New Battery technology for upcoming MC-E?

[easilyled]

I have become very used to using just 1 li-ion to power most of my led flashlights.
ie. RCR123 for smaller sized lights and 17670 or 18650 li-ions for sizes equivalent to 2 conventional CR123s.

With the imminent advent of the Cree MC-E, this would not be possible if its wired in series as
the total Vf of the 4 dies would exceed 12V.

I was wondering if anyone had any insight into what battery configuration could be used in these small sized flashlights
to achieve this voltage.

 

[TorchBoy]

Two words: Boost driver. Pick one you like the colour of.

 

[easilyled]

Two words: Boost driver. Pick one you like the colour of.

Could you give me an example of a driver that would work in an 18650 light
in order to light up an MC-E wired in series?

Lets say it would boost the voltage of a single 18650 li-ion to about 12V
whilst supplying a current of 750ma to each of the 4 dies?

I'm afraid my knowledge of drivers is rudimentary and the only ones I'm
familiar with like GD/NG/Wiz drivers wouldn't do the job in this case AFAIK.

Would it be possible to make an Aleph L.E that could incorporate a driver
that would work in the example above?

 

[TexLite]

Check Tigerhawks thread here to get an idea of different possible configurations:
http://www.candlepowerforums.com/vb/showthread.php?t=204439

A boost driver wouldnt work with a single cell in that setup.With 700ma to each die the input current is going to be well over 3 amps,too high for a boost driver.

-Michael

 

[TorchBoy]

I just worked out it would take 10.3 watts to power a 90% efficient boost driver running 4 dice in series at 700 mA (the maximum rating of the MC-E). Then I thought it would be good to check if that's reasonable for an 18650. The first datasheet I downloaded (yay Google) specified a maximum discharge rate of 20 amps... so I admit I fail to see any problem even with a single 18650 torch if using the right driver. :shrug:

You might have heat and space issues, though. Some of the high spec drivers aren't on 17 mm diameter boards.

I'd like to see the variety of suggestions from others. There are bound to be some radical ideas posted once people in other parts of the world wake up (literally).

 

[TexLite]

Its not the battery you have to worry about,its the driver.

Output power to an MC-E would be 11.53 watts with 85% efficiency.

Input power at that level with a single Li-ion would be 3.84 amps.Too high for even the best boost driver.

-Michael

 

[easilyled]

.

 

[TorchBoy]

Re: New battery technology for upcoming MC-E?

TexLite, you may be estimating the Vf of the MC-E a little high there, if 3.1 V at 350 mA is a starting point. Or are you running it over-spec? That aside, you'd need a smallish driver with "Maximum input current 4 A." Surely that wouldn't be too high for some of the best drivers (no matter how many times you say it would be). :thumbsup:

 

[TexLite]

TexLite, you may be estimating the Vf of the MC-E a little high there, if 3.1 V at 350 mA is a starting point. Or are you running it over-spec? That aside, you'd need a smallish driver with "Maximum input current 4 A." Surely that wouldn't be too high for some of the best drivers (no matter how many times you say it would be). :thumbsup:

I'm getting the Vf from the Cree,3.5Vf @ 700ma typical.

Heres how it works out:

Power Output:4 dies x 3.5Vf x 700ma = 9.80 watts

Power input:9.8w / 85% efficiency = 11.53w

Input current:11.53w / 3.6v = 3.20a

The input current is over spec for any driver that I'm aware of,by memory,Fatman is 2 amps,BB/NG are around 1.5 amps,Shark is ~3 amps.Not sure about the new Cool Shark,will have to check,regardless,it wont work with any of the others.

I'm not estimating or overrating anything,I'm using real numbers.

.

Sorry,I understood your question,I should have been more specific with my response.The figures I gave were for Series configuration,not parallel.It dosent matter if you run the emitter in series or in parallel,input current will be the same for either configuration,if run at the same level,input power will be the same.

Oddly enough,DD in parallel is the only configuration that would work with the configuration your talking about,albeit a resistor would probably be needed.

Edit:There was a post there when I typed the answer above.I also had time(between crashes) to check the max input current on the new Sharks.Wayne has worked these up to 4 amps with improved thermal vias,but the bad news is they are .750 inch,which is too big for the Aleph can I think.

-Michael

 

[TorchBoy]

Re: New battery technology for upcoming MC-E?

I'm not sure you are using real-world numbers. Have you measured the Vf at 700 mA? McGizmo measured it at 350 mA, so 3.3 V sounds very reasonable, which would make it 3.02 amps. I've asked if he can get a figure for 700 mA.

But that's not really a big issue. Where did you get the maximum input currents of those drivers? Not from my list, nor from the driver specs from the sounds of it. Doesn't "4 dice can be driven as long as the boost input current does not exceed 4A" ring any bells?

I still see no need to use new battery technology. :twothumbs

 

[easilyled]

................
Sorry,I understood your question,I should have been more specific with my response.The figures I gave were for Series configuration,not parallel.It dosent matter if you run the emitter in series or in parallel,input current will be the same for either configuration,if run at the same level,input power will be the same.
-Michael

Thanks, I realised what you meant after I posted, which was why I deleted it.

So basically any driver that drives the SSC-P7 should also drive the MC-E
if they have similar efficiencies?

In which case a tri-flupic would probably work?

 

[TexLite]

I'm not sure you are using real-world numbers. Have you measured the Vf at 700 mA? McGizmo measured it at 350 mA, so 3.3 V sounds very reasonable, which would make it 3.02 amps. I've asked if he can get a figure for 700 mA.

But that's not really a big issue. Where did you get the maximum input currents of those drivers? Not from my list, nor from the driver specs from the sounds of it. Doesn't "4 dice can be driven as long as the boost input current does not exceed 4A" ring any bells?

I still see no need to use new battery technology. :twothumbs

Cree,the manufacturer,states Vf of 3.4-3.5 typical @ 700ma,which is the current asked about specifically by the OP,you are the one who mentioned 350ma.Read the datasheet:
http://www.cree.com/products/pdf/XLampMC-E.pdf

http://www.cree.com/products/pdf/XLamp_MCE_cutsheet.pdf

TaskLED,Max input for the Fatman 2 amps or less,bottom of the page:
http://www.taskled.com/techfatman.html

Sandwich Shoppe:
Max input for BadBoy "The critical fixed factor is the input current which is around 1.5A."
http://home.comcast.net/~theledguy/cpf_store/sandwiches/techical_data.htm

Max input for the Nexgen:"Note1: Safe mode is ~1.5A input current."
http://theledguy.chainreactionweb.com/product_info.php?cPath=48_49_61&products_id=560

I dont know where the quote "4 dice can be driven as long as the boost input current does not exceed 4A" is coming from,but if its about the Shark,I've already addressed that in the post above.Those are the new design with improved thermal vias,the old were less.Its too big anyway.

All of my numbers came from their respective sources and my math is correct.

It seems to me your more interested in arguing than stating fact.

-Michael

 

[TexLite]

Thanks, I realised what you meant after I posted, which was why I deleted it.

So basically any driver that drives the SSC-P7 should also drive the MC-E
if they have similar efficiencies?

In which case a tri-flupic would probably work?

Yes,thats correct,as long as your driving the MC-E in parallel.Driving the MC-E in series would have different voltage requirements for the driver.

And yes,the Tri-Flupic should work with the emitter wired in parallel.Mac said he got ~3 amps on burst with with a single cell,so it should be fine.

Thanks,
Michael

 

[easilyled]

Yes,thats correct,as long as your driving the MC-E in parallel.Driving the MC-E in series would have different voltage requirements for the driver.

And yes,the Tri-Flupic should work with the emitter wired in parallel.Mac said he got ~3 amps on burst with with a single cell,so it should be fine.

Thanks,
Michael

So could you explain to me then how the requirements for the driver
differ when the MC-E is wired in series please (as opposed to parallel).

(I know that the output voltage would need to be much more and the
output current much less)

So which drivers would you use in which circumstances?

This is what I really want to know.

 

[TexLite]

So could you explain to me then how the requirements for the driver
differ when the MC-E is wired in series please (as opposed to parallel).

(I know that the output voltage would need to be much more and the
output current much less)

So which drivers would you use in which circumstances?

This is what I really want to know.

When wired in series the voltage the driver is required to output would be higher.A driver may be capable of outputting the current requirements,but not the voltage.

I have a couple of HL-V's waiting on the MC-E,I am planning on setting the outputs @ 350ma.Vf in series will be 12.40,in parallel 3.1 Vf.

I could drive it in series from the 9v supply using a boost driver capable of outputting more than the required 12.4v, and 350ma.Thats 4.34 watts total power at the emitter.Each die would see 1.09 watts.

I could drive it in parallel from the 9v supply using a buck driver as long as the Vf(3.1) was below 9v and the driver was capable of outputting 1400ma.Thats 4.34 watts total power at the emitter.Each die would still see 1.09 watts.

Both of these scenarios are perfectly real and doable.

As far as which one I would use in the configuration you asked about,the Tri-flupic would work in parallel,and it looks as though the Cool Shark would work in series.I think there both too big for the Aleph e-can though.

Hope that helps,
Michael

 

[easilyled]

When wired in series the voltage the driver is required to output would be higher.A driver may be capable of outputting the current requirements,but not the voltage.

I have a couple of HL-V's waiting on the MC-E,I am planning on setting the outputs @ 350ma.Vf in series will be 12.40,in parallel 3.1 Vf.

I could drive it in series from the 9v supply using a boost driver capable of outputting more than the required 12.4v, and 350ma.Thats 4.34 watts total power at the emitter.Each die would see 1.09 watts.

I could drive it in parallel from the 9v supply using a buck driver as long as the Vf(3.1) was below 9v and the driver was capable of outputting 1400ma.Thats 4.34 watts total power at the emitter.Each die would still see 1.09 watts.

Both of these scenarios are perfectly real and doable.

As far as which one I would use in the configuration you asked about,the Tri-flupic would work in parallel,and it looks as though the Cool Shark would work in series.I think there both too big for the Aleph e-can though.

Hope that helps,
Michael

Thanks. I take it that you are referring to Hyper-Lux with 3 CR123s?

So you are saying that with one Li-ion (18650), it would be impossible
to find a driver, that would work with the MC-E in series
because the voltage couldn't be boosted to 12V from 4.2V?

This is what I was getting at in my original question at the start of the
thread:-

ie. If the MC-E is wired in series, but it is still desirable to keep the flashlight
compact and/or powered by only one cell, would one need different battery technology?

 

[TexLite]

So which drivers would you use in which circumstances?

This is what I really want to know.

I know I answered this above,but I thought for prudence and safety's sake I should expand on the answers.

Your looking to drive an MC-E at full power from either a 18650 or 17670.To know if a driver will work with a particular battery,and whether or not it wil be safe,you have to determine the current input requirements.(not speaking of voltage issues mentioned above,those )

Heres the formula:

Power output=number of chips/dies/emitters x Vf x current
Power input=Power output (watts)/efficiency
Input Current=Power input/battery voltage

Series Example:
Power input=4 dies x 3.5Vf x 700ma(.7)=9.80 watts
Power Output=9.80 watts / 85% (.85) efficiency=11.53 total watts out
Input current=11.53 watts / 3.6v (nominal batt voltage)=3.20(3200ma) amps in

Parallel Example:=1 die(because of parallel) x 3.5Vf x 2800ma(2.8)=9.80 watts
Power input=9.80 watts / 85% (.85) efficiency=11.53 total watts out
Input current=11.53 watts / 3.6v (nominal batt voltage)=3.20(3200ma) amps in

An AW 18650 is 2200mah,that means the max safe discharge is 4400ma(2C),so that configuration would be within allowable limits.

An AW 17670 is 1600mah,that means a max safe discharge of 3200ma(2C),At the max safe allowable discharge current from the battery.

The battery voltage will drop a little lower than 3.6v during the full discharge cycle,when this happens,the current requirements will increase.With the 18650 you have some headroom,the 17670 is maxed out.Whether or not this will work and be safe could be better answered by someone other than myself.

Also,at 3200mah discharge,your looking at a half hour runtime.To be useful as an EDC multiple levels would be required to stay out of max for normal situations.

Just wanted to give you the numbers so you could see whether an application would be safe and what runtime to expect.

-Michael

 

[TexLite]

Thanks. I take it that you are referring to Hyper-Lux with 3 CR123s?

So you are saying that with one Li-ion (18650), it would be impossible
to find a driver, that would work with the MC-E in series
because the voltage couldn't be boosted to 12V from 4.2V?

This is what I was getting at in my original question at the start of the
thread:-

ie. If the MC-E is wired in series, but it is still desirable to keep the flashlight
compact and/or powered by only one cell, would one need different battery technology?

Oops,got me while I was typing.

Yes the Hyper Lux-V with 3 Cr123's.

That configuration is going to hit the battery pretty hard.But in that package,I dont know of another battery solution.It should work,just make sure you go with a Multi level driver for practicality.Your single 18650(2200mah) will have the same capacity as two 17500's(1100mah x 2).Both 7.92 watt hours.Your voltage with a single cell will be 3.6v,with two cells 7.2v.

Amperage draw with 11.53 watts from 3.6v is 3200ma.2200mah / 3200ma is .69 hours.

Amperage draw with 11.53 watts from 7.2v is 1600ma.2200mah / 3200ma is .69 hours.

I guess it all evens out.Silverfox or LuxLuthor or someone more knowledgeable with batteries could better tell you how the batteries will behave at those discharge currents.On paper I dont see much of a difference.

I think it will be harder to drive the MC-E in pure series with lower voltage,in those situations I think the preferable configuration is serial-parallel,or straight parallel.With the dies in the MC-E being individually addressable,the Vf of each die can be checked to see if it will present a problem.

In something like an ultra-simple 3 cell DD,I think the only advantage the MC-E will have over the P7 will be the package size,i.e. closer die spacing and a smaller dome/overall package for better throw.With higher voltages,series will be preferable.Every situation is different,with different requirements,and for that reason the MC-E is a serious breakthrough.

-Michael

 

[easilyled]

Thanks for your patience and explanations.

I'm not too keen on using multiple small batteries to reach a higher voltage, because of the threads that elude to "vent with flames" which is why I wanted to know if one could achieve driving the MC-E in series with only one cell like an 18650.

 

[TorchBoy]

Re: New battery technology for upcoming MC-E?

... if its about the Shark,I've already addressed that in the post above.Those are the new design with improved thermal vias,the old were less.Its too big anyway.

Really? Where do you get that from? :shrug:

All of my numbers came from their respective sources and my math is correct.

It seems to me your more interested in arguing than stating fact.

I'm not questioning your maths; it's fine in as much as it depends on spec figures rather than real world figures, and your thoughts in post 17 show you understand what's going on (although do remember that milliamps is mA and milliamp-hours is mAh). So whatever - the Vf is not a biggie. However, it seems to me that you're still assuming you remember the specs of those drivers correctly. It seems to me you're not interested in checking that you've quite got that right. You haven't even read that link in post 2, have you? Hm?