Poorman Mutli-Lux setup method

[krutzbeuazen]

great idea, exactly what i need! want to power a handful of those cheap 1w leds for fixed illumination (had those leds left). had some 7135 boards left from some lights. got me a 15v converter from fleamarket. soldered it all together, easy thing:
3 leds in series connected between a single am7135 and the powersupply. one led connected to the 7135's out. that setup two times. powered it up, and, nice & bright. surprisingly bright and cool color, for those 1w leds.. *scratches head*
and while thinking this, the first leds begin to dim. begin to glow very dim, but blue. begin to glow brownish. the first led explodes. the second setup (in parallel to the first, now dead string) shines normally. a bit later, the first led of this string goes dim, a dark spot forms in the phosphor.

ok, i admit it, i turned it off immediately, and plugged it in again, to see what will ultimatively happen.. the leds were damaged anyway.

so.. all three "voltage dropping" leds died almost instantly. one from the other string died. the led connected to the output survived, in each setup.

never had a led explode, before.. i guess the transformer either has much more than 15v, or i connected the 1735 wrong. will check the voltage of the transformer now..

manuel

 

[krutzbeuazen]

i checked the trafo: 60v with no load, 11v when connected to a halogen bulb.
..i better watch out to short it before connecting anything.. ;-)

i checked the 7135 datacheet, connected everything again, here goes another led.

i then used another 7135 chip, bare, not connected to its board. works, drawn 220mA, stays stable, eventually (a minute) it starts flashing because of overheating.

hint: check the boards you use, even when you removed anything, there might be surprises left..
cost me 5 leds to find out.. i have exactly 12 left for three strings, better be careful now!

manuel

 

[download]

krutzbeuazen,

Basic rule: 1x Li-ion or 3x NiMH to 1x LED.

7135 can't handle the voltage range excessively. (3-6V only)
Suggested to use 110/220V driver directly to your leds.

 

[Changchung]

I just finish my triple P7, I am really happy with it, I am using this driver;

http://www.kaidomain.com/ProductDetails.aspx?ProductId=1845

With the 2 Leds in serie technique



THANKS

 

[lolzertank]

Just a quick question...

The diagrams say the drivers are multi level. From what I know, the levels are implemented through PWM on the supply pin on the 7135 chips. Once the microcontroller turns the 7135s off, doesn't that drop the forward voltage of the LEDs to zero because there's almost no current (couple uA of leakage)? So let's say I used 2x fully charged Li-ion. Wouldn't the 8+ volts fry the AMC7135s (8 - 0 - 0 = 8 :shakehead)?

I can think of three (crazy?) explanations.
1. The microcontroller draws a little current through the first LED, causing the vF to increase and protect the rest of the circuit.
2. The AMC7135's internal transistor can handle more than 7 volts, despite what the datasheet says about the maximum voltage on the OUT pin.
3. There's so little current it really doesn't matter.

Maybe I answered my own question. Still want a 2nd opinion though before I burn something

 

[download]

lolzertank,

When the 7135 driver turn off 100% to the master LED (which connected to driver) It will be no light.
But the 1x slave LED will get a few mA, so it will light up barely.

Battery 8.4V ---> (Driver will turn extra 1V to heat + Master LED 3.7V ) + ( Slave LED 3.7V )

Battery 12.6V ---> (Driver will turn extra 1.5V to heat + Master LED 3.7V ) + ( Slave LED 3.7V ) + ( Slave LED 3.7V )

 

[lolzertank]

I'm not sure you understand my question. Obviously, a LED driven at a couple microamps/nanoamps will have a forward voltage far less than 3.7. Basically I'm trying to figure out what happens when the LED's forward voltage drops when the AMC7135 switches off, mainly whether it'll destroy the AMC7135 chips if the forward voltage drops so much that the voltage across the regulator goes over 7v.

Now that I think of it, I have this huge 10 pack of 1.4A boards ready to have traces cut, improperly used and possibly killed in the name of science... :devil: One chip less won't hurt, right?

 

[download]

3. There's so little current it really doesn't matter.

Just wire it, you will see how it works.

 

[lolzertank]

Thanks for the help! I'll try it!

 

[jar3ds]

so could we use this same idea with the taskled HipFlex? I want to run 5 P7's w/ 6xLiON's... since 6xLiONs=25.2v (the HipFlex has a max of 24v)...

would running the battery right into 1 P7 to drop the total voltage down a Vf to ~21.7v, then put it that into the HipFlex and connect the remaining 4 in series just like normal?

Thanks!

 

[download]

georges80
is the creator of HipFlex. :rock:
He sure have a better answer than me.

 

[jar3ds]

thanks download for your help.... this is what i got from george:

25.2V exceeds the max input spec of 25V. You could put a series diode into the VIN+ line to drop the voltage down 0.5V or so.

cheers,
george.

so I'd assume I could do it with an LED instead of just a normal diode? If this would work, it would be the slickest setup...

Theres something telling me though "its too good to be true" with this... i sware I have to be missing something.... i feel like this setup breaks the laws of physics or something

 

[download]

George's driver is one of the best for flashlight mod. :rock:
You got the right choice. :thumbsup:

 

[TorchBoy]

so I'd assume I could do it with an LED instead of just a normal diode? If this would work, it would be the slickest setup.

Yes, of course it would work, but you'd get a greater difference in current than the 200 µA difference in the AMC7135 setup. A quick calculation (assuming 90% efficiency) suggests it would be around 2,000 mA, because that's what the driver would draw to power four LEDs at 2,800 mA.

No laws of physics broken, just not the result you may have been looking for. Driving all five LEDs together (in series) would have them running on the same current.

 

[The Dane]

Yes, of course it would work, but you'd get a greater difference in current than the 200 µA difference in the AMC7135 setup. A quick calculation (assuming 90% efficiency) suggests it would be around 2,000 mA, because that's what the driver would draw to power four LEDs at 2,800 mA.

No laws of physics broken, just not the result you may have been looking for. Driving all five LEDs together (in series) would have them running on the same current.

Errrrr no!
If the board delivers 2800mA to it's LED's and have a loss of 10% then it uses 2800mA+280mA or 3080ma wich the series LED would see. So if anything one LED will be ever so slightly brighter than the rest, but nothing noticeable to the naked eye.

 

[TorchBoy]

Hej Henrik. Sorry, I thought the Hipflex was a buck driver, so it'll draw less current on in the input if the output is at a significantly lower voltage. I suppose I should have been a bit more meticulous and worked out the power on each side of the driver instead of doing it in one step...

Out: 2.8 A @ 14 V = 39.2 W.
In: 39.2/90% = 43.5 W @ 21.7 V = 2.00 A.

:thinking: Still works. I hope I'm not spoiling the thread with that, eh.

 

[javiole]

Hi,

I have got a question about this desing. I hope someone can help me out. First I am going to redraw Download diagram in a way that is easier for me to understand:

So far I'm using a board with 2 amc7135 and another one with 8 amc7135 with a switch in between so i can have a low (700ma) and a high(2800ma) modes (bicycle use). It works great but, it maybe an easier and simpler way to do it.
It is possible to feed the AMC7135 board with a different voltage reference so the whole set of LEDS will dim. As in this mod of Download diagram:

The resistor R1 would put the voltage down to about 3.6volts, then the potentiometer, used as a variable resistor, would let you choose the voltage from about 3volts to 3.6volts. I understand that the amc chip has a very low comsuption, so R1 won't need to be bulky. Even the pot can be replace by several resistors to be switched to get different modes. Another advantage is that only two cables need to be pass throw the heatsink into the leds room. Can someone clarify if this would work? cheers

sorry for my poor skills drawing diagrams.

 

[download]

javiole,
Just use this 8x 7135 driver, it has MCU (multi-level control IC). Like 3 or 5 mode.
Power on & off to change the mode you like, eg: High 2.8A, mid 1A, low 300
mA ......

Wiring is same as my original diagram. It will be fine.

 

[TorchBoy]

javiole, I think you would get a very twitchy output. The AMC7135 drops out of regulation at 2.7 V. With a polarity protection diode that becomes about 3.3 V (not 3.6 V). Unregulated the AMC7135 may be a little unpredictable and I suspect a very small change in the resistance of your potentiometer would make big changes in the output current, so you might have trouble setting it as you want. The dropout voltage graph in the datasheet is about the only help it offers. Having said all that, it sounds like a very interesting way to use them. Give it a go and let us know if it's practical.

download, where is that driver from? I list it as "unsourced" since I've only seen a photo posted before (from Alibaba) but no retail links.

 

[Aircraft800]

javiole, I think you would get a very twitchy output. The AMC7135 drops out of regulation at 2.7 V. With a polarity protection diode that becomes about 3.3 V (not 3.6 V). Unregulated the AMC7135 may be a little unpredictable and I suspect a very small change in the resistance of your potentiometer would make big changes in the output current, so you might have trouble setting it as you want. The dropout voltage graph in the datasheet is about the only help it offers. Having said all that, it sounds like a very interesting way to use them. Give it a go and let us know if it's practical.

download, where is that driver from? I list it as "unsourced" since I've only seen a photo posted before (from Alibaba) but no retail links.

That was the 3-Mode Regulated Circuit Board for Cree MC-E and SSC P7 sold by Shiningbeam.com SKU 1217. It has been removed from the site :thinking: