Using 2 UBECS in parallel?

[Tim Hooper]

I have a new, 10S powered, project in the offing - a 1/4 scale Nieuport 28 bipe from the Balsa USA kit.

I want to arrange a dual power system for the radio gear for safety's sake, and I've been through the related threads - particularly those relating to Timbo's big PC9 - so I'm aware of the pros and cons of using BECs, UBECs and seperate rx packs.

So what I'm contemplating is this;

The model will use two 5S packs connected in series to yield the required 10S.

It seems a reasonable idea to use a 6v UBEC on each of the main packs, each leading to an SM Services Battery Backer Monitor unit;

http://www.smservices.net/acatalog/Large_Model_Units.html#a41

The backer will draw current from both batteries in normal use, but will switch out a battery in the event of a failure in either flight pack, thus still leaving me with power to the radio for a deadstick landing (as opposed to an out-of-control crash).

Anybody else tried this, or can anybody see any potential pitfalls?

tim

[Steve W-O]

HK also have a device for dual supply, I diodn't look at detail

[John Cole]

Interesting idea. The ground line to your first UBEC will be at zero volts. The ground line for your second UBEC will be at (3.7*5 = ) 18.5 volts. You will then put the outputs in parallel. Depending on how the UBECs are wired internally then either all will be well or you'll get a big flash and a puff of white smoke!

I would put a digital volt meter across the two output ground lines (i.e. the two black wires) before connecting them together in parallel.

Why don't you just use a (highly reliable) NiMH battery as your second feed?

[Tim Hooper]

Steve,

Thanks for the heads up, but it looks as though that particular unit will only accept a 2S power supply.

I'd post a link, but it locks up the page......sigh.....

tim

[Tim Hooper]

John,

I'm not a techy, so I can't quite grasp your reply I'm afraid, particularly regarding the voltage at the negative side of the 2nd UBEC.

I did consider using a 5-cell NiMH as a second feed, but the Battery Backer draws from both units in normal use, so I'm guessing that the voltages would try to equalise each other as a consequence?

tim

[Seamus O'Leprosy]

[url]http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=14464[/url]

fingers crossed

[Seamus O'Leprosy]

I cant post links or photos without crashing either

 

Edited By Seamus O'Leprosy on 19/02/2012 20:29:04

[Tim Hooper]

That's the one I found. Limited to a 2-cell input though.....

tim

[Peter Beeney]

I’m 99.99% positive?? that the negatives will definitely be common throughout. This is the reason why you can’t use standard regulators as balancers for lipo’s. The output has to be isolated completely from the input otherwise you tend to start shorting cells.

We’ve had a little odd situation with a UBEC in another thread, something completely different, though, but it actually proved that the negative was certainly common there. In fact, I’m pretty sure that usually unless any unit is specifically labelled as having an isolated output one pole will be common, to avoid such things as voltages rising above a given level, perhaps.

If you use a five cell NiMH battery as backup, the fully charged voltage voltage will be somewhere between about 7 down to 6.7 volts, say. It you parallel this with the UBEC output in the backer this voltage will shut the UBEC down. There will be diodes in the backer, I suspect, one to each supply, so the slightly reduced voltage will be the same for both supplies. As the battery discharges it’s voltage gradually reduces to that of the UBEC output, which will then start to take some of the load. The battery voltage will eventually reduce to 6V, when the UBEC supplies all of the current.

If you have any multimeters which includes an ammeter you can soon cobble up a little experiment on the bench and watch this happening; as the battery voltage falls the regulator starts to do the driving. The battery would be a bit depleted, of course, but it will never really fall much below 6V so there’s always going to be some left in the event of a UBEC failure. If you did this experiment you could then separately discharge the battery from this point to see how much capacity is remaining, so you would have established knowledge. In some ways, not much difference from the standard 2 battery set-up, if one battery fails when they are both nearly flat, then there’s not much capacity left here, anyway.

If I were going to do this, I’d use the separate battery configuration. That way I’d consider that I’d not put all my eggs in one basket.

PB

[Bob Cotsford]

I'm considering something similar on an IC model, using two LIFE packs to supply the Rx in parallel, each with it's own switch and a diode to isolate it in case of failure or loss of charge. I don't see why a UBEC setup wouldn't work this way. Each can have it's own on-board monitor by wiring them between the switches and diodes.

 

The HK units are pretty chunky, something like 2 1/2" *1 1/2" * 1 1/2" needing a 2"*1" hole in the fuselage side to fit them.  I've got one but can't bring myself to cut such a big hole in the side.

Edited By Bob Cotsford on 19/02/2012 22:32:33

[Martin Harris - Moderator]

Posted by John Cole on 19/02/2012 20:06:34:

Why don't you just use a (highly reliable) NiMH battery as your second feed?

Do you have a source for the highly reliable variety? I'm still looking...

(but not very hard since discovering the joys of A123s)

Edited By Martin Harris on 19/02/2012 23:13:54

[Codename-John]

Electronics isnt really my strong point so I`d check with SM Services to make sure but i dont see a problem with it, electricly anyway

Ive done a little sketch to work it out in my head so let me know if its wrong, what John says about negative voltages makes no sense as electric works on the potential difference so even tho neg on battery 2 is 18.5V as pos is 37V there is still only 18.5V potential difference across each BEC (green lines on drawing) if nothings commoned in the backer (which their page doesnt show), and it states if one supply fails the other is unaffected then it must be separated somehow, then all you have is 2 separate 6V supplies which unless theres a problem share the load, and if there ever is a problem with one battery, you`d have no motor power but theres still 6V supply to control the model as a glider ?

[Chris Bott - Moderator]

Tim I'm with Peter, somewhere along the line the supplies will come together and one of the lines (almost certainly the negative) WILL be commoned together, while the other line carries the regulator and the switching/combining. This instantly puts a short across one of the batteries.

The only way this could be avoided in the battery backer would be to have the supplies converted to AC, passed through isolating transformers and then back to DC again. Efficiency losses and the cost of this system would mean it's extremely unlikely that SM have done it this way.

Can you buy a UBEC for 42V (10S?) If so I think I'd use that one side and a Rx battery the other.

Personally I think you'd be better off with "just" a 2S A123 pack as a Receiver battery. Very reliable/robust, will supply any amount of current that the servos might want and can be charged quickly. Just check the Rx and servos are rated for a 5cell nimh (similar voltage just off charge)

[Steve Hargreaves - Moderator]

I agree with Chris here....I think leave the big batteries to supply the motor & fit a seperate radio power supply......you can still feed two seperate batteries into the battery backer if you wish....2S LiPos or the like.

The only "disadvantage" I can see is that you will need to make sure the radio battery is fully charged as well as the motor battery....

Might be worth pointing out that 10S is around 40 volts too & at DC thats quite a high voltage so be careful Tim & don't let the battery cause a circuit through you (i'm sure you already knew that.....)

[Peter Beeney]

When I first read Codename-John’s post I assumed that the Backer may have been some sort of break-before-make double pole changeover switch, which would also work ok in this configuration. However, on reflection, and something that Tim said in the OP, about the packs discharging together, made me go back and check, and it certainly does seem as though the Backer is very much just two parallel paths with a diode in each, to ensure isolation in the event of a failure.

So I think we will have to safely assume that the negative is common throughout. If I may, I’d like to borrow John’s excellent layout drawing for a moment and just add to it to try and make the picture a bit clearer. We can now continue the neg. line straight through UBEC’s 1 and 2, and we can also make a little loop in the Backer between the two negs in there.Whilst we’re at it we can continue the neg from here straight on up through the ESC, then on to the rx, no break here either, and connect it back to our loop in the Backer. Now we have a complete ring circuit.

If we now disconnect the two BEC supplies underneath the Backer, so that they are not connected to anything, connect everything else together, and check the output of number one UBEC we find that’s 6 volts, lovely jubbly, but the output from number two may give us slight cause for concern. I have a suspicion that the voltage on the two output wires here may be only 6 volts, fine, but if we check from this positive to anywhere else on the common ring it will somewhat higher. This is because it’s neg. pole is the positive end of battery 1, and it’s already raised 18.5 volts with respect to the 0 volts at the left hand side of battery one. I have a bit of a feeling in me water that this reading might give us around 31 volts, that’s the battery voltage less the 6 volts of the regulator. To read this voltage properly, we would need to connect the voltmeter red to battery 1 red and the voltmeter black to the red output of UBEC 2.

Moving on apace, lets now abandon all caution to the winds and plug both regulator outputs into the Backer. Here we can follow the adventures of the positive supply along the positive wire from battery number 1, along to the right and now as the negative connection into UBEC number 2, straight down through this and then up to the Backer, hurrying round the little loop, back out, down and around to the left along the bottom, turning up into UBEC 1, it won’t pause or dally here, either, but rushes out of the top and on to the negative pole of battery 1. To complete the circuit in record time! There is also the additional path via the Backer, then the rx and through the ESC, just to make sure the job’s done properly.

So, the question we might now be asking ourselves, …whilst we are trying to clear the smoke from the burning insulation…... to say nothing of any flames…...have we indeed put a fairly low resistance short across battery number one…
I think if I were doing this, I would be very careful indeed before I connected anything together. Paralleling devices together on series connected cells can be fraught, because it’s easy to short the cells.
Plus I have to say, I’ve not verified any of this in any way, I’ve more or less made it up as I’ve gone along, so I could easily be mistaken. On all counts…

However I would consider, all in all, it at least wants some careful consideration.

Good Luck.

PB

[Chris Bott - Moderator]

Maybe this will help Peter's description? Check out the thin black line that shorts out battery 2.

[Peter Beeney]

Nice one, Chris. Slightly mirror image, but exactly so. The instantaneous dead-short fault current can be a lot of amps, things tend to warm up a bit sharpish, even at these relatively low modelling voltage levels. The internal resistance of a fully charged pack can be quite low.
Never personally had one myself, touch wood, and I’d be the first to admit there’s a first time for everything, but I’ve sorted out the aftermath of one or two bigger scale whoopsies, when someone’s got it wrong.
I'm very cautious…….and scared!!

PB

[Chris Bott - Moderator]

Sorry can't help it. I always think left to right with positve top and negative below.

You are right there is likely to be an extra path too, from Rx negative back through the ESC. Unless it's a proper Optically Isolated ESC of course.

[Tim Mackey]

Tim, can I suggest the following.

Buy a genuine A123 systems LiFe 2s pack, wire it through a standard switch harness, switch it on and fly.

KISS

[Tim Hooper]

Brilliant! Thank you for your input gents!

I'm not an electro-techy by any means, but Chris's diagram makes a lot of sense to me - especially the bit about dead-shorting Battery#2.

I was just being lazy about not carrying a seperate rx pack, I suppose.....

tim

[Peter Beeney]

Chris, now I think you you’re trying to wind me up. Surely it’s only the signal connection that’s optically isolated? I don’t have any opto ESCs, but won’t they have three wires to the receiver, a red, black and signal? If they have a black I very much suspect that will be connected to the main black which goes back to the battery. If this is the case then I guess we must still have that common path.

I’d certainly agree with Tim, I’d just use a single pack of two 123s, and I’d give them a little bit of tlc occasionally. They won’t fail, unless I totally ignore them. To that end I’d add one of the Hextronik onboard LiFe receiver voltage monitors, just in case I should happen to forget to charge them one day.

I’m also just lazy….

PB

[Martin Harris - Moderator]

Having seen the discharge curve (a Sherman tank over Beachey Head comes to mind) for an A123 I wouldn't advise going by a voltage monitor - far safer to do regular top-up charges and occasional capacity checks which, if the information I have is correct, won't do them any harm.

[Chris Bott - Moderator]

Peter nope, not pulling your leg, a proper OPTO isolated ESC has it's own BEC that supplies the ESC circuitry only. The receiver and servos require a different supply of their own, usually a Rx battery.

The ONLY connection between Rx and ESC is optical. The Rx pulses drive an LED, the resulting PPM light pulses are picked up by a photo detector and are passed on to the ESC circuits. (LED and detector are hidden inside one chip).

The idea was to completely isolate any electrical interference generated in the power circuit from entering the RF parts of the installation.

Martin - absolutely. The A123 battery tends to hold it's voltage constant until it is flat, then it is er.. flat. This has been found when it's used as a flight battery to power a motor, I'm not as sure this is the case when servo type currents are drawn, but it's safer to assume it is so.

[Tim Mackey]

Correct on the Life volt drop monitor - its the one drawback with em, they seem to get down to around 3V per cell and stay there for ages, then suddenly BANG they are flat empty in seconds. I did consider a microscream type alarm modded to 2.5V per cell, but again...same problem would occur.

I now simply do a typical flight, from a fully chrged fresh battery, and then see how much goes back in on a charge. That says how much the flight used.... and if, for simplicity sake, the figure is say 100ma for a ten minute flight, then I know I am OK to fly for around 3 hours.

10 mins = 100ma

100 mins = 1000ma

200 mins = 2000ma ( 200 mins / 60 = 3.3 hours )

and the cells still have about 7% capacity left ( 300ma )

YMMV.

 

Alternatively, I also have used an eagle tree logger to determine the exact power consumed during the flight.

Edited By Tim Mackey on 20/02/2012 17:41:44

[Steve Hargreaves - Moderator]

Slightly off topic but still relevant I hope.....

For LiFe cells the A123 cells seem to be the "Gold Standard" & I understand that these are extremely robust & can be charged at a zillion C & will probably survive a nuclear strike but are the foil types of LiFe cells like these any good does anyone know? Obviously they aren't an A123 battery but the chemistry is the same so should retain all the good things about LiFe batteries like zero self discharge, voltage stability, fast recharge etc

I'm using a 3 cell version in my Tx at the moment & it has been most excellent.....

I wondered what the panel thought??