Switch power between 2 Li-Pols

[pabloantoine]

This is a 15ft blimp with auxiliary power source...
I am searching for a simple switch or relay solution to bring online the back-up battery while the main battery is isolated before it goes dead (they are both Li-Pols, 12V 8000/4000 mAh).
The Amps are high (30A) and the Rx must be kept online during this switch.
For the moment I came up with a simple solution using 2x 30A switches in oposition, actuated by a servo turning an eccentric plastic circle.
Isn't there any better way to do this? The components I can practically find limit my options - I couldn't find anything better than these 30A Auto switches.
A latching relay will do from your experience? Or what else?
Thank you for help

[Chippie]

I guess you could do it using a comparator and a mosfet to switch the standby battery over when the main pack is capacity limited...thinking aloud here...

Ermm maybe two mosfets....One switched on when the power is initally applied, the other off, then as above, switch to standby and turn of the main pack by turning off its mosfet..Probably do-able...with a bit of thought

[Chickenthief]

Whether You use them one at a time or run them parallel, Your runtime will be the same, right?
So switch or not You get the same amount of airtime, before You're forced down.
Set Your alarm in Your wristwatch and go by that, or am I missing the boat?

[pabloantoine]

No, time is not the same. The 8000 MAh serves the lift/descent motors, rudder and Rx, so it's taking a lot of abuse. The 4000 mAh serves only a booster motor which can be used or not depending of the wind strength. You can say that time is limited by the ammount of maneuvers on the 8000 mAh, then the 4000 mAh can take over for a short time at full load and help return. A weak 8000 mAh (or a parallel pack with both batts for that matter) will fail you when you might need the whole power to return safely.
The voltage is monitored - let's say you see that a cell in the 8000 mAh is going dangerously low, but you need to return and at full power the motors will max the Amps so badly that the voltage of that cell goes south - kill the pack actually.
It's impossible to predict the flight time -it's around 30 mins at least - because it depends of how much wind you encounter at higher altitude.

The issue here is not whether I need a back up circuit - that I already know from experience- but how to switch reliably between the batteries.
My low tech approach is working now, but the servo really struggles with the switches, which have stuborn springs to carry the 30 Amp per circuit. It might fail when it's needed the most in my nightmares...

[Calmdown]

If I understand correctly, the main issue is that your 'switching' servo has difficulty activating your existing arrangement of stiffly sprung auto-switches?
The simplest solution to me is to use your 'switching' servo to activate a micro-switch and have the micro-switch turn a relay on/off and the relay to switch your 'main motor' between the two batteries.
This way the physical load on the servo is almost nothing, and the 30A power switching is done by the relay contacts.
It doesn't have to be a micro-switch, any small push button switch that takes a gentle/light pressure to activate it will do, and the current rating of the switch doesn't have to be very high because its only switching the relay on/off, not the main 30A.
Let me know if you want a quick circuit diagram to explain things more clearly, or if you have any questions.
All the best, Andy.

[Calmdown]

Forgot to mention a couple of other things...

1: As the 4000mAh is the backup battery '..and may or may not be used during the flight..', I'd suggest you power the rx from that battery, not the 8000mAh one. Reason being is that if the 8000mAh is the battery under the most strain, you are better off powering the rx from the 4000mAh battery which is not doing so much work - gotta ease those nighmares!

2; I see from your diagram you have disconnected the '+' wire from ESC 3, but you have both the '+' wires of ESC 1 and ESC 2 connected. Unless the ESC manufacturer has said its ok to have these two ESC's '+' wires connected, you will need to disconnect one of the '+' wires from either ESC 1 or ESC 2. Unless.... you go with my sugestion in point 1, which is using the 4000mAh battery as the rx supply, in which case, disconnect the '+' wires on ESC 1 and ESC 2, but reconnect the '+' wires from ESC 3.

Hope it makes sense!
Andy.

[pabloantoine]

I totally agree with every point you are saying.
I can use a 5A mini roller switch with 3 contacts in opposition, very reliable (common at Radio Shack). I am pretty positive that the relays would do -they use them on large scale in cars- only here we get to the tricky point. As low tech as my crude sollution is, it doesn't consume any amps. Correct me if I am wrong, but the only relays that would do as well in consumption terms will be latching relays - or impulse relays - I don't know much about my options here because I am rusty in relays. Do they need any special circuitry?
Maybe you know any relays that might do without consumming too much to keep the armatures closed?
Thank you so much for the suggestions.

[Calmdown]

OK, I had a quick look on the Radio Shack web site but couldn't see a relay that would be suitable.
You need a 'Single Pole Double Throw' (SPDT) 12v relay which has its contacts rated at more than your maximum load (I know you mentioned using 30A auto-switches, so I just used 30A as the starting point for the relay specification).

As a reminder of a SPDT relay connections, see the attached diagram of the relay:
C = Common
NC = Normally Closed
NO = Normally open

I found an example of a suitable relay (40A) at DigiKey:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=255-1825-ND
You may be able to find a suitable one at a local auto electric supplier...

The way the circuit works (see the second attachment) means that the relay is only active when you switch to the 4000mAh battery. e.g. when the 8000mAh battery is providing the power, the relay is not energised and so is not consuming any power. When the relay is energised, it will only consume about 120mA anyway (the relay specifications will tell you what current it consumes when energised - the Digikey one I linked to consumes 117mA).

When you switch from one battery to the other, you will temporarily disconnect the supply to ESC 1 and ESC 2 so:
1: You MUST have ESC 3 supplying the power to run the rx.
2: You will need to lower the throttle controlling ESC 1 and ESC 2 to re-arm them after the switch over.
3: It would be a good idea to lower the throttle controlling ESC 1 and ESC 2 to zero BEFORE switching between the batteries, otherwise you could be switching a 'full load' from one battery to the other, and while the relay will handle it, it maybe a shock for the battery, plus it will maximise the life of the relay contacts switching a zero load rather than a full load.

Points 2 and 3 most likely apply to your existing setup, but I thought I'd just mention it anyway...

Hope this explains things a little more clearly, sorry for the quick and dirty circuit diagram, I'm at work now!

Any questions, just ask
Andy.

[pabloantoine]

Andy,
Your sollution to the problem really obliges. It is the most pratical I could think of, especially because it doesn't waste current in the basic state, and the key is the SPDT configuration. In my present solution the switches are positioned to actually overlap circuits for a split second and the Rx is never left "dry".
I also think that your suggestions for decreasing the throttle during switch are common sense, although again I don't expect the ESCs will "notice" anything because they are buffered for fraction of second glitches (I think I read that in the specs) -as it takes for the relay to do the switch. But this remains to be seen in the real application.
Anyway, there might be a possible a little extra connection that will do the trick to avoid the re-arming of ESCs 1-2 (could become an issue, more seconds lost in maybe critical moments to recover from high winds!) - I am thinking to add an "anticipatory" switch in the circle around the servo, besides and in front of the main switch that will trigger the relay.
The anticipatory switch will actually put the batteries in parallel for the duration of the switch. It will only handle 5A for a split second, then release the contact.
I have to see practically how it could be achieved - these roller switches are very flexible in what they can do when positioned around servos.
In practice there will still be needed to lower throttles before switching, to decrease the Amps, but maybe the re-arming can be avoided.
What I gain mostly here is reliability - and that was the whole point into looking for an alternative.
You found it. Great solution -thank you again so much.

[Calmdown]

pabloantoine,
ok, I see what you are saying, a few further thoughts/suggestions:

1 - If you use ESC3 (which is permanently supplied by the 4000mAh battery anyway) as the supply to the rx, the rx would not be left 'dry' at any point during the switch over. Even if you go with my suggestion below, or something different, you need to make sure that at no point do you switch over the supply to the rx from one battery to the other. Obvious I know, but easily forgotten :D

2 - If you wanted to have the batteries running in parallel during the switch over (and thus prevent the need for re-arming etc.) you could add a second micro-switch and relay to temporarily parallel the batteries. I wouldn't use a 5A micro-switch on its own to do the 'paralleling' even for a second or so, as the results could be unpredicatable in the short or long term use of the system...

The simplest way I can think of, to add a second micro-switch and relay and to syncronise it with the main 'switch-over' of the batteries, would be to use two circular disks mounted on top of each other (on the switch-over servo), but cut into eccentric cams - similar to the way you have a single cam at the moment activating the two auto-switches.

The 'paralleling' micro-switch would turn on the 'paralleling' relay, which would link to two battery '+' terminals together. You can just use the same specification micro-switch and relay for the 'paralleling' that you are going to use for the switch-over system.

It would sequence as:
1 - At rest, neither micro-switch activated.
2 - As the servo starts to move it activates the 'paralleling' micro-switch.
3 - A little further in the rotation the 'switch-over' micro-switch activates.
4 - Just before the end of the servo travel, the 'paralleling' micro-switch de-avtivates.

As the servo rotation is quite fast you would just have to test it and play about with the activation points on the cams for the two micro-switches, to make sure that the system operates as required.

Hope my ramblings have prompted some ideas for you.

Have fun!
Andy.

[pabloantoine]

I just ordered 2 relays from Digikey, so you definitely gave me a convincing approach - don't worry. I am very handy at materializing the small details and I will work the finals out on the montage itself.
The second relay I ordered was as a spare, but I am ready to go to the double switch/relay if needed. Still, from my experience, there is a lag time of my dualsky ESCs before they consider themselves "left dry" - so it might be enough the simple switch/relay solution - we are talking about fraction of a second.
By the way, I also found impulse latching relays at digikey, but the Amps were too low. The 117 mA of a simple SPDT is more than OK compared to everything else on this circuit to worry about.

And because your attention obliges, you have in the picts the sort of animal I am working with.
It's a 15 ft blimp carrying a Sony HVR-Z1 camera at the heaviest payload (with a bigger envelope that charges 3 tanks of He, or a smaller Sony HC-7 or photo camera on a slimmer 15 ft envelope -that charges 2 tanks of He). The gondola and booster are the same for both configurations. The station is a Futaba 6EX and the Rx a 606RF with six ch, all used now.
I already added a safety device based on an optical circuit that triggers the release of a 300ft recovery spool at radio link loss -this back up circuit is my second fail-safe.


What are you busy with?

[Calmdown]

pabloantoine,
Thank you for sharing the photo's, that is a really impressive machine!

I'm teaching my two sons to fly electric foamies at the moment, and also teaching myself PIC microcontrollers for both work and some RC ideas I have in mind...

I'm pleased my thoughts helped give you some ideas, let me know how you get on with the final design.

Have fun,
Andy.

[peter273]

Wouldn't it be much simpler to
- have low power part only using the 4000mA batt
- have hi power part only using the 8000mA batt
- have one diode in series with the hi power batt
- monitor hi power batt, controlling a spst relay with the monitor
- connect the low power pack to the cathode of the diode in series with the hi-power pack by means of the (normally open) relay
==> no interruption of low power part
==> simple setup
==> simple control
OK, I'm aware of the voltage drop over the hi power diode, so use a thyristor instead and fire it up at startup by using a differentiated pulse superimposed on its kathode voltage (advantage: lowerer voltage drop and it switches off when the low-power pack takes over).

Just my cents worth...

[pabloantoine]

The last solution sounds like an overkill...
I just finished the montage and it works flawlessly. As I hoped the switching time is so short none of the component "noticed" anything.
I even tested switching at high for both throttles and there is no difference. After the switchover one can remove the 8000 mAh from the circuit physically and the rest continues to work -Rx, rudders, winch, main and secondary throttles...
Connect back and switch to parallel, the 8000 mAh takes over for it's circuit and the 4000 mAh gets back to it's previous load.

The Rx in my choice stays on the 8000 mAh circuit by default - and is switched on the back-up circuit without problems because of the same lag that protects the ESCs and the Rx (I observed the Rx red/green LEDs and it doesn't notice the switchover either).
The relay primary is on the 4000 mAh circuit.

My next stake now is to add a telemetry package and keep an eye on the
voltages. Initially I was thinking about a really low tech solution because this is a video rig (notice the LEDs -they are digital gauges showing sequentially the voltage in each cell of the LiPols) and could regularly tilt up and point the camera in flight to the gauges - but practically it distracts from the shooting.
There are telemetry packages with overlay from a secondary mini camera on the main video downlink - and that's what I am going to use.

Thanks again!

[Ron W3FJW]

How about something like this to monitor batteries...

http://www.hobbycity.com/hobbycity/store/uh_viewItem.asp?idProduct=7217&Product_Name=Wireless_battery_tracker_(6s_&_up_to_300mtr)_869.5Mhz

[pabloantoine]

The range sucks a bit -from the reviews (100 m - must be joking).
Try instead
http://www.rangevideo.com/index.php?main_page=product_info&cPath=49&products_id=135&zenid=672e4ca7ff60c9a20cb728cb1e624324
There is also a competing eagletree product with more sensors.
For a price under $300 both are really significant upgrades - if the contraption can take the additional load - which I humbly assume it can in my case.
Thanks anyway for the interest shown in the subject