have the variable LVC and the good 100V caps. I have the heat sinks, heat sink compound, screws, insulators, cases, gaskets, covers, programmers, and a bag of chips.

Everything you need to program.

You supply the wires, you supply the fets.
If they dont sell I will go ahead and finish modding them with 4110's and wires and then I know they will sell. . . But I am going to charge for my time and I make $60/hour at work so.....

I got them very cheap.... Way cheaper then the ebay price so you will get them cheap as well. I dont want to post the price.

-methods

 

[swbluto]

Well, a modded controller can put out a much higher current(What was the last figure I saw - 80 amps?) and at 100V, that's a nice 8kW. I think that's pretty comparable to what you could reasonably expect from an RC controller without blowing it up(50 volts*120 amps = 6 kW).

And, uggg... I can understand your pain. The only thing from making your particular controller viable is its code! A simple stream of 1s and 0s. Maybe you could get someone else's original controller, do a memory dump and then reflash your chip... somehow. Otherwise, you could possibly ask keywin(Who?) for the compiled hex for the original version which seems less liable to IP theft(and possibly more likely to garner a response).

Actually, hmmmm, I'm wondering if anyone might be able to do a memory dump on their original not-soft-start microcontroller and just post it to this thread or something?(After searching google for copying microcontroller information, it appears I might be a little too optimistic. It doesn't seem like microcontrollers offer a "dump memory" feature, usually.)

 

[GGoodrum]

Patrick, I hope you don't get stuck with these. There's got to be a way to fix this without having to send them all back to China. It would be nice if it turns out they used hardware to implement this, but I'm guessing it's hard-coded, which sucks. If it can be sorted out somehow, I'd love to take 2-3 of these off your hands.

What I'm having trouble understnding is why hard code it? What was he thinking? It could just have easily been made a programmable option. Incredibly short-sighted, in my opinion. :x

-- Gary

 

[swbluto]

http://endless-sphere.com/forums/viewtopic.php?f=2&t=5713&start=135&st=0&sk=t&sd=a

It appears knuckles had a hold of some programming software at one time. If the most updated form of that software allows editing of the "soft start" parameters, than it seems like it might be changeable. He does say "flash" and I'm under the impression you can reflash flash memory, but he also does quote "flash" in the sense of doubting the word's technical accuracy, so it might be a 1-time "hard code" type of deal.

(If that's not what you meant by "hard code", then I don't understand what's so hard about it. I'd normally think hardware would be "hard" in a sense, but it does seem like it could be more easily changeable than software now that I'm more aware of electronics.)

 

[GGoodrum]

This is the PC program that is used to change the programable settings. Near as I can tell, there's no setting for anything related to soft start. "Hard-coded" means that the function is part of the main program, and as such, can't be changed.

-- Gary

 

[recumpence]

Wow, this is all so frustrating!

It seems like every turn we take on controllers, we hit some funky limitation. That is the reason I went to dual motors and dual HV110s. But, that is a complicated mess and too expensive for most people and using motorcycle motors and controllers is way too heavy and also expensive.

It will be nice to figure this all out.

Aren't there other E-bike controllers that can be used with the halls that do not have this freaky delay issue?

Matt

P.S. Methods, good to see you again. How's it going?

 

[Paul.............D]

Hi to everyone new to the forum well been sat quietly watching for a long while like a few years Ha Ha. i have had a few of these infinion controllers before the chinese used to use the cap and resistor for 1 to 2 second delay on throttle ramp as it was cheaper to impliment i hope this helps . luke the adc kelly controller will work on your demon as it can see 25000 motor pulses, hall sensors need to be 120 degrees around the stator as their is to much flux leakage for infinions found this out while using an alternator as a motor PLEASE SEE LINK CAP AND RESISTOR http://www.faqs.org/patents/app/20080218236

Paul

 

[swbluto]

the cap and resistor for 1 to 2 second delay on throttle ramp as it was cheaper to impliment i hope this helps .

Is that true? Then it seems it should be as easily fixable by swapping a component or jumping a component.

 

[Jeremy Harris]

No, it's not true on this controller.

The throttle delay is not as a consequence of any passive components,it's coded into the controller firmware. The only passive components on the throttle connection are a safety ground resistor (keeps the throttle connection held low if it goes open circuit), a 2k2 series resistor and a tiny chip cap that both provide a degree of noise suppression.

Provided the programming fuse in the XC846 (assuming it has one) hasn't been blown to stop code changes, then it should be possible to re-flash the chip with modified code. The lack of available documentation on this chip makes this tricky, so we really need some help from Keywin Ge on this.

It may be that the XC866 is similar, so it might be worth one of our resident software/firmware geeks having a look at doing a code dump from the existing boards. It'd be hard to reverse engineer the source code from 8051 machine code, but might be doable. I've been away from the microcontroller game for many years now, so am pretty rusty on this stuff.

Luke, remember that changing the motors from delta to wye lowers Kv, so also increases the maximum rpm that the Kelly will spin the motors at. Might help.

Jeremy

 

[liveforphysics]

I jumpered over the whole throttle saftey and filter circuit. No difference. The cap and resistor are there just for what Jeremy said they are there to do.


On the bright side, if I re-terminate the motors to run in WYE, I can use a pair of Kellys and not sacrifice any powerband. I will definately be re-wireing them into WYE, if for no other reason than to only need a 5:1 reduction, which will let me run a much smaller lighter weight sprocket. That alone should be good for shaving off a full pound or more.

 

[swbluto]

In the infineon thread in ebike-technical at http://endless-sphere.com/forums/viewtopic.php?f=2&t=5713&start=165 -

Does anyone have an idea what resistor/capacitor on the current model of the infineon implements the soft-start throttle ramp? Apparently some do not like this soft-start "feature" as it hinders wheelies and causes traffic-critical problems.

hi I have enough controlers here to do tests on to see if i can do something about this problem as there are a lot of "soft start" controllers out there that need the "soft start" removing there has already been some ideas on docs 18 fet thread as discribed above i will try them when i have some time free.
with all of the talk there has been around about the soft start not being liked I decided to have a word with keywin, I was on skype chat with him for over 2 hours included in the talk was thaty we want hard start infineons, he has promised to get 50 12 fet boards in stock of the "hard start"type they will be the same as the infineon 12 fet controllers sold at the moment there will just be no time delay on the throttle knuckles can have his old "sawblade of death" as he called his front wheel spinning on throttle up.

Geoff

So it appears there's some hope for the future.

 

[swbluto]

I now just completely read his response and I wonder... who does the 18-fet version?

 

[GGoodrum]

Same place. Not sure why Keywin can get an updated 12 FET version out quickly, but not the 18 FET. The bigger question is can we re-flash the firmware in existing controllers?

 

[liveforphysics]

Good news! Methy is getting 10 chips set with the old program, rather than the get-you-hit-by-a-car feel-like you're-stuck-in-the-mud POS start program that mine have.

I'm going to mask the board, use a heat-gun to lift this chip off, then touch up the pads, and re-solder the new chip. Piece of cake.

Then this chip is going to get hit with my Remington super STW 7mm. It's pieces will break the sound barrier in a radial outward pattern. Maybe that will will give it the "hard-start" on life it was missing.

The super 7 is the big one on the far left. It shoots at 2,520mph (3700fps). About mach 3. I think it will get the chip started quickly.

 

[GGoodrum]

I talked to dennis, at Comcycle-USA (dirtdad here...), who sells a lot of Infineon controllers, and while he agreed that you certainly don't want a throttle delay, once the bike is moving, he said that without the soft start it is impossible to go slow, like on a pedestrian or beach path, without a terrible "jerking". I think what is needed is like what the CC controllers have which is a programmable ramp for the soft start function. Also, there needs to be a some sort of override if the bike is moving. With hub motors, the controller would know this. For a non-hub application, this would be trickier for a freewheeling motor setup.

Dennis said that on his current bike he has one of the 18 FET versions, the 5304 I sold him and a 24s/72V LiFePO4 setup (YESA, I think...). Without the softstart it was pretty much unusable below 10 MPH. He doesn't remember noticing any lag while underway, though. He's going to try it again and specifically look for that.

-- Gary

 

[liveforphysics]

From thinking about it, this effect would be magnified by the amount of gearing reduction you have.

It's effects would also feel magnified by the amount of power you have available. If you can pin the throttle, wait 1-2seconds, and lazily start to roll forward, it wouldn't really matter, because it's not going to be launching you into the street in a wheelie when your power hits.

I'm completely used to bikes with touchy throttles, bikes that snap into a wheelie at the blip of throttle. That's how I like my bikes, and I have no safety issues with that. What is completely unsafe, is giving throttle, getting a mild torque response initially, then getting hit with 10x the power 2 seconds later, while your throttle is in the same spot.

 

[Jeremy Harris]

I'm completely used to bikes with touchy throttles, bikes that snap into a wheelie at the blip of throttle. That's how I like my bikes, and I have no safety issues with that. What is completely unsafe, is giving throttle, getting a mild torque response initially, then getting hit with 10x the power 2 seconds later, while your throttle is in the same spot.

Sounds like the old Husqvarna CR250 moto X'er that I stupidly converted to a "road legal" bike many years ago. It was only designed to run at 7000rpm upwards. Trying to ride it on the road was hilarious. The engine would bog off the line if you tried to pull off smoothly, four stroking like a bastard. The revs would slowly climb if you held the throttle wide open, then at about 7000rpm all hell would break out and it'd pop a wheelie right out from from nowhere. Riding on the beach was fine, just keep going like hell and it was almost as much fun as sex.

I remember the look on the face of the smart-arse kid I sold it to when he rode it away. He thought he was an ace,and totally ignored my health warning about it's peaky torque curve. He got 50 yds up the road from my house before he flipped it. Laugh, I nearly cried.................

 

[recumpence]

When I was 12 my dad bought me a used 79 YZ80. When we looked at it, the guy said "I bought this for my nephew. He took off on it, got about 100 yard away when the bike just flipped out from under him. So, I thought, hell, I will just sell it!" Hmm, that answered the next question I had "Why is there no rear fender?" So, I rode it. I got about 100 feet away and it came on the pipe, the front wheel shot skyward as I fought to hold onto the insane thing. (this all happened around the corner of the building, out of my dad's sight). When I got back from my test ride my dad asked "So, do you want it?"

I said yes all the while thinking "Hell YES!!"

In my book, there is no such thing as too much power as long as you know how to use it.

What was it Jeremy said about sex? :wink:

Matt

 

[liveforphysics]

Wheelie over backwards? Umm... I don't know what you guys are talking about...

http://www.deviantmethods.com/liveforphysics/Lukes/Lukecrash.mov

Notice that I carefully place my foot under the bike before slamming it down onto it. LOL That's my weakest bike too! My other dirtbike is 2x the power, and only 15lbs heavier.

 

[liveforphysics]

I have to do the right click, save content as thing to make it play for some reason. Works fine if you do that though. Enjoy My mom was the camera-woman for this wreck

 

[methods]

Man, we really got off topic on this thread.

A few corrections needed none the less:

1) I do NOT have any chips coming, luke is huffing glue.

2) The programming interface is NOT blown on the soldered controllers and they CAN be reprogrammed IF you can get your hands on the software. You are not going to get the software from Keywin any time soon. You could possibly get your controller reprogrammed. More likely you will wait to get a new one.

3) I have talked with Keywin about emulating the Kelly programmability when it comes to throttle range and throttle response. Dont hold your breath.

4) Just because Dird Dad's implementation is unrideable for him below 10mph does not equate to fact. Different throttles, different motors, different battery voltages, different current limits, different riding style will dictate the sub 10mph performance. A test is worth a thousand opinions. I am not about to write off "Instant Start" due to one opinion about 1 setup.

5) I have already described in a different thread how to add external slow start to an immediate start controller. Capacitor, resistor, and a switch. $0.50

6) I am dumping mine regardless. I am sick of the BS. I will order 10 when Keywin gets the immediate start in hand.

18 fet Infineion boards w/programmer - $72 shipped

-methods

 

[liveforphysics]

Man, we really got off topic on this thread.

A few corrections needed none the less:

1) I do NOT have any chips coming, luke is huffing glue.

Oops! I misunderstood what you were saying. After re-reading, I see what you were trying to say. Now I'm bummed again.