LED electrical question

[tink2011]

Looking back, I'm not sure how I got that either! I hadn't had my coffee yet? My calculations later in the day told me 4.0 ohm, 2 watt. So, this morning reading your calculations matched mine made me a very happy girl! I will keep you posted on my progress. I plan to order the step down DC/DC units and resistors today. First plan is to light up the nose cone with the ribbon lights. I'll keep you all posted, and give anyone a ride on playa...just flag me down! I'll be driving an alien holding the world atop a space ship. Got stereo system installed over Memorial Day weekend, so it'll be a fun ride this year! We burners know how to have fun over holiday weekends...THANKS!

[tink2011]

Hi 5280MeV,

I am not sure how you got 7.14 Ohms - maybe you could post your calculation or assumptions?

I'm not sure how I got that either! Your calculations actually match mine that I did later in the day yesterday, so I am a happy girl! I think maybe that was my calculation if I went from the full 12v input? Not sure.

I will be ordering and wiring in the weeks to come and I will keep you posted on progress. I will be available on playa for transport and rockin' tunes, and just a plain good time. Flag me down.

[tink2011]

GOOD NEWS! I got the LED ribbons all wired up from 12 v deep cycle battery through step down converter (Bought the exact ones suggested earlier in this thread) to 5 v, then wired four 1 ohm resistors in series. IT WORKED! The step down converter gets pretty warm, but it works great. The 1 ohm resistors get warm, but not hot. I plan to encase them in a tupperware container to keep them safe from dust and injury...I have 8 ribbons to wire, so the only problem will be the bulk of all that wiring...hmmm...but I'm getting excited and confident with soldering.

Got the stereo wiring squared away and put in new switches for headlights and fan for head unit. I'm feeling happy! Thanks to all of you for your support during this process. Flag me down for a ride on the alien vehicle...the man burns in 73 days.

-One more question. With the wire wound resistors there's a solid wire at each end. I'm used to soldering with copper wire with strands. When soldering the solid wire should I tin and then solder just like the stranded wire with the two wires parallel or should I try to twist them together and then solder?

[gyre]

Twist and solder

[BBadger]

Sounds like a good setup! Yeah, those electronics will get pretty warm at those high currents, but that's their nature. Imagine how much hotter it would be if you were using less efficient methods to step-down your voltage; I've started roasting PCBs doing that.

If you can swing it, maybe instead get a metal box and seal it up with caulk and rubber lining instead. The reasons are that it'll be easier to mount stuff inside, it'll be sturdier, the contents inside can cool off from the metal walls, you can access the stuff inside like a fusebox, and you can bolt the box to your MV where it won't break off like the plastic could.

For soldering solid wire I would just twist them together and then solder them. No need to pre-tin as you can load up the braided wire pretty heavily with solder as needed. Then cover the joint up with some heat-shrink tubing.

If you're really feeling like making a secure wire-to-wire connection you can use the ol' Western Union Method:

western-union-or-linemans-splice.jpg

[tink2011]

Thanks so much for the great idea on the box. What would you suggest I get? An acutal fuse box? That might be a good idea, but pricey? Also, how to secure the stuff inside? I was thinking velcro...but I am a seamstress at heart and this metal and welding/soldering is all new to me! Any advice would be helpful...thanks again!

[BBadger]

Thanks so much for the great idea on the box. What would you suggest I get? An acutal fuse box? That might be a good idea, but pricey? Also, how to secure the stuff inside? I was thinking velcro...but I am a seamstress at heart and this metal and welding/soldering is all new to me! Any advice would be helpful...thanks again!

Hmm, how big do you need the enclosure? I was doing some thinking and a junction/breaker box might be nice for the job. They're easily mounted, have doors, and you can buy mounting plates for them (or just use some wood). They also have knock-outs (holes) on the side where you can attach aluminum conduits. The conduits are good for managing cables and such, and relatively cheap as well. The enclosures cost from $15 - $40 or more depending on what you want, and the size. There are some at Home Depot or places like that, or even on eBay. As an alternative, you can also try using an ammo can or even metal cash boxes, but they may be harder to mount or drill into.

For mounting, I'm not too sure about the specifics of the boxes themselves, but you can usually put a piece of wood or plastic inside the box (like plexiglass, but don't crack it) and then drive the mounting screws through that piece of wood (drill the holes first) and into the surface you're mounting the box onto. Then mount whatever you want to the wood. You can also just attach screws/posts to the box itself. I would also hot-glue stuff down, just to reduce vibrations and such.

Something to consider is strain relief for your cabling. That means fixing your cables so that if there is motion or pulling the cable or its connections won't be mechanically damaged. So for example, you might want to give the cabling inside the box some extra slack and then zip-tie it to a post so that if the cable gets pulled it'll pull on the zip-tie instead of the soldered mount inside. You don't have to purchase specialized stuff; thick zip-ties work nicely so long as the cables you use have a good covering and you give it some slack to work with. Velcro would work too, but you need to make sure the cables are held firmly. Fixing the cables to a stronger cable or wire can also help reduce the strain.

Finally, you can also put your cables into the aluminum conduits until they're close to the destination point. The benefit is that you won't have wires flying all over the place and the conduits provide some support for the cables. The conduits also attach nicely to junction boxes using conduit box connectors. Who knows, maybe those conduits will add to the space-shipness of your car. Both conduits and the connectors are pretty cheap in bulk at a place like Home Depot. If you need to bend aluminum conduit there are various guides on the net. If it's too much of a pain to bend the conduit, you can also just mount them to other electrical boxes (something you can get cheaper in bulk like this). Just something to think about.

[gyre]

As I mentioned before, buy a box like this or nema aluminum used.

Prices start from scrap and up, depending on demand.

[tink2011]

Alrighty then, great info! I have already run all the wires and have them zip tied to the main underbody (chicken wire over welded metal) and have begun the fabric top covering, so the conduits (although a great idea) won't work. I like the metal box idea, and will head to Home Depot later this week. I'm a hot glue junkie, so that sounds good to me! I will have eight step-down DC-DC converters (a little bigger than a pack of gum) and 40 one ohm resisters split between 8 separate lines. I can mount the box near the deep cycle battery, but it'll be cramped. I will make it work. The man burns soon! THANKS!!! Hope to see you on the playa for a ride on SpaceGrab!

[BBadger]

You can also mount the converters to the side of the metal box for heat-sinking especially if you're running an air-tight box (which you should).

Looking forward to seeing it on the playa!

[tink2011]

Hi Everyone! It's me again with the "ribbon" light strings ordered from China that everyone helped me get changed over from a 4.5 volt (3 AAs) to the 12 volt deep cycle battery. The first LED string that I wired with the step-down voltage reducer (12-volt to 5-volt model posted earlier in this thread) and four 10 ohm resisters has had some problems. The first LED in the string burned out before my eyes tonight (after about 30 minutes of use) and then I noticed a random few others were out, numbers 5 and 8 in the 32 string. The ones that are out are getting warm, and I'm noting some slight brightness difference in the others still lit. I have not wired up all of the light strands yet, and I'm thinking maybe I will leave these on the smaller 4.5 volt packs for this year and "suffer" with the little turn on/off switches rather than run them all to the deep cycle. The EL and CCFL lights should work great off of the deep cycle, but a little afraid to run the LEDs since they are failing already and a big part of the look of the car. Any suggestions? It's that time of year when I need to make a change if it's going to work on the playa. Thanks so much to all of you who have helped with this saga, and so it continues! )'(

[BBadger]

Hi Everyone! It's me again with the "ribbon" light strings ordered from China that everyone helped me get changed over from a 4.5 volt (3 AAs) to the 12 volt deep cycle battery. The first LED string that I wired with the step-down voltage reducer (12-volt to 5-volt model posted earlier in this thread) and four 10 ohm resisters has had some problems.

It sounds like there is too much current flowing into the LEDs. This could be from the voltage regulator not being exactly at the voltage you need, or maybe the resistors have tolerance issues, or even the LEDs might change resistance as they heat up. It could also be from miscalculating the needed resistance. All different things can happen. This is all compounded by the fact that every LED that dies in the strand will cause the the current to be distributed to the others--causing a chain reaction of LED deaths.

Running through the calculations again (it looks like you used 5280MeV's calculations, though I'm not casting blame), 4 ohms may be too little resistance for the strand:

Code: Select all

(5V-2.2V) / 4 ohm / 32 LEDs = 21.9mA / LED

Usually you should shoot for 20mA/LED, and that's assuming the LEDs drop exactly 2.2V. My forward voltage estimates for red/amber is 2.0V not 2.2V, so that would require:

Code: Select all

(5V-2V) / (20mA * 32) = 4.6875 ohm

However, before we try to work for that number, it'd be better to actually tune your regulator/resistors to the measured parameters of your system. Do you have a multimeter? If not, please obtain one, as the next procedure uses it. They're very useful for troubleshooting, and they're damn cheap for such a useful tool. Bring it to the playa too!

Hint: never measure current using the amp-meter setting of the multimeter. Instead measure voltage across a known resistance and calculate it. Then your multimeter won't affect the circuit (much), and you won't have to manually range it from the 0-200mA range to the 10A range, or accidentally blow the fuse (very common). Now back to the subject...

If you purchased an adjustable voltage regulator...

I think you did buy one of these. If you did, we can adjust the step-down regulator's output voltage to reduce the amount of current flowing into the resistors. The amount of current you need in a strand of N number of LEDs is:

Code: Select all

Current for whole strand = N * 20mA

For 32 LEDs, that is 32 * 20mA = 0.64A of current.

For your semi-burnt-out strand, it's fewer LEDs and less current, but we'll deal with that later.

What we're going to do is start from a low voltage, say 2V, and then incrementally increase the voltage until you've measured 0.64A of current passing through one of the current-limiting 1ohm resistors.

This procedure also means you can even remove some of the current-limiting resistors; the lower voltage will compensate. Removing resistors, and reducing voltage in this manner will reduce power waste. You can probably get away with only using a single resistor per strand with the lower needed voltage you'll be determining through this procedure (assuming all the strands have the same LED forward voltage drop).

First set the voltage regulator to a low value like 2.0V. Now connect it to your strand + current limiting resistors like normal:

Code: Select all

Vpp  -> Current limiting resistor(s) -> LED Strand -> Ground

Measure across one of your 1-ohm current-limiting resistors. Because the resistor is 1 ohm, the voltage you read should be directly proportional to the current flowing through it.

Now increase the output voltage of the voltage regulator until you measure your target current: 0.64A. When you reach this value your strand should be receiving exactly the right amount of current. If you wanted to play it safe you could even drop it to 18mA per LED (that would give you 32 * 18mA = 0.576A target current), or whatever level of brightness seems good within the bounds.

Your semi-burnt-out strand...

What about your semi-burnt-out strand? The number of LEDs, N, is fewer (29?). You have the option of trying to replace them with LEDs you can find on the net (like eBay), or assume they're lost and compensate with more current-limiting resistance (you can't change the voltage as that is shared with the other 32-LED strands). I would actually suggest buying LEDs on eBay as replacements so you can do "field repairs" on the playa if needed. It also means you can balance the current back to normal as needed instead of compensating with resistance that can affect brightness if not properly adjusted (described subsequent).

To adjust using resistance, your new target current is: 29 * 20mA = 0.58A. This time you'll need to know the voltage drop of your LEDs. Call this Vf. You should measure, rather than estimate, Vf by powering on your LED strand and quickly measuring the voltage across your LED. You should probably do this before you adjust the voltage regulator's output voltage in the earlier section.

Now with the Vpp from the previous section, and the measured Vf:

Code: Select all

(Vpp - Vf) / (20mA* N LEDs) = the resistance you need

So for example, with Vpp = 4.2V, Vf = 2.0V, and N = 29 LEDs:

Code: Select all

(4.2V - 2.0V) / (20mA * 29 LEDs) = 3.79 Ohm

Oh look, pretty close to the 4 ohm you used before, so you could probably use that if that were the case. If you calculated something like 3.5Ohm, you could produce a more accurate resistance (to balance brightness) with something like (3 x 1-ohm series resistors) in series with (2 x 1ohm in parallel = 0.5Ohm).

If you have a fixed-voltage 5V regulator...

For this you'll need to use the calculations for your semi-burnt-out strands above for all your strands, with a fixed Vpp = 5V.

If anything is confusing, feel free to ask.

[tink2011]

Hi BBadger! You receive the bonus prize of a trip around the playa with a chocolate bar! Seriously, thanks so much for the input. As I laid in bed last night I actually dreamed about electrical and soldering on playa. It was a weird dream! Anyway, I am wondering if I may have caused a problem when I was attaching the wires to the battery last night. I accidentally mixed up the red and black wires on the battery terminals (in my dark garage) and think maybe this might have caused an issue? I say this also because I had run that same strand for at least an hour before and had no problems.

Yes, I got the fixed 5-volt regulator (the exact one listed earlier in the thread from Hobby King). I did not use the meter in this strand, but think that's first on the list. I plan to wire up another strand and do the checks you suggest. To measure "across the resistor" how do I physically do that? One tip on wire coming out of resistor and the other end on the ground wire? Your calculations are fantastic and I am learning a lot, as well as you are confirming what I had suspected last night.

I hope you are having an electrifying day! Looking forward to hearing back from you. I'm going to wire this up this afternoon and I'll post again. THANK YOU! Can't say that enough.

[BBadger]

Hi BBadger! You receive the bonus prize of a trip around the playa with a chocolate bar! Seriously, thanks so much for the input.

Awesome! I'll definitely have to catch a ride on the MV!

I accidentally mixed up the red and black wires on the battery terminals (in my dark garage) and think maybe this might have caused an issue?

So on the battery terminals going into the SBEC (voltage regulator) or the output going to the LEDs? The SBEC should have reverse polarity protection (on the inputs I believe), and not fry if the wires are switched for a short while. If you switch the outputs going to the LEDs it might have damaged the LEDs, but I think your problems were really from one LED dying and then causing strain on the others.

I say this also because I had run that same strand for at least an hour before and had no problems.

Sometimes the LEDs can give the impression that they're working fine even if they're being fed too much current/voltage. Only after some time will they show some degradation or damage. I remember as a kid hooking up LEDs straight to 9V batteries without any resistors (didn't even know what those were) and they worked fine until I connected too many up and it burned out $8 worth of red LEDs--a lot of money in parts back then.

I think the problem is that there is slightly too much current going into your strands (and the general design of the LED strands in general), and this caused some damage over time. Then when one LED died, it increased the current the others had to endure, which causes more stress to them. With each LED that dies, the others suffer even more, causing a chain reaction. In effect, even one dying LED can spell the end for the entire strand.

To combat this, perhaps shoot for a slightly lower per-LED current like 18mA to account for differences in component manufacturing. It'll still be bright, but the LEDs won't be pushed to their limits. I've found that the cheaper LEDs from China as less tolerant of exceeding their specifications. It is easier to do this if you can adjust the voltage rather than mix and match resistors to get the correct current, but I think you'll still be able to do it with your 5V fixed regulator. 5 ohms as your current limiting resistance should work fine for 32 LED-strands ( (5V-2V) / 5ohm / 32LED = 18.77mA/LED). I would also buy some loose LEDs and repair your damaged strand rather than change the current limiting resistance. You'll need spares anyway for LEDs that may die in the field--especially as you have to make repairs quickly to avoid losing the entire strand.

An electronics "field kit" (fishing tackle boxes are great) can be helpful for repairs, and you can store the parts in your MV. Bring the multimeter, portable soldering iron such as this, pliers, wire snips, wire strippers, extra resistors and LEDs, wire, solder, sponge + water for sponge (or better, those brass tip cleaners), and also a solder sucker (I love these things, but remember to properly dispose of the solder), headlamp, and finally some electrical tape. You'll be the electrical field surgeon!

Yes, I got the fixed 5-volt regulator (the exact one listed earlier in the thread from Hobby King). I did not use the meter in this strand, but think that's first on the list. I plan to wire up another strand and do the checks you suggest.

Some things to check:

- Is your SBEC set for 5V, not 6V? It has two settings from the site's description.

- SBECs are used with servos and RC applications, so the voltage may not be spot on. So you should measure the actual output voltage and make adjustments for that. Also the SBEC you're using is rated for 5A so if you're using one for multiple strands shoot for that maximum current usage per SBEC.

To measure "across the resistor" how do I physically do that? One tip on wire coming out of resistor and the other end on the ground wire? Your calculations are fantastic and I am learning a lot, as well as you are confirming what I had suspected last night.

To measure voltage across a component, such as resistors, set the multimeter to measure voltage and place the ends of each of the probes on either side of the component. You'll read a positive or negative voltage depending on which side of the the probes are on, reflecting the direction of the flow of current. Make sure you're not in current measuring mode, or your meter will act like a bare wire in the circuit, and you'll probably blow your fuse or destroy the meter if it has no fuse.

Measuring-the-voltage-drop-across-the-unknown-resistor.jpg

[tink2011]

I set the multimeter on voltage and got the following numbers:

Voltage leaving 5-volt reducer was 5.08 volts
Each 1 ohm resister had a voltage drop of 0.63 consistently.

So, I should add one more 1 ohm resister (to make a string of 5) and that should lower the voltage at the LED to a good level, right? It's strange, because it seems so easy until you start trying to change things up! Sorry to hear about your LED loss as a kid. The learning curve starts early!

[tink2011]

Okay, I just put 3-AA batteries in a brand new light string (the ones that ran beautifully on the playa last year using AA batteries) and measured the voltage drop across the itty-bitty resistor inside the battery pack as it came from the manufacturer. It measured 1.9 drop. Would it be correct to say that the 4.5 volts from the batteries are reduced to 2.6 volts to the LED strand? We know that this combination works well on playa for the entire week.

If so, the original plan of 12-volt to 5-volt reducer wired to four 1-ohm resistors (0.63 drop each) would give me:

5.08 V (actual measurement leaving Hobby King voltage reducer) - 2.52 (across four resistors of 0.63 each) = 2.56 volts reaching the first LED. This lines up exactly with the measurements I took from the historically well-running string straight out of the package using the AAs.

Should I try wiring up a new strand and leave it for an extended period of time? Could the strand I had issues with be defective? Is the Hobby King reducer putting out surges of higher voltage? Or, maybe just go ahead and add another resistor to lower it a bit more because the behavior of the LEDs was consistent with too much voltage? ... Hmmm...

[BBadger]

Yeah, those measurements do seem in line. Still, try measuring across one of the actual LEDs themselves (the two leads of the LED) just in case to see if their drop adds up. That way we have the actual Vf for the LEDs.

It might also have something to do with the power supply. While batteries have no ripple, the voltage output of the SBEC does from the voltage conversion. Some forum posts on other sites report some Turnigy SBECs having ripples up to 250mV, and this can possibly stress your diodes. There are also reports of large amounts of heat being generated. I know you've already purchased power supplies, but you might consider replacing the SBEC(s) with a more robust desktop power supply especially at the currents you're running. Here are some PSUs you could look into depending on your amperage needs, for example:

- $28.95 - 50W 5V output 10A - Meanwell DC/DC Power Supply Single Output 5 Volt 10A 50W 5-Pin

- $22.95 - 25W / 5A /5V version

Those types of power supplies are nicely mountable, slightly voltage adjustable (+/- 0.5V, so you can use 4.5V), are easy to wire, low-ripple (8mV in test conditions), well-protected from shorts/overloads/etc., and have heat-dissipating cases. Meanwell also does a full burn-in of the power supplies. Plus, you can power multiple strands from the same PSU given the large current they can supply. It might be worth the investment.

Regardless, I wouldn't do a burn-in test for your LEDs quite yet until all the measurements are performed and the bases are covered. You can also purchase some capacitors to help clean up the DC voltage too (at Jameco or other part shops like Mouser/Digikey/AllElectronics/Sparkfun/etc.).

[5280MeV]

Would it be correct to say that the 4.5 volts from the batteries are reduced to 2.6 volts to the LED strand? We know that this combination works well on playa for the entire week.

No, the issue is that AA batteries do not give you 1.5V per cell under load. I think that this may be a major reason why you have been having trouble - and a problem with my calculations that I didn't think about.

In reality, you can expect an effective terminal voltage of around 1.1-1.3V under load with good AA batteries. When they are dying they get down to 1.0V each or less. You have to account for the internal resistance of the AA batteries which varies from 0.15-0.3 Ohms per battery. So your three batteries in series add between 0.45 and 0.9 Ohms of resistance.

Actually, this makes more sense, as an effective 3.9V source with 2.2V forward LEDs getting 32 x 20 mA or current would require a 2.7 Ohm resistor.

But there is something funny about the measurements. You also report a 1.9V drop across the 2.7 Ohm resistor which means that there is 700 mA of current running through it. The 1 Ohm resistor with a 0.63V drop means that you were only running 630 mA of current.

I trust your voltage measurements - these are typically right, but I wonder if that is really a 2.7 Ohm resistor. 700 mA is a lot of current draw from a AA battery! Your multimeter should have a resistance setting so that you can directly measure the resistance of that 2.7 Ohm resistor. You can also directly measure the current, but keep BBadger's warnings in mind, make sure not to blow the fuse, and also connect the multimeter in series.

In any case, I definitely think that BBadger is right that you need to add at least one more 1 Ohm resistor.

[BBadger]

But there is something funny about the measurements. You also report a 1.9V drop across the 2.7 Ohm resistor which means that there is 700 mA of current running through it. The 1 Ohm resistor with a 0.63V drop means that you were only running 630 mA of current. I trust your voltage measurements - these are typically right, but I wonder if that is really a 2.7 Ohm resistor. 700 mA is a lot of current draw from a AA battery!

It is a lot--especially for an Alkaline cell--but if it's a brand new battery the voltage will be at its peak even under load and that could be measured across the resistor in that case. I'd like to know what the actual voltage drop of the LED is too.

Your multimeter should have a resistance setting so that you can directly measure the resistance of that 2.7 Ohm resistor. You can also directly measure the current, but keep BBadger's warnings in mind, make sure not to blow the fuse, and also connect the multimeter in series.

Yeah, it would be a good idea to just measure the resistance of that resistor itself. I never bothered learning (rather memorizing) the resistor color bands when I started out because I never trusted them (and the color of the bands varied). So I typically just measured the resistance directly; plus I wanted to rule out the differences due to manufacturing tolerances. At this point I'd actually measure everything (besides the current, as that can be calculated) in the original strand circuit to get solid numbers to work with.