My e bike lighting woes

[FiremarshallDave]

Our family has 4 e bikes (my builds) and I'm building a 5th. We use 6 cell 18650 lithium ion packs to power our lights ( 10 cree xml t6 diodes)(some are 13 diodes) in each light. The lens looks like a flashlight called a SkyRay but are built for bicycle use without the batteries inside the light and instead uses a battery pack you strap to the bike . These throw lots of light like a car head light. We are out in the sticks and don't want to outrun our lights. My problem is the inside of the lights keep breaking. There is a small steel doughnut with wire wrapped around it that breaks off the driver/ control board inside the lights due to road bumps. I cant see where it was connected on the second contact point. No empty solder points evident. I assume it's a choke? (A current regulator)? I now have 4 lights like this. I repaired one and took the driver chip out after there was a small puff of smoke from experimenting with possible solder points.... (oops) and soldered the choke into my dc+ supply. Does anyone know if it is necessary to use a choke? It seems to work. It powered on for 15 seconds on my workbench before I shut it down. I will take it outside for a longer test tomorrow. The pack supply is 7.2-8.4 volts. I will repair more of these if the first one holds up. I read somewhere the t6 diode is current limited to 3000ma that's why I used the choke. I assume it's necessary? Does anyone know about cree diodes? Is each XML T6 diode drawing 10 watts? So each light draws 100 or 130w, 10x each diode? Can I use 12v on those lights using a step down converter off my 58V battery? I don't understand the usage of the driver term when referring to led diodes. Are we talking about limiting the amps and that's all a driver does? Hence the doughnut (choke). If anybody knows and can share some knowledge with me great if not and things go bad I know how to put out a fire.

 

[Ebiker01]

how may lumens 10cree diodes put out ??

 

[indianajo]

Coils inside modern devices, especially toroids, often mean a switcher power supply to change voltage from one to another. I would abandon these failure prone devices. Switcher supplies have to be extremely simple and fully documented to be repairable. LED's have a voltage drop from 1.5 v (red) through 1.7 (green) to 2.0 (white). Usually the designers string multiple LED in series to match the voltage of the power supply, then parallel enough strings to get the lumens the buyer wants. Modern white LED's can be intensely bright but require a heat sink with fins to waste the excess heat.
I'm using a 400 lumen cateye AMPP 400 I paid $50 for. I like it, it is rechargeable from a micro-usb charger. It only takes one long button push to turn it on or off. I had a Niterider500 previously, but you have to cycle through 7 functions every time you turned it on to get 150 lumen steady. That one drowned in the bottom of the pannier bag. I keep the cateye in a peanut butter jar in the bag. I don't leave the lamp out on the handlebar when parked because I have had some lights stolen. None of these lights bolt on, they have a rubber band that wraps around the handlebar.
If you want a really bright automotive class light, there are converters that go from 48 vdc to 12 vdc. A reliable US brand is vicor, that costs $350 each from electronic distributors. I managed to buy some used 48 to 24 vicor supplies from ebay for $65 but the burglar carried them off to the copper scrap yard 9/20. Check ebay for bargains. There are ****ese switcher supplies on ebay for less, but they don't last very long. The wattage of the switcher supply has to be higher than the wattage of the load. The vicors are built to be mounted on a finned aluminum heat sink with heat sink compound between the two.
There are a number of reliable vicor dc dc converters on ebay today. I saw none that took 7.2 to 8.4 v input. There are a number of 48 v input, 24 v input and 12 v input models available. There were no 58 to 62 v input models. Possible outputs were 3.3 v , 12 v and 18 v. There were a number of other output voltages useless for LED use, like 5 v. The 12 v output ones could be used for ordinary automotive lights. There were even a few 15 v output models compatible with modern automotive voltage. Be sure the wattage of any converter provides the watts you need. output voltage times maximum output current equals watts. A number of ads did not specify input or output voltage and wattage and the pictures were too fuzzy or showed the wrong thing to read the label. Those vendors are spraying product at the blind & stupid. Missed me. Don't buy vicor converters from *****, real ones are made in Massachusetts. A product labeled "made in USA" but shipped from ***** is obviously counterfeit. Happy shopping or repairing.

 

[FiremarshallDave]

I know I'm spending more time on this than I should but I like to learn and understand things. Was I correct that the steel doughnut wrapped with copper wire is a choke and it limits current flow not voltage? Does a choke get hot? Next question... If i match the input voltage the light was using (the led light) but take it from my main battery which has more amperage than the small 6 cell packs...and I use an adjustable buck, to match the 8.2V is that a problem? Are the LEDs only going to use the flow (amperage) they need? I'm going back to what I read about the 3000ma limit or 3amps before they burn out. I'm thinking that could only happen in a series configuration. w=A x V so max A=3 3x8v=24W each diode is 10W so my brain says only two can be in series? Obviously I'm doing something wrong because the units have 10-13 diodes. Using what you said, 4 leds in series will = 8v but the draw thru the first one would exceed 3A . I will buy more cree lights but I wonder how many have the Choke in them and vibration will stress fracture them in a few months? there must be a better way to regulate flow. When the term switching driver is used does that simply mean switching dc voltage to match the number of diodes? So in essence a switching driver is a converter or a buck? I hate not understanding something so simple.

 

[indianajo]

There is nothing simple about modern switcher supplies.
If a ferrite toroid has two wire connections it is a choke. Not very likely in a bike light. Chokes are used in switcher supplies like PC's and flat screen TV's to stop voltage surges on the AC input from lightning strikes & Air conditioning motor shutoff. 1300 vdc surges are common on 120 vac household outlets. None of that occurs in a local battery. Chokes resist a change in current and are usually followed up by a MOS supressor to short the input voltage to some safe value like 300 that won't blow the filter caps after the rectifier.
If a ferrite toroid has 4 or more wire connections it is a transformer. These are usually driven by a fet transistor at high frequency to convert one voltage to another. If the 6 cell packs are lithium Ion batteries they could be 6*3.3v or 19.8 nominal, or 3 series 2 parallel at 9.9 volts. If the 6 cell packs are Nimh at 1.2 volts each in series they would be 7.2 volts nominal.
White LED's take 2 volts each in the smaller sizes I buy but could go up to 3 in the high intensity ones. Lamps are usually an arrangement of series and parallel LED to reach a certain voltage. Old tech the LED stack are then combined with a resistor to limit current coming from the battery supply. Modern devices eliminate the resistor and have a switcher supply to produce voltage at exactly that required by the led, even though the battery sags from 130% nominal to 80% nominal over the discharge cycle. My $1 LED flashlights have a 4.5 v battery, a single or paired LED's, and a resistor. My $50 500 lumen bike lights have a tiny switcher supply inside and one LED. If you want to see a tiny switcher supply cut the base off a dead CFL (compact flourescent) bulb for household use. Very tiny, very cheap, very high frequency.
To use your existing LED's with another power supply you would have to measure the voltage drop across the lamp assembly after the broken power supply. One way to do this is use a 12 v supply or battery, a 1000 ohm resistor series the lamp. Then you could measure the actual voltage across the lamp assembly with a DVM.
Then perhaps you could buy a reliable vicor power converter made in Massachusetts to replace the shock sensitive garbage you have been buying. There were a number of used vicor converters on e-bay yesterday, but none with 8.2 v input. 3.3 v was one common output, but inputs for those tended to be 24 v (industrial control voltage) 48 v (telephone voltage) and 300 v (rectified filtered 220 AC). 5 v and 12 v were other common output voltages.
You can buy more varied voltage sources and outputs in converters from ***** off ebay or alibaba, but the reliability of those products is not likely.
You will notice headlights for cars still are tungsten. If you want that much brightness, perhaps an automotifve bulb & reflector would suit you,. You could buy a 48 v to 15 v converter. However you said your bike battery is 60 v. That is another input voltage not supported by used vicor products.

 

[FiremarshallDave]

The packs are 2s3p and I say 8volts, I guess it's 8.4 full charge down to 7v. I will soldier them as there is no reason not to. I shot a few pics and will try to attach them. Thanks for your info. I will try motorcycle lights next. One pic you can see the wire on the (choke? or ferrite toroid) that has detached and then a pic of another light that is still attached. I have a pile of these and have taken out the green board/driver? and soldiered in the (choke or ferrite toroid) thinking it limited flow. Like I said it works now with no driver.

 

[indianajo]

The black box lower right appears to be a surface mount transistor. The 6 leg DIP IC is control. The black brick is probably a film capacitor, labeled with 3 digits & letter suffix , like 474J. A rudimentary switcher supply. If each LED anode is connected to each LED cathode, then they are wired in parallel and probably the source is 2.5 to 3.0 v. There being 13 LED they are unlikely to be connected 6 parallel in series with 7 parallel.
If you can mount the coil where it doesn't bounce around, and keep the 2 wires connected, you may be able to keep using this converter. Think tie-wraps and epoxy. You don't want to inhibit air flow much, the coil will get hot. If the actual ferrite toroid fractures, the life would be in seconds.

 

[Ebiker01]

2 vdc. A reliable US brand is vicor, that costs $350 each from electronic

Seca 2500lumens is great ! Not very good for torrential rains ;
also the warranty for this company is amazing they ship me a brand new one in two days ; didn't had to ship back the broken one so I've usedthe spare parts for a new projects

 

[FiremarshallDave]

Joe I thought the ferrite was still soldered in place but on closer inspection it was just hanging as well and the wrap looks melted. Looks like it was a direct short. My wife had a battery pack she melted and I thought she put a battery in the pack upside down. Now i think this light was the source. Trying to use what you said the Ferrite reduces the voltage to two or three volts? . Can the pack run without the driver if the voltage is reduced to 2 or 3v? I'll probably throw these away but learning how they work is the way my brain functions. The rest of the light is solid but the little crap driver is the fail point. I would love to use them with a better converter. I've looked for the solder point for the broken ferrite wires and there is not one evident on either unit. Can the ferrite on it's own, reduce voltage from 8 to 2 or 3v? Then I would do what I did with the 3 diode light and just solder it inside between the supply+ wires. It worked on the small 3 diode light for 15 seconds before I pulled the plug. I cut out and tossed the board and used only the ferrite as I couldn't tell where that ferrite should have been soldered. That experiment will be finished outside when temps allow. It's bloody cold outside right now in Wisconsin. Too bad I can't show these to the people who made them.

 

[indianajo]

Trying to use what you said the Ferrite reduces the voltage to two or three volts? . Can the pack run without the driver if the voltage is reduced to 2 or 3v? Can the ferrite on it's own, reduce voltage from 8 to 2 or 3v?

No. No. the ferrite toroid is a coil that when oscillated at 2 mhz or 10 mhz or something, changes the excess voltage from the battery pack into greater current at lower voltage for the 13 lights. A buck converter, probably. If you use a resistor to do that, that will use 3 times the energy or so and waste 2/3 of it as heat in the resistor.
There are do it yourself switcher supply forums. I chose not to participate. Too much calculation; too many components you have to buy available only in lots of 100000 with the specification sheet in chinese. I don't even have a meter capable of measuring inductors. I'm thinking of buying an L meter for $150 to repair an electronic organ sitting in my living room with dead D# keys, but the cost of moving the organ out to the dump would be ~$200.
If you really want to diy, buy a 48 v to 3.3 v vicor converter. Then use a resistor to convert the 52 v of your battery to 48 v. Less wasted energy. Your lights may still draw 2 v instead of 3.3, so you could need another resistor there. Running 2 v LED @ 3.3 v could burn them out in tens of hours. Measurement for the lamp LED voltage detailed in post #5. If the LED are stacked series to make 4 v load, then a 48 to 5 v converter would be more the thing you need.

 

[FiremarshallDave]

No. No. the ferrite toroid is a coil that when oscillated at 2 mhz or 10 mhz or something, changes the excess voltage from the battery pack into greater current at lower voltage for the 13 lights. A buck converter, probably. If you use a resistor to do that, that will use 3 times the energy or so and waste 2/3 of it as heat in the resistor.
There are do it yourself switcher supply forums. I chose not to participate. Too much calculation; too many components you have to buy available only in lots of 100000 with the specification sheet in chinese. I don't even have a meter capable of measuring inductors. I'm thinking of buying an L meter for $150 to repair an electronic organ sitting in my living room with dead D# keys, but the cost of moving the organ out to the dump would be ~$200.
If you really want to diy, buy a 48 v to 3.3 v vicor converter. Then use a resistor to convert the 52 v of your battery to 48 v. Less wasted energy. Your lights may still draw 2 v instead of 3.3, so you could need another resistor there. Running 2 v LED @ 3.3 v could burn them out in tens of hours. Measurement for the lamp LED voltage detailed in post #5. If the LED are stacked series to make 4 v load, then a 48 to 5 v converter would be more the thing you need.

Thanks Joe, I am going to experiment with the ferrite. You did elude to the purpose of the ferrite so I will test the voltage up and down stream to see how much it changes the voltage. I follow the greater current lower voltage idea. That seems like exactly what I would want. A quick (test on a working light) to measure the final + wire voltage to the leds (after it goes thru the driver from a 8.4v pack) and I will know if the ferrite inline alone idea will work or if I should just create a new 18650 pack running at 3.6-4.2 volts rather than the 2 batteries in series pack. It's a 2s3p pack that could be a 6p pack.