+ + + The Problems with Overvolting + + +

[safe]

Overvolting

Overvolting is generally a good thing to do.

However, there are some problems that occur that need to be looked at so that people don't try to overvolt with disastrous consequences. In order to improve the presentation "experience" I'm going to load all the charts here first and then reference them as I go.

 

[safe]

Power Output - 30 Amp Controller

I'm starting here with a small 250 watt motor that normally runs at 24 volts. From there I'm increasing the voltage all the way up to 96 volts in three 24 volt steps. The result is an increase in power that is very large. The "naive" person might stop here and say "wow, my machine has more power, this is great" and begin to ride without knowing the likely fate he faces. So this chart is of the increase in power that overvolting can bring. This is the "good news".

 

[safe]

Efficiency - 30 Amp Controller

This is still more "good news" in that the efficiency of this motor actually improves with an increase in voltage. The "naive" person is thinking "wow, this is great!"

 

[safe]

The Bad News...

Heat Losses - 30 Amp Controller

Isn't it always true that there's always a trade off in some way when you do anything? In the case of overvolting the trade off is a nearly equal rise in motor temperature that equals the rise in power output at the lower rpms. This is bad... very bad... in fact unless you have a very massive motor that can act as a heat sink you will in all probability burn up this motor when you find yourself in a situation of low rpms and high load. This might happen if you were to climb a steep hill. So this heating problem needs to be taken seriously if you have a small motor... like the 250 watt motor I'm starting with for this example.

 

[safe]

Attempt at a Remedy...

Power Output - 96 Volt - Limited

So here's an idea...

What if you lowered the current limit on the higher voltage configuration?

Then you could still get some of the increase in power and get the heat down a little. Okay, so let's do that, we take the 96 Volt configuration and then run it at lower current limits like 20 Amps and 10 Amps and compare it to the initial 30 Amps we started with. We see that the power output looks like this..

 

[safe]

Heat Losses - 96 Volt - Limited

Now we look at the heat losses sitatution and see that if we go all the way down to 10 Amps we get back to right around the same place as where we started as far as heating related risks. It appears that if you went all the way up to 96 Volts that a current limit of 10 Amps would give you heating that is about the same. (and maybe even a little bit better)

But before we go on...

Take a quick look up at the the Power Ouput of the 96 Volt at 10 Amps in the previous posting. At 10 Amps the power isn't all that much better. In fact we've gained almost nothing doing this because the low end needs to be limited so heavily in order to prevent a meltdown of the motor.

Hmmmm.... could things get any worse?

 

[safe]

From Bad to Worse...

Efficiency - 96 Volt - Limited

Something else is going on at this extreme voltage level. All permanent magnet motors have a certain charactoristic such that they have a "built in" ideal peak efficiency CURRENT for any given VOLTAGE. If you use a current level above the "ideal" you lose efficiency, but also if you use a current level below the "ideal" you also lose efficiency. So at really high voltage and really low current the efficiency actually goes into decline. We've done all we can with overvolting... and it does have it's place... but it sort of hits a wall of effectiveness because of this efficiency collapse problem. The best you can do with overvolting is to align your peak efficiency and your peak power to be one and the same. That's as far as you can go with overvolting.

And we still never really solved the heating problem... if you modestly overvolt and stay within the peak efficiency "ideal" constraints you still have that nagging heat problem that rises as the voltage rises.

 

[vanilla ice]

So the other axis is motor rpm?

 

[TylerDurden]

MCL is a way to get past the overvolting problems for a small motor that has limited heat dissipation abilities.

Only if the current limit is within the motor's rating.

 

[safe]

So the other axis is motor rpm?

Yes.

However, it's actually "scaled" so as to shrink the higher voltage down to size so that you can make direct comparisons. Without the "scaling" you lose the ability to see how the "shape" changes and can easily get lost in the data.

It's the shape that matters...

Think of it this way. If you start with a 24 volt motor that is geared to the rear wheel at a ratio of 1:1, then the other voltages would map to:

24 Volt - Gearing - 1:1
48 Volt - Gearing - 1:2
72 Volt - Gearing - 1:3
96 Volt - Gearing - 1:4

If you applied those gears to the motor then you get the charts as presented with the bottom axis presenting "final gearing" rpm. You would normally be forced to gear down anyway.

You brought up a very good point that I didn't previously discuss...

 

[TylerDurden]

if you modestly overvolt and stay within the peak efficiency "ideal" constraints you still have that nagging heat problem that rises as the voltage rises.

Heat increases due to increases in current, not voltage.

 

[safe]

Heat increases due to increases in current, not voltage.

PRECISELY

And one of the nasty byproducts of our controllers is that "current multiplication" occurs. The more you overvolt the motor, the more of a "current multiplication" that takes place. At very high rpms "current multiplication" is not a factor, but at less than peak power the heat increases in proportion to the overvolting!!!

You're getting it... heat is created when you overvolt because of the "current multiplication" and not the voltage directly. That's the real problem. And lowering the standard controllers current limit as a way to try to remedy this problem only creates different problems because the efficiency drops when you lower the current below the "ideal" value. (the "ideal" value being based on the current that occurs at the peak efficiency rpm)

On a small motor that has limited heat dissipation capacity overvolting can only go so far... then you hit a metaphorical "wall".

 

[Doctorbass]

Your post is very interesting safe!

It appear that Crystalytre should develop the X6 serie with high temperature supraconductor option :lol:

I worked on physics department of university of Sherbrooke where research about that is done.. it seems that supra at higher temp then 200Kelvin should be more availlable soon...

I just imagine an ebike with liquid nitrogen or simply carbonic dry ice to cool down the motor.. lol.. the efficiency would be maximum.. exept for the wires between the controller the motor and the battery...

I need to found a tesla guy somewhere!..

doc

 

[safe]

Your post is very interesting safe!

Thanks.

 

[dirty_d]

you're still not getting it safe, current division on the battery side not current multiplication on the motor side, nothing magic happens, the motor and controller obey ohms law, the motor current = (controller output voltage - motor backemf) / winding resistance. the smaller divided current on the battery side approximately = motor current * PWM duty cycle.

 

[safe]

...you're still not getting it safe, current division on the battery side not current multiplication on the motor side.

"Current Multiplication" was a name I first heard Fechter use.

I'm not wild about the name... but as long as everyone is aware that the battery current and the motor (armature) current are two separate things then thats all that matters.

The controller measures "whatever" current comes out of the battery. It doesn't know what's going on downstream in the motor and has no idea what relationship exists. All we need to know is that compared to what the controller "sees" (a constant current on the battery side) there is an increasingly larger current at lower rpms on the motor side.

It's this relationship that gives us "excess" current at low rpms and "excess" heat because of it. This "excess" forces limitations on overvolting because most motors have heat limitations. You can only overvolt as much as the motor can handle the extra heat.

 

[safe]

Hitting the Overvolting "Wall"

There are two areas where our little 250 watt motor will run into troubles with overvolting:

The first area has to do with the high rpms that the motor produces at the higher voltage. These high rpms start to test the limits of how well built the motor is. Failure due to bearing overloading and brush limitations start to pop up. The chart below shows the rpm that the motor needs to be able to run at in order to attain it's "peak efficiency" and this is actually lower than the maximum rpm.

The second area that can get you into trouble is the increasing need for a higher current limit in order to satisfy the minimum "peak efficiency" current demands. If you use a controller that does not deliver the necessary current limit at the higher voltage then you start to see the powerband lose efficiency and it's peak power begins to shift to lower rpms. The net effect is that you are forced to run lower and lower into the high heat areas and that's going to stress the motor. So you never would want to set a current limit BELOW the "peak efficiency" current.

Realistically the maximum level of overvolting for a typical 24 volt 250 watt motor is probably going to be about 48 volts...

 

[fechter]

What's the peak efficiency *efficiency* at those voltages?

 

[safe]

What's the peak efficiency *efficiency* at those voltages?

Unite Motors are all around 80% at best... (they claim 78% or better) The voltage doesn't have an effect on the efficiency peak since that's mostly dependent on the batteries basic resistance. If anything I'd figure that losses would increase with higher voltage because you are going to have higher rpms. Windage losses would go up and bearing losses would go up. There are probably other losses that make the higher voltage not deliver on it's full promise.

 

[fechter]

Yes, the efficiency could really go to crap at very high voltages due to higher core losses and windage. At really low voltages, it would also go to crap because it would be running near stall. Somewhere in the middle, the efficiency would peak. Probably somewhere near the rated voltage.

 

[cadstarsucks]

you're still not getting it safe, current division on the battery side not current multiplication on the motor side, nothing magic happens, the motor and controller obey ohms law, the motor current = (controller output voltage - motor backemf) / winding resistance. the smaller divided current on the battery side approximately = motor current * PWM duty cycle.

Actually I think the confusion comes from the *battery* monitors and their users equating the average battery current with actual motor current.

A 30A battery current will burn out a 30A motor at low speed and high load. I do not know how the controllers are working, but the thing that needs to be maintained is the motor current and reading the battery current gives you an idea about battery life but tells you nothing about motor current unless you can do a lot of math on the fly: max speed/speed * battery current.

Dan

 

[safe]

The fancy programmable controllers like the Sevcon have an adjustable setting called "Armature Current Limit" which sets a limit to how much current reaches the motor. The typical controllers we use only pay attention to the battery current and fail to be aware of the armature (motor) current.

The biggest problem with overvolting isn't the voltage, but the twisted way that our cheapo controllers measure the current limit. If we could increase the voltage without suffering a corresponding "accidental" armature (motor) current increase then we could overvolt without any heating problems.

That's why MCL (motor current limiting) is really more like a "correction" than a new idea. It's just a way to be sure that what you get for a current limit at the motor is really what you asked for...

 

[Roy]

Safe,
Can you translate some of these findings
into practical (for me) terms?

I have #408 crystalyte that I'd like to run at 60v instead of 48.
The change involves a fair bit of work: mounting the battery, upgrading
the controller, another charger, wiring, etc. Will it be "worth it"?

Safe,
Seeing your chart re the change in frequency of maximum efficiency
for the model 250watt motor:
48v == ~6100 rpm
60v == ~7800 rpm

Well a typical RPM for the 408 is 300.
at which point max power has, I think, already begun
to decline. So my seat of the pants guess is
that peak for #408 at 48v is about 20mph and
260rpm.

So I'm asking, how much you think the rpm of maximum efficiency
of a #408 would increase when batteries go from 48 to 60 volts?
Would you expect significant increases in controller losses?

I have been thinking that higher voltage would mean I would
be able to produce the same power (eg sustain cruising speed)
with lower resistive loses and lower battery current (5 bats instead
of 4). Bottom line would be getting bike with nice increase in
battery life for same rate of travel (cruising at 19-21 mph).

Comments?

Thanks,
Roy

 

[fechter]

I have been thinking that higher voltage would mean I would
be able to produce the same power (eg sustain cruising speed)
with lower resistive loses and lower battery current (5 bats instead
of 4). Bottom line would be getting bike with nice increase in
battery life for same rate of travel (cruising at 19-21 mph).

Comments?

Thanks,
Roy

If you increase the voltage by adding a battery to the existing pack, then yes, you would see an increase in battery life for the same rate of travel.

You will, however, find it nearly impossible to resist the temptation to go faster

 

[safe]

So I'm asking, how much you think the rpm of maximum efficiency
of a #408 would increase when batteries go from 48 to 60 volts?
Would you expect significant increases in controller losses?

This is an easy one...

Your peak efficiency rpm increases in direct relation to the voltage so:

Starting Voltage = V(a)

New Voltage = V(b)

Starting Peak Efficiency Rpm = Rpm(a)

New Peak Efficiency Rpm = Rpm(b)

The formula would be:

Rpm(b) = ( V(b) / V(a) ) * Rpm(a)

So for your case:

We assume.... 260 rpm, 48V to 60V

And we get:

Rpm(b) = ( 60 / 48 ) * 260 = 325 rpm

So your new peak efficiency rpm would be 25% higher than before because your voltage is 25% higher than before.

The only possible exception would be in a case where you overvolted so much and kept the current limit very, very low and the motor basically wasn't allowed by the controller to pull what it needed to run at peak efficiency. In the real world that doesn't happen much, but if for example you bought a PMG 132 and tried to run it with a 40 amp current limit at 48 volts you would actually be choking the poor thing because it wants 75 amps to breath normally. (it's a big motor)

As for the controller... they are usually about 95% efficient, so I wouldn't worry too much about that. However, I'm sure that products can vary a bit.