Driving a Seoul Lamp from Lithium Ion

[koopa]

Ok, I have had no sleep and have been scouring this site for what seems like 3 days straight, and have found about 50 dead ends, so I am going to post a question (or 5)!

Seoul LEDs (at least the Type U - 240 Lumen variety) seem to have the same voltage range as a Lithium Ion battery (Max 3.8V, min 2.7V). Very annoying. Perhaps this means they can be directly driven off Li cells, but if you wish to have voltage/current regulation for the life of the battery, then you're caught between a buck/boost converter... -> low efficiency.

Now I have emailed Maxim, Texas, National, and sought out their solutions to these problems, and what I've gotten back is some fairly complicated and unsatisfactory solutions (the closest being a National 2700 I think).

My idea is to use a BOOST converter only, drive the LED to 4.1V for the entire lifetime of the battery (at 1.3A), and set a maximum duty cycle of 85% to bring it back into safe limits. That way, I will always have the battery at least 0.3V less than the V-out of the LED?

Now is that possible? Does PWM control allow me to maintain such a high voltage even when I am only driving the LED to an average of 150ma (4.1V), or is it in practice lowering the voltage in a way that would cause problems for a boost converter trying to regulate a 3.75V lithium ion battery.

PLEASE someone tell me I'm an idiot and that I've missed a converter board of >1A, that runs off lithium ion, with >80% efficiency, that drives a SEOUL LED, with PWM control and can fit on a circuit board of diameter 23mm or less. And if it doesn't exist, WHY NOT, and any ideas?

Thanks people!

 

[Curious_character]

Ok, I have had no sleep and have been scouring this site for what seems like 3 days straight, and have found about 50 dead ends, so I am going to post a question (or 5)!

Seoul LEDs (at least the Type U - 240 Lumen variety) seem to have the same voltage range as a Lithium Ion battery (Max 3.8V, min 2.7V). Very annoying. Perhaps this means they can be directly driven off Li cells, but if you wish to have voltage/current regulation for the life of the battery, then you're caught between a buck/boost converter... -> low efficiency.

Now I have emailed Maxim, Texas, National, and sought out their solutions to these problems, and what I've gotten back is some fairly complicated and unsatisfactory solutions (the closest being a National 2700 I think).

My idea is to use a BOOST converter only, drive the LED to 4.1V for the entire lifetime of the battery (at 1.3A), and set a maximum duty cycle of 85% to bring it back into safe limits. That way, I will always have the battery at least 0.3V less than the V-out of the LED?

Now is that possible? Does PWM control allow me to maintain such a high voltage even when I am only driving the LED to an average of 150ma (4.1V), or is it in practice lowering the voltage in a way that would cause problems for a boost converter trying to regulate a 3.75V lithium ion battery.

PLEASE someone tell me I'm an idiot and that I've missed a converter board of >1A, that runs off lithium ion, with >80% efficiency, that drives a SEOUL LED, with PWM control and can fit on a circuit board of diameter 23mm or less. And if it doesn't exist, WHY NOT, and any ideas?

Thanks people!

No, it doesn't work that way. Unfortunately, people not too familiar with electronics tend to think only in terms of voltage, and have the mistaken idea that batteries have a fixed voltage and LEDs behave electrically like light bulbs. They don't, and they certainly don't, respectively.

The first thing to do is take a look at typical LED voltage-current graph, such as the one at http://www.candlepowerforums.com/vb/showpost.php?p=1761783&postcount=46.

Look at what would happen if you tried to put 4.1 volts across the Seoul production part -- the pink one in the middle of the bunch. I'd have to guess, since the curve goes only to 2 amps, but you'd get something like 4 or 5 amps of current. Give it an 85% duty cycle and you'd have an average current of 3.5 - 4 amps. Of course, your battery won't supply 5 amps at 4.1 volts -- it has internal resistance which will drop the voltage and reduce the current. But I hope you can see this isn't what you want to do.

What you need to do is regulate the LED current, and let the voltage be what it is. Your supply does, of course, have to be capable of producing the required voltage at the required current. If you want, say, 150 mA of average current, you can either regulate it at that level, or you can switch a higher current off and on, for example 1 amp with a 15% duty cycle. The LEDs get less efficient at higher current, though, so the latter method won't give you as good efficiency -- assuming, that is, that the efficiency reduction at high currents is due to something other than chip heating.

Only after you begin thinking of supplying the required current to the LED will you be able to make reasonable choices as to a good way to do it.

c_c

 

[matrixshaman]

How about the FluPic board from goldserve? I think his board will do around 1 amp on burst mode. Not sure if that fits all your requirments but they are very cool setups.

 

[koopa]

Thanks Curious!

And thanks for the graph, told me a lot.

I am very much aware of the fact that current limiting is what you have to do with LEDs (because they have non-conventional internal resistances I'm told), I have built a few setups before using Luxeon V.

What confuses the issue is that boost converters DO very much care about the voltage, and I knew that if I tried to run a Seoul at 100ma, the corresponding voltage would become 3.25V, and if I have a fully charged Lithium Ion (3.7V or more), then my boost converter is going to have problems (I think most fail to direct drive, not a good thing).

Now the solution might be a boost/buck converter, but all the ones that I've seen have terrible efficiency (<70%) when the Vin is +/-0.5V from the Vout, and thats going to be most of the life of my torch. I could live with that, but to make matters worse, I can not find one that is <21mm wide.

To illustrate my point, here is a promising circuit I have been using for my Luxeon V. http://www.taskled.com/techfatman.html the FATMAN driver.

My other idea (that you've just made me realise is inefficient -> I thought P4s were more efficient up to 1A, not 300ma), was to drive the LED to something my lithium ion battery would never reach (3.85V?) and then PWM control it down to the apparent brightnesses I need, thereby eliminating the need for a BUCK converter.

So sorry for the confusion, thanks for the input, and again, if you have anything in mind that can drive a SEOUL through its intire current range (150ma to 1A at least) from a lithium Ion (3.7 to 2.7V) battery, I would appreciate it.

Thanks Matrix for the Flupic board, I had come across it but had dismissed it on the basis that I didn't need fancy strobe effects and its current-usage graphs didn't seem very constant throughout the life of the battery, but, I will now take a look again. Thanks!

 

[Curious_character]

Sorry, I don't have a good solution for you. But now you have a clearer idea of what the general problems are. I think you've just chosen a difficult battery type for the LED. Lights using a single lithium primary or Li-ion cell seem to be either direct drive, perhaps with a small resistor, or use a boost regulator like the P1D CE. Boost is no problem with a primary cell. When you put a Li-ion cell in one of the boost regulated ones, it apparently quits regulating and reverts to direct drive. Or it goes ahead and boosts the voltage to something fairly high but limits the current to the LED in some way -- a resistor, linear regulator, or internal circuit resistance. They're not terribly efficient when operating in that mode, though.

c_c

 

[wasBlinded]

I wouldn't get so caught up in the inefficiency of boost-buck circuits. The Wiz2 from the Sandwich Shop will be efficient in range of 80% - proabably better - give true current regulation, and get the most from your Li-ion cell without endangering it with overdischarge.


In terms of light output per watt over the life of the Li-ion cell, you are going to have a hard time beating the efficiency of that circuit. You can of course drive your LED pure PWM, but you will lose some of the efficiency of the circuit with the inefficiency of driving the LED at very high currents during the on part of the duty cycle.

If the Vf of your particular LED is low enough, and you don't need to drive at high currents, you can use a pure buck circuit with your Li-ion cell. I've done this successfully with a TWOH LED and a Downboy 500. It stayed in regulation for more than 90% of the useable cell capacity, and its dimming warned that the cell was all used up - but not too far gone.

 

[koopa]

Thanks for the help Curious! I know you're right, Li-ion + P4 is difficult to current manage, but impossible is my middle name! (No its not, I just wanted to impress you).

Blinded, yes! I just ran across the sandwich shop in my CPF travels, and the GD Buck/Boost converter seems perfect for my needs (the just released sequel to the Wiz2): http://theledguy.chainreactionweb.com/product_info.php?cPath=48_49_61&products_id=961

Only problem is, I can't seem to get in touch with the guy who runs the website. I need to order one and I need to know what potentiometer/resistor set I can use to vary current on-the-fly between 100ma and 1.5A. I think the person who sells them is Dat2Zip, does anyone know if he's out of town at the moment / how to contact him?

Thanks guys!

 

[wasBlinded]

If you are going to be using Li-ion exclusively, I would still try to get the Wiz2 instead of the GD. The Wiz2 has a higher Vin cutoff (2.5v vs 1.8v) and will be safer for your Li-ion cells.

I suspect, but don't know, that it will be difficult to use a potentiometer with these convertors. The current sense resistances are extremely low. Can you find a potentiometer that operates smoothly in the .06 to .12 ohm range?

As far as contacting Wayne, you will have trouble nailing him down for advice. He runs a small operation and supplies his goods pretty much without technical support.