24 Batteries - 4S6P Configuration - Trouble?

I have a light in the planning stage, I’ll spare the details for now except that I’m thinking MT-G2 driven on an adequate heat-sink at around 4.0A. Driver with some sort of step-down to 30% or less after a couple of minutes to extended run time.

The light will be intended for long shelf life then extended power outage use.

The thing about this host is that is has enough room for a small motorcycle battery (but I don’t need to fill all that space with batteries).

I thought about Li-Ion initially but self-discharge and charging issue shifted me to Lithium primarys, specifically Energizer Ultimate Lithium AAs.

If I were to use 24 batteries in 4S6P configuration I would end up with approximately 7V open circuit and almost 15000 mAh (if I understand HKJs review properly) at 3.0A load and 18000 mAh at 1.0A.

An MT-G2 would provide the light I’m looking for and the batteries would provide both the shelf life and run time.

My question though is am I playing with fire (literally) by wiring these batteries in 4S6P configuration?

Sheesh, lithium primaries?! That’s gonna get expensive reeeaaallly fast.

Why not go with some of the A123 cells? Not the highest mAh batteries out there, sure, but massive current draw potential. Build in a nice little balancing connector to your light, get a $20 hobby charger…charge that sucker at 4A and be fully charged in under an hour? That sounds much better to me.

Centura D's 5s2p for 16AH?

I understand your point about the expense but this light is for a friend and in reality will be nothing more than a conversation piece.

It’s a 30s Lightmaster Lantern Searchlight that will sit on the mantel above the fireplace in his early 1800s farmhouse. It will be turned on a few times then set aside.

I want it to be practical in the unlikely event that it has to be used, maybe 5 or 10 years from now.

I don't know the answer to your question, but it seems reasonable that it should be fine. Otherwise, there would be issues with them being used on all the series-type devices out there.

I would parallel at the cell level, not overall. That way the individual banks of cells will be able to keep each other in balance should a cell decide to go south. That is how I plan to arrange my scratch build cells. Somthing like this. Horizontal lines are contact plates. Vertical lines are the cells.

__

||||||

__

||||||

__

||||||

__

||||||

__

I don’t see a problem with series/parallel lithium primaries. For example the Nitecore TM26 allows 2s4p 123A lithium primary cells, although they are not wired in the way recommended by ImA4Wheelr (which I also recommend).

You probably need to change your math a little bit though. Since you have a 6P configuration, divide your current draw by 6 before using HKJ’s graphs (which were generated for individual cells). So for a 3A draw from the pack you should follow the green line on the graph (3A / 6 = 0.5A). This nets you approximately 3.15Ah * 6 = 18.9Ah. The discharge ranges from 1.45v down to about 1.25v before falling off rapidly, so you are only generating 5.8v at the beginning of the discharge and 5v at the end. The LED won’t light below 5v and at 5v it barely lights up. At 5.8v you only get a 1A drive current.

You just don’t have enough voltage to make this project work properly with a 4s config. Better do 5s5p and use a driver to regulate the current.

I’m using djozz’s MT-G2 crash test graph for reference on this as well as HKJ’s Energizer Ultimate Lithium AA review:
https://budgetlightforum.com/t/-/19600#comment-417559
http://lygte-info.dk/review/batteries2012/Energizer%20Ultimate%20Lithium%20AA%20UK.html

What about just using a 12V SLA battery and a led like this?

http://www.leddna.com/10w-square-3000k-3500k-warm-white-led-light-850lm-120-degree-10vdc-12vdc/

or this?

http://www.ebay.com/itm/10W-Round-LED-warm-White-High-Power-900LM-LED-Chip-9-12V-Bulb-Lamp-brightness-/321395932001?pt=LH_DefaultDomain_0&hash=item4ad4b0bb61

It would make for a warm white, like the light would normally be and also very floody like the original.

12V battery direct drive, also old school. Or 8 "C" Alkalines?

Just a thought...

I just noticed that if the stock bulb is in focus you’re going to need a massive podium to get the LED out there.

I believe one of the OPs main goals is to have a load and forget light that could still be used in 5 to 10 years without any battery maintenance, hence his desire to use lithium primaries. I’d opt for a driver that runs on fewer cells and use the extra space in the host as spare cells storage.

KuoH

Things got a little crazy around here last weekend, thanks for the responses.

  • ImA4Wheeler, Good point about the battery configuration. I had to think about it for a minute but makes sense. Regardless of what I end up doing with this light I’ll consider it for future use.
  • wight, Another good point, I was only considering the OC voltage and total mAh rating at 3A.
  • OL, I thought about using SLA but that needs to be maintained.

My first thought was to use an XM-L2 on a copper star, nice and warm in the 3K range, I think I’ll head in that direction.

An XM-L2 helps with a couple of issues.
First off power source. I’ve built a couple of XM-L2 lights using Q-Lite or KD V2 drivers driven with 3X lithium primaries and that’s what I would prefer to use.
Second, dealing with the reflector. I haven’t measured it yet but the opening in the reflector is less than 20mm but larger than 16mm, I can use a 16mm Noctigon without modifying the reflector.

As wight mentioned, the focal point of the reflector is fairly high. I know I’ll never get a solid spot out of the light but I plan on testing a bit to find the best height for an LED. Once I have that measurement I’ll make a pedestal from copper rod and mount it to a heatsink in the lamp body.
I have a bunch of CPU heatsinks from my PC building days, I was looking through them last week and came across a likely candidate. It’s about 2”x2.5” with fins about 1.5” high, as I cleaned all the dust and crap off it I realized it has a 1” copper core pressed into the aluminum body. Nice! I’ll have to cut it down but it will work fine.

Another cool thing about this light, it’s focusable, like a Mag-Lite. There are springs behind the reflector, screw the bezel out and the reflector moves forward with it.

Thanks for the info and suggestions. This will be a long term project due to lack of time but I’ll try to remember to take photos and post when done.

personally would not use lithium primaries in series if left in the light connected over a long period.

You will have a lot of cells and all it would take is one cell to misbehave and there could be problems. You could make a protection circuit to prevent reverse charging I suppose.

As you have the space could you not fit a charging circuit, and monitoring circuit with a warning light so you friend could just plug in a say a wall-wort and top it up when needed?

What specifically are you worried might happen with that situation?

Even if you kept lithium primaries new in their boxes I would not bet on 24 staying the same/equal over the wanted time period. On several occasions I have had lithium primaries develop disparities between individual originally matched cells both new and in lights over time.

Lith primaries do have a very good shelf life but even from new if used in multiples they should be checked before use.

Connected in a series/parallel pack, if a cell/s malfunction = reverse charging = worse case scenario Booom.

It’s easy for me to believe that a bunch of different Lithium primaries will degrade differently over time. It’s harder for me to get worried about safety issues. While idle you won’t get reverse charging, only parallel cells pulling each other down to the lowest resting voltage. That energy would bleed off as heat as far as I know.

You wouldn’t see the reverse charge happen until the light is in use. You won’t get the reverse charge until one parallel set hits 0v. By the time that 1s has gone flat you’ve lost a a significant fraction of your total pack voltage - I figure your output current will be pretty low unless you had a regulated setup with plenty of headroom. Assuming you didn’t, it wouldn’t be a high rate of charge, and if these things are anything like other chemistries which I more about then a slow and small reverse charge won’t cause an explosion. Assuming you did have a lot headroom over your output voltage the current would stay high resulting in the battery(s) trying to take a reverse charge at a rate of several amps. That doesn’t sound good, but in practice I’m not really sure what would happen.

As far as reverse charging, do we know that the Li/FeS2 chemistry acts badly when taken flat and reverse charged? I’d be happy to get educated on why we should watch out for this.

Plus: who cares if a cell or two explodes anyway? If it happens due to a reverse-charge, that will be during use - eg the explosion/fire will be an attended one. I don’t really worry about those, it’s the unattended explosions/fires I care about… like the ones that happen when a battery charging setup goes haywire.

Also, have you considered a hotwire mod? You could use a typical medium voltage bulb, look at LuxLuthor’s CPF thread some inspiration. I think a high voltage incandescent bulb might be a great match for 24s1p but you’d have to put in some leg work to figure that out. Of course this would change a few factors:

  • 24s1p might be enough to burn up the PTC if a dead-short happened.
  • You’d lose that “check out this modernized lamp” angle…
  • …but you’d gain that “see this old incandescent lamp? It’s still brighter than those modern LED doohickey’s!” angle.
  • Thermal management needs would change, I assume you’d build a simple stalk that lets a ceramic bulb base screw into it on the end. Primarily you’d just want to isolate the bulb from anything that melts I think…
  • You may find that an incan filament will focus better in that reflector than an LED. YMMV
  • regulated driver not required

I think the OSRAM 64458S bulb mentioned in the link above might be a decent choice driven on 8s3p. You’d need to use a hotwire calculator (find on CPF, hotwire.xls is one possibility) to figure out the wattage you’d get from that, but maybe like 60-70w? If so you’d have a little under 9A draw, should be OK for 3-parallel.

:zipper_mouth_face:

with that I’m out