BJT DC current Gain contradiction

Hello,

I am was reading the datasheet for MMBT3904L, I will be using it to drive a 10mA LED. I wanted to find out what is hFE at IC=10mA.
I was confused by the information presented in the datasheet in 2 places. The first one is the on characteristics that shows hFE( large signal DC gain) as a function of collector current that is 40 for collector current 0.1mA to 100 at 10mA

However on another page of the datasheet they show that hFE( large signal DC gain) is constant from 0.1mA to 40mA at 25C.


I got a bit confused, as to why hFE is presented differently in two places? or is it?

The first chart is a minimum specification over a large number of units. The graph is a typical characteristic.

Thank you Analog. Would you go by the minimum spec or by the typical characteristics graph when calculating the base current?

rocky79 said:

.................... Would you go by the minimum spec or by the typical characteristics graph when calculating the base current?

Click to expand...

If you are using the transistor as a switch then neither applies, since those values are for operation in the active region (Vce = 1V), not for operation as a saturated switch. To fully turn on the transistor an hFE (Beta) value of 10 is typically used [look in the ON CHARACTERISTICS table at the conditions for the Vce(sat) measurements].

Checkout the 10mA curve on Figure 16 of the data sheet. To insure the transistor stays fully saturated when "on", the circuit should be well into the flat region of the curve. crutschow recommends 1.0 mA which is at the end of the curve and a good conservative design. I would be comfortable with 0.5mA in this application. In any case, the idea is to have a circuit which is insensitive to the exact value of hFE and the value of whatever signal voltage is turning the transistor circuit on and off.

Thank you Analog and crutschow.
I looked at another BJT like this 2PD602AQL,215 from NXP it's the cheapest general purpose BJT@ 5 cents a piece. The information about the base current vs VCE is not available, can you tell what the base current should be for a collector current at 10mA? ( using it as a switch-Saturation)

I was surprised that anything was much cheaper than the MMBT3904L. I found at Digikey/USA the MMBT3904L is about 1/2 the price of the NXP part in 3,000/reel quantity (~$0.02 vs ~$0.05). Even less in 10,000/reel (~$0.015). One of the attractions of the 3904 is that it is well specified and a workhorse part. Tough to answer your question about base current. You would be best to go with the 1:10 rule for NXP part.

As a general rule, using a Beta value of 10 (base current equal to 1/10 of the maximum collector current) works for just about any BJT when used as a switch.

I always thought of using BJT for small current application. Considering their low price and there low bias current/voltage to do the job. Plus it's relatively more immune to ESD compared to a Mosfet.

However it looks like my statement is not so valid anymore considering you can buy a Mosfet ( NX3008NBK,215) ( with a turn on voltage of a 0.9V that can perform better than a BJT for the same price($0.05)
I will be driving the gate at 2.5V. According to Figure.8 of the datasheet, the resistance will be around 1.4 ohms. If I am driving the LED at 10mA then a voltage drop of 14 mvolt will form across the drain source, that's much better than the 0.1Volts Vce for the BJT.
Are there any advantages in going with the BJT for this simple application? It looks like this mosfet is a clear winner. What do you guys think?

If you are counting pennies, see #7 above. MMBT3904 is <$0.02 in quantity. With a MOSFET, you will not need the current limiting, series resistor which goes from the control signal to the BJT base. However, check your "off" control signal and make sure it is always low enough to keep the MOSFET turned off under all conditions. The voltage drop of the BJT vs MOSFET is probably swamped by the LED and current limit resistor drops. Probably not important.

I think we are getting to the point where it is necessary to know more about the overall circuit before anything is a "clear winner". What is the source of the control signal? What is the LED type and the voltage source, etc.? How important are pennies of savings? Are there other parts of the circuitry which need a similar part? Is it possible or desirable to reduce the number of items in the Bill Of Materials, etc.?

This is a sub-circuit and consists of microcontroller 3 pinouts driving 3 LEDs via MOSFET/bjts. We won't be ordering a 3000 pieces reel. Most likely it will be lower quantity. In any case $ saving is always good specially when I don't see any advantage going with the BJT. BOM reduction is also a plus. A pull down resistor at the gate will be there on all 3 inputs and I might include a 50 ohm resistor in series with the gate to reduce ringing on switch on.
The LEDs are surface mount indication LEDs (10-30ma) low power.

rocky79 said:

..............................and I might include a 50 ohm resistor in series with the gate to reduce ringing on switch on.
..................................

Click to expand...

If you are driving LEDs you don't need to worry about any possible ringing. You don't need the series resistor, especially if you are worried about pennies.

For three LEDs, have you considered driving the LEDs directly from the microcontroller output? Most micros can sink 10mA on a pin and a total of 30mA output low current. Do the micro and LEDs run off the same supply voltage?

I am not that worried about pennies, my point earlier was that if a mosfet turned out to be cheaper and better than the BJT for that simple app then why pay more ( low quantity)
The system will eventually go through EMI testing so I was thinking an inline resistor with the gate may help reduce the turn-on/off EMI.

More current might be needed to drive the LED, so I wanted to leave that option open.

Thanks for all your help, I appreciated your input on this