Dimming for add-on gauges.

It may help you to know that the Miata dash dimmer works by varying the duty cycle of the *ground* to each bulb--the "dimmer" is actually an oscillator, not a potentiometer. All bulbs are always hot on the +12V side.

Unless your aftermarket gauges have dedicated ground wires that do not ground anything but the bulbs, you'll need to add a separate pot to dim them by reducing the hot-side voltage.

Lanny Chambers wrote on 7/28/2004 13:21:

Or add some electronics, e.g., an inverter with a capacitor. Basically, just a transistor, and a couple of resistors and capacitors. But you gotta know about electronics a bit, of course. I've done that to dim my aftermarket radio/cd player. Works fine.

-Joe

I know. I'm trying confirm which of the lines on the stereo connector is the active-low PWM line, and which is the switched

+12V line.

My other sources indicate that they're not. The +12 side is controlled by the "TNS" relay connected to the headlight switch (it's what turns on the dash lights when the parking lights or headlights are on). It's the labelling of "Pulsed Dimmer (TNS+)" that is confusing. There should be one switched +12V line and one active-low PWM line. The pinout I found has two pins labelled "dimmer" or "dimming":

"Pulsed Dimmer (TNS+)" and "Inverted Dimming"

I'm trying to confirm that the former is the +12V line controlled by the headlight switch and the latter is the active-low PWM line.

They do.

Ick. If I did have that problem, I'd put an NPN transister on the high-side of the bulb, and connect the base of the transistor through a resistor to the active-low PWM line.

Why is a capacitor needed? Here's what I would have tried:

+12V | | ___ |/ PWM --vvv--| PNP |\ | Bulb | | gnd

I think I said NPN in my previous post, but it should be a PNP

Right, that's what I meant to say. Sorry.

Then you can simply parallel each wire off an existing instrument illumination bulb, no?

You engineers...always aiming for "elegant", when "easy" would get you back onto the road quicker. Elegant is for winter projects, man! :-)

Grant Edwards wrote on 7/28/2004 14:44:

To make the output smoother. The input is a square wave, and I wanted not just invert the input, but flatten it. A plain inverter would work in your case because of the slow reaction times of light bulbs. With my radio illumination all LED-based, I didn't want the LEDs flickering, though... I don't have an oscilloscope anymore, so I went more with whatever I remembered from my college electronics class.

-Joe

I assumed that the frequency of the PWM was above that which was visible -- the same way you don't notice fluorescent lights flickering at 100Hz or 120Hz.

Wow... I hadn't realized the U.S. had upgraded their electric supply system. Here in Canada, we're still at 60 Hz.

Pete

94BRG

-- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system

6.0.719 / Virus Database: 475 - Release Date: 7/12/2004

Heck, in Washington, DC, we're so far behind the times that we're still at 60 cps.

-- Larry

Maybe fluorescent lights light at both the positive and negative voltages?

Leon

Exactly. A sinusoid such as AC mains voltage has two "offs" per cycle: the postive-going zero crossing and the negative-going zero crossing.

Perhaps it's the educational system in Canada that needs upgrading? ;)

perhaps... but we need to see the whole cycle, not a half cycle.

Pete

On 2004-08-05, Pete Breemhaar wrote:

Each cycle of 60Hz AC power has two "off" points at the zero-crossings and two "on" peaks (one positive one negative). That's an on/off rate of 120 times per second. Light output is a function of the absolute value of the voltage -- the light doesn't go to "negative brightness" during the negative half of the AC mains sine wave.

Here's the voltage plot:

1 ++--------+--******-+---------+---------+---------+---------+--------++ + +** *** + + + sin(x) ****** + .8 ++ *** ** ++ | * ** * .6 ++ ** * *+ | ** * **| .4 ++ * ** ** ++ | ** * * | .2 ++* ** ** ++ |* * * | 0 *+ ** * ++ | * * | .2 ++ ** * ++ | * ** | .4 ++ ** * ++ | * ** | .6 ++ ** ** ++ | ** ** | .8 ++ ** ** ++ + + + + + *** +** + +

-1 ++--------+---------+---------+---------+----******---------+--------++ 0 1 2 3 4 5 6 7

And here's the light-output plot:

1 ++--------+---*****-+---------+---------+----*****+---------+--------++ + + ** **+ + + ** **abs(sin(x)) ****** + .9 ++ * ** ** * ++ | ** * * ** | .8 ++ * * ** * ++ | * * * * | .7 ++ * * * * ++ | * * * * * .6 ++ * * * * *+ | * * * * *| .5 ++ * * * * *++ | * * * * * | .4 ++ * * * * * ++ | * * * * * | .3 ++ * * * * * ++ | * * * * * | .2 ++* * * * * ++ |* * * * * | .1 +* * * * * ++ +* + + +** + + + * * + 0 *+--------+---------+---------+---------+---------+---------+--*-----++ 0 1 2 3 4 5 6 7

Notice that the waveform of the latter has twice the frequency of the first? If the first one has a frequency of 60Hz, then the second one has a fundamental frequency of 120Hz (there are some higher order harmonics due to the fact that it's not a sine wave any longer). I'll spare you the FFT plots of both...

Grant, I thought that going to "negative brightness" is how the "RiceBoys" got the lights to glow under their cars. Gee, I guess I was wrong again ! :-)

Bruce RED '91

aw, come on... quit joking around. I was trying to have a serious discussion here !

Pete

:)

It is pretty cool that gnuplot can still generate ascii-art plots, eh?

"aw...come on..." ??? A serious discussion to impress who ? Yah see, we all know that some of us have Vast information to pass on to others here. It is far out of the way to teach school here on a technical subject. If you can't take a joke about things like this, just maybe off line is a good place to impress the person that you are schooling. "aw...come on, lighten up." :-)

Bruce RED '91