Maximum frequency limit of any Power Transistor

[manishanand14]

Hello Guys

I am using TIP31C in a PWM circuit at a frequency of 500Khz.

By seeing the datasheet how can I find what is the maximum switching frequency TIP31C can withstand without any problem.

There is a parameter called High Current Gain — Bandwidth Product
fT = 3.0 MHz (Min) @ IC = 500 mAdc.

What does it tell?

 

[WimRFP]

Hello, maximum switching freqeuncy depends on many factors.

This transistor comes under many brands and datasheets show different values for same property.

http://www.datasheetcatalog.com/datasheets_pdf/T/I/P/3/TIP31C.shtml shows you many datasheets, several of them specify ft and switching behavior under various conditions (for example Mospec).

Once you have a transistor, turn-off and turn-on times vary significantly on how you drive the transistor and whether or not you use the transistor in a hard-switched or soft-switched application. You may know that increasing the frequency gives more switching loss, so also the accaptable switching loss determines the maximum frequency for your application.

To be honest 500 kHz is very high for this transistor in a hard-switched application. I expect relative high switching loss.

Probably you have these at hand, If not a mosfet would be better.

- - - Updated - - -

ft (transition frequency) is that frequency where the AC hfe of transistor reaches 1.

They may specify an AC hfe at certain frequency. When hfe = 10 at 1 MHz, then ft = 10 MHz.

The DC hfe can be high (for example 80), but when you increase frequency your reach a point where hfe start to drop with 6 dB each doubling of frequency. This is mainly due to base emitter diffusion capacitance.

There is no straight relation between ft and switching frequency as other transistor parameters play a role in that. Therefore they may specify switching times.

You will see that the turn-off time is always longer then the turn-on time. This is because of charge storage in the base region when the transistor is switched into saturation. This charge must be removed first before the transistor comes out of saturation and begins to turn-off.

 

[mtwieg]

Like WinRFP says, switching times aren't closely related to transition frequency. But certainly you can't push the switching frequency anywhere near f_T. Even 500KHz is far too fast to get decent performance out of that transistor (or any power BJT in general).

 

[dick_freebird]

You need to look at the storage time at whatever your
forced beta and worst case temp and worst case process
will give you. Many BJTs don't even spec that for any
condition, if they aren't aimed at switching applications.
Let alone the variety of conditions that would apply to
a converter with wide load range, temperature swings
and so on.

Forget switching speed, think about losing control of the
duty cycle fundamentally and having this vary with load.

I've seen plenty of 40V bipolars with microseconds of
base storage time. Which won't play at 500kHz unless
you clamp them, in which case you will eat even more
efficiency.

This all is why nobody uses bipolars in power conversion
(other than the real slow, high current guys with IGBTs -
IGBT being sort-of Baker clamped, and still slow as hell).

 

[manishanand14]

Is 100K okay for a BJT in general.

 

[schmitt trigger]

100 Khz is achievable with BJT, but it has to have a very specific base drive to reduce storage time.

Google "Baker Clamp".

Additionally, you have to select a transistor which is specifically designed for fast switching. I don't think the TIP31C was conceived as a fast switcher.

 

[mtwieg]

Baker clamps can help with fast operation, but in general a mosfet will always give better speed than a BJT. Back in the days, engineers figured out little tricks to improve switching speed of BJTs, but such methods get pretty complicated (current transformers, RCD networks, etc), and even then they're inferior to mosfets.

I'm pretty sure devices like the TIP31C were meant more for audio amplifiers.

 

[WimRFP]

@manishanand14: If you tell us something about the application (type of switcher, voltage, current, duty cycle range, soft or hard switched, loss/efficiency, etc), we may help you better. Just as an example: in a Royer oscillator (frequently used for MF power generation or DC to high voltage converters), such transistors can perform well at 50..100 kHz, but in buck converter at 100 kHz, efficiency will be low compared to what you can expect from a (cheap) mosfet.

 

[schmitt trigger]

Like other posters have mentioned, BJTs were used in the early 80s when there were no power Mosfets available. In those days, the highest frequency would be 25, perhaps 50 Khz.

The BJTs also required emitter ballasting to match the currents thru parallelled devices, which meant further losses.

But with the advent of power Mosfets, the BJTs were abandoned altogether for power switching applications.

 

[manishanand14]

I am using this circuit for LED dimming at 200kHz.





Please tell me how I should connect a n-channel MOSFET at VF2. I mean for connecting a BJT we calculate the base drive resistor.
Here how shall I connect the MOSFET if i use it instead of a BJT [obviously for high switching speeds].

Can I use a totem pole driver.

Please give me the connection diagram.

 

[schmitt trigger]

As a matter of fact, to properly drive a Mosfet at 500Khz, you DO require a Mosfet driver.
There are many available, inverting, non inverting, low side, high side, single, dual..........check the following link from Microchip and download some datasheets which will give you all the info that you require.

https://www.microchip.com/ParamChartSearch/chart.aspx?branchID=9010&mid=11&lang=en&pageId=79

 

[FvM]

I wonder why it's necessary to use 200 or even 500 kHz for LED dimming?

Apart form the many right things that have been already said about MOSFET's versus BJT for fast switching, there's another rather simple point. BJT need to be operated with sufficient base current to achieve low saturation voltage. Do you know where you get the base current for the TIP31? Using a base resistor? That's really inconvenient and will cause a lot of additional losses.

 

[manishanand14]

Dear FvM and schmitt trigger

Thanks for reply

Do you know where you get the base current for the TIP31? Using a base resistor? That's really inconvenient and will cause a lot of additional losses.

Thanks for pointing out that considering the fact that TLC555 can source max 10 mA of current. So in this case TIP31C will most probably be operating in active region resulting in losses.

I just wanted to modify this PWM circuit which was originally used with NE 555 and TIP31C BJT for say <10kHz output PWM ; to generate high frequency PWM till 1Mhz because TLC555 can go till 1.2Mhz.

LED dimming is just one example I have given.I know I can do PWM dimming of LEDs with lower frequency like 10-50 kHz or even lower. Can you suggest me a suitable frequency for LED dimming in general.

So for replacing TIP31C with a IRF7406 [MOSFET] I must use a MOSFET driver like UCC27518 or UCC27519 single-channel, low-side gate driver device.

Out of UCC27518 and UCC27519 which one shall I choose.



Out of inverting and non-inverting logic which one shall I use and why?

- - - Updated - - -

 

[schmitt trigger]

use non-inverting.

Non-inverting means exactly that... whenever the 555 output is high, the Mosfet driver's output will also be high, but capable of a much higher drive.
A simplified analogy would be an emitter follower transistor.

 

[manishanand14]

I found this amazing blog by Tahmid.

https://tahmidmc.blogspot.in/2012/12/low-side-mosfet-drive-circuits-and_23.html



I guess i can also use this totem pole driver. Can I use:transistors BC548[NPN] and BC558[PNP]

Please explain me the use of R27 and R25 in this circuit.

 

[schmitt trigger]

To ensure that, under no signal conditions, the nodes are pulled down and the main mosfet remains solidly off.