liveforphysics
100 TW
Well shoot Miles! That settles it! They've found the key to over unity! It was in that cogging causing vibration and noise without consuming any energy! 
**Important** reality check on motor, voltage, current etc.
- Thread starter liveforphysics
- Start date
liveforphysics
100 TW
Here is one of the many pieces of propaganda slotless/ironless/coreless motor guys like to make:
http://www.designnews.com/article/2168-Brushless_slotless_and_cogless.php
I do think there is something there (though maybe not as they explain it), and its' interesting they mention the copper increasing the loss, this would agree with Justin's findings of the lower resistance wound motor having higher drag torque.
http://www.designnews.com/article/2168-Brushless_slotless_and_cogless.php
I do think there is something there (though maybe not as they explain it), and its' interesting they mention the copper increasing the loss, this would agree with Justin's findings of the lower resistance wound motor having higher drag torque.
John in CR
100 TW
Doesn't it have to do with the unmatched number of poles on the rotor and stator, creating a net force of attraction to a certain number of positions? A brushed motor doesn't cog. When I remove the windings from my burnt bafang for a rewind, I'll see if it still cogs with no copper.
Yeah Luke, you towing me was a perfect example that the resistance was constant, so the only increase of my pull on your shoulder as you sped up was only the increase in my rolling friction and a bit of wind resistance.
That ebike tow using the larger diameter wheel bike with the Methods controller as the tow truck, ties right into this thread too, because we must have been dancing close to controller failure with the current limiting that had to be going on with the extra load and partial throttle.
Yeah Luke, you towing me was a perfect example that the resistance was constant, so the only increase of my pull on your shoulder as you sped up was only the increase in my rolling friction and a bit of wind resistance.
That ebike tow using the larger diameter wheel bike with the Methods controller as the tow truck, ties right into this thread too, because we must have been dancing close to controller failure with the current limiting that had to be going on with the extra load and partial throttle.
John in CR
100 TW
Here are the assumptions, real life I might add:
1. Difficult to drive low turn count motor where most riding is in current limiting (I assume the point of max power or 1/2 the no-load speed is where LFP's red line resides, and if it's not there, it's very close), where the cutoff is 37mph at this voltage.
2. Identical battery and voltage.
3. 2 different current limit settings via shunt modification so the ratio of battery and phase current limits have a constant ratio using the same controller.
4. Very low speed riding, where very low duty can still result in failure from narrow spikes in phase currents due to accidental repetitive small pulses of the throttle, though important, is not a topic of this post.
5. Acceleration and speed are identical. With the lower current limit, acceleration is generally WOT up to 30mph, and with the modified shunt an equal acceleration is accomplished through moderation of the throttle. Throttle position at cruise is identical, since speed is the same and neither current setting results in max speed = cruising speed.
My question is, will the heating of the FETs be identical along with efficiency and the PWM. I believe the answer is yes, since I just modified a controller for active air cooling, but I took the shunt modification farther than I have before with the same controller. If I'm hard on the throttle, then I get significantly more performance, but the controller gets failure range scorching hot without the fan turned on. However, just by going easy on the throttle during acceleration with equal cruising speed, the controller stays just warm to the touch.
Now the big issue, hills. I've got regular hills that I cruised up at WOT many many times without issue with the lower limit controller. The speed was right around the current limit cutoff, so I was at or very near full duty. How should I approach the same hill? Higher speed will require higher power and higher battery current, but if I try to go the same speed as before, then I think I'll be at partial throttle and chopping the current via PWM at high load could put the controller at risk. I believe the right answer is WOT is better on hills, period, and I may have an identical top speed anyway unless I was actually in current limiting before. That means with active cooling of the controller and WOT there should be no way the same hill kills an identical controller unless I add enough extra load to push me back into the rpm range of current limiting.
Did I pass the test?
1. Difficult to drive low turn count motor where most riding is in current limiting (I assume the point of max power or 1/2 the no-load speed is where LFP's red line resides, and if it's not there, it's very close), where the cutoff is 37mph at this voltage.
2. Identical battery and voltage.
3. 2 different current limit settings via shunt modification so the ratio of battery and phase current limits have a constant ratio using the same controller.
4. Very low speed riding, where very low duty can still result in failure from narrow spikes in phase currents due to accidental repetitive small pulses of the throttle, though important, is not a topic of this post.
5. Acceleration and speed are identical. With the lower current limit, acceleration is generally WOT up to 30mph, and with the modified shunt an equal acceleration is accomplished through moderation of the throttle. Throttle position at cruise is identical, since speed is the same and neither current setting results in max speed = cruising speed.
My question is, will the heating of the FETs be identical along with efficiency and the PWM. I believe the answer is yes, since I just modified a controller for active air cooling, but I took the shunt modification farther than I have before with the same controller. If I'm hard on the throttle, then I get significantly more performance, but the controller gets failure range scorching hot without the fan turned on. However, just by going easy on the throttle during acceleration with equal cruising speed, the controller stays just warm to the touch.
Now the big issue, hills. I've got regular hills that I cruised up at WOT many many times without issue with the lower limit controller. The speed was right around the current limit cutoff, so I was at or very near full duty. How should I approach the same hill? Higher speed will require higher power and higher battery current, but if I try to go the same speed as before, then I think I'll be at partial throttle and chopping the current via PWM at high load could put the controller at risk. I believe the right answer is WOT is better on hills, period, and I may have an identical top speed anyway unless I was actually in current limiting before. That means with active cooling of the controller and WOT there should be no way the same hill kills an identical controller unless I add enough extra load to push me back into the rpm range of current limiting.
Did I pass the test?
johnrobholmes
10 MW
I call the cogging force "detent force". Both terms are used seemingly interchangeably in technical writings, such as this http://books.google.com/books?id=ena3AH2rHFoC&pg=PA137&lpg=PA137&dq=detent+force+motor&source=bl&ots=BROcaRztBx&sig=nEp1ZfbpgdK9a-u6HeiIcuq3pNE&hl=en&ei=pdxZTJXvBouLnQfOkdCiCQ&sa=X&oi=book_result&ct=result&resnum=6&ved=0CCcQ6AEwBQ#v=onepage&q=detent%20force%20motor&f=false
I would say that minimizing the detent force is certainly a prime concern for hub motors in failure modes. Not so much of a concern to me with a non-hub setup, unless low speed tractability were of large concern. In this case, multiple speeds would help the low speed resolution and also get relieve the controller of strain at low speed high torque situations.
I would say that minimizing the detent force is certainly a prime concern for hub motors in failure modes. Not so much of a concern to me with a non-hub setup, unless low speed tractability were of large concern. In this case, multiple speeds would help the low speed resolution and also get relieve the controller of strain at low speed high torque situations.
jbond
100 W
- Joined
- Oct 1, 2010
- Messages
- 204
On the Infineon controllers (ecrazyman) the parameter designer software lets you set a battery current limit and a phase current limit. Various people's comments above suggest that the phase limit is done by an approximation from the controller watching the same shunt as used for the battery current limit sensing but knowing the current PWM and possibly sampling it at specific times in the PWM cycle.
The question is what happens when these limits are exceeded. Does the controller smoothly cut PWM until either or both drop back under the limits. Or is it a much cruder limit that cuts power completely until everything recovers on some much slower cycle? With all the talk about working in the current limiting region, it wasn't clear to me how the limiting is done.
This relates to two things.
1) Strategies to set the two values. Logically, the battery limit should be set according to what the battery/motor can cope with. So say 2C on a 10AHr battery and so 20A. The Phase current limit should be set mainly by the FETs and heat sink. The 2.5 times rule of thumb is just a recommendation based on a circuit layout, FET choice and heat sink. So a typical 6 FET, 20A controller can tolerate 42A Phase limit scaling up for more FETs and higher rated controllers.
2) Normally these controllers relate throttle position to average voltage via PWM. This makes the throttle mainly a speed controller as the motor will try and speed up to near whatever it's no-load rpm is for that voltage. Can we turn the throttle into a torque control by having it set the battery current limit? eg Use an external circuit or a modified CA to set a battery current limit from the throttle control and then drive the controller's throttle input to set the PWM to stay below this.
This is all in terms of thinking about an Infineon type controller with a brushless geared motor like the Bafangs or BMC/MAC.
The question is what happens when these limits are exceeded. Does the controller smoothly cut PWM until either or both drop back under the limits. Or is it a much cruder limit that cuts power completely until everything recovers on some much slower cycle? With all the talk about working in the current limiting region, it wasn't clear to me how the limiting is done.
This relates to two things.
1) Strategies to set the two values. Logically, the battery limit should be set according to what the battery/motor can cope with. So say 2C on a 10AHr battery and so 20A. The Phase current limit should be set mainly by the FETs and heat sink. The 2.5 times rule of thumb is just a recommendation based on a circuit layout, FET choice and heat sink. So a typical 6 FET, 20A controller can tolerate 42A Phase limit scaling up for more FETs and higher rated controllers.
2) Normally these controllers relate throttle position to average voltage via PWM. This makes the throttle mainly a speed controller as the motor will try and speed up to near whatever it's no-load rpm is for that voltage. Can we turn the throttle into a torque control by having it set the battery current limit? eg Use an external circuit or a modified CA to set a battery current limit from the throttle control and then drive the controller's throttle input to set the PWM to stay below this.
This is all in terms of thinking about an Infineon type controller with a brushless geared motor like the Bafangs or BMC/MAC.
Jeremy Harris
100 MW
The XieChang controllers (they stopped using the Infineon chips a year or so ago and switched to a Chinese made 116 chip) don't measure phase current and can only make a very approximate guess on what it may be based on an assumed value for the motor time constant, LR. The A/D isn't fast enough to sample at the PWM frequency, which is why there is a secondary protection circuit that shuts the controller down in the event of a high current spike (it typically shuts down at 120A for a 6 FET, 240A for a 12 FET and 360A for an 18 FET controller).
The way that phase current limiting works is that the controller measures battery current and then makes an assumption, based on the PWM on time, it's preset value for motor LR and the set phase to battery current ratio, as to what the phase current might be. It then cuts back the PWM on time to limit the phase current to what it thinks is the set value, but which might easily be out by a factor of 2 or more. It's not fast enough to cut by PWM on time on a cycle-by-cycle basis, hence the need for the current spike protection circuit.
We don't know what the preset value of motor LR the controller uses. It's not programmable directly via the Parameter Designer software, although I have a strong suspicion that this may be the variable that's actually changed when the 'phase current' setting is changed.
You can get an idea of how crude this is by looking at the very wide range of motor time constants that are around. Without the means to change this in the controller, and not knowing what the preset value is, we can't even begin to guess what the real phase current limit might be, particularly if using a low inductance, low resistance, motor. I would hazard a guess that the controller assumes that the motor will be something like a Bafang or other common type, so anything with a lower resistance or inductance will mean that the phase current limit under-reads by an amount proportional to the LR difference.
Torque control, or more accurately power control, is possible by intercepting the current signal and using it to provide feedback to a PID loop, with the demand signal being the throttle input. I looked at doing it a few weeks ago and discovered it wasn't easy! The small microcontroller I was using wouldn't run the loop fast enough, it needed something with a bit more oomph. It's not really a problem, but in this case it might be better to do the PID loop in hardware - my guess is that a couple of op amps and a handful of passive components would do the job, but tuning it for a good response might be a bit of a fiddle.
Jeremy
The way that phase current limiting works is that the controller measures battery current and then makes an assumption, based on the PWM on time, it's preset value for motor LR and the set phase to battery current ratio, as to what the phase current might be. It then cuts back the PWM on time to limit the phase current to what it thinks is the set value, but which might easily be out by a factor of 2 or more. It's not fast enough to cut by PWM on time on a cycle-by-cycle basis, hence the need for the current spike protection circuit.
We don't know what the preset value of motor LR the controller uses. It's not programmable directly via the Parameter Designer software, although I have a strong suspicion that this may be the variable that's actually changed when the 'phase current' setting is changed.
You can get an idea of how crude this is by looking at the very wide range of motor time constants that are around. Without the means to change this in the controller, and not knowing what the preset value is, we can't even begin to guess what the real phase current limit might be, particularly if using a low inductance, low resistance, motor. I would hazard a guess that the controller assumes that the motor will be something like a Bafang or other common type, so anything with a lower resistance or inductance will mean that the phase current limit under-reads by an amount proportional to the LR difference.
Torque control, or more accurately power control, is possible by intercepting the current signal and using it to provide feedback to a PID loop, with the demand signal being the throttle input. I looked at doing it a few weeks ago and discovered it wasn't easy! The small microcontroller I was using wouldn't run the loop fast enough, it needed something with a bit more oomph. It's not really a problem, but in this case it might be better to do the PID loop in hardware - my guess is that a couple of op amps and a handful of passive components would do the job, but tuning it for a good response might be a bit of a fiddle.
Jeremy
jbond
100 W
- Joined
- Oct 1, 2010
- Messages
- 204
How fast do you think the current limiting cycle is? Is it varying the PWM to stay under the limits once per motor cycle or longer or during a single phase. Just trying to get a feel for factors of 10. :wink: Would this also be affected by the block time?
One thing that bothers me. The discussion above about people blowing controllers under high load, low speed, low PWM width suggests that reducing PWM width can increase the phase current demand. But if the controller is trying to get under the phase limit by reducing PWM isn't it going to make things worse rather than better until PWM actually drops to 0%
One thing that bothers me. The discussion above about people blowing controllers under high load, low speed, low PWM width suggests that reducing PWM width can increase the phase current demand. But if the controller is trying to get under the phase limit by reducing PWM isn't it going to make things worse rather than better until PWM actually drops to 0%
Jeremy Harris
100 MW
jbond said:
We have no way of knowing how fast it is, but can assume it's pretty slow (in the tens, or even hundreds, of uS range) because of the extra protection hardware. They wouldn't add cost to the controller by fitting this unless they needed to for a good reason, and the only reason I can think of is that the current sampling loop time is just too long to protect the controller from high phase current spikes.
The high load, low speed, full current limiting case is an example of the preset motor LR coming into play. The controller assumes (as it must) that the rate of rise of phase current is limited by a particular motor time constant. If we then fit a motor with a much lower time constant the rate of rise of phase current will be much faster, so the preset "safe" minimum PWM on time may now not be safe at all - the current could rise to an unsafe value despite the controller cutting the PWM back to what it thinks is a safe time. The controller has no way of knowing that the phase current is really much higher than it thinks it is, as it can't measure it directly. The controller won't cut the PWM back further, as it doesn't think it needs to, at least in the time it takes for the main current sense loop to measure the battery current, realise it might be a bit high and then act to limit current further.
The block time may well have a significant adverse effect, as it deliberately allows a momentary over-current situation to exist, for a time set by the programmer.
The bottom line is that we're using these controllers outside their design limits much of the time. If we didn't hook them up to motors with a lower LR than they are designed, and then tweak the settings to get more performance, we'd probably not see them fail. We're only really using these controllers because they are available cheaply and we now know enough about them to be able to mod them to do pretty much what we want. They do have some serious limitations though, not least of which is using a FET package that is inherently poor for high power use.
Jeremy
317537
10 kW
- Joined
- Oct 19, 2008
- Messages
- 939
Hi Jeremy.
I was just reading your last post a few times. You seem pretty focused on all this tech stuff.
Mind if I comment. Please correct me if I get anything wrong
My knowledge in controller workings is ever increasing and I hope to the point one day I could design my own.
EM spikes have got me intersted in a lot of things around induction of motors recently which lead me to your last post here.
A lot of chaos goes on when an operating wheel in the real world especialy with sensorless operations. Big bumps in the road sends + - energy shocks through the phase lines very much so with direct drive motors and IMO caps and fets can take a beating off the road.
And the recovery of diodes in fets is not light speed so high energy can get out on the skin of conductors and find a path through the fet silicone.
How much EM can be allowed to punch out the N P barriers can up to the voltage generated by EM and the resistance of the fet in on and off states.
A fet may have a longer life if em spikes are properly filtered. The types of spike I think that we do not concern ourslves with is high energy low current spikes. They are very short in time and very high in voltages. Energy is ionic on a sub atomic level, I certainly do not have a meter to measure this, Lord! I could be blind to component failure and not know it. Is anything fast enough? EM protection limits failure in the short term but not over longer terms of frequent abuse.
High energy spikes are phase inducted and if you really hit the throttle and sitting back on the motor you are sinking a lot of energy into the motor cores to be released with high energy.
I think if we look deeper in magnetic moments concept and the speed at which dense magnetic intensity decays off a switch at some stage in the thousands of phase events and hall events needing to be so perfectly timed on a bumpy road, your coils are going to let off a few zaps along the way. The more you push the motor IMO the more damage you can do by means of EM to the controller.
I agree with you Jeremy, the controllers we use are be adequate enough to power a leagal ebike but going off road one may need to seek a better solution for reliability.
So much goes on in this small box that makes our ebike sun shine if you could help me understand this better I would be much greatfull.
I was just reading your last post a few times. You seem pretty focused on all this tech stuff.
Mind if I comment. Please correct me if I get anything wrong
My knowledge in controller workings is ever increasing and I hope to the point one day I could design my own.
EM spikes have got me intersted in a lot of things around induction of motors recently which lead me to your last post here.
A lot of chaos goes on when an operating wheel in the real world especialy with sensorless operations. Big bumps in the road sends + - energy shocks through the phase lines very much so with direct drive motors and IMO caps and fets can take a beating off the road.
And the recovery of diodes in fets is not light speed so high energy can get out on the skin of conductors and find a path through the fet silicone.
How much EM can be allowed to punch out the N P barriers can up to the voltage generated by EM and the resistance of the fet in on and off states.
A fet may have a longer life if em spikes are properly filtered. The types of spike I think that we do not concern ourslves with is high energy low current spikes. They are very short in time and very high in voltages. Energy is ionic on a sub atomic level, I certainly do not have a meter to measure this, Lord! I could be blind to component failure and not know it. Is anything fast enough? EM protection limits failure in the short term but not over longer terms of frequent abuse.
High energy spikes are phase inducted and if you really hit the throttle and sitting back on the motor you are sinking a lot of energy into the motor cores to be released with high energy.
I think if we look deeper in magnetic moments concept and the speed at which dense magnetic intensity decays off a switch at some stage in the thousands of phase events and hall events needing to be so perfectly timed on a bumpy road, your coils are going to let off a few zaps along the way. The more you push the motor IMO the more damage you can do by means of EM to the controller.
I agree with you Jeremy, the controllers we use are be adequate enough to power a leagal ebike but going off road one may need to seek a better solution for reliability.
So much goes on in this small box that makes our ebike sun shine if you could help me understand this better I would be much greatfull.
