Revisit a age old debate, Esc's mounted on the arms.

[Pumpkinguy]

I've read countless forum posts on this, I've heard all the opinions, Ive seen the fixes. Still, I feel the need to reopen the debate.
Having a little more conductor between the battery and the ESC's, allowing active cooling vs keeping them close to the nucleus of the UAV.
We know that long power wires to the esc can cause premature failure or that's what they say anyway.
I would bet mounting my escs on the top of my arms under the prop wash will keep temps near ambient whereas mounting them near the center of the UAV will have them see temps of 120-140F even with the additional heat sinks Ive added.
We know heat is electronics worst enemy so what Is worse?
Furthermore many manufacturers mount the escs under the motors. D800, 900, 1000. Freefly Alta.
The manufacturers claim that have made special allowances in the applications but have they? Are their esc's really any different?

@Hexacrafter

 

[Motopreserve]

I think the problem with this debate is that the only way to truly have a real-world answer is to run 2 setups, with the ESC location (and wiring) being the only variable. Run non-stop, to see if there is pre-mature failure rates on the ESC at the motor mount.

 

[John Doecliff]

Funny you should bring this up. I'm getting ready to start a new build and I just became aware of this issue and read up on it and I think I understand it. I'm going to mount my ESC's under the motors and add additional capacitors just before the ESC. They sell this little 6pack of caps and they solder up in line with the power leads to the ESC and buffer the voltage spikes. Supposedly the problem is that the cap(s) on the ESC will eventually fail so if you add these additional caps you eliminate that risk. I'm also going to twist my power leads going to the ESC because it helps to cancel out the increased inductance with long wires. Not sure if its right or wrong but it's my understanding of the issue.

 

[FerdinandK]

I think, there are a lot of things mixed up in this topic.

First (and to me most important) is, that the copter is a flying vehicle. So flying vehicles depend heavily on their weight. If you set an action (or have to take a decision) you always have to think if this action will increase weight (increase power consumption, increase structural load, increase heat production, ...) or if this action will decrease weight (decrease power consumption, decrease structural load, decrease heat production).

As the ESCs are concerned, a lot of "fear of failure" because of long power lines comes from EDF-Jets, which run for 5-7min on full load, ESCs inside, long power lines, here often the ESCs burn in flight, and one of the reasons is because of the long power lines (sometimes with wrong dimensions).

On Multirotors typically you do not run the ESCs "full throttle", and the ESCs are also most of the times massively overrated, since you want to cover the peak-loads which will be used in flight for seconds only. In addition you have already a lot of capacitors on board, since they are on the ESCs all over the copter, and not all ESCs are at the same throttle level, and are just in one fixed moment pulsing full throttle (that is how part load is done). That is also damping the electric system.

So keep your wires short, as short as possible, keep the amount of solder low, as low as possible, avoid power distribution boards (as they are not needed). The position of the ESC should be in the airflow of the propeller, as "under the hood" of center plates is not optimal, as well es under the motors, as there the airflow is minimal.

Also you not only need to plan a build, you also need to check when the build is finished. If the temperature somewhere on the copter is 20° C above ambient, it should get your attention, you should change everything possible, to prevent heat from being produced, to cool things down, to exchange things or to rebuild things. It is not only critical for electronics and motors, also if you search for efficiency, the loss is right there where something is getting hot.

Think twice. Some ESCs for example have nice heat-sinks, but in addition there is a heat shrink tube around, which transforms the heat sink to be a decoration. In this case you can remove the heat-since, save the weight and add a heat shrink tube again, or cut free the heat sink. (a sticker on the ESC also does not help).

best regards
Ferdinand

 

[Motopreserve]

Ferdinand,

Thanks for weighing in - I have always appreciated your builds and expertise. Quick question: what's your preferred method of power distribution of you steer clear of a pre-made PDB?

@econfly had some good info on this subject a while back - maybe he can update is on what he's found more recently.

 

[FerdinandK]

This are two of my wire trees:

Here a drawing of the wire tree of my first X16:

So my solution is soldering and crimping.

best regards
Ferdinand

 

[Motopreserve]

Thanks!

 

[ProfEngr]

Having the correct wire size for the load is critical; I*I*R is the power loss through the wire. With that many motors, I adds up quickly and then you square it. So even a very small resistance in the wire becomes significant. If we take #12 wire with 1.58mOhm/ft DC resistance (pulled from here) and use a running current of 100A (probably unrealistically low for 16 motors at hover) that gives:

100*100*0.00158=15.8W/ft lost to heat just in the wiring. The voltage loss would be 0.158V/ft of wire.

These are just numbers I pulled out of the air and the first resistance/ft I found online so real-life will vary greatly. But you get the idea.

 

[FerdinandK]

@ProfEngr
That is why it is important to think and calculate before and not after the build, and this if were the efficiency comes in, and if you have quite a number of motors, you do not need to power all of them from one source.
This small X16 with 14" propellers, 4S setup (where you can see the schematics of the wiretree above) has in "standard" loading 7,4kg AUW (2kg payload 2kg battery 3,4kg copter) in this setup the average current per motor is <4A (the flighttime is 25min in the setup). As you can see in the schematics, the motors are powerd from two sides, always 4motors directly (wire shorter than 10cm) and 4 motors/ESCs with 50cm wire distance. So the average max amps to these distant (50cm) Motor/ESCs are <16A in this setup. The overall connection is only good in case of one-sided battery failure, then the other side will take them over with 12x4=48A average load. The thickness of the lines in the schematics correspond to different wires, the thickest is 6mm2 then 4mm2 and then 2,5mm2.

Even in the max. load case I have been flying 15,6kg (3,4kg copter, 4kg battery, 8,2kg payload which still gives 16min of flying time) the average current per motor is <11A, so the distant 4 motors (50cm) are powered from one battery-side with <44A on average (using a 6mm2 wire). That should do the job. Even in case of a one sided power loss, the wires can hold as long as it takes to bring the bird down. 12x11 is 132A on average for about 40cm wire length using 6mm2.

As I once did a frame-change I have a picture of the wire-tree of this X16:

to get an impression of the copter:

best regards
Ferdinand

 

[ProfEngr]

Agreed, I didn't bother to check the requirements of a typical large-prop motor. Seems they draw less than I'd assumed (shame on me). Anyway, it just reinforces the need to do one's homework when picking parts and sizing components. My comments weren't meant specifically for you as you seem to push past the bleeding-edge in your builds. Just to the general audience for a did-you-consider-this-aspect-of-the-design. I have more of an electrical background than mechanical so stuff like this will usually get at least a glancing thought.

You strike me as the type to reject the concept of "but... but... you can't do that."

Not to get too far OT, but are the motors on each hobbywing controlled independently by the FC? That would be 16 throttle signals.

Bet the FPV from the look of your flight style is awesome.

 

[FerdinandK]

It is 16motors with 16esc, but 2 escs are controled by 1 FC-motorchannel each.
http://multirotorforums.com/threads/tiger-motors-u8-28x9-2-props.12966/page-19

best regards
Ferdinand

 

[violetwolf]

Back OT: I have a heavy hexa that I've run outboard ESC's on for three years without any problems.

I use ESC's one size bigger than required, and add a couple of inline caps... Although I'm not convinced it is necessary.

The ESC's always come down cool at ambient temps.

PS: I'm with Ferdinand regarding PDB's. A harness is quick to make and 100% reliable.

 

[Pumpkinguy]

To everyone that has posted, thank you. It seems like at this point the old theory had been debunked. What exactly does twisting the power wires achieve?

 

[ProfEngr]

Wire twisting helps to prevent interference between the power wires and the other electronics on the craft.

 

[violetwolf]

Yes, mainly reducing EM (Electromagnetic) Radiation from the wiring. (Which may affect compass etc). You can either twist them or bundle them closely together. One wire (-) will tend to cancel out the other wire's (+) tendency to become an electromagnet.

Same thing when building harnesses to replace PDB's. Bind the positive and negative as tightly together as possible.

 

[Old Man]

This EM radiation, or cross talk, is usually more prevalent with smaller gauge wires, generating the need to twist wires. Larger gauge do fine in tight parellel.

 

[John Doecliff]

This has all been very helpful as I am getting ready to build a 960 class multirotor.