NPN BE diodes in parallel

[treez]

Hello

Is it OK to put NPN BE diodes in parallel as follows..


[mod:external link replaced with the image, please attach images in the post rather than using external servers]

...i mean, even if the BJTs were the same, there may be batch differences in BE voltage, and at high temperatures, one BJT may be on, and may disallow the other one to turn ON? (due to one not having enough BE voltage)?

 

[keith1200rs]

Definitely not a good idea. The only time it would be acceptable is with matched devices on the same piece of silicon.

Keith

 

[treez]

but i thought the Vbe ON/OFF voltage is not a sudden ON/OFF..but gradual , especially when the base current is low, as in the schematic?
So surely it will be OK?

 

[betwixt]

Yes it is gradual but unless they are perfectly matched, the one with lowest turn-on voltage will conduct more base current and prevent the other one from conducting as much. Remember that the BE junction behaves like a diode and clamps the voltage at Vf, the transistor conducting most will clamp the voltage and prevent full conduction in the other.

Don't confuse the gradual increase of collector current as the base current increases with the BE voltage. In a linear condition you control the base current to set the collector current, if you simply apply voltage between base and emitter it will sink as much current as it can.

Brian.

 

[keith1200rs]

True, but with two different transistors or even two samples of the same device, you can find one will be turned on hard at say 0.6V and the other only manages a few microamps of collector current. The datasheet may not give enough detail to determine how bad it will be. A resistor in each base solves the problem. Depending on the collector current you need, a small emitter resistor can be used.

Keith

- - - Updated - - -

Brian beat me to it while I was typing (slowly)

 

[FvM]

Essentially, the thread misses a clear question

Is it OK to put NPN BE diodes in parallel as follows..

Means exactly what?
-OK to achieve collector currents within a certain tolerance?
-Does at least guarantee both transistors to be turned on?
-Should even work with transistors of very different area?

It turns out, that typically both transistors are saturated with a "forced Beta" of about 2. It will work even with considerable temperature differences and some type variantions. But good design tries not to rely on typical parameters or clear weather conditions.

 

[treez]

It turns out, that typically both transistors are saturated with a "forced Beta" of about 2

..that's also what i was thinking....that the NPN's would each be pretty well saturated anyhow, so it doesnt really matter.

Certainly, on our 20 prototype pcb's its not been a problem, but i think it should be changed before going into full production.

 

[FvM]

Is it a problem of saving one resistor?

 

[treez]

no the problem is that i implemeted the one npn, then later the other, then realised that the npn vbe diodes are in parallel....if i'd have seen it straight away i would never have done it.

The circuit isnt quite liek the above, -there are two base resistors, (also in parallel) and again this being an accident because i was in such a rush