The length of time a transistor can be on

[Raymond31415]

Is there a limit to how long a transistor, like in a cpu, can be constantly on?

 

[Phrak]

The time limit for a transistor to be turned on is determined by the product for the voltage drop from collector to emitter (or drain to source) times the collector (drain) current.

This is in units of power. If it is powered on for over this time interval, the device may be fried or degraded. The scenario I've given is for a single turn-on pulse called the non-repetitive peak power.

 

[Raymond31415]

I see how that can be used to calculate a power unit but how would that translate into real time, eg: ms,s or minutes.

 

[pantaz]

If the transistor is operating below peak, and power dissipation is held within spec, then it should last a very long time.

There is a http://www.google.com/search?q=free+mtbf+calculator" for component life estimates.

 

[Phrak]

I see how that can be used to calculate a power unit but how would that translate into real time, eg: ms,s or minutes.

It would be useful to know what application you had in mind.

However, bipolars and FETs have a safe operating area of operation (SOA), limited by maximum rated voltage and maximum rated current. In addition bipolar transitions are limited by a maximum power rating, whereas FETs are limited by R_DSon, drain to source resistance with gate turned full on.

As for your application--if you have one--if your on time is in excess of the order of a 10 millisecond, depending on the size of the transistor, this is nearly equivalent to continuously turned on, so it would be sufficient to see you are within the SOA of your transistor.

I would point you to example plots on data sheets but my adobe reader is crashed and refused to be de-crashed.

 

[Raymond31415]

I was planning on using a transistor on for periods of hours at a time, or a day at the most, operating at the lowest voltage and current possible, but I haven't been able to find anything on the internet about operating tolerances for npn and pnp transistors.

 

[Phrak]

I was planning on using a transistor on for periods of hours at a time, or a day at the most, operating at the lowest voltage and current possible, but I haven't been able to find anything on the internet about operating tolerances for npn and pnp transistors.

OK, then you just care about operating within the SOA 100% duty area. Are you operating in the linear region, or are you going to turn the transistor full-on like a switch? What load are you driving and what is your supply voltage?

 

[Raymond31415]

I only need a transistor to turn on another transistor, but at times it maybe turned on and off, at other times it may be left on for hours if possible, I can adjust the load resistor and source voltage to what ever is necessary to make it possible.

 

[Phrak]

Several people on this forum that could find you a transistor and then describe the specifications and charts that dictate it's selection. If you are only using your transistor to drive another transistor, it's all the simpler.

How do you intend to purchase it. There's radio shack and Digikey...

 

[Raymond31415]

I am talking about the type of transistor like in the cpu of a computer, the smallest technology is able to produce today, I wanted to design a circuit but I wasn't able to find anything on the internet about how long one could be continuously, on operating at the lowest voltage and current and still function, if it could stay on for a few hours that would be good, indefinitely would be ideal.

 

[Phrak]

OK. You got me. I have no idea what you are up to. Is this a theoretical question about minimal operation, or are you building something.

 

[Raymond31415]

Well, it's a theoretical question, but if the smallest transistor can stay on for a few hours or indefinitely, I will go forward and have the circuit built.

 

[Phrak]

Your question has evolved into a very different regime. You are questioning the maximum sustainable power for a transistor to drive another transistor in a switching mode, turning on and off in the order of gigahertz, such as the microprocessor your computer uses.

This is beyond the domain of my profession. However, chip makers try to make transistors operate at smaller voltages with smaller capacitance, and therefore less current hungry, so they can operate at less power without overheating.

 

[Skiddaw]

If this is a general question then I believe the general answer is "there is no limit to turn-on time" as long as the power/temperature parameters are kept within specification.

 

[Raymond31415]

thanks everyone for your answers.

 

[Phrak]

If this is a general question then I believe the general answer is "there is no limit to turn-on time" as long as the power/temperature parameters are kept within specification.

no, on time, not turn-on time.

 

[rcgldr]

Getting back to the cpu example, imagine a simple tiny program that loads a value into a register then either loops forever or halts. Somewhere in the cpu are a set of latches that are holding that value in the register for as long as the cpu is running, which could be years if that was the goal.