Separated VS parallel. Which is better?

Power System
21

I personally like the TI chipsets. I wouldn’t even install the FET safety circuits… you could, but it would be somewhat lossy, and you’d need a bunch of FETs to deal with the whole load.

I’d suggest a chipset like this:

If we bridge all the balancing lines, one chip can monitor it all. If something fails or the ESR changes on one cell, you can detect it and then find which if the parallel cells is defective and remove it. Most importantly, you can find the under-voltage condition easily during flight.

22

Paul I could be missing something, but what I was referring to was one battery charging another if a user accidentally plugs in a battery with lower charge. This current wouldn’t go through the ESCs at all in a parallel configuration.

diagram
In this diagram you can see a complete circuit without the ESC in the loop. I don’t see how the ESC could impact current flowing between these batteries.

One way this risk could be mitigated is by doing what Davek said and keeping all the batteries together in a pack. However Davek, consider that some users might want to use more batteries for longer flights, some might want to use fewer batteries for lighter takeoff, short flights, and some might want to land, swap batteries and take off again. I would prefer this system be stupid proof whether it’s being used with one battery pack or 10.

23

Yeah. If you want multiple battery configs for different Take-Off configurations, you either a) be really careful that you’re hooking similarly charged packs up together, or b) use a BMS safety circuit (with MOSFETs) to prevent fires/unintentional charge/discharge. Either way, you’ll need to bring packs to same level before you can connect them and fly…

24

I guess it could be as simple as detecting differences in voltage and sounding an alarm (audible or on the controller) if the voltages are too far apart, or if the cells are sufficiently imbalanced. I’m liking the idea of just using that BMS chip, fusing the batteries, and letting the user deal with it if the batteries aren’t charged right.

This is starting to sound a lot simpler than putting in switching, diodes, etc.

One thing that might be nice instead of fuses though could be a marine/automotive circuit breaker. I’ve found them really useful on robots just to have as an on/off switch for the batteries, but maybe people prefer fuses for being faster/cheaper.

25

So let’s say you have to different batteries one has a charge of 3.6v and the other has a charge of 4.20v. If you put the batteries into parallel and run it into your main power. What will happen is the systems draw the same power from each pack (the voltage does not distribute evenly between the packs/ the higher voltage battery doesn’t charge the lower voltage one) So the pack that was fully charged (4.2v) drops down when power is drawn by the system to say 50% (3.6v) and the battery that was already at 50% drops down to 0% (3.2). This can be a problem if the user doesn’t charge up there packs and they will ruin the one battery that was halfway charged because it voltage will drop to low but it won’t create a backflow when initially plugging in the two batteries.

I hope that helps clear things up, but the system we are developing will let you know if there are major voltage imbalances.

27

@Pdwhite Hi Paul. I wonder if you could talk a little on the power demands on the battery pack.
So far, it looks like it’s being designed for 12S LiPo voltages.
Can you tell us (ballpark) what sort of current is being pulled in your tests so far. 1. On takeoff and 2. during level cruising.
This info will assist in picking the right battery combo that doesn’t sag too much under load and optimise cycle count

28

Sure, around cruise the system pulls 120A and at full throttle it pulls around 300A.

30

A couple of questions please:

The multi star battery packs are rated 10c. To me it seems better to have 8x 10k packs easily sharing the motor loads with their 10c rating. 2x packs per motor Vs 4x 20k packs working harder for their 10c rating 2x packs per 2 motors, double the load?.. or does it not calculate that way?

Next question. Why can’t we leave an auto cell balancer connected as a simple bms and also then just charge through the main lead? I saw a video on YouTube of this sort of thing being done. It’s like a poor mans balnce charger, but could have the advantage of a single plug in to charge. ??

I’m out of my depth here, hence the question which may seem stupid with an obvious disadvantage. Thought it’s worth asking though.

Or

31

@PDWhite I’ve a question which may be dumb but if I don’t ask I’ll be forever puzzled. You said:

Is that load divided between four ESCs? So each unit would have to be rated at a minimum of 75A, call it 100A for a safety buffer?

Regards

32

Yep that correct the ESC are 80A but they are newer tech ESC so the can handle much higher currents (even the manufacturer says they can handle 100A with some air flow).

1 Like
33

If parallel system with a fully charged battery and a dead battery plugged in together there is a chance for a burst of current of 117A. I=VR

Fully charged 12s Lipo around 50V , fully discharged 12s around 36V difference =14V
Cell internal resistance =.01ohm times 12 cells in series equals .12ohm.
14V*.12ohm = 117A

Am I doing the math right? Looks manageable to me for just a short burst until the cells equalize. Not a good thing to happen but shouldn’t be catastrophic with adequate wiring.

With a BMS to make sure individual cells don’t get out of safe range shouldn’t hurt anything.

Kirk.

34

At the moment we’re not considering a full BMS for each battery - it’s lossy from a power perspective, will add cost and complexity. We’re just considering using a BMS for cell voltage monitoring and fuses for protection. Without a BMS or fuse, connecting full and dead packs will definitely start a fire!

This would mean we’d use an external charger and external balancer. It also means that people will need to be careful when connecting/disconnecting. I’d like us to build a config that can be charged without disassembly - which will eliminate that risk from day-to-day usage.

35

Is the indicated thrust being measured with a full 12s battery (50V) at 100% duty on the ESC (so Iphase = Ibridge)? If so, what does the maximum thrust drop to as the battery charge reduces (or does the ESC have current regulation and field weakening to have a more constant max RPM?).

36

@Pdwhite sorry to bring this topic back after one year but im just starting to study and review all the electrical part, i ask in advance for an apologie if im using old information and this has been addressed.

Im a 47 year old with 10 year pg, ppg and tandem ppg flyer and also electronic engineer and and for me this affirmation you make here its wrong. I definitely agree with @DanielArnett that if you connect two batteries with different voltages in parallel, yes or yes the higher voltage battery will charge the lower voltage battery unless there is a Diode to prevent charging or a fuse that triggers when this high amp charge goes to the other battery.

image
image346×548 76.8 KB

Once said that i will like to add also that you mention in another post on this same topic that the ESCs will prevent one battery pack to charge the other, and this is true but ONLY if you connect on battery pack to 2 ESCs and the other battery pack to the other 2 ESCs. This wring will be like the HIBRID PARALLEL diagram that @DanielArnett post but with only 2 packs instead of the 4 he puts here, but the idea its the same.

image

I think this is the best advantage of using this configuration, because its a way to prevent a fire in two situations:
1)By connecting a full charged and a not full charged pack in parallel.
2)By a cell damage in one pack during flight, and having the other pack start to overcharge the damaged pack, because of the imbalance that the bad cell create. I think this last reason alone its worth having this wring configuration.

Sorry if i have outdated info but have made my best to look to latest wring for batch 3.5 and i found that the 2 packs are wired in parallel and then just have on main to all 4 ESC’s.

This is my 2 cents for safety.

3 Likes
37

Hi kitepoc,
I am asssembling my batch 3 kit, and also prefer the hybrid parallel configuration for the reasons that you mention. But on top of that it also reduces the impact of the startup peak current of the ESCs. In the case where all 4 ESCs are connected together and each set of two batteries connects to it through a main switch, the first set to be switched on will get the startup current of all 4 ESCs. The second main switch will join a circuit which is already powered up, and has no worries (when both battery sets have same voltage :slight_smile:).
With only 2 ESCs connected to a main switch, the startup current will be half of that of 4 ESCs. And the damaging power dissipation over the internal resistance of the main switch will only be a quarter (P=I^2xR).
I presume that the controller will be powered from one of the two parallel sets.

Disadvantage is if one set dies, then 2 motors continue (if it is not the set that feeds your controller). This will give you asymmetrical thrust, or a torque if you connect the motors diagonal (I think the torque version is the best). But I am used to turbulance, so prefer that above the risk of fire….

I don’t think it will impact on the battery charge reading of the controller, would it?

In below picture I crossed the connecting plus and minus lines which I would skip, to get the hybrid parallel configuration.

Do I see it wrong?

Corrected%204-way%20layout
Corrected 4-way layout.jpg2048×1536 795 KB

38

I dont like it - bad switch your running 2 motors. One set of bats depletes slightly faster than that other then you going out of balance again…no? You should see a bad cell declining during a balance charge cycle.

If you do go that route you need a precharge on the other switch as well.

I did this – if one switch fails for some reason that power is transferred from other pack to the other switch - I want to set something up that tells me this has happened – maybe run some wires from small voltage step down circuit boards that go across each switch out to some LED’s on the current display on the arm. If one LED out your on one switch

image
image.jpg932×610 310 KB

Cheers

39

Turbulence is one thing but suddenly losing two motors during full throttle could line twist pretty severely unless wired diagonal. I think the best way to protect the batteries (as well as prevent fire) is to have a good monitor that can watch each individual cell.

That bottom red wire is redundant… did you mean to draw both?

What do you mean by this? Are you talking about the arching that occurs when the circuit is closed?

40

Hi @Ralph,

I have not yet start to build mine because im traveling, but for the startup peak current that its making arcing on the switch, im planning on testing with a PTC (fuse that changes resistance in order to maintain a maximum current ) if this works OK, i will not need to have a separate switch for precharge, i will only need to wait some seconds after connecting batteries and turning main switch on. Im planing on using 2 PTC of about 100-200mAmps, 1 for each switch.

Your statement about peak current reduction its true so maybe the arcing problem might not be a problem with this configurations we are choosing that reduce 2 Esc’s per switch instead of 4. Need to make some test’s

The disadvantage you said about one Bank dying its NOT true, because im planning on putting on one bank, one Right motor and one Left motor, so you don’t twist on your paraglider if the other bank dies, and also one will rotate CW and the other CCW so you will maintain the torque free advantage of the openppg. Only disadvantage will be half power.

About power to the controller im planning on using power from the 2 banks, for this you only need 2 diodes, with this the controller will be always powered as long as one bank its working.

Finally on your picture you need to maintain the same ground for all the systems, so dont cut that black wire, just the red one. If you don’t put the ground the controller will not be able to talk to the 2 escs that are not on the same ground. Hope this is clear. (im traveling and don’t have to much time to make a diagram for you, but if needed i could try to make one). A small black wire would work, but i prefer to use the same #8 because otherwise you would need to protect that cable with a fuse in case it takes a short circuit. (ak fire :wink:)

Hello @E-pusher, i agree with you, but it’s just a safer way to do things, you can not see easy a bad cell during charge if you don’t monitor the charging process. A bad cell will be lower when you start to charge and will end sooner than the others, but at the end you will see all cells top at 4.2 so i think its hard to always analyse each pack you charge looking for this, but not impossible. i just prefer to go safer because its an option on this configuration of batteries.

About the pre-charge switch, i mention above i will not use it and will be using a PTC on each bank/switch for that job.

One final note about safety, its that i will be using a 250 Amps fuse in the serial connection of each bank, that is where you connect the positive of on 6s bat to the negative of the other 6s bat to make a 12s bat. This way any short circuit in any #8 cable will blow it. I still have to make the fuse with the XT150 connectors, and later after some flights and testing i might add thermal 250amps fuses with integrated switch, for in case of fire have the switch cut off by its own.

p.d. Sorry for my english and hope you can understand.

1 Like
41

No I was just making the black wire red - it was confusing me every time I looked at it :grin:

42

Sorry, I ran out of Red wire. Haha