Let the debugging begin

Threw together 6 LEDs using an experimental construction technique. The new technique has promise, but I need to experiment some more. Here are the results...

1) One LED had the green/red sections wired in backward, so had to cut that out...down to 5 LEDs in the string.

2) Connecting the brake signal wire turns red on full brightness and the other colors off. (At least SOMETHING works as planned).

3) Red and green are adjustable, but for some known (as of yet) reason, the blue pot adjusts red as well. Blue does not light up in any LED. this is quite odd behavior.

4) Red and green are always on with at least 2-3 mA flowing through each (out of 20mA), regardless of pot position, so there's too much light showing even when turned down to the minimum. This is VERY wierd behavior. I'm hoping this is tied directly to problem #3.

Other than that, I'm reasonably happy. I'm not looking forward to the amount of work ahead of me as it involves the construction phase of the screens, but I promised.

I will advise when I have more debugging info...

 

<sigh> ...and this is why I want to debug things before telling people to start sending in their screens...

OK, it seems that there is a bad connection/loose wire/whatever along the LED string...a few presses here and there and all of the colors start working. The control box seems to do what is expected of it, but I can't really chase down the "constant on" issue until I resolve the loose wire/bad connection issue.

I'm going to leave the box on for a couple of hours to see how things run over the long haul. I will say this, though...at full power, the little critter cranks out a LOT of heat (so much so that I can't touch the resistors/transistors for more than an instant without burning myself). I turned it down to a bit over 2/3rds power for testing purposes, and it's still too hot to touch. After a couple of hours I'll see how hot the box gets...after 10 minutes or so it has warmed up significantly. I may turn the current down to half...it's pretty frikkin' bright, even at half power, so that may be satisfactory.

 

thanks Mac!

i know i enjoy hearing how things are going... and good to hear that the loose wire thing was a easy fixer.

 

Heat, sound like something we can mount over the fuel tank.


EDIT: This was just a light-hearted joke, something to bump the thread back up top. Almost all the trunk mounted amps generate heat to the levels Mac described.

 

Mac, in this day and age I think you are going to need to find a way to FORCE a lower setting. People are dumb, and they never read the directions. This will in now way prevent them from sueing you when their car burns up. I am sure that 99.99999999% of the people here would have not problem, but it just takes one.

 

there are a couple layers of rather thick metal between the interior and any fuel.
Unless you decide to zip tie it to the fill lines or something.

Everyone please note... Please do not zip tie electronic devices to the fuel fill pipes or tubes. thank you.



It gets hot... Is it hot enough to damage the unit? If it just gets hot and cannot be touched while in operation I am fine with that.
I'll just make a little bracket and install it somewhere where it won't touch anything and no one will touch it.

Is the case metal or plastic? I'll bet I could find a heat sink to solder on.

 

Sorry kids, haven't had any chance to work on it today...the 'rents leave Sunday morning.

The components themselves get really hot, but the plastic box just gets annoyingly warm from the radiated heat (enough to keep your hands quite toasty during the winter months). I will certainly limit the maximum current to about half of the originally planned setup (down to 10 from 20mA)...this is an easy fix, just a value change for 3 resistors. I'm still hopeful the extreme heat is from the loose connections in the LED string, but I doubt it...it is seemingly just a property of the design (things would have been much better if we were running off of a 5V supply instead of 14V).

 

I'm getting depressed working on this one...

OK, to avoid the whole short circuit issue, I put 6 of the LEDs on a breadboard and connected it up to the control box. The colors are hard to perfectly match between all 6 LEDs, but I think in the finished product the color difference will be minimal (at least from what I've seen so far).

The main issue, however, is the heat. At 10mA/color, the brightness is quite good, but the heat is still high. I'm not very comfortable with how this will affect the longevity of the components in the control box, even though they're only running at 50% or less of their rated capacity (power resistors at <50% and transistors at <10%).

Regardless of my previous promises, I'm reluctant to sell this as is due to the above issues. I have an option or two I'm going to try. I have new parts coming in (specifically, a new LED type) that will allow me to design/install the circuit in a way not currently possible with the parts on hand. The new design would significantly lower the power necessity, but would increase the cost modestly (about $20). I believe I can use the current PCB design, but some part values will need to change (acceptable to me).

So, I leave it up to me fearless clientelle to make some decisions. You can accept the circuit in it's current form, with its potential downside of high heat and POSSIBLY reduced circuit lifetime (I say POSSIBLY since the components ARE running at MUCH lower levels than what they are rated for, and the LEDs (the difficult part) will be fine, regardless of what happens to the control circuit). modifry is also working on a digital version of the control box, so if it worries you, you could upgrade to his box in the future. Or I can make the minor design changes, use the more expensive LEDs, and call it a day.

I leave it in your hands...

 

Surprisingly, the new parts showed up today along with the original order of project boxes...

 

here's a vote for quality. I'd wait until you're satisfied Mac.