Is this transistor datasheet correct?

Question

I was about to post a question asking for a schematic critique, although while checking things I've noticed a few things on the transistor datasheet that strike me as being a little strange. The datasheet in question is for an NPN, TO-92. The second page states the electrical characteristics, and they're making things a little harder to understand how transistors actually work.

The Vceo and Vces are marked down as minimum values, which would seem to imply that there must be a minimum voltage between the collector and emitter of 45V to 50V depending on the total current passing through the collector. Should these two values be marked as maximum rather than minimum?

Similarly, the Emitter Base Voltage, as I understand is the maximum voltage that can exist between the emitter and base, assuming that the emitter ended up with a positive voltage in reference to the base voltage. So again, should this be marked as maximum rather than minimum?

Finally, the DC Current Gain has a minimum of 100, but this appears for an Ic of 100mA. Am I able to assume that I'll always have a current gain of at least 100, irrelevant of Ic? If not, how am I to know the minimum current gain if I'm not passing in 100mA?

EDIT: After looking at a different transistor, I've found the 2N5551 from Fairchild Semiconductors, and the respective datasheet. It's far easier to understand, as the parameter names are far less ambiguous. For example Collector-Emitter Breakdown Voltage etc, the keyword being breakdown in the name. Also, they also provide a spice model and plenty of graphs. Seems there's a large difference in the quality of datasheets.

Answer 1

It's saying that the transistor can be operated up to 45V. This has to be specified as a minimum value because if it was specified as a maximum value then you wouldn't know the lower limit that might cause it to fail.

It's a guarantee on performance - if you bought a car that was guaranteed to do 150mph you'd want the MINIMUM guarantee to be 150mph. It can't be expressed any other way.

On the other hand if the car manufacturer specified acceleration from 0 to 60mph as 4 seconds, you'd want that to be a maximum value i.e. you can always guarantee to do 0 to 60 in 4 seconds (max). Think about it.

Regarding hFE, use a better transistor that has a full spec. The spec you linked doesn't even have a part number other than the generic (but suspicious) name "T0-92". Here's what the current gain for the BC847A transistor looks like in its data sheet: -

This is the sort of data I'd expect to see in a data sheet.

Answer 2

Usually the columns min/typical/max on datasheet refer to the spread of the mentioned parameter among a manufacturing batch. This means that, for example, Vceo is guaranteed to be at least 45V for any specimen in the batch (at the given conditions).

In other words, if you buy 100 such BJTs and determine their Vceo you'll discover that each one has a different value, but all the values will be at least 45V. Could they also list the maximum value? Sure! But that won't be useful in designing a product: who cares if one out of 100 BJTs has a Vceo of 60V? The designer need to know what's the guaranteed minimum value so that its design doesn't break when a random sample is placed on the PCB!

The confusion stems from the use of the words "minimum/maximum". Here you use it twice, with different meaning: Vceo as reported is the minimum maximum Vce value, i.e. the minimum value (among the batch) that you can get for the maximum allowable Vce voltage.

On the other hand, Icbo is reported as a maximum value, because that is the maximum guaranteed value among any sample in the batch. Why not reporting the minimum? Again, the relevant parameter for reliable design here is the maximum: Icbo is a leakage current, ideally you would want it to be zero! So when you design a product the worst case scenario is when Icbo is big, hence you need a guaranteed maximum value.

As for hFE, in most design it is useful to have a guaranteed minimum value, because most of the time you use negative feedback circuits which work well as long as the gain is at least a specified value. Here they report also the typical value, since this gives you an idea of the parameter spread among the batch. In the specific case you have min 100 and typical 400, so you could guess that for 100 samples of that BJT most of them will have an hFE around 400, with a minimum of 100 guaranteed and, by symmetry probably you won't have samples with hFE greater than 700=400+(400-100).

Answer 3

Imagine you (using a curve tracer) increase the collector voltage (with zero base current) until the collector current is 1mA. That voltage will be a minimum of 45V. Similarly for Vces, collector current of 0.1mA and emitter current of zero.

And again, increase the negative voltage on the base until you see 10uA. That won't occur until a minimum of 5V (as guaranteed by the manufacturer).

As far as current gain goes- the short answer is that you don't know for sure. You cannot assume the gain is >100 for other currents. You can look at the typical graph of gain vs. current and make some reasonable assumptions of what the minimum gain is for other currents, but there is no guarantee (so make conservative assumptions if you want a reliable design). If you are at a bit less than 100mA usually the gain will be similar. At 50nA or at 500mA you're likely to see quite different behavior.

Answer 4

The Vceo and Vces [and Emitter Base Voltage] are marked down as minimum values ...

Yes, and if the voltages go above those then the transistor is likely to be destroyed. Those are therefore the minimum dangerous voltages.

If not, how am I to know the minimum current gain if I'm not passing in 100mA?

Look at the charts later in the datasheet. One will undoubtedly relate gain to collector current.