I have received some power transistors that may or may not be fakes. Capacitance measurements (with a multimeter, i.e. roughly 4x too high results compared to data sheet) löök fine, but workmanship is really rustique.
Any idea of how to measure f_T reasonably accurately? I suspect one needs a low pass filtered oszillator with a frequency of single-digit MHz, The transistor would probably be operated in common emitter mode. The voltage gain at say 5 MHz is a rough measure of f_T, with the probem that V_Ce is anything but constant. Added to this is the questiion of how much bias current and what kind of a modulation to allow.
How does one measure c_ob with some significant bias voltage, by the way.
I believe there were some standard measurement circuit in an earlier edition of Tietze-Schenk, but it was thrown out with the focus of the newer editions on integrated circuits. Newer give away copies of vintage editions of textooks or databooks!!!
Any idea of how to measure f_T reasonably accurately? I suspect one needs a low pass filtered oszillator with a frequency of single-digit MHz, The transistor would probably be operated in common emitter mode. The voltage gain at say 5 MHz is a rough measure of f_T, with the probem that V_Ce is anything but constant. Added to this is the questiion of how much bias current and what kind of a modulation to allow.
How does one measure c_ob with some significant bias voltage, by the way.
I believe there were some standard measurement circuit in an earlier edition of Tietze-Schenk, but it was thrown out with the focus of the newer editions on integrated circuits. Newer give away copies of vintage editions of textooks or databooks!!!
I posed the same question to Randy Slone some time ago. His answer: You can't -at least not with a DMM and other typical hobbyist equipment. Appearently not even with the somewhat more sophisticated equipment someone like Slone has accumulated.
This is only what I was told, maybe someone else has more encouraging info.
This is only what I was told, maybe someone else has more encouraging info.
Well, with a DMM you can measure hFE at some very low collector current and with minimal bias voltage. And you can measure cob without any bias, so values will be 3-5x too high.
With a fast scope and a singal generator you should be able to measure fT. I am just trying to remember the best setup to use. I really shouldn't have given away the older edition of this textbook....
With a fast scope and a singal generator you should be able to measure fT. I am just trying to remember the best setup to use. I really shouldn't have given away the older edition of this textbook....
Which means you'd need a sweepable signal generator that goes out to several 100 MHz and an equally fast scope.
I think the standard way is to use some fixed V_CE bias like 10 V, a fixed frequency source at 10-20% of expected f_T and then measure the current gain. Current can be measured by including small resistors in the base and emitter leads and measuring the voltage drop across them. fT is then the measured current gain x oscillator frequency. I am just trying to remember the setup that minimizes the influece of parasitics such as C_ob.
Also, f_T is about small signal current gain, so it is not h_FE (f) but beta (f). For comparison to data sheets, one needs to set the specified bias current and then use a current modulation that is about 10% of the bias.
I wonder if this is the spec that is really useful for class AB output transistors...
I think the standard way is to use some fixed V_CE bias like 10 V, a fixed frequency source at 10-20% of expected f_T and then measure the current gain. Current can be measured by including small resistors in the base and emitter leads and measuring the voltage drop across them. fT is then the measured current gain x oscillator frequency. I am just trying to remember the setup that minimizes the influece of parasitics such as C_ob.
Also, f_T is about small signal current gain, so it is not h_FE (f) but beta (f). For comparison to data sheets, one needs to set the specified bias current and then use a current modulation that is about 10% of the bias.
I wonder if this is the spec that is really useful for class AB output transistors...
take a look at the Intersil HFA3102 high speed transistor arrays -- 10GHz -- you could build a high speed dif-amp and take the zero crossing point as unity current gain. (i am using them in the Phasemeter).
i'm not into characterizing RF transistors, but I would suspect that a mixer is used to draw those curves we see in the texts.
i'm not into characterizing RF transistors, but I would suspect that a mixer is used to draw those curves we see in the texts.
Technical university.
If you really want to measure that, you may go to one of technical univercitis. Maybe there you can measure all of this parameters with a special equipment. There is a lot of univercitis with electronics speciality .
C_ob and other capacitances is not a direct capacitance. It is a model. If you have time, you can read some books about semiconductors, then you can find why bipolar transistors isn't good in audio power applications.
If you want to use them as a followers, then you don't need to worry about ft and capacitances. Drivers with low output impedance is all you need.
If you really want to measure that, you may go to one of technical univercitis. Maybe there you can measure all of this parameters with a special equipment. There is a lot of univercitis with electronics speciality .
C_ob and other capacitances is not a direct capacitance. It is a model. If you have time, you can read some books about semiconductors, then you can find why bipolar transistors isn't good in audio power applications.
If you want to use them as a followers, then you don't need to worry about ft and capacitances. Drivers with low output impedance is all you need.
- Status
- Not open for further replies.