6X16650 wired parallel to direct drive P7

[killforfood]

I haven't been real impressed with the available P7 driver boards and will probably just make some simple direct drive 3D NiMh Maglites.
I also have this battery and wonder what the Pros and cons would be if I re-assembled all six cells in parallel and used it to direct drive a P7?

 

[qwertyydude]

Make sure you have some real good heatsinking. I tried a 3 cell 4600 mah pack made from 3 rc car sub-c cells. On a full charge fresh off the charger, my P7 test light was all the way up to 4.5 amps.

 

[killforfood]

Is there any problem with charging 6 batteries in parallel?
This is the charger I'll be using.

I'm not sure what I was thinking when I typed up this question. I must have had Lithium-ion on the brain.
If I wire all six of these batteries in parallel they will be about 1/3 the needed voltage to energize my emitter. What's really needed to properly direct drive a P7 is join in series two sets of three parallel wired batteries. DUH!

 

[Kremer]

I think you mean 3S2P for those six cells to get 3.6V and 7.6Ah, which should work great for the P7. But unless I'm mistaken that charger says 4-8 cells, you will need one for 3.

 

[Mr Happy]

And bear in mind that charging parallel configurations of NiMH is not a good idea. You will need to split them and charge the two 3S sets separately, or charge all six in series.

 

[IMSabbel]

About the driving in parallel:

Do not underestimate the current you can get with a low Vf P7.
Freshly charged you might very well push >6A through the led with that much cells in parallel (and thus low internal resistance)

 

[mdocod]

Have you considered an MCE? You could wire the emitters in the package 2S2P which would work off a 7.2V pack *reasonably* well, setting it up for direct drive may require some experimentation with your particular MCE, I would suggest using a variable resistor, in conjunction with a pair of DMMs measuring current/resistance to figure out if you are going to need a resistor of sorts inline with that LED to keep it under control on that pack (and to determine what resistor value you need if needed). With a 2S2P configuration and higher overall circuit voltage, combined with lower current, you'll have less voltage loss from resistance and probably won't be able to do a DIRECT drive, but might be able to get close to it with a low value resistor. Remember, with it wired like that, you're not aiming for 2.8A anymore, you're aiming for ~1.2-1.4A instead.

 

[killforfood]

I think you mean 3S2P for those six cells to get 3.6V and 7.6Ah, which should work great for the P7. But unless I'm mistaken that charger says 4-8 cells, you will need one for 3.

Yah thanks for clearing that up. I knew what I wanted to say but just didn't know the proper nomenclature. Is a 3S2P configuration considered to be 2 cells by the charger?

And bear in mind that charging parallel configurations of NiMH is not a good idea. You will need to split them and charge the two 3S sets separately, or charge all six in series.

This should be doable with a simple charging harness.

About the driving in parallel:

Do not underestimate the current you can get with a low Vf P7.
Freshly charged you might very well push >6A through the led with that much cells in parallel (and thus low internal resistance)

I hadn't considered this issue. I'm thinking that Mdocod may have the elegant solution by using an MCE 2S2P

Have you considered an MCE? You could wire the emitters in the package 2S2P which would work off a 7.2V pack *reasonably* well, setting it up for direct drive may require some experimentation with your particular MCE, I would suggest using a variable resistor, in conjunction with a pair of DMMs measuring current/resistance to figure out if you are going to need a resistor of sorts inline with that LED to keep it under control on that pack (and to determine what resistor value you need if needed). With a 2S2P configuration and higher overall circuit voltage, combined with lower current, you'll have less voltage loss from resistance and probably won't be able to do a DIRECT drive, but might be able to get close to it with a low value resistor. Remember, with it wired like that, you're not aiming for 2.8A anymore, you're aiming for ~1.2-1.4A instead.

Good information, thanks. Do you have a link that explains the proper DMM hookup locations? What amount of resistance are we talking about here? I'll have to dig through some old parts and see what kind of pots I can come up with that will handle the current loads.

Thanks for all the info guys. This project is just beginning to get fun. I'm going to attempt to tri bore my 2D today. I'll do it using a custom made swedging tool. Should prove interesting, cross you fingers.:sweat:

 

[mdocod]

Here's how I would do it:

Mount the MCE to a heat-sink, and setup whatever switch (be it the stock mag switch or whatever) is going to be used in the final design, do any resistance fixes before the test.

Solder/wire the legs of the MCE to be series/parallel as called for.

Set one multi-meter to the "10A" or 20A" setting. (depending on what yours has, I know it varies from meter to meter, one of mine is 10, the other 20, hehe)..

The other multi-meter should be set to ohms. It should be arranged to measure the total resistance of the pot combined with the multi-meter that is being used to take the current measurements.

here's what I think you need to set up for this test:

Start with the POT set to maximum and turn down the resistance (turn up the "current") until you reach your desired current. Do the test on a fresh from the charger pack, and aim a little over spec IMO. I'd shoot for ~1.5A on a freshly charged pack, since the pack will only over-drive it this much when you are fresh from the charger, and you'll want to balance that performance with more normal times, when the pack has been sitting a few days settling.

I would also consider doing a few more tests, like one with the pack charged, but rested for 24 hours, to see what resistance is required to get an ideal running current, and compare that to the previous tests. Basically, in the end, you are going to need to find a compromise of over-drive on a fresh pack, and normal running on "rested" pack that you are OK with. Most importantly, need to make sure that the maximum possible initial current is within reason, and will not trigger thermal runaway.

Let me know if that helps

I've never done it myself, but this is the way to optimize a particular LED to a particular battery pack without the use of a regulated driver.

[edit in] To be honest, I'm not 100% certain the exact ballpark of resistance you are going to be shooting for as I'm not too incredibly familiar with these setups. I know "of" them, and understand the basics, but haven't set anything like this up myself. The effective resistance of an LED in a circuit is a fast changing thing, it drops as voltage rises, and drops as heat rises if I understand correctly, which creates a bit of a catch 22 that I don't know how to process. If I'm doing my math right (please someone double check me on this), the total circuit resistance would need to be somewhere in the ballpark of 5-6 ohms for the operating current to be ~1.4A from a ~7.2V source. So I'm guessing that the resistor you will be adding to the circuit in the end will NOT be a very high value resistor, like something in the neighborhood of 0.5 - 2 ohm would be my guess. Hopefully an LED expert will chime in here and give further input.

-----------------------

OTHER IDEAS:

Having said all that, regulation would still be best... have you looked at all your options? What about a MaxFlex driving a MCE with all the emitters in series! This would give you perfectly equal current across all the dies, and flat regulation through the discharge. You could set it to 700mA and it would be just about perfect if you ask me. (equivalent to running ~2.8A on a P7 or MCE wired parallel)
Check this out:
http://www.taskled.com/maxflex.html

If I did my math right, you'd be drawing ~1.66A from the battery pack after efficiency losses, which should give you a solid, regulated 2+ hours of runtime.


Eric