# Hfe of NPN at extremely low base current

1. ## Hfe of NPN at extremely low base current

Hi,
The attached is a delay circuit….when the PRIBIAS1 rail gets its 14V voltage, then Q20 holds Q19 off for some 40ms.
However, the base current of Q20 can only be 50uA max.
The Ic in Q20 must get to some 450uA in order for it to turn Q19 off.
The BC846ALT1G datasheet does not describe what happens to hfe at these extremely low currents.
Do you think the circuit will be OK?

BC846ALT1G NPN datasheet
https://www.onsemi.com/pub/Collateral/BC846ALT1-D.PDF

•

2. ## Re: Hfe of NPN at extremely low base current

Hi,

50uA base current isn't extremely low.

And the datasheet tells a lot of low curents:
Look at
"ON CHARACTERISTICS" -> "DC current gain"
Here they use 10uA of I_C (collector current) .... this is about 1/50 of what you need
And the typ specified gain is 90 .... means a base current of just 0.11uA (only 1/450 of your base current)

Also figures 3, 4, 5 and 6 include the collector current you ask for.

Klaus

•

3. ## Re: Hfe of NPN at extremely low base current

Therotically IBQ20=65uA due to high value of 100K Resistance.Decrease that Resistance to lower value..OR use a Zener diode with a lower breakdown voltage..

•

4. ## Re: Hfe of NPN at extremely low base current

Hi,

To the circuit.
I think it should work.

Some ideas that come into my mind:
* the current through the zener is much lower than specified --> Expect a much lower zener voltage.
* maybe a diode to fast discharge the capacitor (at powr OFF) is useful.
* a critical point is, that the first bjt needs to go well into saturation (V_ce below 0.55V) to switch the second bjt safely OFF.
For this low V_CE the h_FE will be lower than given in the datasheet.
You may relax this by dividing R214 into two resistors, maybe 15k each. Connect the C of the first bjt to the center of these resistors.
Add a R_be to the second bjt. Maybe 2k.
Then V_CE of the first bjt at switching OFF the second bjt will be at about 4.5V ... @630uA.

Since you have a two bjt solution you may add a small capacitor from second bjt "C" to first bjt "B" to speed up switching edges.
(Positve feedback)

For a more precise solution I personally would use a low power comparator...there are some with built in voltage reference.

Klaus

5. ## Re: Hfe of NPN at extremely low base current

Thanks, without doing a new pcb, i am wondering if we can just replace R215 with 22k, to give more base current...then we can achieve the low enough vce_on of the q19?
We didnt in the end need all the dealy that we get with the top post's cct

6. ## Re: Hfe of NPN at extremely low base current

Hi,

Without additional information I see no reason why not to decrease R215 .. maybe increase C214 to get the desired timing.
It should improve performance.

You could also replace one or two bjts (preferred the second one) with a Mosfet ... the gate current is about zero, while the output current may be much higher....with less voltage drop.

Klaus

7. ## Re: Hfe of NPN at extremely low base current

a critical point is, that the first bjt needs to go well into saturation (V_ce below 0.55V) to switch the second bjt safely OFF.
For this low V_CE the h_FE will be lower than given in the datasheet.
Thanks, i now wonder if we can make R215 = 22K, and then surely the hfe will be enough to make the second bjt's vce <0.2v?

It turns out we didnt need so much delay as the micro can be compus mentus quicker if we clear the PWRTE bit in config

•

8. ## Re: Hfe of NPN at extremely low base current

Hi,

Before we talked about V_ce of the first bjt (to safely switch OFF the second bjt), now you talk about the V_ce of the second bjt.

We have no information about the second bjt's collector current .... with 0.1mA there wil be no problem ... with 100mA it's about impossible.
Why do you want it to be < 0.2V?
Still no information why not use a mosfet, which easily could satisfy all the given reqirements, because it has a huge current gain.

It seems the requirements change from post to post...timing, currents, voltages...
You know, I'm a friend of deciding the requirements first, then do the design.

Maybe you could tell us what the whole circuit is good for..

Klaus

--[[ ]]--