# In your experience, how well do NPN SPICE models replicate real DC current gain

Title, I'm currently running some simulations and noticed that my DC current is in the very rough ballpark of datasheet parameters. Is this normal or does it indicate there may be something wrong with the circuit itself. This is not a question on any particular circuit I am just trying to expand my knowledge on simulations vs. reality.

• What transistor? If you identify it and post the SPICE model it would be useful. – Peter Smith Jul 31 at 14:25
• Say what.. the simulation matches the datasheet and therefore you think something is wrong? Or is this some alternate-universe meaning of "in the right/same ballpark"? – pipe Jul 31 at 14:30
• @pipe I mean the DC current gains are wrong for the biasing but never exceed or dip below the max and min expected. Like at X bias it'a 300 and at Y bias it's 30 but they aren't really correct at the bias points I set it to. But never leave that range. – Jaywalk Jul 31 at 14:32

## 1 Answer

Look at the datasheet of a typical transistor, for example the P2N2222A

Notice look at the specified value of the DC current gain.

See how only minimum values are specified. Only for Ic = 150 mA, Vce = 10 V is there a typical value specified. For that condition the DC current gain is guaranteed to be higher than 100 and is expected to be typically around 300.

That's a factor 3 change !

This is "normal" for most transistors. The DC current gain is largely determined by Base thickness and Base / Emitter doping concentrations. These are all somewhat difficult to control in manufacturing!

Only if the manufacturer would measure each transistor and sell it as having a DC current gain being within a certain range could you get a more predictable value. This "binning" isn't usually done as it is expensive resulting in expensive components.

Furthermore: it is simply not needed !

It is much easier to just design the circuit such that it can support even the minimum DC current gain. If the transistor then has a higher gain the circuit design would be such that this will not change anything.

I'll be blunt: a circuit which relies on one or more transistors to have a specific (a value within a narrow range) DC current gain isn't a good design. Almost all common circuits only need a DC current gain above a certain value, as long as the transistors meet that requirement, the actual value of the DC current gain is irrelevant.

Most transistor models just model the typical value of the DC current gain. There might be an option in your model and/or simulator to simulate the extreme ("corner") values as well. But again, if your ciruit, even in the simulator is dependent on a transistor's DC current gain, you should consider changing the design such that it does not rely on this transistor property.

Example: you could amplify a current using an NPN, then the amplification factor will change "all over the place". A current mirror using many transistors in parallel (with emitter resistors) will be much better. Next step better: an opamp based circuit using resistors to do accurate current amplification.

• I see thank you, this is the sort of confirmation I was looking for. This particular situation wasn't critical for any design just something I've noticed and I've been told time and time again to be wary of simulation results as they don't always reflect reality. I was just looking for some insight into how I should interpret the datasheet. – Jaywalk Jul 31 at 14:40
• @Bimpelrekkie Well said. – analogsystemsrf Jul 31 at 14:42
• Consumer electronics often uses binned transistors, to save a bit of money. For example 2SC1815Y where the Y indicates hFE of 120-240 ('only' a 2:1 range). They have good enough control over the parameters that there is not a significant price premium for that level of selection. – Spehro Pefhany Jul 31 at 15:32