Henry111's free tech support thread

[mark5]

Personally, I don't think anyone of us should waste our time or knowledge giving this guy the time of day.... let alone give him any help. But then again likely just misunderstanding what he's saying........ But if I am so is everyone else on this forum.

Bob

Everybody?????????

People here enjoy solving problems and answering questions more than they care about a question asker's ungrateful behavior. Just like sooner or later you'll again try to dictate the exact way you want your free, unpaid for answers or you'll snap at someone who gave you an answer you didn't like. You've done it before. Regardless people will still be helpful answering your questions and for that you're lucky. After all, you do need a source of tech support for the ebikes you sell.

 

[dnmun]

did you have thread before where you talked about reversing the polarity on your controller? that is what happened to this controller. someone recently asked so maybe it was you.

 

[windtrader]

This form of commutation is the simplest; you can implement the control electronics with a few $1 chips and three $0.50-$1 Hall effect sensors to tell you where the rotor is at any given time; it's also fairly efficient at high speeds. However, at low speeds you start running into serious problems with torque ripple; each time there's a state change the rotor goes from feeling almost no torque (because it's right on top of the energized winding) to feeling the full torque exerted by the next winding as it comes online. Instead of exerting a steady torque and moving at a roughly constant angular velocity, the rotor is effectively flinging itself from one torque vector to the next. At high rpm you don't feel this much at all because the torque pulses are coming on and off too fast for the physical system to react to each one individually (and so all you see is the average torque, which is fairly steady). At low rpm (especially on a small vehicle like an electric scooter) you can feel the torque ripple quite well, and it can be quite disruptive.

Are you describing "cogging" or is cogging more a timing/phase issue?
How do control systems smooth out the torque ripple at low speed?

The low speed cogging/timing symptom seems to be more common with RC motor, what is different?

 

[spinningmagnets]

The shunts look like they melted. It could have been from running them with too much current (i guess), or...as another poster suggested, the polarity was reversed. I don't know what reversed polarity would do, exactly, but...each component that electricity flows thgrough has a certain amount of resistance. the resistance limits the amount of flow.

If you take a wire (which is a good conductor, but a very poor resistor) and touch it to the two posts of a high-current battery, the wire will immediately get very hot and quickly melt. So, I'm guessing if current flows in the proper direction through the controller, the line-up of components keep the current down. But, when current is flowing in the wrong diretion...apparently...no resistance=high flow, and a quick melt-down.

Concerning "torque ripple", I'm not familiar with that term, but from the way he's describing it...I has read that high-pole count motors have a smoother operation at low RPM.The on/off electromagnets on the stator are poles. Here's a pic below comparing a BMC and a BPM, two popular motors. A low pole-count can be run to high RPMs with less eddy current heat losses, plus costs less. High pole count is more expensive, but smoother torque at low RPM, plus runs on sensorless mode easier at low RPMs.

 

[mark5]

I need for this bike to be absolutely legal as to speed limit, preferrably less than 20mph.
It is a Golden Motor (Magic Pie II) with 36V Lithium battery.
Is there a speed govenor on this motor?
If not, your suggestions please.

There is an active, intelligent system of control already intergrated into you Ebike.
It's called your brain and right hand.
Where do you live that you are in such fear of exceeding the "limit"?

Did that answer feed your ego sufficiently? Or do you need more opportunities. I'm sure you can find plenty around here. The bike is to be ridden by a young person who I want to be sure rides at slower speeds.

Oh, I see, take a beach cruiser with a Magic Pie II frt. mounted to a ridged fork and cap the speed to 20 mph to make it safe for a "young person"
Who's ego is being massaged here?

The Golden Motor MP-II can be programmed for max amps.
I found the answer at: http://www.ElectricBicyclesMagazine .com

Read back though his posts, he's affiliated with that E-rag.
The whole thread smells Trollish.

That e-rag where Henry111 "found the answer at," i.e. that he was shilling, is his own publishing business. But that wasn't all. That e-magazine issue he linked to also happened to have an ebike review where his staff writer shills an ebike cruiser sold by a company that turns out to be Slick111's own ebike sales business.

Yeah, I know this thread is two years old.

 

[dumbass]

Personally, I don't think anyone of us should waste our time or knowledge giving this guy the time of day.... let alone give him any help. But then again likely just misunderstanding what he's saying........ But if I am so is everyone else on this forum.

Bob

Everybody?????????

People here enjoy solving problems and answering questions more than they care about a question asker's ungrateful behavior. Just like sooner or later you'll again try to dictate the exact way you want your free, unpaid for answers or you'll snap at someone who gave you an answer you didn't like. You've done it before. Regardless people will still be helpful answering your questions and for that you're lucky. After all, you do need a source of tech support for the ebikes you sell.

Think and feel as you wish just as others will. I guess your opinion is we should except being dumped on if the person doesn't like our replies. And I'm sorry I don't remember me disrespecting someone that was trying to help me. Even if they misunderstood my question. But well I call out someone acting like they are in someway better then the rest of us? Yep, I sure will. But if you noticed I wasn't the only or even the first one to point out the attitude. Now was I? And yes I am very thankful for all the great advise I have received here. Just as I have tried to give a little back. And I will continue asking for others advise and experiences in the future. And will also continue to share mine. But I fail to see why anyone needs to have an attitude toward people that are trying to help them. Even if they misunderstood the question (which wasn't clearly asked) And remember, many of those he called out are at least in my opinion well respected long term members of this forum. And some have since bothered to respond to him again. So.., I guess we will just have to agree to disagree.

Bob

 

[mark5]

Wasn't talking about you at all Bob. Looks like Henry111 is getting a great deal. Free, unpaid for tech support from E-S people to help run his ebike sales business, can boss his employees around saying "I thought I told you I wanted my answers this exact way," insult them without consequences. Even after all that there will always be some E-S people that will answer his questions. They're getting played.

 

[Henry111]

did you have thread before where you talked about reversing the polarity on your controller? that is what happened to this controller. someone recently asked so maybe it was you.

Yes, the polarity on this controller was accidentally reversed.

How did we drift off from my original question on controllers into this involved discussion of motors?

 

[dnmun]

there was no drift. i asked a direct question.

 

[dumbass]

Wasn't talking about you at all Bob. Looks like Henry111 is getting a great deal. Free, unpaid for tech support from E-S people to help run his ebike sales business, can boss his employees around saying "I thought I told you I wanted my answers this exact way," insult them without consequences. Even after all that there will always be some E-S people that will answer his questions. They're getting played.

No problem Mark. And I certainly meant no disrespect to you ether way.

Bob

 

[Henry111]

Thanks everybody for all your help.
I have repeated the original photo here for clarity.
Extrapolating from the answers received from you folks, here are the part names that I have so far:
Please correct me. I will come back to what each part does later.
Apparently capacitors are the major components.
What is Number 12?

Number 1: (12 in all) Mosfets.
Number 2: Capacitor
Number 3: Capacitor
Number 4: Capacitor
Number 5: Capacitor (gate drive circuitry)
Number 6: Capacitor
Number 7: Capacitor
Number 8: Capacitor
Number 9: Capacitor (gate drive circuitry)
Number 11: Resister (BJT (bipolar junction)
Number 12: ????????
Number 13: Resister
Number 14: Resister
Number 15: Capacitor
Number 16 Capacitor (bypass capacitor)
Number 17: Capacitor

 

[999zip999]

Can't we all just have a war of words and forget Henery's ???? You know that old fashion E.S. go round of name calling dirt throwing and some sissy kicking, I remember the good old days. Lol. Hell with the Mods.

 

[d8veh]

No. 12 is the 12v regulator (actually 14v or more), which supplies the power for the mosfet gates, and also to the 5v regulator. You missed the 5v regulator, which is very important because it supplies the power for the CPU and all the sensors. It's just south of capacitor #10.

The two big resistors #13 and #14 cut down the battery voltage to the right level for the 12v regulator.

Another important feature is the voltage divided that the CPU needs to determine the battery voltage. Look for two black resistors with four digit markings somewhere top left.

The chip under #7 is the microprocessor, which reads the signals from the pedal, brake, throttle, temperature, battery voltage (divider) and current (shunt/s) and speed sensors, then uses its algorithms to compute a pulse width and timing for the mosfets. The output goes to little transistors that switch the 12v to the mosfet gates.

 

[cycborg]

Since the electromagnets that are turned on and of are connected in groups of three (3-phase), the MOSFETs come in groups of three. 6-FET, 9-FET, 12-FET, etc. The more FETs you have, the more amps of current you can flow

Something I've been wondering about, and maybe this is a good place to ask -

Each of the three phases requires a high-side and a low-side switch, which explains the 6-FET configuration. And I assume 12-, 18-, 24-FET etc. configurations just parallel the FETs to allow higher current. But how does 9-FET work? Does the high side pass more current than the low side, or vice versa? Or do the extra FETs provide a current sink for the inductance in the motor windings during some stage of the switching sequence?

 

[cycborg]

Apparently capacitors are the major components.

Well, the capacitors are definitely the biggest components, and they are important, but I would say that the major components are the parts in ARod1993's schematic - the microprocessor ("brain"), the gate drivers, and the MOSFETs. You've identified the MOSFETs, and as d8veh points out, the microprocessor is the chip that appears below the capacitor labeled 7. It looks like this controller doesn't use gate driver ICs like in the schematic, but rather discrete transistors. I think you can see these in your closeup of the burnt-out shunts - they're the 3-terminal components labeled Q4A, Q1B, etc (if anyone knows this for sure, feel free to verify or correct).

Think of the capacitors as "smoothing out" the power transfer between various sections of the circuit. They do this by storing electric charge, which requires a large volume, which is why they appear so prominently. But the real action happens in the microprocessor -> gate driver -> MOSFET path.

In addition to the components, you might also want to label the wires for others to identify, since understanding what goes in and out is as important as what happens in between.

 

[Henry111]

there was no drift. i asked a direct question.

Sorry dnmum, I was not referring to you.

 

[Henry111]

Here are two photos (front and back) of a speed controller. It was sent to me as a replacement for one to be removed from the inside of a Golden Motor (Magic Pie). Would anyone care to discuss the differences between this speed controller and the one already under discussion?
Side note: To replace this speed controller there is a lot of tedious soldering to be done. I've heard (don't know, but have heard) that GM has upgraded it so that it is an easier-to-replace simple plug-in.
By the way: I have no intention of installing this speed controller.

 

[cal3thousand]

Is there anyone in the Los Angeles/Long Beach, CA area who knows how to trouble shoot a Lithium?
Who knows how to take one apart and put it back together again?
Who truly understand the chemistry of the various Lithiums?
Who would like to earn some extra $ bucks.
If so, I would like to meet him in person.

I'm not a 'battery tech' by any trade, but I can take one apart and put it back together. I understand the chemistry of various lithiums and I would like to earn extra bucks.

I'm available in Downtown LA or on the Westside depending on the weather. :wink:

 

[ARod1993]

This form of commutation is the simplest; you can implement the control electronics with a few $1 chips and three $0.50-$1 Hall effect sensors to tell you where the rotor is at any given time; it's also fairly efficient at high speeds. However, at low speeds you start running into serious problems with torque ripple; each time there's a state change the rotor goes from feeling almost no torque (because it's right on top of the energized winding) to feeling the full torque exerted by the next winding as it comes online. Instead of exerting a steady torque and moving at a roughly constant angular velocity, the rotor is effectively flinging itself from one torque vector to the next. At high rpm you don't feel this much at all because the torque pulses are coming on and off too fast for the physical system to react to each one individually (and so all you see is the average torque, which is fairly steady). At low rpm (especially on a small vehicle like an electric scooter) you can feel the torque ripple quite well, and it can be quite disruptive.

Are you describing "cogging" or is cogging more a timing/phase issue?
How do control systems smooth out the torque ripple at low speed?

The low speed cogging/timing symptom seems to be more common with RC motor, what is different?

Technically speaking, it's all the same phenomenon; in both the trapezoidal control case and the misaligned timing case the issue is that the torque being exerted on the rotor is not constant. In the case of trapezoidal control the torque vs. position function looks rather like an inverse sawtooth; as the rotor moves past a stator slot the torque exerted steadily decreases until the rotor pole is almost perfectly in alignment with the slot, at which point the control circuit energizes the next pole and the torque jumps back up to peak as the rotor follows the control circuit around the stator. At high speeds on a vehicle you don't feel it all that much because the torque ripple is happening at hundreds to thousands of hertz, and your mass (along with that of the vehicle) acts as a mechanical low-pass filter smoothing out the stator ripple. At very low speeds the ripple occurs at only a few hertz (and if you're pulling a pile of current on startup) it's going to be fairly high in magnitude, and that's generally when people start to notice cogging.

If the sensor timing is off such that state transitions happen too early then the motor torque will look more like a badly discontinuous triangle wave rather than a sawtooth, and will probably run a bit rougher; if the motor loses synchronicity with the controller then all bets are off because brushless motors can't slip the way squirrel-cage motors can. If the timing is off such that state transitions happen late, then the rotor will experience momentary back torque at the end of every state, followed by a surge in forward torque at the very beginning of the next state. That will be even more noticeable and can manifest at any level from a bit of rough running at moderate to high speeds up to a series of rhythmic jerks that are close to forcibly removing you from the vehicle if you're going slowly.

RC motors aren't, strictly speaking, more predisposed to cogging than hub motors are (the only exceptions being cases where specific hub motors have either their magnets or their stator slots set at a bit of an angle); they're all fractional-slot permanent magnet outrunners; my guess would be that the issue has more to do with the controllers. Hobbyking model airplane/boat controllers are sensorless designs optimized for non-inertial viscous loads; with such loads you can basically just start playing the trapezoidal commutation pattern at an arbitrary frequency (limited only by battery voltage and motor KV) and the motor will be able to snap to and follow along almost instantly. That's not really possible with inertial loads like a vehicle; if the controller commands a startup frequency corresponding to several mph and you're standing still, the motor will not be able to get you up to speed instantly, and if/when it loses sync with the controller it's going to cog and jitter and quite possibly make a number of very sad noises. When you use RC motors with controllers designed for vehicles (Kelly, Lebowski, Infineon/Xie Chang, some of the eLifeBike stuff), they're either sensored (Kelly, some Lebowski controllers, and I think the Infineons) or they use a control algorithm designed to handle inertial loading well (eLifeBike stuff, Lebowski controllers) and so it's much less of an issue.

 

[windtrader]

RC motors aren't, strictly speaking, more predisposed to cogging than hub motors are (the only exceptions being cases where specific hub motors have either their magnets or their stator slots set at a bit of an angle); they're all fractional-slot permanent magnet outrunners; my guess would be that the issue has more to do with the controllers. Hobbyking model airplane/boat controllers are sensorless designs optimized for non-inertial viscous loads; with such loads you can basically just start playing the trapezoidal commutation pattern at an arbitrary frequency (limited only by battery voltage and motor KV) and the motor will be able to snap to and follow along almost instantly. That's not really possible with inertial loads like a vehicle; if the controller commands a startup frequency corresponding to several mph and you're standing still, the motor will not be able to get you up to speed instantly, and if/when it loses sync with the controller it's going to cog and jitter and quite possibly make a number of very sad noises. When you use RC motors with controllers designed for vehicles (Kelly, Lebowski, Infineon/Xie Chang, some of the eLifeBike stuff), they're either sensored (Kelly, some Lebowski controllers, and I think the Infineons) or they use a control algorithm designed to handle inertial loading well (eLifeBike stuff, Lebowski controllers) and so it's much less of an issue.

This is a wonderful explanation of the fundamental design objectives of the controllers and their suitability (or lack of) for use where inertial loads are substantial. It explains why the vast majority of off the shelf RC controllers do not perform well in ebike applications and maybe the differences in performance between ones designed for ebikes.

Sensor vs algorithmic designs. It seems the controller needs feedback to know how to control the application of power to the motor. I'm not even sure how they work without such feedback loops. Common sense seems to indicate more sensors allow controllers to be smarter and provide improved, if not optimized, motor control.

 

[fechter]

Each of the three phases requires a high-side and a low-side switch, which explains the 6-FET configuration. And I assume 12-, 18-, 24-FET etc. configurations just parallel the FETs to allow higher current. But how does 9-FET work? Does the high side pass more current than the low side, or vice versa? Or do the extra FETs provide a current sink for the inductance in the motor windings during some stage of the switching sequence?

It took me a while to figure that one out, but here's my theory:

During partial throttle, one bank of the high side FETs is switching on/off at the PWM frequency, which is typically 10-20kHz. When the high side turns off, the collapsing magnetic field will try to circulate current through the body diode of the opposing low side FET. Since the voltage drop of the body diode is much greater than the voltage drop across a FET that's turned on, it will generate more heat. By using 2 parallel FETs on the low side, the heat dissipation will be a little more even.

More advanced controllers will turn on the opposing low side FETs during the 'freewheel' circulation to reduce heating. This is called synchronous rectification. In this design, the heating of the high side and low side will be much more even (and far less than using the body diode). It's somewhat tricky to make the FETs turn on at just the right time to catch the freewheel pulse without having both high and low side on at the same time, which casuses instant destruction (shoot-through) as it shorts out the battery. Any controller that has field oriented control has synchronous rectification. This also makes regeneration during braking much more efficient.

 

[dnmun]

Here are two photos (front and back) of a speed controller. It was sent to me as a replacement for one to be removed from the inside of a Golden Motor (Magic Pie). Would anyone care to discuss the differences between this speed controller and the one already under discussion?
Side note: To replace this speed controller there is a lot of tedious soldering to be done. I've heard (don't know, but have heard) that GM has upgraded it so that it is an easier-to-replace simple plug-in.
By the way: I have no intention of installing this speed controller.

it is the same. all these controller designs are essentially identical. if you don't like soldering you have a really long row to hoe.

 

[ARod1993]

RC motors aren't, strictly speaking, more predisposed to cogging than hub motors are (the only exceptions being cases where specific hub motors have either their magnets or their stator slots set at a bit of an angle); they're all fractional-slot permanent magnet outrunners; my guess would be that the issue has more to do with the controllers. Hobbyking model airplane/boat controllers are sensorless designs optimized for non-inertial viscous loads; with such loads you can basically just start playing the trapezoidal commutation pattern at an arbitrary frequency (limited only by battery voltage and motor KV) and the motor will be able to snap to and follow along almost instantly. That's not really possible with inertial loads like a vehicle; if the controller commands a startup frequency corresponding to several mph and you're standing still, the motor will not be able to get you up to speed instantly, and if/when it loses sync with the controller it's going to cog and jitter and quite possibly make a number of very sad noises. When you use RC motors with controllers designed for vehicles (Kelly, Lebowski, Infineon/Xie Chang, some of the eLifeBike stuff), they're either sensored (Kelly, some Lebowski controllers, and I think the Infineons) or they use a control algorithm designed to handle inertial loading well (eLifeBike stuff, Lebowski controllers) and so it's much less of an issue.

This is a wonderful explanation of the fundamental design objectives of the controllers and their suitability (or lack of) for use where inertial loads are substantial. It explains why the vast majority of off the shelf RC controllers do not perform well in ebike applications and maybe the differences in performance between ones designed for ebikes.

Sensor vs algorithmic designs. It seems the controller needs feedback to know how to control the application of power to the motor. I'm not even sure how they work without such feedback loops. Common sense seems to indicate more sensors allow controllers to be smarter and provide improved, if not optimized, motor control.

The difference between sensored and sensorless is another interesting topic; in this case the sensors being referenced are specifically position sensors; they output various voltage combinations based on the position of the rotor. Sensorless controllers aren't being run without feedback; they instead have voltage and current sensors on the phases and they can tell roughly where the rotor is by watching for the zero crossings of the back EMF on each of the phases. Hall sensors are nice, but there are a few issues with using them depending on the application. A lot of RC motors don't come with Hall sensors attached, and aftermarket external Hall mounts like the ones on Equals Zero Designs are at least somewhat prone to damage depending on how the motor is installed; a friend of mine ran an Equals Zero adapter with the motor installed such that the sensor board had lower ground clearance than the main vehicle body, resulting in the sensors getting trashed every time he hit a curb or went through a puddle (although that specific failure mode is easily avoidable if you install the motor such that the place where phase wires exit the motor is oriented at 12 o'clock rather than six). Internal Hall sensors are better than that in some ways, but they're small and will most likely cook well before the motor phase wires do; a controller dependent on Halls becomes much harder to use if one or two sensors are damaged.

Trapezoidal sensorless controllers are out there, and not that bad, although a lot of the commercially available Chinese ones aren't of the same quality as Kelly, nor do many of them have a lot of nice features. They have decent robustness, and if you swap out the FETs and put a solder glob overtop the current sensing shunt then you can get 1 to 2 kW out of them for not a ton of money. On the higher end, there are a number of non-trapezoidal control setups that don't use sensors, the best of which (by my reckoning, at least) is the Lebowski controller. His system works reasonably well at low to no speeds; he puts a high-frequency current pulse through the motor at startup that gives him position information based on inductor nonlinearity. Once the motor is moving, the controller can keep track of the phase voltages and currents on the motor windings, which allows it to keep running track of where the rotor is and then continuously vary the phase currents and voltages to produce a smooth torque vector at all times.

 

[Njay]

When the high side turns off, the collapsing magnetic field will try to circulate current through the body diode of the opposing low side FET.

Called assynchronous rectification

Since the voltage drop of the body diode is much greater than the voltage drop across a FET that's turned on, it will generate more heat. (...)
More advanced controllers will turn on the opposing low side FETs during the 'freewheel' circulation to reduce heating. This is called synchronous rectification. In this design, the heating of the high side and low side will be much more even (and far less than using the body diode).

Much, much more heat. I have experienced the difference 1st hand in a selfmade controller, and is abysmal. I don't know why there are any BLDC controllers doing assynchronous rectification, the hw is basically the same as for synchronous, if not exactly the same. For DC motors you can replace a FET for a diode and control is simpler, but for BLDCyou already have the hw there.

It's somewhat tricky to make the FETs turn on at just the right time to catch the freewheel pulse without having both high and low side on at the same time, which casuses instant destruction (shoot-through) as it shorts out the battery.

Even in a controller with synchronous rectification there will be a tiny amount of time (the dead time) that the body diodes conduct, exactly because one must be off before the other turns on (Texas Instruments actually has a driver chip that takes a predictive approach and is able to prevent diode turn on by allowing a tiny amount of shoot through, which apprently still gets better efficiency; there's a topic on that technic from teh stork here in the forum).

 

[windtrader]

The term "sensorless" is misleading. I have read various methods for adding Hall Effect sensoring to motors. Seems the majority of the RC motors discussed are of the sensorless variety. Are there fewer sensored motors since most assume non inertial loads?

There seems to be two main controller camps. One camp designs controllers for driving RC motors with no inertial load. The other target running motors used in ebike applications where there is inertial load. Is there any consensus or experience base where ebike type controllers are used with RC motors or does this combination not make much sense since most RC motors are sensorless?