Armature Current Limiting

[safe]

Armature Current Limiting

For those who have been reading here for a long time I had originally called this technique "Motor Current Limiting" since I didn't know whether it was possible or had been done before. Now I know better and the standard phrase is "Armature Current Limiting". This idea refers to the practice of a controller limiting the amount of current that can flow through the coils of the motors armature in order to set a limit on heat creation. The standard controller limits the current on the battery side and due to the bizarre inductance effect that takes place in a motor this standard technique actually produces more current on one side of the controller than the other because of PWM. (Pulse Width Modulation) If you are new to this just trust me that "it's weird".

Okay... so that's a review of what the thing "is"....

 

[safe]

Fechter

We've discussed a programmable circuit and in the process of solving for that problem we used a Frequency-to-Voltage chip in the design to convert a square wave signal from a gear tooth sensor to a steady and smooth voltage.

So here's the thought...

If you really wanted to create a high quality Armature Current Limiting circuit then why not use a Frequency-to-Voltage chip to smooth the PWM current you get from either a shunt or a Hall Effect Current Sensor?

I'm thinking that for my Project #003 and Project #004 I'll likely be needing the most control and so I'd like to use the programmable circuit, but for my Project #002 I'm thinking I'd use the Armature Current Limiting circuit since I don't want to have to bother with modifying the motor to fit the gear tooth sensor. The Project #002 would likely be run with a 1200 watt motor, so my heat budget will be high enough to begin with that a little extra heat is okay. (if we ever got a race together I'd slap a 750 watt motor in the bike for race day and hope to not burn it up... the net power output for the two is about the same)

The Armature Current Limiting is a simpler thing to do because it doesn't require modification of the motor. (and might be "good enough" to risk it on the smaller motors) By the time we got done designing the Programmable Circuit it was pretty sophisticated... (maybe more that I want do deal with)

How much better might the Frequency-to-Voltage chip be compared to something simple like a capacitor and a resistor? Is the chip a luxury that is not needed? (your original design used just capacitors and resistors for smoothing)

Using a Rectifier and a Capacitor

 

[fechter]

The F to V converter is not really what you want. For measuring the motor current using a shunt or hall current sensor (preferred), the simple RC circuit will work fine. The PWM frequency is much much higher than rate at which the motor current will change (or needs to be controlled at), so it should be easy to get a good filtered signal.

There are other circuits that can provide a more accurate measurement, but those would be overkill, since the limit is going to be adjustable.

 

[Jozzer]

Don't controllers like this (http://www.alltraxinc.com/Products_AXE.html) already have armature current limiting?

 

[TylerDurden]

Don't controllers like this (http://www.alltraxinc.com/Products_AXE.html) already have armature current limiting?

Yep. Armature current limiting is common in other (non-ebike) applications.

But it might be possible to build some add-ons for cheaper controllers. They won't substitute for a properly sized motor, but it may help some folks keep from toasting an overloaded motor.

:?

 

[safe]

Bigger Motors Weigh More And Are Illegal

The central idea here is to get more top end out of a smaller motor. When you overvolt a motor it tends to raise the heat production across the entire rpm spectrum. With Armature Current Limiting the heat increases, but since you've crimped the low end heat down a lot lower than stock it reduces the heating that you would get with a normal controller.

To really understand the motivations for this you have to understand the Federal 750 Watt Law. In that law they establish what a "legal" electric bike is based on something very bizarre ... the heat rating of the motor. You are not allowed to sell an electric bike that passes as a "bicycle" unless it has:

750 watt rated motor (which effectively means a heat rating)
Operable pedals
20 mph top speed (as sold)

...in order to take a motor that is designed to handle a specific heat rating (thus giving it a rated load of 750 watt) and then increase it's power you need to make sure that you don't burn it up. Simple overvolting can burn up a motor, but a combination of overvolting and Armature Current Limiting can give the extra top end power while trimming the low end heat enough to keep the motor alive.

So while a bigger motor is "better" (because it can handle more heat) it's not "legal" and so we can't even consider it. Otherwise all the electric bikes would be built with Eteks or PMG 132's. (which are without doubt illegal)

This chart shows a classic case of taking a 750 watt motor running at 36 volts and 40 amps that is overvolted to 48 volts and keeping the 40 amp limit. The difference is that the 40 amp limit is now on the Armature side rather than the Battery side. You can see that you get a 40% increase in top end and it looks like you might take a chance and run a higher amp limit and still get away with it. (look at the red heat lines) Setting a higher heat (amp) limit would increase the power even more. Then throw in cooling techniques and... well you get the idea...

It would be nice if there was a moderately priced Armature Current Limiting electric bike controller. Then for $50 you could overvolt and switch to Armature Current Limiting at the same time. In many cases it would be a direct plugin where someone might jump from 36V to 48V.

250 watt goes to 350 watt
350 watt goes to 500 watt
500 watt goes to 700 watt
750 watt goes to 1000 watt

...which means a new "rating" effectively is established.

 

[safe]

Fechter's Armature Current Limiting Circuit

This is the Armature Current Limiting circuit that Fechter posted a long time ago. I'm thinking that this is so easy to do that it's worth it to start with this and work forward from here. Baby steps...

The positive qualities about this circuit:

Uses no motor power. (Hall Effect sensors are better than shunts)
Doesn't need a 5V battery. (uses the existing throttle voltage)
Is cheap. (maybe $20 by the time I'm done)
Is pretty simple. (good for circuit design beginners like me)

 

[safe]

Mounting?

Do you simply press the #4 and #5 leads up against the motor wire(s) without removing the insulation?

I'm guessing you more or less just tape this sensor to the appropriate motor wire and don't change anything. Which motor wire? Or do you use both motor wires? It seems like the way they designed it they've left it open as to how you can mount it. Or are these mounts really thick so that you can solder directly to them?

$5.54

 

[fechter]

You have to cut a motor wire and solder the wires onto the tabs on the sensor. Since motor wires are heavy and stiff, you don't want the tabs to break off the sensor. Normally, the sensor would be soldered to a piece of circuit board, along with the wires. That's what I would do. The mechanical stress from the motor wires would be transmitted to the board, not the sensor.

You could solder the motor wires directly to the sensor and maybe glue a piece of stiff plastic across the whole thing to make it strong.

 

[safe]

You have to cut a motor wire and solder the wires onto the tabs on the sensor.

So it's ONE wire... that's what I figured. You essentially need to CUT a wire and create a gap so that each end is soldered to the two leads of the sensor. And which one would I choose, the positive or the negative motor wire or does it even matter?

 

[fechter]

It won't matter. Whichever one is easier. The current is the same everywhere in the loop.

 

[safe]

So far I have:

http://digikey.com/scripts/DkSearch/dksus.dll?Detail?name=OPA2344PA-ND

Op Amp - OPA2344EATR-ND - $2.57

http://digikey.com/scripts/DkSearch/dksus.dll?Detail?name=620-1110-ND

Hall Sensor - ACS755LCB-050-PFF - $5.54

Any advice on the diode?

 

[fechter]

Almost any diode will work. A 1N914 would be standard. I'd rip one out of an old radio or some junk.

 

[safe]

Oooooh... 5 cents...

http://digikey.com/scripts/DkSearch/dksus.dll?Detail?name=1N914BCT-ND

1N914BCT-ND - $0.05

 

[safe]

Pot you can't Smoke?

http://digikey.com/scripts/DkSearch/dksus.dll?Detail?name=P3G7103-ND

P3G7103-ND - $1.84

 

[rkosiorek]

so the purpose of this circuit is to measure the current in the stator winding of the motor (variable up to 50A) and when the limit is exceeded the throttle signal is pulled low to reduce power going to the motor? That effectively this will act as if the throttle input had been reduced but you will still get power.

that would be unlike the shunt current limiter built into most controllers those cut the motor off abruptly. this would allow you to throttle up to full power but not exceed it.

do i have that right?

of course if the limit set t by this device is higher than the limit set by the internal shunt current limiter. the motor will shut down hard.

rick

after thought - would this not measure the same amount of current as the shunt, but just apply the limiting in a different fashion?

 

[safe]

It's New Years Eve and I'm up to beer number five, so if my posting strays a bit you have to forgive me. :lol:

Basically the core of the matter is that the standard controller limits the current that it sees coming from the battery side of the controller and not the motor side. But what does the motor side look like? Well, it's different because Pulse Width Modulation (PWM) creates a choppy flow of electricity that is a combination of the full battery voltage and the ability of the motor to allow the "bursts" to flow. The motor has an inductance behavior which is the ability of it to allow current to really move energetically. When the pulse is cut off the "inductance pipeline" becomes empty and that allows the extra voltage surge to flow more energetically than if there were a constant flow.

The best analogy I think is automobile traffic...

If the freeway is open then traffic can flow at 75 mph, but when there is a traffic jam things slow all the way down to 25 mph or less. The PWM allows the traffic of electrons to thin out and so you actually get more current flowing than you might expect. Some traffic systems have designed metering systems to accomplish this same idea.

The bottom line is that Armature Current Limiting COMPENSATES for all the effects that the Inductance makes and brings it back to a sort of "undeformed state" where the current going to the motor is a constant. (which also means that heat is a true constant)

Armature Current Limiting is the kind of limiting that MOST people think is really happening in a regular controller but isn't.

It took me about six months to understand it...

 

[rkosiorek]

so the next question is:

is this a continuation of Mr. Fechter's current based throttle design? as i remember it was left off with the suggestion of the use of a hall current sensor but the thread ended there.

rick

 

[safe]

...is this a continuation of Mr. Fechter's current based throttle design?

No.

The Current Based Throttle design makes the Current Limit something that is adjustable by the throttle. That's a nice idea, but very different.

Armature Current Limiting effectively eliminates the distortion that the motor Inductance causes. It's basically correcting a behavior that some people find desireable because excess current flow means excessive torque in places that you normally wouldn't expect torque to be. In fact, if it weren't for this "defect/behavior" hub motors would never be practical because they rely on extra torque showing up at low rpms in order to get their motors moving. This "defect" has a price in that the motor heats much more because of it.

The central reason for doing Armature Current Limiting is when you have a small motor and you want it to behave like a bigger motor without overheating.

 

[safe]

However...

What if you took the Inductance correction of Armature Current Limiting and COMBINED it with the current limit resetting charactoristics of a Current Based Throttle?

That's a new idea!

Normally the Current Based Throttle resets the current limit in the same manner that a standard controller would do it. If you crack the throttle wide open you would (like with a standard controller) get a current limit that created monstrous amounts of motor heat.

However, if you combined the Armature Current Limiting into the design when you cracked the throttle wide open at LOW RPMS it would only deliver as much current as would be permitted to reach the absolute armature current limit cap. At the higher rpms the "undeformed" current limit rises so you still get the powerful top end. This would actually make the Current Based Throttle far more practical because it would provide a little better protection in the low rpm area against wild current surges. In the case of full throttle all the time it would become the same as an Armature Current Limiting circuit.

This is kind of a nice idea...

Fechter... when you get over your hangover take a look at this one...

 

[fechter]

Alcohol shoud be used as a motor fuel, not a drink...

I think it would work great to do the current based throttle and just put the current sensor on the motor side instead of the battery side. This way you would get a true torque based control.
The testing I've done with the sensor on the battery side shows that you can really shave off those nasty current peaks on startup with a little wirst control. With the sensor on the motor side, it would behave in a similar manner, but cut the low end amps even more.

In this configuration, the motor torque would be a function of throttle position. This would behave much like an ICE. Driveability would be good.

 

[safe]

I think it would work great to do the current based throttle and just put the current sensor on the motor side instead of the battery side.

Well I agree... it should work really well.

Do you have the circuit posted somewhere that we could look at again?

An argument could be made to make a Current Based Throttle and combine it with the Armature Current Limit idea since there are many throttle positions with the Armature Current Limit circuit that gives "does not matter" results. In other words with ACL there are times when full throttle or 3/4 throttle will produce the same actual outcome. Using the CBT you will always have some actual effect on the motor when you vary the throttle.

Combining these two might be a really nice design...

However, for the really extreme cases of overvolting and overheating of smaller motors the programmable circuit discussed before still seems the ideal. I was looking at the AllTrax programmable software and they have:

Linear Ramp Rate
Progressively Increasing Ramp Rate
Fast Rise, then Flat, Then High Peak
...and a couple other shapes.

So the idea of messing around with the way that the power comes on is obviously something that is being taken seriously.

 

[safe]

Current Based Throttle

Armature Current Limiting

I'd like to see the Hall Effect version redone as a combined Current Based Throttle and an Armature Current Limiting circuit. It can't hurt. There needs to be a maximum current limit knob, but other than that it should be possible to not need any other knobs. Since the Hall Effect sensor gives a more precise voltage to compare against it should be possible to not need so much trimming to get it right.

Just by looking at the circuits the biggest difference is that the Armature Current Limiting circuit sort of just attaches itself to the throttle signal and modifies it or let's it go unchanged. With the Current Based Throttle it looks like you always change the outcome of the throttle signal by forcing it through the circuit... so it's has a higher degree of "intervention" in the behavior than the ACL circuit does.

Armature Current Limiting is a passive circuit.

The Current Based Throttle is an active circuit.

 

[safe]

Study, Study, Study...

I've been studying the idea of the Current Based Throttle being applied to the Armature Current Limiting circuit idea. The chart shows some interesting things and take your time trying to understand it's implications. Basically the highest OVERALL efficiency occurs at somewhere around HALF throttle. In this case that translates to a 20 amp current limit using the ACL side limitation technique. If you use FULL throttle you lose a little efficiency at low rpm (which is to be expected) but up at high rpm it gives the best efficiency that is possible. However at 1/4 Throttle you get a rather strange result... the peak efficiency NEVER ARRIVES because at 10 amps the motor is STARVED for energy and delivers sub optimal performance. However at low rpms the efficiency goes UP and is quite good. This fits the way someone might actually want to ride because when you are just "putting around" you tend to run a lower rpm, but when you are going flat out you tend to run in the higher rpm.

It's an interesting system...

The biggest difference is that the Current Based Throttle would always seek to achieve the maximum rpm no matter what throttle setting you use. You end up adjusting the current limit and not the voltage.

 

[safe]

Not Such A Good Idea?

Comparing a "regular" Armature Current Limiting (ACL) circuit with an ACL circuit with the Current Based Throttle (CBT) added doesn't show an improvement at 1/4 throttle.

Sometimes "K.I.S.S." really is the better way to go.

Looking at the chart we see that not only can you get more power at lower throttle settings (which translates into a more useful throttle) you also get better efficiency down there.

Basically the CBT theory falls apart below the optimal current limit for the motor.

You would need to figure out some way to have a lower threshold where the current limit could never drop below some point and then any lower throttle setting would translate into a lower voltage setting and NOT a lower current level.

Comments from Fechter?

Do you see my point on this one?

(it makes that weird AllTrax dual ramp rate profile make some sense... if you start with a high level current that equals the minimum current level then ramp up to the peak then you would be better off than with a pure Current Based Throttle)

So the logic would have to go:

From No Throttle Up to the Minimal Current Limit would involve an increasing voltage.

From the Minimal Current Limit up to Full Throttle would involve the current based throttle and the fixed maximum voltage.