driver circuit designs?

[hank]

I'm not sure if anyone's collecting driver circuits, if so I missed it.
There are new ones appearing all the time, here's one:
http://www.edn.com/article/CA6470830.html?industryid=47042

note the comments and mentions of how to find other info there

 

[VidPro]

i collect them, on the desk there after removing them from my lights :thinking:
got some charred ones, a few fried ones, a few that lasted 2 weeks, some that dont have squat in output, and some very inneficient ones :shakehead some that had Stupid modes, like 2 mode ON and Strobe? whats that about.
Some that heated up my leds more than the led itself does, and some that drained my battery when the light was off. and some that are about as regulated as the battery you put in the device. some that think the drive current of a "high powered" led is 120ma. some that pulse so badly with PWM you would think it was in strobe mode.

mabey i should start a Cruddy Driver Museum

 

[Illum]

I've collected all the ASCII/Schematics that appeared on the forum...but not actual drivers though
I've wanted to build this....can you imagine running CREEs off as low as 1V?

Source: Maxim's APPLICATION NOTE 1029

 

[VidPro]

off as low as 1V?

1 volt of what? a reverse charging set of nicads? a expended about to leak alkaline? a now completly ruined li-ion ? or a set of lithiums that are out of balance expended and about to blow up in your hand

no i cant imagine, unless for say something like parellel 1.2v cells, that would be great, like a single D nimhy.

thats the thing about fully completly regulated current control devices, they tear a SET of batteries to shreds when they attempt to maintain full current at the low end of the dishcarge.
And the opposite
anything that allows running off a adequite capacity and rate, SINGLE or parrellel set of cells, has always worked great.

 

[MrAl]

Hi there,

Interesting circuit, the original one talked about at the beginning of this
thread and can be found here:
http://www.edn.com/contents/images/6470830f1.pdf

What i find strange however is that whoever designed this thing didnt seem
to realize that when you put a 270 ohm resistor in series with a 20ma LED
that resistor uses up a full 0.1 watt while the LED only uses 0.07 watt.
What does this mean? It means very inefficient circuit and if it is patented
that just proves that you can patent anything no matter how well it works

Ok, so what can be done? Maybe change that 270 ohm with a 10 ohm and
take it from there...increase R1 for example.

Here are some readings with the original circuit

[R3=270 and R2=10k as original design was]

Vin=2.0v, Iout=22ma,IinAvg=178ma, Eff=22 percent
Vin=3.2v, Iout=57ma,InAvg=300ma

Now change R3 to 10 ohms and R2 to 37k and get this result:
Vin=2.0v, Iout=22ma, IinAvg=60ma, Eff=64 percent
which is a great improvement in effciency already.

 

[Illum]

off as low as 1V?

1 volt of what? a reverse charging set of nicads? a expended about to leak alkaline? a now completly ruined li-ion ? or a set of lithiums that are out of balance expended and about to blow up in your hand

no i cant imagine, unless for say something like parellel 1.2v cells, that would be great, like a single D nimhy.

thats the thing about fully completly regulated current control devices, they tear a SET of batteries to shreds when they attempt to maintain full current at the low end of the dishcarge.
And the opposite
anything that allows running off a adequite capacity and rate, SINGLE or parrellel set of cells, has always worked great.

well, My idea was to have some sort of an circuit that uses the 2V 100ish AH batteries as ideal fixed lighting...but the idea was expensive so I left it at that...apparently I misread the thread title

 

[VidPro]

2V 100ish AH batteries

ahhhh, you mean like the Hawkings Rolled electrolyte lead acid cells?

 

[Illum]

2V 100ish AH batteries

ahhhh, you mean like the Hawkings Rolled electrolyte lead acid cells?

http://www.germes-online.com/direct/dbimage/50024146/2V_170Ah_Sealed_Lead_Acid_Battery.jpg
something like that, I saw it in Homepower magazine concerning photovoltaic battery banks...thought it would be a good idea if LEDs are used as residental lighting fixtures [similar to brlux's excellent example of indoor LED lighting] a 2V 170AH [shown above] could potentially last near forever for a couple LEDs in parallel drawing less than 350ma each...but since I dont have the mathematical concepts well understood, I didn't know what the resulting runtimes might look like...and how would I be able to charge a 2V cell?

 

[VidPro]

http://www.germes-online.com/direct/dbimage/50024146/2V_170Ah_Sealed_Lead_Acid_Battery.jpg
something like that, I saw it in Homepower magazine concerning photovoltaic battery banks...thought it would be a good idea if LEDs are used as residental lighting fixtures [similar to brlux's excellent example of indoor LED lighting] a 2V 170AH [shown above] could potentially last near forever for a couple LEDs in parallel drawing less than 350ma each...but since I dont have the mathematical concepts well understood, I didn't know what the resulting runtimes might look like...and how would I be able to charge a 2V cell?

charge it with solar, and BIG FAT WIRE, because moving low voltages at high amps is tough, due to small voltage drops being so much of the total voltage.

that is a cool battery, i want one how much are they ?
actually i want 2 (thier small so it would DIRECTALLY match my leds voltage after long wires

you could easily design a solar pannel that was very low voltage, like say 3v (for charging the 2v) because the sperated solar cells come in .5v pieces usually.


i remember using these things http://www.atbatt.com/product/3590.asp
hawker, not hawking, they were still lead acid, but they worked very well.

 

[Illum]

probably why the terminals on the cell is monstrously monstrous compared to the terminals of my car battery:naughty:

according to http://www.unitedpowercorps.com/mti.html
their selection of 2V FLA cells range from 150AH to 3000AH
not sure of prices...but likely to be expensive...plus its flooded so shipping rates are also likely to be :sick2:

now that you mentioned Hawks....
I saw a Hawker Cyclon 2V 2.5AH SLA about the size of a D cell for $8
http://www.atbatt.com/product/3490.asp
which might be of value for those who are willing to create a prototype of an entire residential lighting model running off a couple of "D's"

Its just an idea...

 

[havand]

http://www.germes-online.com/direct/dbimage/50024146/2V_170Ah_Sealed_Lead_Acid_Battery.jpg
something like that, I saw it in Homepower magazine concerning photovoltaic battery banks...thought it would be a good idea if LEDs are used as residental lighting fixtures [similar to brlux's excellent example of indoor LED lighting] a 2V 170AH [shown above] could potentially last near forever for a couple LEDs in parallel drawing less than 350ma each...but since I dont have the mathematical concepts well understood, I didn't know what the resulting runtimes might look like...and how would I be able to charge a 2V cell?

A 'couple' is a very nonspecific number!

I'm not saying this is 100% right, but just some guessing...

Let's say you're using SSC @ 350ma. Gives you an average Vf of something like 3.125 V (binning is 3-3.25V). Let's also assume a driver efficiency of 75% You can find higher, but boosting 2V to 3V could cost a bit.

So, to run 10 LEDs @ 350mA @ 3.125V would require..
10*0.35*3.125=10.9375 Watts.
With efficiency losses in the driver... 10.94W/0.75=14.58W

The load on the batteries would be 14.58 W.

You've got a 2V 170Ah battery. Now, I know nothing of these batteries, so I'll assume a modest? voltagesag. 14.58Watt draw / 1.8V supply = 8.1Ah current draw.

170Ah supply / 8.1Ah draw =20.99 hours of runtime. Give or take ( probably take ).

Easier way would be Supply watts / draw watts. 1.8*170/15=20.4 hours.

So, a single battery will run your LEDs for something like 20 hours. Up your LEDs, up your current draw, the battery gives up faster, etc.



Now, say you built a bank of 5 of these things in parallel and wanted to run 25 LEDs at nominal.

Supply
1.8V*170Ah*5=1530 watts.

Draw
3.125V*0.35A*25/0.75 (efficiency)=36.46 watts.

1530/36.46=41.97 hours.

Now if you can build a rig to recharge a 1530 watt panel in a day or two without taking out a loan to do it, let me know
Of course you would also have a LED lighting system that could run for ~40 hours (with conservative use, what 8 nights, maybe?) with 25 LEDs producing 2500lm total. Seems pretty sweet to me

 

[VidPro]

cool, i already do
i run about 40watts of leds off a li-ion pack, its "self regulated" meaning direct drive , goes up as there is more available power, goes down as there is less available power.
the 850Watt Li-ion pack is charged up with a Mere 9amp solar pannel
the solar pannel is about 2 feet by 4 feet, cost about 300$ to build by hand.
hmm its the storage that breaks the bank on solar stuff.

there are 3 tricks it uses to have this light up ONLY a big kitchen.
its on a motion sencor, so it is only on when needed
its on a light sencor so it doesnt turn on when the sun is out
and of course, cause its self regulated.

its sort of self regulated in BOTH directions, the lower the battery is, the higher the rate of charge.

the STORAGE does need to have BUKU capacity beyond the possible use of any one day, but neither the solar pannel, nor the leds have to USE or charge the whole of the storage every day, no battery would LAST recharging and discharing 100% 365 times a year.

NO drivers
i had to prove that it was possible, cause people can say it cant be done, usually there is adequite juice for a few other rooms, but the wiring is unsightly

there are pics of it around here somewhere, its my LED kitchen, it has run non-stop now for years
it tests LED, it tests Li-ion cells, it tests Direct drive theory and self regulation theory, and it tests solar cells, well pluss it lights the dang kitchen, but who cares about that

 

[havand]

I gotta know...What'd that li-ion pack cost you? I assume they're allin parallel...Otherwise you'd need balancing, etc. You don't have any kind of protection circuit on that? How do you keep it from overcharging?

 

[VidPro]

it cost TOOO much originally they were surplus cells from some failed laptop project, and it only cost about 300$ (hey this will be cheap , , ,NOT)
but then these started failing after ABOUT a year and a half, and i replaced the ones with high self discharge, about 1/3 of the originals are working good, it costs some 300-400$ more to replace the bad ones.
the originals are still dying, the replacements have worked great for years, i wont know more till they also start going bad.

yes it is all parellel, definatly.
i want to replace the little cells with a few mondo AUTO sized lithium, but the INITIAL cost is prohibitive, even though auto ones would handle many more short cycles.

all the "electronics" had to be redone about 3 times, till they worked continuously and correctally, and didnt WASTE stuff so fast.

first the solar cell was regulated with a 317 (type) adjustable regulator in the high temp package t-220 or something? anyways the voltage regulation suffered from huge losses, that 1.5v drop across it was killing the efficiency.
then i tried to make a relayed protection curcuit, because HIGH AMP power protection was not easily available. the relay would disconnect the rack when the voltage went to about 4.15. the problem with that is the relay wouldnt reset untill it was way low again, a cheap 555 timer curcuit would attempt the reset, and it required power to run the curcuit and relay.

after that they made available these 8amp protection curcuits, WHICH i discovered were only good for 4 amps :-( so 2 of them do the job. again they do not reset till the next day, or if i manually trip it.

it is Rare (but it occurs) that it reaches full charge, because of the self regulation, the charging voltage (vrses max voltage) of the solar pannel is very close to the max voltage, when the voltage differential between them is low, the charge is slow. so it tapers itself off. for example when its around 4.1volts the charge tapers to about 3-5amps depending on sun.

low voltage cutoff, which again rarely occurs , is the power relay that switches the LEDs on, it just wont work at around 3+V so the thing just quits working if the battery goes to low. mostly the leds going low draw so little power, that solves the cutoff all by itself.

it might sound like i know what i am doing, but if i did it might have some microcontroller in it that did everything and took only little power to do so. it was just a test.
every possible loss or drain was attempted to be removed, an original setup has SOOO many losses down each chain of items it really ammounted to neer 50%. losses start adding up quick.
so even the diode protection for the solar pannel was replaced with something that has very low resistance.

 

[havand]

I find my most limiting aspect with custom rigs is my own knowledge of circuit design. It frustrates me quite often. I have not yet found a really good online tutorial about how to do it. I can do the simple resistor stuff, but how to efficiently use and pick out chips and regulators is very difficult. I keep revisiting the idea of having large scale, solar powered LED area lighting for either outdoors or indoors, but the solar panels and battery costs keep coming back to tell me no. A bit frustrating that I don't have the disposable income to drop a few k on a off the grid, custom solution.

 

[VidPro]

I find my most limiting aspect with custom rigs is my own knowledge of circuit design. It frustrates me quite often. I have not yet found a really good online tutorial about how to do it. I can do the simple resistor stuff, but how to efficiently use and pick out chips and regulators is very difficult. I keep revisiting the idea of having large scale, solar powered LED area lighting for either outdoors or indoors, but the solar panels and battery costs keep coming back to tell me no. A bit frustrating that I don't have the disposable income to drop a few k on a off the grid, custom solution.

start with a 12V marine deep discharge F-sized battery, put it in a battery case, so nasty acid junk doesnt ever get at the user. 50-80$

get a 100watt solar cell $$$$,
and use 3or4 in series leds,

if your power costs 35c a Kiliwatt, like mine , you might only loose 50%
if your power like in some places only costs 5-7c , its a total waste.

i researched the math of solar finance, and even in california if I didnt change my lifestyle with efficiency, unless I feed the extra power back to the grid, i would lose money untill power costs me even more. (which it will eventually).
IF i had to run , maintain, and replace STORAGE, it would be a money loss

if i increase efficiency, and reduce losses and waste, like some people learn when they go solar, then i would save without going solar. its all the things we waste without realising it , that are taking all the power that solar cant even provide.
so when you "go solar" you find out how quick all that power you make you wasted , and stop the waste.

 

[R11GS]

I have played with a lot of switch-mode regulator designs (both voltage and current controlled) in the past, but most of them did not require efficiency to be near the top of the list... But I do feel I understand very well what it takes to make them as efficient as battery powered designs demand. Most folks consider >90% where you need to be, with 95% being a point of diminishing returns.


Anyway, I'm generally not that interested in "reinventing the wheel". If I can find something off-the-shelf that will do what I need, then I will almost always go that route. But often enough, what I need will have something that the off-the-shelf designs won't do.


I'm in a small predicament like that right now! I wanted a dimmable, single AA (but I would consider a 123) red LED thrower. I modified a L1D by simply dropping in a red Cree. Not much to brag about there. I knew the Vf on the red Cree would only be ~2V (actually at 2V this thing is bright!) and I knew that most single cell lights out there use a current controlled boost converter. So I chose the L1D hoping that with a 1.5V AA, the boost converter in it would drive the red LED nicely.

But it's clear that while the regulator is boosting the voltage to about 2V, it's not regulating and I've got no dimming; although it is pretty cool. It's very likely that the lower output limit is just too high even running off a AA. This is usually because boost mode regulators have a lower pulse width limit, but not knowing the L1D circuit, that's just my best guess. :shrug:

I suppose I could try a P3D head with the red LED on a P2D body with a 123 and hope that the P3D buck regulator will drive the low voltage I need. But I think this is a big crap shoot and kind of doubt the P3D head will like the low voltage of a single cell. :shakehead



But to the point. Now I'm thinking of rolling my own. I've been scribbling on the backs of envelopes with fury! I'm very confident of doing a 123 design, possibly a microcontroller-less design, but the more I look into a single AA the more I appreciate how challenging that low voltage is. I am also thinking of a microcontroller based design, but I haven't done the hands-on thing with a micro for a few years and lack the development systems to even get started here. Besides, I'm much better regarded for my analog designs.


So of you guys out there messing around, who has some MCU skills and tools, and wants to play? I am certain I can make some very elegant MCU based switch-mode designs. I've not only been thinking about LED drivers but incandescent regulators. If we can solve your problem and mine at the same time... Maybe come up with something that will help a bunch of other folks too?

 

[hank]

> I wanted a dimmable, single AA ... red LED thrower

Oooh!.. Anything that would run red would run amber just as well.

I want in on this if anybody's going to build them or help us build them.

There are a few little astronomy red LEDs with rheostats, but basically quite dim at their brightest.

 

[havand]

Not an IDEAL solution, but you COULD use a single li-ion, running 2 red leds being boosted by a Shark with a 10k potentiometer. Only condition is Vout>Vin otherwise that is the lowest dimming level possible, when they're equal.

 

[wakibaki]

hmmm...

Interesting circuit. I see Coilcraft are now specifying their inductors in microfarads.

This is a nice little switching device from Linear, LTC3216, good for 500mA, *no inductor*, 2 levels, tailored for 1 lithium rechargeable.

http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1003,C1039,C1133,P7768

I got a couple of samples, then I realised. They are truly minute. 3x4mm with 12 pins. My layout software and laser printer are really struggling to produce PCBs this detailed. Then somehow I have to get them to solder themselves on my radiant hob.

For control a 6-pin flash microchip PIC. I got a circuit I used to program one of the bigger flash PICs from a PC serial port, I think it will do these too.

Of course the flash PICs are in-circuit-programmable, so it's possible to build the board without reference to the programming.

Can't be more'n a few lines of code to turn a couple pins on and off.

w