transistor switch question

[Othello]

I have to make a magnetically switched 2 color LED, using a reed switch.
To increase the life of the reed switch I used a transistor to reduce the current thru the switch from the 25mA or so (up to 55 mA) to a few mA or so.

See circuit.

Since the circuit will have to work from freezing to hot day temps I am worried that my biasing scheme is not the best, but I can't figure out the values for a better (I suppose) bias with an emitter resistor, which maintains my low supply voltage and minimal parts count.
Any help?

 

[Reloadron]

I don't get it? A simple magnetic reed switch like this one can handle about 500 mA way more than enough for your LEDs and well exceeding the 2N3904 max current. I have used many reed switches that will handle an amp. Why bother with the transistor? Do you have the data sheet on the reed switch you have? A Google of HSI Sensing will bring up countless flavors of switches including ones like this that can switch 2 amps and handle over 5 amps.

Ron

 

[BrownOut]

You don't really need a emitter resistor, and in fact, you don't want one. At 2.3V, your power supply is just barely enough to light your led's. And your LED's are backwards, and only 1 has a current limiting resistor.

 

[Othello]

My initial answer was just eaten by the "cloud".

Sorry about the inverted LED. I use no limiting resistor, this is why the power supply is set soo low. The 39 ohm resistor is for color balancing.

The reed switch can easily switch 30 mA, but it lives longer if it only switches 2mA. Longevity is requirement in this case.

But what about the stability of the circuit, will it work from 30 to 80F, with varying beta of the transistor etc.??

 

[BrownOut]

the transistor is configured as a switch, and not an amplifier, so beta isn't a factor. It's better to use a series resistor for LED current limiting instead of relying totally on Vf, because that's a process and temperature variable, and device current is exponentially related to it.

 

[Othello]

I would like to understand your concern.
The power supply voltage is fixed. And less than the max. forward voltage of the LED.
The forward current of the diode decreases with temperature.

What do you think could happen?

 

[BrownOut]

If you want to blow up your LED's then connect them any way you want. If you want a circuit that is reliable and performs well, then always limit the current with a resistor. As I stated, the forward voltage is a process and temperature variable, don't depend on it.

And forward current goes up with temperature, not down. Ever hear of thermal runaway?

 

[Reloadron]

The reed switch can easily switch 30 mA, but it lives longer if it only switches 2mA. Longevity is requirement in this case.

While I can understand your concerns I would suggest you read this data sheet on a typical Coto reed switch.

Note this part:

Loaded conditions (resistive load: 5 V; 100 mA; operating frequency: 125 Hz)
Life expectancy: min. 2 x 10^7operations with a failure rate of less than 10-8 with a confidence level of 90%.

That is about 10 million cycles of operation at 5 volts and a 100 mA load. Reast assured that switch is not going to care about a 30 mA verse 2 or 5 mA load in the least. Years ago I was part of testing these switches with a 50 mA load. Their life cycle expectancy was one billion cycles (10^9).

Testing involved placing 250 switches at a time on a tray. we used 4 trays (1,000 switches) per load placed in an environmental chamber. We di a bake at 250 Degrees F for 24 hours then a quick pull down to zero degrees F. for another 24 hours. That was over 10 million cycles of operation at two temperature extremes. Average failure rate was about 3 switches per 1,000 and two of those 3 were generally technicians bending the leads and cracking the glass envelope when loading the test boards.

Now if you want to believe the added transistor will somehow increase the mean time between failures of the reed switch and discount reed switch data sheets then go for it. Just my opinion if anything you will see a transistor fail long before a naked switch would.

Just My Take
Ron

 

[Othello]

Reloadron,

I am sure you have much more experience in testing and evaluating these switches than I do, my opinion was guided by Hamlin, who makes our switch.
The tested the model I intend to use and since we aimed at 250 000 000 cycles they determined that at 2mA we would be fine.
And since I don't have their test set up, that's all I can go by.

But you are right, it would be simpler without a transistor, so I will study your material carefully, maybe there is a way...

To Brownout

You seem to get annoyed with me. My question was not about civil liberties (can I blow up my circuit? Yes I can!!), my question was trying to understand your concern. And getting assistance.
I have heard of thermal runaway, but does it apply here?
Maybe I don't interpret the data sheet properly.

 

[Reloadron]

OK, then since you are using Hamlin switches and acting on their recommendation I would run with their suggestion. I was not aware you already had selected a specific switch or the ratings of the switch. That said, absolutely run with their suggestion. They built it and they know their product better than I do. As to life cycle testing and general testing of reed switches that was an experience unto itself.

Ron