(yet more) Automotive design question...

[CJ]

Just copy a schematic from a car with LED instrument lights, if you know which car, and know where to get a diagram.

Anybody need a free schemo to a Sony 27 inch, circa 1990?

Entertaining to watch PhD's re invent the resistor because the end user wants rocket science for a 2 nd grade science fair project.

 

[JohnRoberts]

[quote author="CJ"]Just copy a schematic from a car with ......[/quote]

You sound like management...

Good engineers hate to copy anything even when they should. I managed an engineering group where we had almost 40 years of our own schematics to copy. The senior guys were OK with copying, because they were often copying their own prior work. On their own stuff they always had tweaks they wanted to make but couldn't. When copying somebody else's work they were highly motivated to find flaws and things to improve, so you benefit from a mature design fine tuned by a critical eye.

I'd let the junior guys take a shot on original design before I told them what they did wrong and gave them an old design to study. I rarely made somebody use a circuit they didn't like, but they had to make a strong argument for doing it their way. Time (to market) is money.

JR

PS: the correct answer is what the customer wants....not us.

 

[Guest]

My diploma project was a microwave alarm system. It was a laboratory of thick film ICs, so tantalums like modern SMD electrolytics were not a problem. The management gave me "prototypes", some from Metrovox, some from Italian company I forgot the name. The best of both had 22 electrolytics, three opamps, and many transistors..

Both looked so ugly, like made of parts taken from schoolbooks, so I decided to design mine. As the result, I had only few transistors, one OpAmp, and only 2 electrolytic capacitors, absolutely necessary: one to define a lower limit of captured Doppler frequency, and another one to set a delay time for short targets like birds, mice, etc... The IC was a cover for a waveguide head, and a head was a heatsink for the IC.

...it is always interesting to look at schematics for some original solutions, but training own brain is more profitable...

 

[CJ]

Use the best design possible, so that means exploring all the possibles, so yes, looking at old schematics is a great start, as long as you know what to improve, and what not to improve.
sometimes resistor values they pick because they were already in stock, so you can make improvements due to the economical nature of assembly lines.

I figure if BMW or VW has a circuit, you could "upgrade" to that, as long as its a good circuit.

The first guy does not have a chance, thats what we are trying to pull off by "borrowing" and hopefully "evolving", the capitalist way of our materialistic lives. Thats how we beat the Russians, stimulate development.

Did you see that guy who hid his radiation badge so he could get hazzard pay at Chenorbyl?
Sheesh! He went in four times, everybody is dead but him, looks like a drinker, heck, I would if I took in 3 million RADS or whateve4r.
He probly does not need a flashlight to walk home.

Yes, end user is boss, it is much better to drive at night with the dash off, just that little amount of light can lower yuour night vision quite a bit, usually on long freeway trips, that is.

Noted in the AAA manual for driving tips.
Page 17.

 

[mikep]

[quote author="CJ"]Just copy a schematic from a car with LED instrument lights[/quote]

maybe not. dont assume those circuit designers know what they are doing. this is a little off topic, but lately ive come to really hate LED tail lights. Ive noticed that a couple of newer car models have horrid flickering LED tail/brake lights, most notably: nissan murano SUV and multiple Cadillac models. they must be flashing less than 60-70Hz to be so noticable. they are distracting and I think unsafe. it makes no sense that they would set them up to flash so slow. with a $1 microcontroller you could set the flash rate as high as a few HUNDRED kHz, if you wanted.

IMO you should drive those dash LEDs with a simple one-transistor current source. use the 50 Ohm rheostat in series with maybe another 50 ohms in the emitter of said transistor. experiment with the base voltage to get it in the "right" range. the base voltage bias is also a convienent place to put a photocell to automate the intensity, or provide remote cutoff, etc.

mike

 

[JohnRoberts]

I'm sorry officer... I didn't notice he was slowing down because I was distracted by his flashing brake lights....

Methinks that's the point.

JR

 

[SSLtech]

Ah, but what a few people may be missing is that there's MORE than just the instrumentation lighting being dimmed.

There's also the climate control, and a lot of back-illuminated (incandescent-bulbed) switches for auxiliary lighting, rear refroster, central locking, rear wash/wipe, mirror adjustment etc.

All this is connected from the same dimmer feed as the instrument lighting. PWM should work, as would a variable voltage follower, but a variable current source might not be a universal solution... specially since some people might want to build one and drive more LEDs and fewer incandescents, or vice-versa.

Keith

 

[mikep]

I didnt realize that. So why dont you leave the dimmer circuit unmodified and make a little "glue" circuit to interface to the LEDs? Is that what you were talking about the whole time?

the existing circuit then will exhibit somewhat changing voltage at the output of the rheostat (still loaded by the other bulbs). the reason it stays bright may be that the changing voltage is still always a good bit greater than the LED foreward voltage. it is the current that sets the brightness, set largely by the series resistance, which is the rheostat + whatever series R you put in to keep the LEDs from blowing up. that series R must bes alot bigger than the rheostat, hence adjusting the rheostat doesnt make much difference. you want a circuit that senses the changing voltage and puts out a changing current. the trick is going to be getting it to work so that the LEDs and filament bulbs appear the same brightness at different dim levels. even replacing the whole dimer arrangement with a PWM approach wont get you that automatically.

mike p

 

[CJ]

You want to engineer it so you can just snip a coupler like they use for outdoor lighting fixtures over the wire, so you do not have to re crimp any Molex connectors under the dash.

I hate to say this but the best solution would be to rotary switch in one LED at a time, that way you simplify things, the car does not catch fire from a 555 chip shorting out due to distributor noise on the line, there is zero waste due to heating a regulator, zero circuits to fire a FET, no room under there for a DIY circuit box, and it will not change with time.

Voltage regulation on cars is screwy anyway, so anything you build is gonna move around a bit, as the battery ages, gets cold, etc.

Of course I recommend vacuum technology would be the coolest, that cool blue glow like the old Fluke 8088a, or on the window display, or a led flashlight for the speed zones.

Capacitence
Resistance
input transformer
output transformer
Inductance

oh cool, new boilmaker program, you type "C", it types Capacitence

etc

http://www.phraseexpress.com/

:thumb:

Back on toipic, check this:

http://www.mikesflightdeck.com/led_dimmer.htm

 

[Guest]

Consensus: put 510 Ohm resistor in series with each LED to be happy. Or more, if LEDs are too bright.

 

[CJ]

how many audio geeks does it take to screw in a lite bulb?

Keef, you should go Green, put a solar panel on the roof for nite lighting.

:razz:

 

[PRR]

> what a few people may be missing is that there's MORE than just the instrumentation lighting being dimmed.

Then why even ASK?

And why was this not in the Specifications For Design? (Curiously missing in this thread.)

If there is significant hungry filament, hang R+LED on the filament feed.

IF the goal is to LOSE the filament waste (couple watts on a system which cruises at 14,000 Watts and peaks over 70,000 Watts) and HEAT, yet keep the legacy rheostat, then do the diff-amp against a bridge with the rheostat as one leg. This can even work with some inky lamp in the system, if you find a fat-output amp.

Just working a voltage on a 2N3055 Base is horridly non-linear. The post far up the thread with 1K in series with Base "does" work, sorta, by assuming that 2N3055 Beta is fairly constant (but now it is very sensitive to load current).

 

[SSLtech]

[quote author="PRR"]And why was this not in the Specifications For Design? (Curiously missing in this thread.)[/quote]
Third post in the thread on page one:
[quote author="keef"]Can't use CCS because of mixed LED/bulb combinations... [/quote]

[quote author="PRR"]If there is significant hungry filament, hang R+LED on the filament feed. [/quote]
This is how it's done now... but the 20 Watts or so of filament waste reduction thus far has drastically reduced the dimming range of the rheostat for both the LEDs and the remaining incandescent bulbs... hence the idea to hang a 10Ω/25Watt resistor off the rheostat, although that's putting a concentrated source of heat back inside the dashboard...

Keith

 

[Guest]

[quote author="SSLtech"]
[quote author="PRR"]If there is significant hungry filament, hang R+LED on the filament feed. [/quote]
This is how it's done now... but the 20 Watts or so of filament waste reduction thus far has drastically reduced the dimming range of the rheostat for both the LEDs and the remaining incandescent bulbs... hence the idea to hang a 10Ω/25Watt resistor off the rheostat, although that's putting a concentrated source of heat back inside the dashboard...
[/quote]

So, back to PWM...

 

[CJ]

Tungsten filaments, do they round off the sq wave 555/FET signal?

If the LED's do not, does square wave pulses harm the PN junction?

Just a side thought not related to the task at hand...

You can pulse tunsten filaments also, not as efficient due to cool off/warm up cycle, but thats how I did it on my mountain bike dimmer that I used to sneak past the rangers while doing ilegal single trak at nite...

heres a jpg of a 555 dimmer.
you could try a antilog pot to see if it makes the LED dimmer more linear:

 

[JohnRoberts]

LEDs are routinely driven from square waves when multiplexed so that's is not an issue. The inductance of lamps will not care about edge rates either.

JR

 

[CJ]

OK.
Hey, why not try the old Cougar sequencing tail light circuit?

I have the microfilm for the wiring diagram...

You could sequence different colored lamps in the dash while driving, blinking red, green blu, how cool would that be?

The cops might not like it, but if you were single, looking to get some action...

Then just pay a hooker. Its a lot less stress, believe me.

Jus kiddin

:grin:

 

[bcarso]

Apropos sort of, based on some of the comments on withstanding load dump in the automotive environment:

http://www.powermanagementdesignline.com/showArticle.jhtml?articleID=204804092