Why part throttle is so bad??

[chichisport]

Hi ,

I´m trying to figure why part throttle is so bad efficiency wise . I need 215 watts at 400 rpm and also 400 watts at 2300 rpm with the same motor.

Using the formula :

Power(Kw)= (Torque(Nm)*rpm)/9550

The figures are:

215 watts 400rpm 5.13Nm
400 watts 2300rpm 1.66Nm

Torque is quite linear to current so the difference is substancial, looking at a plettenberg datasheet I found roughly

6v 35 amps for 400rpm loaded kv 66 eficc 80% of max 85
30v 13.3 amps for 2300 rpm loaded kv 76.6 eficc 80% of max 87

I´m that mad to think about lowering so much input voltage and pushing up the amps .I think the only thing needed is good chunk of oversized controller or a previous sinusoidal dc/dc converter in series to a normal sized controller.

All type of help would be greatly appreciated.

Juan

[JetPlaneFlyer]

Juan,
I'm no electric expert but it appears to me that you are making some fundamental errors..

For a kick off for steady state operation the torque produced by the motor must be matched by the torque absorbed by the prop. Torque on a prop is due to drag and drag is proportional to velocity squared. Torque required must therefore generally rise to the square of the rpm. According to your figures you actually need much less torque to spin a prop at 2300rpm than you do to spin it at 400 rpm, which must be incorrect.
Looking at the power calculation, if you really did need 5.13Nm to spin a prop at 400rpm (must be some size of prop) then you would need 2300^2/400^2 x 5.13 = 169.6Nm to spin it at 2300rpm. Using your formula this would mean power = 40.8Kw (not 400w)

And finally your motor is drawing less current at 30V than it does at 6V... again this must be wrong unless you are comparing different motors. I would expect to see current increase with increased voltage.

Steve

[chichisport]

Well, the fact is when you go high (a lot) the air is thinner , the plane needs to go faster (more rpm) but the prop also sees thinner air and then the torque is lower but not the power.
We can discuss it further if you want but I think I have a proof for that.

By the way ,the prop seems to need to be quite large to be extremely efficient ,so the low rpms at sea level.

I´m looking into car drives as the case is similar to a car going uphill slowly vs highway cruising , power wise.

Thanks.

[JetPlaneFlyer]

Juan,
If you go high then yes you are correct, drag and therefore torque will fall... But you would have to be VERY high indeed for torque to fall to your predicted values, certainly too high for conventional RC. My rough calculation is that (assuming the 5.13Nm @ 400rpm figure is at sea level) you would have to be at an altitude of about 100,000ft before torque dropped to 1.66Nm @ 2600rpm.
If I'm not mistaken this would break the world altitude record for horizontal flight in a winged aerodynamically lifting aircraft currently held my NASA's Helios. The Helios had 21kW of power; a little more than your predicted 400W but there again the Helios was pretty big at 247ft span. NASA's budget was also probably a little larger than yours?

Altitude would not explain why the motor draws less current at 30V than it does at 6V????

Perhaps it would help you to get meaningfull answers if you explain what project you have in mind?

Steve

[chichisport]

You are not too far of design considerations :eek:

In any motor rpm follows voltage and torque follows current. High voltage high rpm ,low current low torque.

They call it Kv and Kt :D

The QUESTION is quite simple , what´s its the best way of getting rid of excess voltage :
1 DC/DC converter
2 motor controller
Both are basically the same but I´m a little uncertain in details.

[vintage1]

You are not too far of design considerations :eek:

In any motor rpm follows voltage and torque follows current. High voltage high rpm ,low current low torque.

They call it Kv and Kt :D

The QUESTION is quite simple , what´s its the best way of getting rid of excess voltage :
1 DC/DC converter
2 motor controller
Both are basically the same but I´m a little uncertain in details.


Well now I understand the QUESTION...:D

Look, the practical issues of throttling an outrunner are many and various, and suffice to say, the best way to do it is the way everybody already does.

By chopping the current to - if you like - lower the average torque, and therefore only run at lower RPM. BUT the real question I think you are asking, is 'why is that so inefficient'.

Ok, its a long story.

We COULD in principle use an analogue type linear regulator, but that is TERRIBLE in that all the lost volts end up as heat.

So we use time division - PWM - to throttle. Now in a perfect world with perfect components, that would be perfectly efficient. Our switching devices would be either off, when they cant get hot. Or fully on, with no voltage across them. So no heat their either.

But the components are not perfect. There is always resistance - in the motor, the wires, the pack, and the switching FETS. So you do lose power when they are ON.

The question is, why do you lose MORE power at part throttle than at full?

Cue slightly technical explanation. IF a motor is running at a given speed, it generates a back EMF. In fact what limits its speed is when the back EMF matches the voltage that is applied. At full throttle that is the pack voltage, and the motor slows just enough under load for the back EMF to drop to compensate for the voltage lost in the resistance of all those elements in the chain.

So far so good, but consider what happens at say half throttle. The motor is only generating HALF the back EMF...and so when you give it a throttle pulse, the current through the circuitry is MASSIVE. Its only limited by the resistance, which is fairly low, and the inductance of the motor windings and wires, and, indeed, there may well be parasitic capacitance around in the motor and the ESC that make peak currents even higher..

So whilst your AVERAGE current draw is maybe only a 1/4 of full current draw at half throttle, your PEAK currents may be several times higher. Since power loss is the square of the current, times the time it is on, you can see that even in the simplest of cases, whilst twice the current for half the time is the same average power in, the losses are twice as high .. So that's why at the least, its always better to lower the pack voltage than run permanently throttled.

When you then look at the case of normal throttling, its very complex.

When the power devices first switch on, there is a massive current spike as all the capacitive elements get charged up. That's a short spike. Maybe 100ns or so. Not too bad, but a 100A 100ns spike is a great broadband transmitter..so generally that's why there is another capacitor on the power side of the ESC, so that can dump ITS power into the ESC, and prevent that spike running up the battery wires and screwing the receivers etc.

Once that initial spike is over, we now have the situation where full pack voltage is across a motor winding. A winding that is only generating about 50% pack back EMF. However the winding has inductance, so the next phase of current is a bit lower than the resistance might make you think, and the current starts rising again as the inductance field gets 'charged' . In this sense motor inductance is a Good Thing, as the more there is, the lower the peak currents in the MOTOR wiring and motor will be, and trying to average out currents over the cycle gets us as near best efficiency as we can.

But guess what? the better the motor is, the lower its inductance will be. Take a cheap ESC that works happily with a cheap chinese motor and put it on a dead good Steve Neu special..and BANG. Throttling GOOD motors is harder than throttling bad ones. You need a lot more current rating in the output devices than full throttle current, and a lot less resistance. And a very good capacitor on the input to try and contain the high currents within the motor/ESC loop. Such motors benefit from a higher switching frequency, as you trade off more capacitive losses for less inductive related ones.

Timing also makes a difference. Since the back EMF is not constant..it builds up and it dies away..IF you arrange to adapt timing so that you are only for example applying part throttle current at the time when the back EMF is highest, you will also limit the peak currents.

Now no manufacturer does this, but its possible. As is also artificially adding inductance to motor leads with toroids in the phase wires. I have been meaning to try that out to see if it makes a difference.. but never got a round tuit..

You see, by and large we are not hugely concerned with part throttle efficiency. As long as the ESC doesn't go up in smoke, most people don't care. And, if it does, get a bigger one. In the end all these issues reflect in resistive losses..so big fat wires and big fat FETS will reduce those as well..and that's what is in quality controllers. And motors.

Its probably true to say that using a well overspecced controller will be more efficient at part throttle whatever. But it adds weight. As does using a bigger motor.

I am a great fan of using big motors run well below max ratings. More efficient.

[chichisport]

plas,plas,plas..bravo ,bravo :D

Sincerely ,I´m a big fan of your post .With that one I´m devoted.

In my ignorance I have arrived at nearly the same conclusion . Big oversized motors and esc is almost always a good thing.

The previous example is around an orbit 15 datasheet ,pretty big for 400 watts ,its not its best point but matches both requirements at a good 80 %.

I will look into Tango´s as even Steve Neu call them the more efficient motors,not the most powerfull for the weight .It will be far more easy if they draw the real performance like plettenberg do.

Thanks again.

[chichisport]

The kontronic controllers are something in they own range, in a heli I have a prety good example .
With 6s lipos esc cool , motor 50C ; 10 s a123 esc cool motor 60 degrees(a lot of part throttle when the current is low,thanks to the governor).

[chichisport]

Finally with the aid of a marvelous and expensive tool ,I got a true answer to the how bad? the Zimmer lmg 450 is a 4 channel power analizer with some smart gadgets built in. Obviously it comes at a huge price:
http://www.zes.com/products/lmg450_e.html

Putting 3 channels in the motor wires , one in the input wires and using the scripts we will be able to see directly the efficiency of the controller at different ranges.
Worst 90% at 1/5 throttle with 20v 12A
2ND 95% NO LOAD (2AMPS) full throttle
Best 98.5% full throttle 20v 20A

At the worst situation a Jazz 55-8-32 would thermal in 6 minutes at 95C , added a heatsink and after 20 minutes the temp would not go over 70 C . No forced cooling in the test.24 watts to be radiated!!!!
The motor,a neu 1910 1Y did not go more than 30C with his internal fan , I´m really interested in building a eddy current brake to know the real efficiency of the system in that condition ,I estimate 84*90= 75.6% efficiency.