> I though it'd be entertaining to make the taillights on
> my Caprice light sequentially.

> ... I guess that using a
> PWM-like "turn them on and off real fast" thing will work fine for
> making the lights appear to fade in and out.  The bright side of an
> 1157 should be on the order of 25-30 watts, which means I have to
> switch about 2.5 amps at 12 volts.  I need individual bulb control,
> and there are 6 or 8 bulbs, depending on how ambitiously I try to 
> kill
> my time.

1.    A FET switch is liable to be MUCH easier to implement. Bipolar 
takes significant drive for significant load. Even using Darlingtons 
you tend to have to deal with this.

A FET with Rdson of say 100 milliohm (very poor spec) achieves an 
equivalent bipolar Vsat of 0.25 volts at 2.5 A which is hard to 
achieve with bipolar.

PIC etc will drive a logic FET gate directly.

Even if you use a P channel FET so as to be able to switch the high 
side the drive circuit is still modest and simple.

An N Channel FET can be used as a high side driver if you are willing 
to provide an "above top rail" gate drive supply. using a P Channel 
FET is easier.

2.    Be aware that cold bulb filaments have inrush currents many many 
times that of operating current - typically around 10 times operating 
or around 25A in your case. As you'd probably be using a TO220 FET or 
SMD equivalent anything you'd sensibly use should handle this OK.

Here's a simple but useful primer on driving incandescent bulbs

        http://www.allegromicro.com/techpub2/an/an295012.pdf

Adding either a "warming resistor" or a series limiting resistor will 
increase lamp lifetime and give the driver an easier job. Both methods 
have good and bad points.

__________

An Agilent lab lesson here with some impressive current/voltage/time 
curves.

    http://www.educatorscorner.com/media/Exp17i.pdf
    From    http://www.educatorscorner.com/index.cgi?CONTENT_ID=2188


3.    For PWM above about 1 kHz you are not liable to annoy the bulb 
too much as it's thermal time constants *should* be rather longer than 
this. A photodetector and scope would prove this. Note that a bulb run 
from AC experiences 100 or 120 Hz half sinusoid cycles constantly. 
Anything substantially faster than that, even with a square wave form 
should be OK.

I am currently PWMing halogen bulbs at 1 kHz plus at present for a 
non-automotive application but don't have any lifetime data yet.







        Russell McMahon 

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist