Sensorless Controller Solution?

[Dauntless]

As engineering types go, I make a damn fine Cultural Anthropologist, which is my approach to about anything. So I just keep reading and reading and reading about the Turnigy HXT motors and others like it. I have a definite idea about these exploding controllers. (Which includes the desire to avoid them.)

http://www.hobbypartz.com/07e-c-combo-m3.html

This pairing is intended to run a large RC offroader. Not a plane turning a propeller, a truck, etc., with at least a little of the similar loading as a bike. The motor is 1,350 watts, 18.9-22.5% as compared to 6,000, 6,500 and 7,000 for the Turnigy 8085, 80100kv130 and kv180. And yet the controller is just about the amperage of 160 amp Castle people are swearing by. The smallest controller I'm finding people using for that particlar capacity of motor is 90 amps; those using it are told that's not big enough by others using 120 amps. Who in turn, are warned of their own deficiency by those with 150 amps. I just think these people are as close to being the experts in the sensorless motors that we'll find. An RC car just has to be considerably more responsive than an electric bike, so the bike needs even more latitude. Hobbyking as a new "Monster" that says flat out it's not to be used for a bike/scooter.

Sounds kind of like houses of straw, sticks and stones. The 8085 might hold up with a 200 amp controller, to my understanding as high as Castle goes. The two 80100's might like something bigger. http://kellycontroller.com/kbl4830124v-48v300abldc-controllerwith-regen-p-257.html

But you don't have to listen to me, maybe you LIKE blowing up all those controllers. I notice the giddiness with which it gets discussed around here.

 

[spkpn]

Hi,

this is a topic that I am very interested in recently. As I see it at the moment (and I could very well be wrong), there are two problems with sensorless motors at the moment.

1) sensorless controllers seem to often have trouble to turn an electric motor at low rpm and high load. A possible solution here would be to use a flywheel clutch, which can be bought for comparably little money.

2) There seems to be a lack of out of the box sensorless controllers that offer torque controll. Here unfortunately I dont see an easy work around for people (like myself for instance) that are not skilled enough to build their own controller.

 

[Lebowski]

1) sensorless controllers seem to often have trouble to turn an electric motor at low rpm and high load. A possible solution here would be to use a flywheel clutch, which can be bought for comparably little money.

This raises the question, what's the definition of 'low rpm' and 'high load' ? You need to relate these
fuzzy terms to real world ratio's of back-emf voltage, phase current and winding impedance. (you
probably mean low back-emf, high current and high impedance, but how at what ratio of back-emf
to phase current times phase impedance is 'low rpm high load' ?)

 

[spkpn]

This raises the question, what's the definition of 'low rpm' and 'high load' ? You need to relate these
fuzzy terms to real world ratio's of back-emf voltage, phase current and winding impedance. (you
probably mean low back-emf, high current and high impedance, but how at what ratio of back-emf
to phase current times phase impedance is 'low rpm high load' ?)

I cant really give an answer to that I am afraid. To many technical terms that I have no idea what they mean. Hopefully there will be time in the future so that I can read some stuff explaining all that . However, what I meant to say was that I heard (also here on the forum, I think) that some people have problems with using sensorless motor/controller combos when they use it together with a fixed transition (no clutch) ratio on a motorbike. It seams that these problems particularly concern accelerating hard from 0 rpm, hence my "low rpm-high load" statement.

rgrds,
spkpn

 

[Kin]

I was under the impression there is no technical way to get around the low back emf problem besides intelligent algorithms for starting up. Or, more sensitive detection of back emf, but that goes with more noise.

 

[Lebowski]

I was under the impression there is no technical way to get around the low back emf problem

I ran my motor on 0.3 V back-emf with my homemade sinewave vector sensorless controller (0.3V and 0.1A got me a few rpm,
this is where a motor with no iron core and a bicycle axle shines but otherwise pretty useless though)

 

[Kin]

Sweet. To be honest I do not know too much about the specifics physics/electronics. I suppose I should clarify, though, that

I meant to say, :"I was under the impression there is no way to get around the huge current spike of from-zero acceleration in conditions of low back emf" Meaning, I was under the impression that the biggest issue will be acceleration from a stop, and I did not think there was an easy way around that response problem.

I do acknowledge that low back emf is something that can be solved by being sensitive, but that the solution might not be very helpful for a dynamic range and for the acceleration from stop/slow situation. But, pan Lebowski knows WAY more about controllers than me .

 

[spkpn]

and I did not think there was an easy way around that response problem

perfrctly possible that I am completely missing the point but wouldnt a flywheel clutch be an easy mechanical solution to that problem?

 

[Kin]

Well, here is what I understand your suggestion to be:

Keep the motor running at a stop, so that the rpm doesn't go down, by having some clutch system. Please correct me, if this is not what you mean.


The problem is that I don't know how this will help. Because, when you reengage the motor, you'll have that same huge spike, since there is really almost no energy stored in the rotation of an unloaded rc motor.


There is an alternative interpretation of what you're saying. Maybe you're suggesting some system that stores up energy in a flywheel that can be applied for the initial acceleration. Thats kind of cool. I think it might be complicated, I'm not sure.

What you said probably is more clear than I make it out to be. My uncertainty is a product of the fact I'm not very experienced with anything.

 

[spkpn]

sry for my late answer.

what I meant is neither

it is just a system that if the motor turns below a certain rpm (say below 3000rpm) it is not loaded. As it accellerates the clutch begins to "grab" and the motor starts to do some actual work. I assumed that it is problematic for the controller to "find out" the position of the rotor with low rpm and I assumed that it was that which causes cogging and low efficiency at low rpm, which should not be a problem if work is only done at above 3000rpm.

However, I feel quite unconfortable with writing up my guesswork, as I really dont know how that stuff works and my "thoughts" are very unfounded (infact, they rely on how I was interpreting posts and not on knowledge on the matter)...

 

[spkpn]

the confusion my posts provocated are probably due to a translation error. What I meant when I wrote "flywheel clutch" was actually a centrifugal clutch sorry for the confusion.

 

[John Bozi]

a simple solution is adding a small sensored motor with a basic granny gear for 0-10kmh.

I was running bbs02 and hub motor until my halls failed again. I will probably go sensorless controller this time, and the take off from zero won't be a problem.