Separated VS parallel. Which is better?


#1

Comparing our prototypes with Paul has shown that the main difference between the two systems is the way in which the energy of the batteries is distributed to the escs/motors.
Since my English is not the best I made a scheme to explain it better.


The most interesting things to me are the handling of the inconveniences.

If a cell is at risk of undervoltage:
With the DISTRIBUTED system you can’t do anything else but choose to continue flying by permanently killing the cell or stopping ALL motors.

With the SEPARATE system it is possible to stop only that motor in order not to damage the cell and continue to fly

If a motor or esc breaks down:
With the DISTRIBUTED system the energy of the batteries is distribuited on all the other motors, so in theory no range is lost.

With the SEPARATE system it is impossible to use the residual energy in the non used batteries so the range is reduced by 1/4.

Of course all this if there is software that will handle the inconveniences, perhaps at two levels, if the inconvenience is harmful only to the bank account (undervoltage batteries, esc burned) the software warns the pilot. If it can be dangerous for the life of the pilot(overheating batteries), it acts by limiting or cut power to the single sector of interest.

I try to summarize the pros and cons that come to my mind:
DISTRIBUTED (pro)

  • If a motor fails the autonomy remains unchanged (theoretically)
  • You can decide how many batteries (and therefore weight) to fly without major changes to the battery pack
  • It is possible to connect the system with a single switch / connector

DISTRIBUTED (cons)

  • If only one cell fails, you must land
  • Longer cables

SEPARATE (pro)

  • If a cell fails, you can continue to fly

SEPARATE (cons)

  • you must add each connector / switch for 4

Do you come up against other pros and cons? For now I prefer the separate management, but I can not figure out which of the 2 systems is more redundant. I think a problem with the ESC or the motor if well sized is very rare. On the other hand, practically every flight happens that one cell is discharged before the others.

What do you think? Feel free to add pros and cons.

  • Separated
  • Parallel

0 voters


#2

That’s a tough one. I’ll have to think about this for a while before I contribute a vote. It is very nicely explained with those diagrams though!


#3

Likewise, im going to try out both ways on my next build.


#4

Sorry but I can’t think without images :rofl:
These are my first thoughts on possible usable battery packs. The version with 8 packs of 16.000mAh is the one I’m using now.

Oh in my experience the best results (in term of flight time) are obtained with a Roadster2 26 at 107kg flying at 3kWh.
What is your experience?


#5

Hello. Very good job guys but i wonder how the behavior of the paramotor changes when only 3 motor work?


#6

Hi Nacho and welcome!
So when a motor stops in flight you almost don’t notice it.
But when you’re at low speed (like during the take off) a motor stop can twist you, I have a video somewhere in my old PC of this test.
For this reason we are developing a software that handles this type of inconveniences, for example by redistributing the power in the same way on the right and left side.
The scheme (on the paper for now) is this but I’m not very sure with the last one. I would like to test it in flight.


Will motor failure cause serious yaw?
#7

I’ve had LiPo fires using parallel battery harnesses. My recommendations are as follows:

  1. Ensure the batteries are disconnected individually and cannot be disconnected while remaining connected to each other (You seem to be good here).
  2. Fuse the batteries first. See the image below for my recommended wiring.

This fuse prevents the batteries from charging one another at high enough currents to cause a battery fire. For instance if the user accidentally plugs a discharged battery in with a group of charged batteries, the charged batteries will charge the discharged battery at very high currents- melting wires and starting fires.

The above diagram would also be improved by adding circuitry to prevent backflow where the fuses are.


#8

Actually how about a hybrid option?

If one motor stops getting power then its counterpart on the opposite corner does as well.


#9

I see what you’re saying, the ESC does prevent any current flowing so one battery can’t charge another when in parallel. Fuses are always nice, you would need around a 100-200A fuse depending on battery configurations. Which is one thing to consider… battery configurations and having the freedom to change them. So whatever configuration that ends up being chosen it has to be flexible in accepting different battery configurations and not be a wire mess.


#10

I am liking the idea of fuses between each battery and ESC. It’s a good starting point in power system safety.
My greatest concern is a potential fire on my back. Though the risk is minimal in this application, I have experience with it on my e-bike (pulling 200A @ 12S with a backpack full of lipo’s)
Personally, I will be looking at this and drawing on all my mistakes to minimize the risk as much as possible. Can post more on this as the topic matures.


#11

I’ve asked around the lab and our consensus is

  • Fuse every battery pair.
  • A diode that can handle the amperage can prevent backflow into any battery. This will prevent cell damage and fires. Possibly one of these?
  • With low voltage cutoff regulators, having one battery go out will not cause the other batteries to stop working when in parallel. Parallel definitely makes for a better user experience, but we have to be careful that the user doesn’t plug in too few batteries, drawing too many amps. Here’s a DIY version I’m considering modifying for our needs.
  • We want to make/buy circuits to do all of these and communicate with the controller so the user can see all the batteries voltages.

@Pdwhite it would be great if you could measure the amperage your Motors/ESCs are pulling, so we know what amperage to spec for. One of these should work.


#12

With a Diode you’re losing a not negligible amount of power due to the forward voltage drop.


#13

Putting relays or diodes inline probably isn’t a great idea for this sort of configuration. The best solution would be to use a proper BMS w/ in-line MOSFETs. That would give you all the battery management safety, charge tracking, etc - but it would be expensive and not sure it’s really worth it.

Fusing is a good idea, however :slight_smile:

For cell voltage measurement, we’ll probably do best to use the balancing port and simply monitor for cell under-voltage conditions and report that back to flight controller and ultimately pilot. Balancing will happen on the charging side.


#14

In my experience I never found a -not bulky- but reliable BMS that can handle over 150A so I decided to balance charging and monitor on discharging.
I use 8 frsky sensors (FLVSS) for single cell monitor and they are reliable, now there is a even more affordable version (MLVSS) without the oled screen.

You have to buy also a Frsky ID selector board to get them works together.
The FrSky protocol is decoded by an arduino so then you can do everything you want with the data.
Now I use also other sensors from FrSky but I would like to get Current, ERPM and Temperatures directly form the ESC like the VESC (but this is a bit too much off topic)


#15

Oh man this is a brilliant and cheap concept! I love it


#16

instead of Frsky sensors, I’d suggest we a BMS chipset from TI, and just using it for its cell voltage measurement capabilities. I’ve been using them for years, they’re cheap, easy to wire up and talk SBUS/I2C - which will reduce the wiring dramatically.


#17
I’d suggest we a BMS chipset from TI

Have a link? Also why not connect it to a relay for low voltage protection. Do you have any schematics or board files for how you use it? I’d love to try it out.

I was also considering monitoring the cells for charging. It’s a matter of time before a user inexperienced with LiPos mistakenly plugs a discharged battery in with some charged batteries, and when that happens it could cause a fire. We either need backflow protection on the battery or we need to monitor cell voltage and detect charging.

Displaying voltages on the controller would be cool too, but the user shouldn’t be worrying about cell imbalance while flying.


#18

The packs will be charged using balance chargers as that common practice and not charges in parallel so there won’t be an opportunity for backflow while charging or plugged into the paramotor as the ESC prevent it.

Here the diagram dave and I were discussing


#19

yeah, it’s way smarter to combine them to one big pack and leave them that way. The only reason to take them a part is if a defective cell develops. So the architecture above could actually be reduced to ONE BMS from the N shown above.

As Paul says, balancing only happens during charge.

There will be way too much voltage drop across anything other than a fuse (unless you use 2 big mosfet as a switch - but that’s expensive and also lossy).

So I’d suggest we just

  1. do inline fusing
  2. Leave the packs connected always (unless repair needed)
  3. Monitor voltages during discharge with BMS (and monitor temperature)
  4. Balance during charge with external balanced
  5. Use FC/screen to alert pilot of any issues in flight, etc

#20

You mean bypass the relay/fet circuit during the discharge only?
SBUS/I2C so every bms can be addressed on the same I2C right?
Give us a link