1
\$\begingroup\$

I'm trying to work out the gain of a transistor, the datasheet is here, although I have extracted the relevant images here:

enter image description here

The datasheet gives the gain for Ic=5mA and Vce=2V as 25, and the remainder of the results seem to indicate something of a bell-curve of the gain plotted against Ic, which with some rough interpolation would allow you to find a very rough and approximate gain for any Ic value in between 5mA and 500mA; although this is all for a Vce=2.

The graph gives a shape expected of the results, but the tabular results don't seem to match the graph (i.e. 150 mA should be 40, but graph value seems closer to 160 - i.e. the maximum). It seems the graph is plotting the maximum gain rather than minimum - doesn't make the graph useless?

enter image description here

The tabular data, while useful is strictly for the stated values of Ic=50-500mA and Vce=2. Obviously, 2V is quite low, so if this was part of a circuit running at Vce=5 and the Ic pulling say 100mA, then how could this datasheet ever be useful, considering those values can't even be found in the datasheet?

\$\endgroup\$

4 Answers 4

4
\$\begingroup\$

As they've not labelled the different curves it looks pretty dubious to me. Normally the graph would apply to typical values so you would not necessarily find them in the tabulated values (this one does not show typical values).

I might guess it's showing you either temperature (highest trace would be highest temperature, lowest trace lowest temperature) or unit to unit distribution for the high gain version, with the middle trace the typical value. For what it's worth, the graph is unchanged from the Motorola BC635 datasheet. In any case, the relevant curve is probably the middle one.

hFE is not very dependent on Vce once Vce is high enough, so typically the gain would be around 150 at 100mA and 5V, but it might be as low as a bit less than 40 or as high as a bit more than 160 (assuming 25°C, and assuming it could be any of those transistor types.

So you should design your circuit so it will function for hFE between (say) 35 and 180 and you'll be fine. If you need higher gain, or tighter beta specifications, other transistors have higher gain at 100mA and in some cases you can specify the beta bin to reduce the range to more like 2:1 than 4:1 (best avoided if you can).

Edit: To clarify where the numbers came from- what I get from the graph (taking the middle curve) is that the gain does not typically change much between 150mA and 100mA, so we can reasonably assume similar limits. Those limits (not the graph) are what is guaranteed for the transistor. Only those. It's dropping pretty fast above 150mA so I would not make the same assumption if you said 200mA. The 40 and 160 (from the tables) are hard guarantees of transistor performance (you can complain to the supplier if the transistor does not perform within those limits). The curves illustrate what usually happens under some conditions and if the transistor does something different, you have no cause to object.

Gain vs. Vce, as you get much above saturation (when Vce >> Vbe, say above a volt or so) the gain hardly changes. 2V is well above 1V, as is 5V, so we can expect the gain to be almost the same. To illustrate this, consider this set of 2N3904 curves:

enter image description here

enter image description here

If the gain stayed constant the lines would be horizontal with voltage- they are not quite horizontal at higher base currents, but they are reasonably horizontal, and (more importantly in your example) the gain is only higher at higher Vce, so we're fine on the low end, but had better add a bit to the maximum gain guarantee.

\$\endgroup\$
1
  • \$\begingroup\$ Do you think you could clarify the last two paragraphs please - I'm not really understanding you. I understand that hFE isn't too dependent on Vce, but what is "high enough". Also, I'm unsure about the "a bit less than 40, or as high as a bit more than 160", where are those figures and how are they found on the graph (assuming that's what you used). \$\endgroup\$
    – R4D4
    Jul 3, 2015 at 16:48
3
\$\begingroup\$

It's a badly written data sheet.

Sometimes you must just give up on a transistor if you can't trust the data sheet. For instance, there are three curves in the graph and they are unknown i.e. On-Semi have failed to give relevant information about what those curves mean or represent.

The graph next to it has several curves and referred to is a transistor called a BC635 yet, this data sheet is for the BC637 and BC639.

Unless there is something great about the transistor that is embodied elsewhere in this data sheet I'd give up and find a BJT with a data sheet I could trust. If you need it resolved because you have (say) just bought a kilo of these parts then start digging around other manufacturer's of this device to see what their data sheets say.

If that fails contact On-Semi but don't hold your breath. It also might be worth looking at the DS of the BC635.

\$\endgroup\$
4
  • \$\begingroup\$ Hi Andy, thanks for the quick response... Feels like I can't win! Now I'm wrestling with incompetent technical writers!!! But no there's nothing special about this transistor. Would you be able to expand your answer, on how I would a) find the gain for Ic values other than those specified, and b) what guidelines should be followed when working with different Vce values, i.e. 2V in datasheet, but 12V in circuit for example? Thank you. \$\endgroup\$
    – R4D4
    Jul 3, 2015 at 16:26
  • \$\begingroup\$ It's bad luck what with this question and the previous one but it's a good learning curve and hopefully you are in helpful company here!!! Once Vce is out of the dirt, hFE pretty much stays constant for a given value of Ic. Sometimes you just have to keep diggging! \$\endgroup\$
    – Andy aka
    Jul 3, 2015 at 16:35
  • 1
    \$\begingroup\$ Maybe try the ZTX651: diodes.com/datasheets/ZTX650.pdf \$\endgroup\$
    – Andy aka
    Jul 3, 2015 at 16:37
  • \$\begingroup\$ That's the best transistor datasheet I've seen! :) \$\endgroup\$
    – R4D4
    Jul 3, 2015 at 16:49
1
\$\begingroup\$

The curves appear to correspond to the min-typ-max of the BC639-16ZLT1 (third entry in the table at 150mA).

Gain for individual BJTs varies wildly. If you need a better number than 100-ish, someone's going to have to test individual components.

\$\endgroup\$
0
\$\begingroup\$

β (beta), the gain or amplification factor of a transistor, normally is given when solving a circuit equation. However, if it is not given, it can be calculated if the currents, Ib (the base current) and either Ie ( the emitter current) or Ic (the collector current) are known.

If Ib and Ic are known, then β can be calculated by the formula:

Beta= Ic/Ib

Beta=(Ie/Ib) -1

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.