Sensorless controller question

Jeremy Harris

Jeremy Harris

100 MW
Joined
Oct 23, 2007
Messages
4,208
Location
Salisbury, UK
I've bought one of the surplus TongXin motors and controllers from Justin at ebike.ca to play with (please don't hit me with all the reasons why the TongXin stuff is bad - I've already read it all........).

What puzzles me is the way that the controller starts the motor from a standstill, as it doesn't have any Hall sensors, just the three phase wires to the motor. It must work the same way as the RC model motor controllers, I guess, but having read all the threads on here relating to sensorless controllers I am still not sure. The motor isn't a "pedal first" type, as it has a freewheel, so it can't be spun up to speed before applying power.

My interest lies in looking for a possible substitute controller, maybe using a RC model type or perhaps building one from scratch as a project. The existing controller is potted in resin and looks pretty near impossible to reverse engineer without destroying it.

Has anyone ever tried to get inside one of these controllers to figure out how they start up? Any clues would be useful.

Jeremy
 
Something wrong with the Tongxin? first I heard of it. All I know about them is they are smallish but reliable.

An RC controller should be usable on any 3 wire brushless, with or without the hall sensors on the motor. Since the sensors don't do anything for the motor, the controller can ignore them

The only big problem with RC controllers is the voltage. most of the small ones don't go above 18.5 volts
 
Thanks for the reply, the bad stuff I'd heard was from a chap on here who hadn't got a good word to say about them. I've just heard of a chap here that's blown a TongXin controller after very little use, which is what got me thinking. I reckon the motor has good potential for power assist, although it's never going to match the poke from my Crystalyte.

I hadn't thought about the battery voltage problem with the RC controllers, but maybe a cheap one would make a good basis for a hack. I guess it'd be easy enough to fit a new output stage, with nice beefy FETs and a high voltage driver, just retaining the RC controller to do the clever stuff. The only other snag I can see with this approach is the need for a PPM input to the controller, but a simple timer circuit should work OK to hook a throttle pot up to, or maybe just a cheap RC servo tester module.

Jeremy
 
So how much did it cost, the website doesn't list the price?
I'm considering getting this (trying to decide amongst a few others) & it's news to me too that there's problems with the Tongxin.
I think it was just the one guy with an axe to grind for some reason is the only negative I've heard so far.
Did you ask Justin why he's declined to carry these?
I'd assumed it was becuz he decided to get behind the Ezee but just curious if there was something specific he didn't like about the Tongxin.
Probably just that it's really small & demand from his customers seems to trend towards always MORE POWER!
 
Not sure about Justins prices or the original controller question, but for more info contact Dianna Lin at Tongxin, they've replied quickly to my emails - xinyue_china@hotmail.com


Their price list is here http://www.tongxin.net.cn/ebike_components0610.xls
Remember prices change quickly when purchasing internationally, and some say you get what you pay for.

Tongxin direct from China seem VERY cheap, freight is the biggest component of the order.
 
I didn't ask Justin why he was selling off the TongXins, but would guess that it's as you say; they aren't fast enough for some.

I already have a Crystalyte 405 in a 20", with a 35A controller running on 48V, and that is really a "no pedal" bike. I bought the TongXin as an experiment to build a bike that just gives assist up hills, so forcing me to pedal a bit. The plan is to build an intelligent crank torque sensor instead of a throttle.

It may be just as easy to build a whole new controller, hence the question as to how these sensorless controllers work (or more precisely, how they start up). I've done some digging around and it looks like there are a few home-brew RC controllers about, so I may opt to start with a design like that and beef it up to work at 36V.

It'd be nice to end up with a really light system that doesn't massively alter the ride of the bike, yet provides enough extra oomph to get up hills without too much sweat. I guess I could go an buy something, but building it seems more fun.

Jeremy

PS: Thanks for the link to the TongXin prices, they are pretty amazing!
 
SolarBBQ was into Tongxin a while back, if i recall correctly i remember something about the company selling him DOA controllers and not honoring warrantee.. and various other factory blunders..

I also like the ida of a low-power bike for real deal exercise.. my Clyte bikes are totally not pedal friendly lol..
 
That's the thread I read, he seemed to be pretty hacked off with them. In contrast, I've heard that the chap over here whose controller packed up emailed TongXin, who have responded very quickly and are sorting his controller out, so maybe they've now improved their customer service.

I've found a couple of circuits for sensorless controllers, so it looks like I may well end up playing about with some wiggly amps for a few weeks. Provided I can keep the smoke in I should be OK.......................

Jeremy
 
although most instant start controllers rely on hall sensors to provide rotor/stator position information to start the motor running, some embedded processor controllers employ an algorithm that delivers short pulses in a sequence to the windings until the motor starts to revolve and normal operation using back emf position feedback can be used.

i did take apart one of justin's Tongxin controllers and there is a nice big embedded processor in there. it was coated in epoxy so i could not read it's part number.

i thought the worst problem with the Tongxin motor was the extremely wide with of 140mm across the flats that it used. tough to find forks wide enough for that.

rc controllers are not instant start and require you to flip the prop smartly to get the motor going.

rick
 
The motor I have isn't as wide as that, it's just a tad over 100mm (if I were to guess I'd say it's actually 4" wide). My front forks are about 102mm wide and it fits with a little tiny bit of clearance.

I'm looking at making a "brute force" add-on card to an RC model motor controller as a starting point. I have bought a cheap RC brushless controller that I've hooked up on the bench with a small brushless motor and it starts the motor consistently in the same direction from stationary, so I guess it must be doing something clever to kick the motor off, then sensing the back EMF zero crossing points to get in sync.

The plan is to use the RC controller as-is, but with the output feeding a higher voltage FET power stage. I'll need to feedback and voltage clamp the motor to the original controller outputs to ensure the back emf zero crossings get sensed, but this should be easy enough to do. One advantage of using an RC controller as a basis is that they have a lot of programmable stuff built in, like the ability to change the timing, control start up ramps, even motor reverse. It also save writing code...................

Jeremy
 
rkhiosek is talking about the rear motor.
I have these two pictures of Tongxin motors, different sizes from what I can see. Not sure why.

Keep us updated Jeremy, I'm keen to hear you thoughts on this motor, I'm thinking about buying one front and one rear 260rpm for a dual setup for torque and 30+ km/h.

dianji1.jpg


dianji2.jpg
 
rkosiorek said:
rc controllers are not instant start and require you to flip the prop smartly to get the motor going.

rick
Click to expand...
Not True. Any RC controller made in the last few years is instant start. They send a single pulse to the motor, wait for the back EMF, then the processor calculates the position and rotation and can run from there. it only takes one clock pulse for the controller to work it all out, and most are clocked at 14,000 to 32,000 HTZ.

The problem for adapting these to ebikes is that the initial pultse has a 50/50 chance of knocking the motor backwards. The controller corrects for it on the next pulse, but it can kick the motor around the wrong way before it gets spinning the right way. It wouldn't be able to move the bike much on only one pulse, but it might be un nerving to someone who wasn't expecting it to have a bike kick backwards before taking off
 
The Tongxin and Puma motors have a built in freewheel, so if it kicked back on startup, you wouldn't feel it. It might cause a slight delay in takeoff.
 
a lot has changed since i first looked at these RC controllers a couple of years ago. that will teach me to go through life with blinders on.

rick
 
Hi Jeremy,

(Nick here, just registered on this board)

In reply to your question - I am working without any inside knowledge of how the controller actually works, but here's where I would start if I had to do it.

I would run it as an AC motor until it was moving enough to give me the signals needed to run it as a DC one. At start up it would only want short current pulses anyway, so I would cycle the pulses round the windings in the direction I wanted to go. With a 3 pole geometry, if the armature is in the worst case position and the first pulse makes it move backwards, then the 2nd one will make it go forwards. With more poles, you could get the first N pulses going the wrong way, but then the next N pull it the right way.

And as has already been suggested, if there was enough movement and signal after the first pulse to determine direction, then that would take over.

Nick
(Somerset, England)
 
See my post:
http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=3111#p44811
on how I fried a perfectly fine Castle Creation 80A RC controller palying with the hub :). Unless specifically designed for it the overspeed can kill the controller because it does not do sync rectification in the condition of back-emf exceeding battery voltage and all that back-emf current flows through fet intrinsic body diodes with the good chance of frying them. Besides I really do not like the start-up in sensorless controllers.
 
The real problem I feel with the rc controllers is the lack of current limiting. They have a couple cheap controllers good to 100 amps and 50v but if everything wasn't match right you could blow it.
 
This miniature controller does concern me reliability wise. I'd like to have a few backups.

Anyone know where you can buy a tongxin controller only?

Thanks
 
Dang! I don't know how you managed to get that thing un-potted like that. Seems like you could repair it at this point. Some nice IRFB4110's should make it bulletproof enough. Looks like the switching inductor took a hit during unpotting, but those are replaceable too.
 
It took a long time to de-pot the controller. After I got started I could not give up!The "switching inductor" is definately toast. The part number is gone. It may have been the thing that blew the controller. I would love to upgrade it to be bulletproof - the fets on this controller look easy to replace - there is not so much solder on them - but I don't know for sure what blew.
 
You must log in or register to reply here.

Similar threads

E
Help understanding / replacing controller please!
Replies
14
Views
1,547
encore-moi
E
DOA
NotAWheelchair / Utah City Trikes controller
Replies
1
Views
524
amberwolf
amberwolf
DomW
Dual drive ebike- controller alternative.
Replies
12
Views
1,948
DomW
DomW
Botvink
    • Like
My Brushless Design for Hall Effect Type Motors, Looking for Suggestions
Replies
6
Views
843
mxlemming
mxlemming
amberwolf
Common Controller Programming Software Project
Replies
0
Views
478
amberwolf
amberwolf
Back
Top