Disadvantages of higher Battery Voltage

hias9

1 kW
Joined
Jul 11, 2018
Messages
422
I used to think that two performance disadvantages of higher battery voltages would be:
- Heating up quicker on lower speeds
- Less range on lower speeds (compared to a same Wh lower voltage battery)

However when I take a look at the ebikes.ca Motor Simulator, I cannot really find a performance disadvantage of higher voltage (when comparing to a same Wh lower voltage battery)

According to the simulator, when maintaining a for example 30kph at 9 percent uphill, both systems in this example would need around 1000W motor power and around 1250W battery power.
Here is the link to this example: http://www.ebikes.ca/tools/simulator.html?bopen=true&motor=MX4503&batt=cust_72_0.05_12&motor_b=MX4503&batt_b=cust_48_0.05_18&cont=cust_150_350_0.03_A&cont_b=cust_150_350_0.03_A&hp=0&hp_b=0&wheel=20i&wheel_b=20i&throt=12&throt_b=18&grade=9&grade_b=9

The 1250W battery power would be 72V*17.4A for system A and 48V*26A for system B. However the phase amps would be exactly the same (around 40 phase amps). Heating and range (of two batteries with same amount of Wh) would be the same.

So is there any performance disadvantage of a higher battery voltage?
 
The biggest disadvantage I can think of is that there is not a big selection of 72V battery packs that are ready-made for sale. If you are building a pack from scratch, then you can choose any voltage you like, any cell brand/model, and any shape...

48V / 52V ebike battery packs can power a common inverter in a power outage. Not a big deal, but it's an option.

Finding a BMS that you like is hard, for a 20S pack it's even more difficult. (I don't know if it's possible to use two of the common 10S / 36V BMS's)
 
Thanks for the reply!

Yes, there are practical disadvantages. Also another limitation would be the voltage limit of the controller.
But the question was meant to be more theoretic if there are any performance disadvantages.

Because many people say that using a higher voltage will cause excessive heat on low speeds, but according to the simulator this is not true.
 
I would think it would be the controller that would potentially lose efficiency (and get hot), as it's what does the conversion from battery power to motor phase power. I know my controller heats up way faster at low RPM high load than it does pushing the same battery amps at higher speed. Whether it's enough to have a meaningful impact on overall system efficiency is another question.
 
There is an advantage in achieving the best power with a controller, to come as close as possible to the voltage limit of the mosfets. Then, since the lower resistance mosfets are limited to 100v, 24s is making the optimal voltage to obtain the highest power per overall size and weight.

Going above 24s, has some serious disadvantages because the mosfets that are rated higher than 100v, are much higher resistance. Thus we need to build bigger and heavier. Some of the effective power is lost in size and weight, so is the expected efficiency of many of the bike’s components. So, above 24s, we are entering an escalade of size, weight, money...
 
If you give a motor voltage way above what it was designed for, it might arc through the insulation of the windings. We are talking 2x+ here.
 
Good points. But is it really a good idea to run 100V mosfets at their limits (24s)?
 
in general
for given like 1000W drawn from battery regardless which controller you use = the higher the voltage the less current for given wattage.
clear adventage
speed of DD motor in general depends on voltage.
 
Disadvantages

Cost, size\shape for placement, weight if there is more Wh.
New controller with appropriate LVC, Capacitors.
Beefier wires if amps increase.
Appropriate Battery Management System.
Obtain a new charger.

*

36V 10Ah pack with a max discharge of say 40A and the controller's max 40A.
72V 5Ah pack gives the same Wh, now with two options
1) 72V 5Ah 40A w\cont 40A
2) or use 72V 5Ah 20A w\cont 40A wouldnt be good b\c the cont asks for 40A (at a certain point in the twist of the throttle) no matter what the voltages are and would heat up the battery.

*

You make a good point there, I was thinking the extremes of 1V and 1000V or 200V, but a motor can run on any voltage, but as you mention any voltage safely and without destroying the windings is the key.
Tommm said:
If you give a motor voltage way above what it was designed for, it might arc through the insulation of the windings. We are talking 2x+ here.
 
amps lowers when you increase voltage for the same wattage drawn,
amps = heating.
thinner wires NO " beefer wires..." when you increase voltage
 
miro13car said:
amps lowers when you increase voltage for the same wattage drawn,
amps = heating.
thinner wires NO " beefer wires..." when you increase voltage
Correct. However, if that higher voltage requires that you run your controller at a much lower pulse width to achieve the same winding current, then you will see higher losses in your controller (specifically higher peak currents and higher switching losses.) So your controller will tend to run hotter if you keep the same motor and go to a higher battery voltage.

Also, if the higher voltages allows you to draw more current (i.e. same motor constant, more voltage available = more current can be pulled with a wide open throttle) you'll need the beefier wires as well.
 
Controller losses are generally very small compared to the motor losses so running a higher voltage won't make a real noticeable difference. Chances of blowing up the controller are greater at the higher voltage though.
 
The majority of current in the controller FETs is motor current. Running higher voltage will reduce battery current but will not change motor current, so higher voltage won't reduce controller heating from current significantly.

As was mentioned, running higher voltage can increase controller losses from switching, but perhaps more importantly if higher voltage FETs are required they have higher resistance and so their losses go up dramatically (i squared times R), especially since they are carrying the motor current, which is higher than battery current and the main source of controller heat.
 
hias9 said:
Good points. But is it really a good idea to run 100V mosfets at their limits (24s)?

I run 4110 mosfets above 100v when my battery is fresh charged, because I charge them lipo 4.25v per cell, sometimes as high as 4.35. I’ve been doing so for years. 18 X 4110 controllers running 24 s up to 105v, and seeing short 25+ Kw acceleration bursts daily. Of course I have had controllers damage, and sometimes they need mosfets replaced a few times before I have a reliable set that last. Yet, once I have one reliable, it is usually good pretty long. I always have a few on the shelves anyway, like all other components and service parts for all of my bikes. That is a basic requirement to ride performance bikes, with strict maintenance schedule of course.

I sm not saying it is the best for everyone to run 24 s, but it is part of the optimal power to weight ratio that you can build a bike.
 
Alan B said:
Following advice of exceeding ratings of components is a poor approach to building a reliable system. If you like systems that break without warning, go for it. For reliable transportation follow a more sensible approach.

Agreed. I akways have a bike that is more conservative for comfortable long range, reliable transportation. Sport is another story, where components are pushed to the limits and frequently modified or replaced. Performance always have a high cost in parts and maintenance. It is the same with any other kind of motor sports. Yet, when I was riding motorcycles, the inventory was much more expansive and the labour much longer.
 
Alright, either I'm taking crazy pills here, or some eminent experts in the field have replied, and have failed to noticed that to get his simulation to work, he throttle limited both bikes.

Higher voltages do not cause motors to heat up faster when you're throttle limiting them.

72v @ 12% throttle = 8.64v
48v @ 18% throttle = 8.64v

Coincidence? I think not. The motor is seeing exactly the same "resultant" voltage from PWM because of the throttle setting. From the motor's point if view, it can't tell what battery is attached.

Try the same thing with a load that exceeds the capability of the controller (with throttle at 100%), and see which motor heats up faster:
 
Luke is incredibly busy, and I don't want to speak for him. However...he has stated many times over the last few years that he prefers lower voltages and higher amps. The lower the voltages, the better the efficiency of the components, which has fewer waste-heat losses. Of course, as Fechter just said, controller losses are small either way.

One of the things that held back using lower voltages in the past was the limited selection of available hubmotors. If you wanted high speed, you had no choice, you had to use high voltage. Now we have many choices. Luke mentioned the low resistance of thick conductors found in low-turn count / high Kv hubmotors. He also mentioned how BMS's were often a weak link in the system, and a frequent source of trouble. By using a lower voltage, we have fewer BMS channels, with each channel being a potential source of trouble. 36V is 10S, and 72V is 20S.

So a pack made from 200 cells could be 10S / 20P, or 20S / 10P, with the difference being that the 10S BMS would be half the size, and possibly much cheaper. Would it be more reliable? Who knows, but...20S BMS's are rare, and there are hundreds of 36V BMS's to choose from. 24V / 7S is an option, if the controller and motor can take 100A, and you only need 2400W (as an example).

The Zero motorcycle uses 105V / 28S with a single stage from the motor to the rear wheel, and the air-cooled motor is fairly large. It works great.

The Alta is a dirt bike, and it is lighter, but also more expensive. Alta uses 355V (no P-connections on cells, only series). High volts, low amps. The motor is small (light) and high RPM...plus liquid-cooled. It also has a high internal reduction before the power reaches the outer shaft to drive the chain to the rear wheel. More complex, more expensive.

Zero does make an enduro "dual sport" version (semi-knobby tires), and Alta does make a street version (smooth-tread tires), but...if I win the lottery, I would buy the street Zero-S, and the dirt-version of the Alta. Those versions are what they do best.
 
Sunder said:
Alright, either I'm taking crazy pills here, or some eminent experts in the field have replied, and have failed to noticed that to get his simulation to work, he throttle limited both bikes.

Higher voltages do not cause motors to heat up faster when you're throttle limiting them.

72v @ 12% throttle = 8.64v
48v @ 18% throttle = 8.64v

Coincidence? I think not. The motor is seeing exactly the same "resultant" voltage from PWM because of the throttle setting. From the motor's point if view, it can't tell what battery is attached.

Try the same thing with a load that exceeds the capability of the controller (with throttle at 100%), and see which motor heats up faster:

Motor heating primarily comes from the current flowing through the motor (phase amps). At different battery voltages, if your controller is limiting motor current to the same amount, the motor heating will be the same.

If your higher voltage setup allows you to feed more current to the motor, it will heat up more, but you're also getting more torque.
 
Addy said:
Motor heating primarily comes from the current flowing through the motor (phase amps). At different battery voltages, if your controller is limiting motor current to the same amount, the motor heating will be the same.

If your higher voltage setup allows you to feed more current to the motor, it will heat up more, but you're also getting more torque.

Correct. For a fixed load (Same speed, same weight, same incline, etc), the controller must maintain speed. Since motors are locked to a KV, to maintain the same speed, they must reduce the effective voltage through PWM.

In equilibrium, Same load = Same effective voltage * same current.

Hence quick calculation above showing that the throttle position multiplied by the throttle position had the same voltage, hence same current.

I'm just bringing this up, because it's the second time someone has used the simulator to "prove" higher voltages/higher KVs are not less efficient. It's actually not true, it's to do with the assumptions they've keyed into the simulator. At any low below the capability of both controllers, there will be negligible difference in efficiency. When controllers are overloaded, the lower voltage/lower KV one will tend to be more efficient.
 
MadRhino said:
Of course I have had controllers damage, and sometimes they need mosfets replaced a few times before I have a reliable set that last.
That's because you are driving the MOSFETs far past their limits. You are eventually ending up with MOSFETs that do not fail immediately under that abuse. Like most components, their failures under overload will fall within a bell curve; you are seeing the small end of the bell curve.

Poor design is not "a basic requirement" to ride performance bikes. Indeed, it is something to be avoided by people who value performance over constant failures and degradation.
 
billvon said:
MadRhino said:
Of course I have had controllers damage, and sometimes they need mosfets replaced a few times before I have a reliable set that last.
That's because you are driving the MOSFETs far past their limits. You are eventually ending up with MOSFETs that do not fail immediately under that abuse. Like most components, their failures under overload will fall within a bell curve; you are seeing the small end of the bell curve.

Poor design is not "a basic requirement" to ride performance bikes. Indeed, it is something to be avoided by people who value performance over constant failures and degradation.
You are pushing this at tad too far, with ‘poor design’ and ‘constant failures’. First, I leave everyone in the dust in the mountain here, and the failures are always the bikes trying to follow me. Once I have a bike tuned and set up, it does require frequent maintenance but it does the ride everytime. If that is poor design, what about all those who didn’t succeed keeping behind me, or didn’t complete the course? That is, all of them who tried.


My street bike tops well above 70 mph, and its last failure was 6000 miles ago. So, it seem that by your standards, almost everyone on the sphere are building poor design.
 
It's just the nature of high performance anything. If you're not breaking stuff, you can probably be going faster.

Race vehicles aren't generally designed with reliability as a priority, sure it's a factor, the thing needs to make it around the track at least occasionally, but there's always a balance somewhere between going fast and avoiding failure.

Finding that balance wins races.
 
dustNbone said:
It's just the nature of high performance anything. If you're not breaking stuff, you can probably be going faster.

Race vehicles aren't generally designed with reliability as a priority, sure it's a factor, the thing needs to make it around the track at least occasionally, but there's always a balance somewhere between going fast and avoiding failure.

Finding that balance wins races.

Exact. Test laps are expansive, when the race is going to be perfect. But, what a feeling when the machine is perfect and you fly with ease and confidence.
 
Back
Top