Project help with Lithium Cells

[james2k2]

Hi All,

I have an LED project that needs to be powered by battery and offer at least 4 hours battery life when all LED's are lit (it has 18 1/2W LED's).

Please be patient with me, lithium powered systems are new to me, and after working in the aviation industry I am well aware of the safety precautions and what happens when cells go bad!

Cost is an issue as I want to make several of these, but happy to spend a little more if it makes it reliable and SAFE!

I have tried using 10x NiMh AA's, and in all fairness this didn't work too badly. These were 2500mah rated batteries, with discharge graphs (done by people on other forums) suggesting they aren't far off the mark. The problem with these is that as the voltage drops, so does the led intensity. They are also quite bulky in reality.

My next idea, was to use Lithium cells. So I bought a ready made pack that came with a charger as well. It is made up of a 2s3p 18650 configuration and is rated at a rather conservative 1600mah per battery thus giving me 8.4V, 4800mah with a nominal voltage of 7.4V. I wired it up and to my disappointment, only got 2 hours 37 mins out of it. Needless to say I think I've been hit by the ebay false battery claims! I've attached a discharge graph, and at 4.3V the battery protection kicks in and stops the pack functioning until it receives charge.



So now I'm looking for a better solution. I've decided on two possibilities so far, but this is where I need help. The Turnigy branded packs (at HobbyKing UK) have had good reviews and in some cases even exceed their rated current. These have no protection though so would need that and a charging system. I found a 2s/3s constant current Lithium charger on ebay, which is fine but as this only handles charging, I need a protection circuit too - can the likes of these: eBay auction: #http://www.ebay.co.uk/itm/1X-2S-Li-ion-Lithium-Battery-18650-Charger-Protection-Board-Module-3A-7-4V-8-4V-/191887828279?hash=item2cad67e537:g:hMMAAOSwnNBXUPcV be wired without the balance connection just to offer undervoltage protection, or is there a better solution?

My next idea was to use a bank of 8 individual 18650's in holders and wire them up in a 2s4p configuration with the same 2s protection board above, but then I'm basically in the same boat, and individual batteries were coming out quite a bit more expensive than the Turnigy packs.

The main problem I have is that I need to be able to power the device AND charge at the same time. My thoughts for this were to use a dpdt relay and simply switch the output terminals off from the battery when plugged in.

A bit more info on the project is that for the Lithium battery option, I've got a DC-DC Boost Converter based on the XL6009E1. I'm not keen on it's quality as the output cap gets quite hot really considering I'm only drawing about an amp when it's rated for 4A. I also only got about 87% efficiency from it (calculated by doing input power consumption vs output power consumption). I have therefore opted for the MT3608 based unit. This runs at 1.2Mhz instead of 400Khz, doesn't need as higher rated parts and should therefore be more reliable and cooler equaling greater efficiency. I'm still waiting for this to arrive. In all fairness, the battery module I have currently could be OK and the efficiency of the boost converter could be the problem, because as the voltage drops, efficiency goes way down.

Any advice is very welcome. Many thanks,
James

[Aodhan145]

What brand were your batteries? Ultrafire are notorious for lying about specs.

[doobedoobedo]

You could use an active low voltage detector to switch an n-channel mosfet for the under voltage protection. Probably about the simplest method.

[james2k2]

Thanks for your replies.

The battery pack brand is unknown. I didn't want to open it up just yet as it may prove useful is another less power hungry application. I have heard that almost all the brands with *fire in the name are pretty much useless. If I went for the self-build pack, I was looking at the likes of LG, Efest and Samsung, although Sanyo has been mentioned as being reasonable too. I've learned to stay clear of any batteries suggesting radical claims like 4000mah, and even 5000mah per 18650 for $2 

@doodedoobedo: Do you know of any low voltage detectors that are ready made on a board please? Yes I am being lazy.

[james2k2]

Ok, so to try and aid where I'm up to, I've done the below block diagram. Basically, it's the wiring up in the middle that I need help with.

I know this is simple, but I'm banging my head against the wall trying to work out how best to do it... 

Whenever I think of a way, I end up with some sort of loop where the battery would effectively be charging itself by looping via the charger.

Link to diagram: https://drive.google.com/open?id=0Bw9EPPuDoxJsWTluUUljb2k3WjQ

[amyk]

I have an LED project that needs to be powered by battery and offer at least 4 hours battery life when all LED's are lit (it has 18 1/2W LED's).

Effectively 36Wh of energy, correct?

I have tried using 10x NiMh AA's, and in all fairness this didn't work too badly. These were 2500mah rated batteries, with discharge graphs (done by people on other forums) suggesting they aren't far off the mark. The problem with these is that as the voltage drops, so does the led intensity. They are also quite bulky in reality.

With 9W of power, I hope you're not using only a current-limiting resistor, as that will waste a significant amount of the energy available. A constant-current switching regulator designed specifically to drive LEDs would be your best option.

My next idea, was to use Lithium cells. So I bought a ready made pack that came with a charger as well. It is made up of a 2s3p 18650 configuration and is rated at a rather conservative 1600mah per battery thus giving me 8.4V, 4800mah with a nominal voltage of 7.4V. I wired it up and to my disappointment, only got 2 hours 37 mins out of it. Needless to say I think I've been hit by the ebay false battery claims! I've attached a discharge graph, and at 4.3V the battery protection kicks in and stops the pack functioning until it receives charge.

4.3V for a 2S pack (2.15V/cell) is far too low for a cutoff and will shorten the cell life significantly; 3V/cell (6V total) is usually the lower limit. According to this test of a 1600mAh 18650, usable energy is a bit below 6Wh, so 6 of those cells will already give less than the 36Wh you require. Add on top of that the losses in any DC-DC converter you're using, and if you really are just using current-limiting resistors for the LEDs (please specify), 2:37 doesn't sound so far off.

[james2k2]

Hi Amy, thanks for your reply.

Point 1) Correct.

Point 2) There's a lot more complex circuitry going on than I've divulged in the LED running department. It has digital dimming of 3 channels (RGB), and a wireless controller and a few other bits n bobs. No need to be concerned about that bit, although I should probably note the control circuitry and wireless receiver uses about a watt. So it is closer to 10W of consumed power when running.

Point 3) I do agree with you here really, but I have no control over it as it was a ready-made pack and I've simply noted my findings. I have been using guides like this one: https://learn.adafruit.com/li-ion-and-lipoly-batteries/voltages

[james2k2]

Sorry to post again, taking into account of what Amy said, I did the following calculation:

5.844W * 6 = 35.064Wh

Approximation of 10W power consumption by light... 35.064 / 10 = 3.5064 Hours

87% maximum efficiency when battery is at 7.4V... 3.5064 * 0.87 = 3.05 Hours.

In hindsight, this battery pack isn't all that bad, especially considering as the voltage drops, the efficiency of the dc-dc module greatly reduces.

There is still a problem though. I would continue use of this style of ready made pack (with a higher capacity) with the exception that the power supply has a tendency to think the battery is charged and cuts out randomly when the battery is low, thus switching the LED's on and off annoyingly.

[doobedoobedo]

All batteries drop voltage under load due to their internal resistance. Li-ion batteries then tend to recover some more voltage once the load is removed. I think you may need some extra hysteresis in you under-voltage cut-off to stop the flashing.

[james2k2]

Sorry Doobedoo... Please forgive me. How would I go about adding extra hysteresis? Am I right in thinking something like a large cap across the input?

Thanks,
James

[doobedoobedo]

The hysteresis will be built into the chip on your protection board. A cap across the input won't make any difference.

If you can find out what the SOT-23 part is, then you can probably change the hysteresis setting by changing one of the resistors on the board.

[james2k2]

Thank you. Before I go opening up the pack, I'll wait for my new dc-dc board to arrive and I'll see if it offers any extra life. 3 Hours may be enough for what I want, and then I can look at the pack in more detail.

I shall upload a new discharge graph when I can..

Thanks again for all your help so far

[Kilrah]

The "intelligence" of cutting power at the right level and preventing the system from restarting when voltage rises again needs to be part of your system design.