Circuit design: protecting LED tail lights from transients

denpo

Carburated Nihilist
Joined
Jul 14, 2010
Messages
2,523
Location
Montreal, QC
I dusted off my LED tail lights project, I'm starting to design a proper PCB for them.
I did a first try with the laser-print-on-cheap-glossy-newpaper-transfert and the result is promising.
circuit.jpg


In current configuration led are in serial, with a resistor to limit the current, and a voltage regulator (LDO LM2940) to get a clean 12v.
Still, the circuit is open to potential transients with voltage spike theorically up to 80v and reverse polarity.
I was thinking about putting a 18v transorb (P6KE18A) and a fast recovery diode.
Thing is, I'm no electronic expert, so I'm not sure where and how I should put them in my circuit.
Anyone with real electronic knowlegde could help me out on this?

Here's the current schema
taillightschema.gif

I can profile the Eagle CAD .sch file on request.
Thanks.

Needless to say, once completed all the datas will be uploaded to Vettemod for everyone's use.
 
Don't need the regulators.....just a 22 mfd cap to ground will suppress any spikes of any concern.....maybe a 50ohm resistor is series before it.....:D
 
I dunno, my decade CMOS counter/divider and NPN transistor drivers for the LED's been there in the vette for 18 years now, never missed a lick, the LED's in my console been there too, got LED marker lights for the camper, they work fine....just 7 in front, 7 in the back in parallel I think it's a 50ohm resistor in series ....no problems....the CMOS input runs directly off the tach integrator/filter....same as the tach....look at any old auto radio schematic, they never seemed too concerned with all that surge crap.....

methinks you making a mole hill outta a ant hill....something like that....

starter spikes appear on the collapse of the mag fields on the OPEN side of the starter/DRAIN side, the battery sees nothing for spikes due to it's very LO impedance.....the rest of the possible inductive loads would by injector firing and maybe residue from spark plugs....

and that not enough to do shit, more chance of a lightening bolt doing you in.....

:crutches:
 
I dunno, my decade CMOS counter/divider and NPN transistor drivers for the LED's been there in the vette for 18 years now, never missed a lick, the LED's in my console been there too, got LED marker lights for the camper, they work fine....just 7 in front, 7 in the back in parallel I think it's a 50ohm resistor in series ....no problems....the CMOS input runs directly off the tach integrator/filter....same as the tach....look at any old auto radio schematic, they never seemed too concerned with all that surge crap.....

methinks you making a mole hill outta a ant hill....something like that....

starter spikes appear on the collapse of the mag fields on the OPEN side of the starter/DRAIN side, the battery sees nothing for spikes due to it's very LO impedance.....the rest of the possible inductive loads would by injector firing and maybe residue from spark plugs....

and that not enough to do shit, more chance of a lightening bolt doing you in.....

:crutches:
That very possible I'm overthinking the thing. See I want this version to be final, making the PCB will be enough of an hassle.

NPN based led drivers, something like that ?
FTFKKU6Z2HEWIFNTP5.LARGE.jpg


Aren't NPN based regulator designed to regulate current only?


Worst voltage spike scenario seems to be Load Dump
 
You can overpower a LED, true, but when power is reduced, there is little variation noticeable in light output.....yeh, some science instrument maybe say so, but when looking at it, or beaming off a garage door, not so much.....

I don't see the point of MOSFETS either.....

:shocking:
 
Denpo, I think you were on the right track with your original approach. IMHO your references sound like they know what they are talking about. Better to be safe than sorry. Getting as clean a 12v power possible from close to battery + terminal is also preferred and may minimize the need for trans-orbs, but why not be safe. Not expensive are they? You will need all the protection you can get if you choose to put that battery cutoff switch in the circuit without that red sense voltage wire and someone inadvertently opens it up while the engine is running. :lol: (Or, that switch gets worn and intermittently looses connection.)
 
Denpo, I think you were on the right track with your original approach. IMHO your references sound like they know what they are talking about. Better to be safe than sorry.
Thanks for chiming in Bob.
The post I included the link is one amongst tens of discussions I've scanned. I have the same feeling, the dude knows its job.

Getting as clean a 12v power possible from close to battery + terminal is also preferred and may minimize the need for trans-orbs, but why not be safe.
About the 12v source, well, it's the "signal" 12v coming from the brake switch, the flashers and the healight switchs, nothing much I can do about these. Each line will see less 600 mA of load. So I see no point getting better voltage source and using the "signal" 12v as actuator. But I can be wrong.

If I read you well, the "further" you are from the voltage source, the more likely are the transient. I thought you would just loose voltage.
I hate how common logic is a failure when applied to electricity.

Not expensive are they?
Regulators are 1.30$ each, trans-orbs 20 cents. No much reason to put them in, when you know how
You will need all the protection you can get if you choose to put that battery cutoff switch in the circuit without that red sense voltage wire and someone inadvertently opens it up while the engine is running. :lol: (Or, that switch gets worn and intermittently looses connection.)
No comments :lol:
 
Ok, guy, know let's suppose for a second the extra protection are worth the couple of dollar i would cost.


-Can someone tell me if the transient suppressors and the voltage regulators I've put in my circuit are in the correct configuration.

-I've looked on how to protect the system against accidental polarity protection, with a diode of course, and turn out there's two design, serial and parallel.
Serial is simple, but eats voltage, parallel, which is preferable, is just beyond my electricity understanding. :sos:
 
Answering my second question....

[ame]http://www.youtube.com/watch?v=IrB-FPcv1Dc[/ame]
 
Following the video I posted, I found this Mosfet to be a good candidate.
http://www.fairchildsemi.com/ds/FD/FDS4935A.pdf
So-8 package, harder to weld, but now I can have real pcb, I'm gonna try to push the quality.


Here's the revised schema/board
Eagle don't have the LM2940, so I put a LM340 instead.

schemav1.gif


boardv1.gif

The leds are on the blue side, le component on the red side.
 
I admire your willingness to do some old-school PCB work. I've got the chemicals and stuff in the garage to do some board work this winter.

I just (quickly) skimmed through the thread (so some comments didn't make sense to me) but the schematic looks like it could be simplified. Couple questions:

How much current are you targetting for each LED string?

Why are there three sets of LED strings? I can see Run and Brake/Turn, but what is the third?

Crap, someone at the door. Gotta run.
 
I admire your willingness to do some old-school PCB work. I've got the chemicals and stuff in the garage to do some board work this winter.
My last year propotype, with pre-drilled board was a proof on concept on the amount of light generated. But with those boards the work always end up crappy. It had to be better
I just (quickly) skimmed through the thread (so some comments didn't make sense to me) but the schematic looks like it could be simplified.
Not sure if they all do.
I'm really interrested to know.
Couple questions:

How much current are you targetting for each LED string?
Each led drawn 20mA, so each line shouldn't draw more than 400 mA.
Why are there three sets of LED strings? I can see Run and Brake/Turn, but what is the third?
I decided that all four tail light would come with light, brake light, and flasher.
Light is a the center, then come brake light, flashers are on the outside.
Crap, someone at the door. Gotta run.


By the way, I realized I need to rework it to accomodate white led higher voltage dropout. Here's lastest version



schemav2.gif




boardv2.gif
 
Each led drawn 20mA, so each line shouldn't draw more than 400 mA.

I don't like your expression of it, hooked in SERIES the current drain will be still 20 mills/mA assuming you have the proper current limiting resistor to make the voltage at the first LED enough to run the bunch correctly....

keep in mind that LED's take an inherent just shy of a volt to crack the diode, but have internal current limiters to make them work on X voltage, read the directions....so if you have a bunch of LED's that are current limited with internal resistor packaging, you hooking series maybe not fire them at all, let along put out light.....

a more typical current for LED's is 25-30 m/A that I know of....

:stirpot:
 
.................
I decided that all four tail light would come with light, brake light, and flasher.
Light is a the center, then come brake light, flashers are on the outside.
......

Refresh my memory. I thought that brake, turn, and flasher were all the same "filament", and running lights would make a total of two filaments/circuits.
 
Each led drawn 20mA, so each line shouldn't draw more than 400 mA.

I don't like your expression of it, hooked in SERIES the current drain will be still 20 mills/mA assuming you have the proper current limiting resistor to make the voltage at the first LED enough to run the bunch correctly....

keep in mind that LED's take an inherent just shy of a volt to crack the diode, but have internal current limiters to make them work on X voltage, read the directions....so if you have a bunch of LED's that are current limited with internal resistor packaging, you hooking series maybe not fire them at all, let along put out light.....

a more typical current for LED's is 25-30 m/A that I know of....

:stirpot:
You're correct, the total current stay 20mA whatever the number, good news so I can use 100mA voltage regulator (smaller package), thanks for pointing this out.

I'm using this tool to get my optimal configuration (ie: the one that dissipate the least energy)

Specs of my led :

Red:
Forward voltage : 2v
Foward current : 20mA

White:
Forward voltage : 3.5v
Foward current : 20mA
 
.................
I decided that all four tail light would come with light, brake light, and flasher.
Light is a the center, then come brake light, flashers are on the outside.
......

Refresh my memory. I thought that brake, turn, and flasher were all the same "filament", and running lights would make a total of two filaments/circuits.
Damnit, you're freaking right. See, I didn't have my Vette in one piece for a long time, and at the time light pattern was least of my concern.

So, I see in the electrical diagram (76), flasher and brake arrive as brown and white at the column and goes back as yellow. I think I could make the brake switch its own line directly to the light, bypassing the column. Jim, are you here? :sos:

I found out hard white leds (understand : blueish) show as orange light through the red bubble covers, how could I not exploit that.
 
Top