Requesting advice: MOSFET or IGBT in place of BJT?

Go To Last Post
17 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi all,

 

I would like some quick advice. First of all, please no "why do you want to do that" or "use a switcher" or any other things that are not what I asked.

 

With that, please see the attached circuit diagram. It is a simple and very nicely performing voltage regulator.

 

The idea behind it is that the LM-317 initially provides the output, but as soon as the current demand increases enough, the voltage across the PNP Vbe makes the transistor begin to conduct and it "pushes in" the required extra amount of current to maintain the LM-317 programmed voltage.

 

By adding one or more NPN transistors to the circuit making a "PNP-NPN Darlington", any amount current can be provided, and always perfectly referenced to the LM-317. The way the 317 is wired allows it to sense any drop along the path (up to the point where it's connected) allowing for remote voltage sense if desired.

 

My question / advice request is: If I wanted to replace the NPN "boosters" or even just the main PNP transistor with a MOSFET or IGBT, which would be preferable, and why? Or, are neither suitable substitutes for BJT parts?

 

My thinking is that maybe MOSFET parts would be more rugged and less likely to be damaged by temporary (fraction of a second) overloads. Of, course, I'm probably way out in left field about this...smiley

 

(click thumb for full size)

JPEG image

 

 

Thanks!

Attachment(s): 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

Last Edited: Thu. Jul 23, 2020 - 12:05 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Igbts are chosen for high voltage applications to overcome the mosfet’s high RDSon.
Mosfets might be challenging due to the voltage to turn it on. Say it is 10V, then the input voltage needs to be around 10V above your output voltage for a p chan or you need to have a boost circuit for an n chan.

Why use a lm317? Try a uA723.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Use an MJ11033 if you need more current.

 

Jim

 

 

(Possum Lodge oath) Quando omni flunkus, moritati.

"I thought growing old would take longer"

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With some value changes, a FET might work. One difference is that a FET usually takes several volts to turn on, while a BJT only takes 0.7V. 

 

I doubt that a FET will give you any better performance, reliability, etc, than a BJT. FETs have their own failure modes.

 

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

That circuit looks much like the classical O/P current boosted voltage regulator found in the much loved National Semiconductor Databook which I no longer have unfortunately.

 

As such I think it relies on the current gain of the pass transistor. Switching to a MOSFET would mean generating a larger control voltage for the gate which in turn reduces the headroom for the LM317.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The circuit is clever and dirt simple.

The regulator is set up as the classic const current nicad charger (I used to build many of  'em around 1980!!)...the current is simply Vref/R , or  1.25/220, this current (5.682ma) flows through the 2200 ohm

 

so vout is 2200*1.25/220 + Vref =12.5 +1.25 =13.75 volts

 

The regulator only cares that the 5.68 ma flows through the 220 ohm resistor; it doesn't care whether that current is supplied by Vout, or the regulator hitting the gas to help out (if the Vout is too low, this current will be too low without the regulator's help).  Any help provided by the regulator is through the Vin terminal, flowing through the 22 ohm resistor.  The 22 ohm Vdrop turns on the transistor.  So as the regulator provides more help, it is helping its own effort by turning on the pass transistor.  Working together, they end up supplying the setpoint current & thus the 13.75 Vout.  Teamwork!

 

Seems like you could use a P-mosfet...you can raise the 22 ohm resistor (say to 220 ohms) to develop a larger transistor control voltage.     Note that this may mean your Vin needs raised (say the mosfet took 3V to turn on enough) One rough thought (have to think some more), for the mosfet is you might be able to move the 2200 to the V+ side & connect ctrl&220 to V-...so the regulator resides down at the bottom rather than up on the top.   The 22 ohm would become much larger, like 2k

 

    

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
Why use a lm317? Try a uA723.

 

LM317 is simpler.

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

N.Winterbottom wrote:

That circuit looks much like the classical O/P current boosted voltage regulator found in the much loved National Semiconductor Databook which I no longer have unfortunately.

 

As such I think it relies on the current gain of the pass transistor. Switching to a MOSFET would mean generating a larger control voltage for the gate which in turn reduces the headroom for the LM317.

 

 

Of course the regulator depends on the pass transistors having gain! smiley

 

But the MAGNITUDE of the gain is unimportant, other than having enough (and Hfe 10...20 or so is more than enough).

 

As far as "headroom", the regulator is a single voltage output, so headroom is of no concern.

 

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:

The circuit is clever and dirt simple.

The regulator is set up as the classic const current nicad charger (I used to build many of  'em around 1980!!)...the current is simply Vref/R , or  1.25/220, this current (5.682ma) flows through the 2200 ohm

 

so vout is 2200*1.25/220 + Vref =12.5 +1.25 =13.75 volts

 

The regulator only cares that the 5.68 ma flows through the 220 ohm resistor; it doesn't care whether that current is supplied by Vout, or the regulator hitting the gas to help out (if the Vout is too low, this current will be too low without the regulator's help).  Any help provided by the regulator is through the Vin terminal, flowing through the 22 ohm resistor.  The 22 ohm Vdrop turns on the transistor.  So as the regulator provides more help, it is helping its own effort by turning on the pass transistor.  Working together, they end up supplying the setpoint current & thus the 13.75 Vout.  Teamwork!

 

Seems like you could use a P-mosfet...you can raise the 22 ohm resistor (say to 220 ohms) to develop a larger transistor control voltage.     Note that this may mean your Vin needs raised (say the mosfet took 3V to turn on enough) One rough thought (have to think some more), for the mosfet is you might be able to move the 2200 to the V+ side & connect ctrl&220 to V-...so the regulator resides down at the bottom rather than up on the top.   The 22 ohm would become much larger, like 2k

 

    

 

Great analysis. That's EXACTLY how it works.

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Current limit? about 3A.

 

 

Does the LM317 have thermal shutdown? May need to keep the 22 Ohm I guess to bias the LM317.

my projects: https://github.com/epccs

Debugging is harder than programming - don’t write code you can’t debug! https://www.avrfreaks.net/forum/help-it-doesnt-work

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My suggestion was a joke! I thought us oldies would get it.

Funnily enough the venerable uA723 turns up in your other thread.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You could use the 0.22 ohm with an opamp across& RC to form a "fuse"..longer lasting overloads would then have lower reaction thresholds than fast pulses.  It could drive an npn (like 2n3904)  to gnd at ctrl thus make it it seem like more current is being delivered than really is (hence causing the output to fall at very high loading).   

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Thu. Jul 23, 2020 - 11:50 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Similar from Linear Technology though 2m ohm instead of 220m ohm :

LT1431 - Programmable Reference (datasheet)

[page 8, bottom]

FET Low Dropout 5V Regulator with Current Limit

[pass power NFET, 2m ohm sense, '431 shunt regulator (100mA max), 5V out at 2.5A limit, 12V drive]

via LT1431 Datasheet and Product Info | Analog Devices

 

edit : improved '431 (reduced noise, increased bandwidth)

Voltage References - MaxLinear

 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Fri. Jul 24, 2020 - 12:07 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ki0bk wrote:

Use an MJ11033 if you need more current.

Jim

 

They have "interesting" Safe Operating Area curves too.  Ask Me How I Know...  devil  S.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Scroungre wrote:

ki0bk wrote:

Use an MJ11033 if you need more current.

Jim

 

They have "interesting" Safe Operating Area curves too.  Ask Me How I Know...  devil  S.

 

Datasheets should show "unsafe operating areas".... i.e. the whole graph! angry

 

It's no fun to watch $60 swirl down the drain because some idiot (me) stuck a toe outside the SOA... 

 

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It's no fun to watch $60 swirl down the drain because some idiot (me) stuck a toe outside the SOA... 

Not a lesson you'll forget in a hurry.......... 

 

Nevertheless, this been an instructive thread

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Krupski wrote:

It's no fun to watch $60 swirl down the drain because some idiot (me) stuck a toe outside the SOA... 

 

Oh, it may be no fun for you, but it's great fun for us!  surprise  devil  Carry on! S.