How to REALLY measure hFE of NPN transistor?

[dusan]

So I googled how to measure hFE using two ammeters and this thread came up: https://www.eevblog.com/forum/testgear/measuring-hfe-of-a-transistor-by-multimeter/ - OK, seems simple enough:



hFE=154, seams reasonable for BC547. Then I got briliant idea to replace the transistor with short and the result:



hFE=144, that's really good for piece of wire. Before I head to the patent office, are there any better methods at measuring hFE?

[Kim Christensen]

You can invalidate the 2nd scenario with an additional step in your test:
Remove A2 (to break circuit) and ensure current reading on A1 drops to near zero. If A1 current stays where it was, this means a faulty transistor, and that the previous hFE reading was invalid.
Note: You don't need R3 for these tests.

If you want more detailed hFE measurements use a curve tracer.

[wasedadoc]

Hint: What is 2756 Ohms multiplied by 3.24 mA?

[IanB]

Before I head to the patent office, are there any better methods at measuring hFE?

Probably any method is better than that one 

That measurement circuit is basically useless, as you showed by removing the transistor (good experiment, by the way).

For a proper measurement of hFE, the value of V_CE should remain constant during the test. With the way R1 is arranged in that circuit, V_CE is never going to remain constant. Always be suspicious of what you find on the internet, and verify with other sources.

[IanB]

Here is a university electronics lab exercise that may give you some ideas:

https://personal.utdallas.edu/~mtinker/EE3110/ElectronicDevicesLaboratory/LabManuals/7.%20Bipolar%20Junction%20Transistor%20Characteristics.pdf

If you don't have sophisticated source measure units to hand you will have to improvise that part, but you can follow the general procedure.

[iMo]

My first beta (or aprox hfe or h21) meter had a switch with a couple of R_base resistors (like 3M9/390k) and a 4.5V battery.
See below - with 390k base resistor the Vbe is always around 0.7V thus the base current is always aprox 10uA with 4.5V battery, npn transistor type regardless.
The ammeter current is aprox hfe*I(R_base), as the base current is negligible compared to the collector current..
In the below example hfe = 2.9mA/10uA=290

PS: added more examples showing you cannot replace the transistor with a piece of wire while getting the same hFE 

[barshatriplee]

Here is also an old thread about how to measure the hFE of NPN transistors. You may wanna take a look: https://www.eevblog.com/forum/testgear/measuring-hfe-of-a-transistor-by-multimeter/

[Sredni]

It doesn't work because you are saturating the transistor, as @wasedadoc hinted.
What you are measuring is beta_forced, i.e. you are no longer on the flat 'horizontal' part of the VCE-IC characteristic, but you are on the slanted leg, instead.
When the transistor is saturated, it appears as almost a short on the collector mesh, and that is why you get the same value for IC: its value is determined by Vcc and Rc alone.

Change the base resistor to 2.2meg and your op will be in the linear region. I get a beta of 450, well in line with the parameter specified by the simulator (Microcap).

[Picuino]

Here my prefered circuit.



Beware of transistor heating!!!

[dusan]

The method with only base resistor works fine but only for ideal transistor:

Code: [Select]

* first line is ignored
.model QMY_NPN NPN(bf=200)
V1 1 0 9
VA1 1 c 0
VA2 2 b 0
Q1 c b 0 QMY_NPN
RB 1 2 47k
.control
tran 100u 20m 19m
meas tran ic max i(VA1)
meas tran ib max i(VA2)
let hfe = ic / ib
echo "hfe = $&ic / $&ib = $&hfe"
.endc
.end

The example above will print:

Code: [Select]

hfe = 0.0346132 / 0.000173066 = 200
But when I use realistic transistor model:

Code: [Select]

* first line is ignored
.model QMY_NPN NPN(bf=200 is=5e-14 ikf=0.3 ne=2 var=24 vaf=155 ise=1.4e-11)
V1 1 0 9
VA1 1 c 0
VA2 2 b 0
Q1 c b 0 QMY_NPN
RB 1 2 96k
.control
tran 100u 20m 19m
meas tran ic max i(VA1)
meas tran ib max i(VA2)
let hfe = ic / ib
echo "hfe = $&ic / $&ib = $&hfe"
.endc
.end

It prints only 177 (the highest I ever get is with 96k base resistor):

Code: [Select]

hfe = 0.0153783 / 8.66167E-05 = 177.545
All methods mentioned in this thread including that PDF shows lower than real value.

[TimFox]

I have a Keithley constant-current lab supply that works at these current levels.
If I supply the emitter through that supply (other end of supply grounded), apply the desired collector voltage from a power supply (other end grounded), and measure the base current through a microammeter (other end grounded), the ratio of emitter current divided by base current is h_FE.
If you don't own a constant-current supply, you can use a regular lab voltage supply and resistor, measuring the emitter current with a milliammeter in series with the resistor.
Of course, both of these meters can be normal 3.5-digit DMMs.
(I use that setup to measure the grid current in vacuum tubes, where I need to add a resistor in series with the cathode since the cathode-grid voltage will be positive and the output voltage of the current supply must be negative for my supply.)

[Doctorandus_P]

With a single resistor such as picuino shows, it is quite easy to overload your transistor if the poteniometer is in the wrong setting, but if you add the "protection" resistors such as in the first circuit, then you have to verify the transistor is not in saturation.

Also note that that Hfe is not a single number. It varies with operating parameters such as base current Emitter voltage and temperature.
There can also be quite big changes between the Hfe's of different transistors of the same type, even when they are from the same batch. When designing a circuit you should always allow for changes in Hfe. Adding an emitter resistor is a common technique. In this case, when there is a lagre emitter current, the B-E voltage is reduced and this reduces the base current and stabilises the currents.

If you have two SMU's then you can simply hook one up to generate a base current (for example 0 to 100uA) and use the other as an collector / emitter voltage source with a current limit.

[mawyatt]

I have a Keithley constant-current lab supply that works at these current levels.
If I supply the emitter through that supply (other end of supply grounded), apply the desired collector voltage from a power supply (other end grounded), and measure the base current through a microammeter (other end grounded), the ratio of emitter current divided by base current is h_FE.
If you don't own a constant-current supply, you can use a regular lab voltage supply and resistor, measuring the emitter current with a milliammeter in series with the resistor.
Of course, both of these meters can be normal 3.5-digit DMMs.
(I use that setup to measure the grid current in vacuum tubes, where I need to add a resistor in series with the cathode since the cathode-grid voltage will be positive and the output voltage of the current supply must be negative for my supply.)

Hfe, hfe, or Beta is defined as Ic/Ib with Vce held constant, you are measuring Ie/Ib, which is close to Ic/Ib, since Ic ~ Ie for high Beta, but not so for low Beta.

Best,

[TimFox]

To be accurate,
I_E = I_C + I_B
Therefore, calculate h_FE = I_C/I_B = (I_E - I_B) / I_B =  (I_E / I_B) - 1 from my measurement.

For both beta and my vacuum-tube grid measurements, if the plate-cathode or collector-emitter voltage be critical to the measurement, one can add a high-impedance voltmeter at the cathode or emitter to ground to correct the supply voltage measurement.
Sometimes, you also need to know the base-emitter or grid-cathode voltage at a given current.

Just as in Dante, about the 1% error in the Julian calendar:
Ma prima che gennaio tutto si sverni
 per la centesma ch'è là giù negletta
Paradiso, canto XXVII, line 143

[Zero999]

Just connect the base to the collector via the ammeter and select a collector resistor to set the collector current.

[TimFox]

A2 will short the collector to the base, and therefore force V_CE = V_BE.

edit:  corrected typo for BE

[wasedadoc]

A2 will short the collector to the base, and therefore force V_CE = V_CB.

Did you mean V_CE = V_BE

[TimFox]

A2 will short the collector to the base, and therefore force V_CE = V_CB.

Did you mean V_CE = V_BE

Oops:  I edited in a correction.

[dusan]

I used Zero999 method and here is the result:
Real hFE is 200 but no matter what value of R1 is chosen it doesn't get close to 200, barely 160.

[IanB]

All methods mentioned in this thread including that PDF shows lower than real value.

The real value is what you measure. If you do not measure it, it is not real.

Consider: you think the "real" value is 200, but you try to design an amplifier circuit around that transistor that assumes the gain is at least 200. Your circuit will not work as designed, because how can your amplifier circuit behave any differently than your measurement circuit?

[wasedadoc]

As has already been pointed out, Zero999's method has the transistor operating with almost no voltage difference between collector and base which means that the collector to emitter voltage is only around 0.7 Volts.  The transistor's current gain in that condition is significantly less than when it is operated  under the typical conditions encountered in a linear circuit.

[Zero999]

As has already been pointed out, Zero999's method has the transistor operating with almost no voltage difference between collector and base which means that the collector to emitter voltage is only around 0.7 Volts.  The transistor's current gain in that condition is significantly less than when it is operated  under the typical conditions encountered in a linear circuit.

So what?

[TimFox]

The good practice I learned for designing circuits with discrete BJTs was to use a minimum value for beta for the biasing, then check with maximum values.
Also, beta does depend on V_CE as well as on I_C.
At higher collector voltages, the depletion region between collector and base narrows, reducing base current for a given collector current, increasing beta.
This called the "Early effect", and is important to BJT distortion, since it makes the input current (into the base) a function of the collector voltage.

[Zero999]

The good practice I learned for designing circuits with discrete BJTs was to use a minimum value for beta for the biasing, then check with maximum values.
Also, beta does depend on V_CE as well as on I_C.
At higher collector voltages, the depletion region between collector and base narrows, reducing base current for a given collector current, increasing beta.
This called the "Early effect", and is important to BJT distortion, since it makes the input current (into the base) a function of the collector voltage.

Thanks for explaining it. The idea is to get a low h_FE, so once can be sure the transistor is suitable for the application.

[floobydust]

I find h_FE is best measured at your circuit's operating point wrt V_CE if you need it tight. Also, always put a ceramic cap say 10nF across E-B to shunt RF pickup as well. Graph from the AoE showing some of their results.