Higher voltage means lower current for a given power requirement. Lower current means less power loss to heat. The power required at the propeller is what is left over after the power lost in transmission is removed. Another way of saying this is that he power the battery must supply in what the propeller requires plus the power lost in transmission. In an ideal world where we could make batteries exactly to fit the application, which we can't, the higher voltage battery suitable for the mission profile will always be lighter.. The trouble we find is that all manufacturers of battery pack play games with C rating and MAH derating certain pack for greater C rating visa versa. What we truly need to know and understand is the total energy capacity of the pack or cell. in general the 18650 cells have a greater energy density than the poly cells. Trouble with them is they are often bigger than necessary for a specific mission. For the mathematically inclined:

      Power supplied by the battery = power required by the propeller + the power lost to heat

      Power supplied by the battery = (Torque x RPM/ 5252) + (Current x Total resistance (Wires, motor, back EMF...) )

      Voltage x MAH = ((Torque x RPM/ 5252) + (Current x Total resistance)) x Time

Note, the MAH on the pack doesn't mean a whole lot. Some manufacturers put the Watt Hour rating on the pack. That's the value we need to compare. Battery A Whr = Battery B Whr. You can get that roughly by the product of the pack voltage time the MAH rating. However it's the chemistry that makes the big difference. The 18650 cells win here as well as they also have the best energy per volume.

Here's how this impacts our model application. The parasitic power, the energy lost to heat that does us no good, is a function of the square of the current. Simply put is is very hurtful to our cause because of the square term. Write the first equation:

     Pbattery = Pprop + Plost

Power in an electrical system:

     Power = Current x Voltage,   where Voltage = Current x Resistance

Then:

     Plost = I x V =  I x (I x Rttl ) = I^2 x Rttl

Now to understand how this impacts us and drives us to always want to use the highest voltage for a specific size and weight we solve for the current:

     I = Square root ( Plost / V )

This is kind of a circular definition but you can see that the current is vary much driven inversely to the voltage. The higher the voltage to lower the current by a very significant amount. Keep in mind this current thing is what drives the heat up and down. Batteries get hotter, wires get hotter, MOTORS get hotter causing issues with plastic mounts....

So, short version. Always use the highest voltage you can. To compare two packs, multiply their voltage by their MAH. IF you know what one configuration is that works, use that V x MAH and then use the comparing packs V x MAH. If the latter is very mush greater than the former, look to see if there is smaller cell pack.

My $0.02 worth what paid for.

   

The Badass 2320 is capable of running with 6s batteries. The trouble is finding a 20 or 30 Amp ESC to run them that is capable of using a 6s battery. For a given flight a higher voltage battery will consume less power over the flight due to the lower I^2R power loss. The trouble is finding proportionally smaller cells. It's difficult combination to sort out effectively. In terms of responsiveness the extra overhead is good too.

I really like that BA 2320 specs and am hunting similarly for the package. I'm about to just give in and go for the bigger ESCs. The BA 2320 "fits" within the original mount rails for a 35 size ICE requiring only minor amount of grinding. The Cobra 2221 does a good job of making the Ringmaster / Flight Streaks fly. We see some mount issues when the Cobra is loaded up and gets hot. The BA 2320 on a 5s or 6s should make more output power and run cooler. That configuration is a few shop hours away from going to trial. The test bed will be a Banshee rescued from some garage rafters after hanging for 25 years.

My mission there is bolt on conversion of these old planes hanging around. I'm very close. There has been three iterations since the first Flight Streak I posted last summer.

If'n it were me doing what you're doing Ken, I'd go 6s and use whatever ESC fit. But then that's just me.

My $0.02, worth the price paid.

The BA 2320's will definitely pull those props on a 6s battery. Spec wise it can run 550 watts continuously. I don't know what the 3 blade 10x5 draws but the 2 blade pulls 250 - 270 watts. I'm guessing the 3 blade will pull 20% - 30% more. I should run a couple tests. I would definitely run the 6s pack. I'm assuming 18650 pack which will be a very good choice. My opinion FWIW. The power and volume density of them are the best.

... For a given flight a higher voltage battery will consume less power over the flight due to the lower I^2R power loss. ...

Only if you don't size the wires and connectors to the current.  If you use wires that are 40% larger in area (basically, 5/6 larger in diameter) and connectors to match, the I^2*R losses will match.  ESC losses may or may not be higher at lower voltage/higher current.  It depends on the ESC.

I don't think that I^2*R losses are going to be an issue unless the wires are seriously undersized.

Only if you don't size the wires and connectors to the current.  If you use wires that are 40% larger in area (basically, 5/6 larger in diameter) and connectors to match, the I^2*R losses will match.  ESC losses may or may not be higher at lower voltage/higher current.  It depends on the ESC.

I don't think that I^2*R losses are going to be an issue unless the wires are seriously undersized.

What you say is absolutely true, one could recoup power lost to the supply by increasing wire size. The opposite is also true, one could reduce the wire size by increasing the voltage. A 40% change in wire size change can go either way. Going from a lower voltage battery to higher voltage battery would save 40% in wire weight. Visa versa. Going to lower voltage battery would, as you suggest, increase the weight by 40%. The planes and power systems we fly are generally oversized and can handle some small inefficiencies.

12000rpm / 18.5=650KV Motor
12000rpm / 22.2=540KV Motor

The mAh and C rating of the pack depends on the demand from the motors.

That's also basically correct but leaves out one aspect, the power required to run a propeller. For instance, it takes 250 watts to run a 10x5 propeller at 10,000 RPM and if we chose a motor than can only run 180 watts continuously, the motor would ultimately fail. So, we need to look also at the power rating for the motor. Generally we need about 20% - 30% more continuous rating, in this example we would want a motor with 300 W to 350 W capability.

There are a number of combinations of battery and motor to run that propeller. But that isn't all the battery has to supply. The battery has to supply the watts for the propeller plus the watts lost to heat. That heat is the current squared term. When we increase the supply voltage, that current is reduced and consequently the power lost to heat is reduced which in turn reduces the power the battery has to supply. The power lost to heat is somewhere around 10% - 20% of the total power, often a lot more. This means that by increasing the voltage you can save weight in both the batteries and the wiring.

This is an exercise we go through all the time in full scale aircraft installations.  It is why you see 28 volt systems instead of 12 volt systems. On one jet I did, we saved nearly 300 pounds in the powerplant starting and control system. We spent a lot of money optimizing the starting system and ended up with 2 28v batteries that could be switched from parallel connection for normal operation to series for starting. There's more to it than the simply wiring in that case though involving the toque and speed of the motor.

Basically it takes a fixed amount of work to start an engine in the same way as it takes a fixed amount of work to fly a model through the pattern. The battery has to supply that work plus what ever is lost to heat. The batteries ability to supply that work and its weight are a direct function of its chemistry. The more work required means a larger amount of stuff the battery has to have. We could package all of that stuff into one package, ie one cell. Or we could package it into many cells. Since the voltage is also a function of the chemistry, we can arrange multiple cells and  increase the voltage which reduces the losses which means we don't need as much stuff to heat the environment. This means our battery can be made lighter. This is a double win. We win because the wiring is lighter and we win because the battery can be lighter.

Practically speaking though, our batteries are limited by cell size. This doesn't truly allow us to fully optimize the power system. An example is for a typical small airplane such as a ringmaster or flight streak. We could use either a 2200 MAh 3s Lipo or a 1800 Mah 4s Lipo with the same exact motor and the performance would be practically indistinguishable and the weight difference minimal. What we would see operationally, is that the total Whr we put back in would be lower in the 1800 MAh battery or as people like to say, how many MAh put back in.

The other aspect we would notice is in the longevity of the batteries. Reducing the current and therefore the heat production includes what each cell sees. Less heating within the cells means the batteries will last longer. For many people this is irrelevant because the batteries often have a different fate that ends their useful service.

12000rpm / 18.5=650KV Motor
12000rpm / 22.2=540KV Motor

The mAh and C rating of the pack depends on the demand from the motors.

I don't see how dividing the RPM in to the voltage has anything to do with KV selection. 
Decide on a prop diameter and pitch and RPM that you need to run.  Pick the KV of the motor to match the rpm range of the desired prop and the voltage of the system through the whole of the flight.  You need enough motor KV headroom to accommodate the ever lowering pack voltage under load and/or at the end of the flight or you will experience a drop in performance.  You don't want to be running at 100% throttle due to insufficient KV headroom.
 
Li-Ion and Lipo batteries have different characteristics under load, as well.  Li-ion have a greater tendency to voltage sag under load, so that must be accommodated with a higher motor KV as well. 

The guys that are flying the 6S Li-ion setups with Igor Burger's full active system and prop are setting their cutoff voltage down around 2.7-2.8v per cell.  The reasoning is to avoid potential cutoff/sag under load.  The planes with the active systems and Igor's 3-blade carbon props spinning at 10500+rpm pull more amps and need the KV headroom in general, and, to cover the momentary sag potential.  The commonly used Badass 710kv 3515 motor is a popular choice in 6S applications due to the rated KV, weight and power.
 
When I have pulled the data logs from my flights with my 6S li-ion system, I have never dropped that far under load.  I think the lowest I have dropped is 3.1v.  The caveat is that I currently fly a governor only 2-blade system in my 64oz plane. The Hubin FM9/Castle ESC, BA 2826/690kv, BA 12x6 2 blade prop running at 9600rpm is a pretty low amp draw setup overall compared to the active systems and 3-blade props.

So if you know or have an educated guess regarding the potential voltage drop in flight, and the desired RPM for your prop, you can do some quick math to figure a safe KV range.  You don't want to be operating fully maxed out for the available voltage at the vertical 8, overhead 8 or clover.