OK, I ran the board for about 2 hours at 10mA, which means it should have hit steady state within 10-15 minutes. The case is warm, but not uncomfortable. This runs the transistors at 5% of their rated power and the resistors at 25% of their rated power, so this thing certainly isn't going to blow up (engineers normally design for running at 50-75% of a components rated value, so I'm WAY below that). I am satisfied with the final outcome, even if it did give me a huge scare.

I'm still looking at beginning full production the first few days of the coming week, so be on the lookout for a new thread (I will post a link to it here).

By several strokes of luck (and some really good ideas by my father while he was here), the design is coming together better in many areas (other than the heat issue) than I had hoped for. I intended to put a fancy label on it all, but I may forego that just so I can get some of them out the door.




While I certainly appreciate the votes of confidence and the willingness of many of you to accept a slightly higher cost due to my poor planning, I couldn't in good conscience charge more for my mistakes. I set a price, I set a cutoff point for those who would receive it at that price, and I will honor that. I wish to make this into a business, and in doing so hope to instill a sense of confidence in my abilities at design and good customer service. I eventually want people from every car club/board claiming that MacGyver is a stand up businessman and his products are top notch. If I can't reach that goal, then there's no point in me doing it for anything other than the fun of it.

 

Mac,
I am perfectly happy to accept an increase in the cost if you feel the design needs re-working. I think all of us would rather have you fully pleased and confident with the product that you are going to be putting your name on (and we are putting in our cars). If it costs an extra bit, then so be it. We all spent over $30,000 for these cars, so an extra $50 is not (in my opinion) worth splitting hairs over

- Gan

 

i've been quietly watching this post for a while now. one ... cuz the product is awesome. and two ... i've enjoyed hearing about mac's frequent updates. it's refreshing to see someone willing to work so hard for a quality piece of work. not enough of that in the world.

now .... just need to get my S2000 before i come begging to get on the list.

 

I don't know how different this is, from what you are doing, but I just saw some pics of another lighted windscreen.

I'm still behind you all the way, but maybe there is some info to be found from this other project:

https://www.s2ki.com/forums/showthread.php?...threadid=137409

-Pete

 

Can you put me on the list Macgyver.

 

moc, done...



Pete, since cox is using just a couple of separate white and red LEDs, he is able to wire them in a different fashion than my 6 multi-color LEDs (his serial, mine parallel). The advantage to serial is a reduced current consumption (i.e., less heat) with the disadvantage of an increased number of individual LED packages (i.e., more cramped for space). I can go to a different package at an increased cost which would allow me to wire them in this fashion, but it also means redesigning/recutting the boards.

One possibility is to reduce my package count from 6 to 4 (don't know why I didn't think of this before), thereby reducing my current requirements and still allowing me to put in a decent amount of current to each package. I originally wanted 20mA to each color/package, but I had to drop those requirements down to 10mA when the heat issue came up. It was still bright enough, but I (being the anal retentive engineer that I am) wanted more. If I drop down to 4 LEDs, I can drive them with 15mA/color for some extra brightness. With my current windscreen, that puts the end LEDs 6" apart form each other (for those with the S2000 logo etched in, the end LEDs are just past the end of the 'S' and last '0'), so it more than covers the entire design (and would adequately cover designs 8-10" wide, possibly more). The big advantage with this change is no new boards are necessary, the LEDs stay the same (just 4 instead of 6, same cost packages), and a few resistors need a changed value (very simple, I just have to order more).

I just popped 2 LEDs out of the breadboard to prove to myself it would look good, and it does, so I'm going to go forward with production using this method. I may space the 4 LEDs out another 1/4" (currently they're 2" from each other) to get a bit more coverage towards the end, but if the designs don't reach that far out, there's no point in moving them.

Needless to say, project #2 won't suffer from the heat issues (with my luck it will suffer from other issues) since I'm using the more expensive LED packages which allow serial connections. Not to mention the driving method will be vastly different, but I digress...

 

Just like Mac said, I am using 8 5mm LEDs (4 red/ 4 white) wired in series. Therefore I only have 20mA running to each string. This really cuts the power consumption.

Mac - how bright are those 3 mm LEDs? Mine red is almost too bright, white is about perfect. I am using 2500 mcd/45 degree viewing angle for the whites and 3500 mcd/45 degree viewing angle for the reds. And that isn't even bright for 5 mm LEDs (I went for viewing angle over brightness).

Oh yeah, those 5 mm LEDs are a PITA to mount....

 

The brightness ranges from <1000mcd for the red all the way up to >3500mcd for the green, but the viewing angle is about half of yours. To increase the viewing angle, I chop the lens off of the LED (a boring process when you're doing 100s of LEDs, I assure you ). The advantage to my circuit, though, is the ability to control the brightness of the LEDs to anything up to and including the maximum.

 

I could put a pot on mine to control the brightness, but I think running them at 20 mA is about right. Like I said, the red is bright, but like you said about being an anal retentive engineer, more is better.

Are you sure cutting the tips off the LEDs increases viewing angle? I believe most of the viewing angle is determined by the parabolic refector inside of the "case". I talked to some of my optics friends about it and they didn't think cutting the tips off would do anything. Of course, a simple experiment would answer this.....

 

Try it yourself to see what I mean. The multi-color packages I've been working with have a pretty darn tight viewing angle, somewhere between 10-20 degrees. You can hold a piece of paper above them to see the different colored cones...the paper has to be held about 6" or so above the LEDs before the cones really start to converge. Cutting the top off close to the die (about 2mm above the highest die wire) changes that to (in my guesstimation) something well above 100 degrees. Ever seen packages that have an inverse cone cut into them (the tip is a concave cone, not convex like the typical LED)? That design is meant to reflect the light to the sides for good overall visibility from any angle, but tends to be somewhat dim at 45 degree viewing angles.