mosfets vs power transistors / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

mosfets vs power transistors

previous 1 2 3 next

Move Thread

LAN_403

Sulaiman

Sun Jan 20 2008, 12:06PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Posted: Sat Jan 19 2008, 02:36AM
Is a mosfet just a switch, or can it be slowly "opened" like a power transistor to allow varying amounts of current through it?

The attached graph (for IRFP460A) shows the story;

IF the junction temperature is kept constant (difficult in practice)
then over a range of Vgs , Ids increases with increasing Vgs.
The relationship is not linear but it is monotonic so feedback can be used to 'linearise' current.
The two lines of the graph show how Ids varies with Vgs for junction temperatures of 25C and 150C.

Notice for Vgs = 7V the current for Tj=25C is the same as for 150C, about 42 Amps.

At lower Vgs the drain current depends on Tj, so for example at Vgs= 5.0 V the Drain current would be about 3 Amps initially with Tj = 25C, as the Junction heats up the current increases until at Tj = 150C , Id is about 8.5 Amps.
This is what you observed!

For Vgs <= about 3V there is negligible Drain current, the transistor is OFF, for Vgs >= about 7V the transistor is ON.
So a mosfet can easily be used as a switch.

Depending on how you use it, a mosfet can be considered as either a (nearly) linear device, or as a switch.

HV Enthusiast

Sun Jan 20 2008, 02:59PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Mates wrote ...

IamSmooth wrote ...

Is a mosfet just a switch, or can it be slowly "opened" like a power transistor to allow varying amounts of current through it?

There are high quality sound aplifiers based on mosfets...

Yes, but most of these are Class-D where the MOSFET *is* acting as a switch.

Dr. Slack

Sun Jan 20 2008, 07:21PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Yes, but most of these are Class-D where the MOSFET *is* acting as a switch.

Many are class D, but that's irrelevant because the rest are class A, AB or B, where the MOSFET is acting as a voltage controlled current source. Think of a MOSFET as a valve if that helps (an N channel depletion mode FET that is, which has a negative Vth), an N channel enhancement mode FET (the common type) has a positive Vth, both types, as well as the valve and the BJT, are inverting.

How does one program the resistance? I've tried to read on this, but I am missing it.

Don't bother, as the Vth is temperature depedant, so will the effective resitstance for any given Vgs. Just aim to control the thing you need to control. If the load vanishes from your generator, presumably you need to control the terminal voltage to avoid it rising. Configure the MOSFET as a voltage clamp. Connect a bias resistor Gate-Source to draw a few mA at Vth, say 1k ohm, then connect X volts of series zeners from Drain to Gate, and viola, it will draw as much current as necessary to limit the Vds to VzenerX + Vth.

A way to reduce the FET dissipation (350W is pretty state of the art for a single small package) would be to connect a resistor in series with the Drain, to drop most of the terminal votlage and hence power, when pulling the full rated current from the generator. Connect the zeners to the top of the ressitor, don't leave them connected to the drain. You could do worse to use a lamp as a resistor, high disspation, and it lights to warn you the clamp is operating.

Electroholic

Sun Jan 20 2008, 07:44PM

Registered Member #191 Joined: Fri Feb 17 2006, 02:01AM
Location: Esbjerg Denmark
Posts: 720

Steve Conner

Mon Jan 21 2008, 11:39AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I described a dump load for a renewable energy system in my PhD thesis. There's a schematic and explanation of how it works.

In case you can't be bothered reading that, yes, MOSFETs can be used as voltage-controlled linear elements, though I'll avoid the term "resistor" as they are more constant current sinks like a pentode vacuum tube.

They do the high-power-linear-thing better than bipolar transistors, because they don't suffer from second breakdown, well at least not to anything like the extent that BJTs do.

My load used a bank of TO-220 MOSFETs and big wirewound resistors on a big fan-cooled heatsink, controlled by a TL431 shunt regulator IC, in the manner of the voltage clamp that NeilThomas described. I also used resistors for the reason that Neil stated: they are cheaper and more reliable than silicon, and you only need one-quarter as much silicon as if you tried to dissipate the same power in MOSFETs alone.

IamSmooth

Mon Jan 21 2008, 05:36PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Steve, can you post the link to that thesis so I can peruse through it. Thanks.

I just purchased some large aluminum heat sinks and 10 175W NPN transistors. I was going to get some MOSFETs, too. What concerns me is that when I did my test the transistors held the current steady at a constant base current, while the MOSFET continued to change value for a given drain voltage. Maybe I need to see what happens when I get a good heat sink on them.

Steve Conner

Tue Jan 22 2008, 10:30AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I can't remember off hand which appendix it's in.

GeordieBoy

Tue Jan 22 2008, 11:16AM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

> I can't remember off hand which appendix it's in.

Appendix B: A voltage-controlled dump load

Lots of interesting stuff there. The section about power measurement using a soundcard is cool! (Appendix D)

As several people said the drain-current of a MOSFET with fixed Vgs can vary dramatically with temperature. For switching MOSFETs in general Rds(on) has a positive temperature coefficient whilst Vgs(threshold) has a negative temperature coefficient. So, depending where you are operating the device, drain current can either increase or decrease when junction temperature rises. The negative tempco of Vgs(th) means that paralleled switching MOSFETs do not inherently share current in linear applications.. (As the junction temperature of one MOSFET rises, the gate-source threshold voltage for that device falls and it turns on harder, taking the lion's share of the load current and getting hotter. The same behaviour can cause hot-spotting and localised thermal runaway in regions across the surface of one device.)

As Steve explains towards the bottom of p.204 (and top of 205) this can be corrected by using a feedback loop to force each MOSFET to draw the requested current. Here is an example of how multiple MOSFETs can be connected in parallel to ensure load-sharing in linear operation:

http://www.microsemi.com/micnotes/APT0002.pdf

"It is not appreciated in the all-pervasive switchmode world that power MOSFETs cannot and do not meet their
published Forward Biased Safe Operating Area (FBSOA) ratings, because the average designer working in this
discipline does not care. Such is not the case, however, among those engineers constrained to use these devices in the â€œlinear modeâ€..."

.

Steve Conner

Tue Jan 22 2008, 12:33PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Well, I must admit that when I designed that dump load, I still thought that MOSFETs shared current by themselves. So I just hooked all four devices in parallel.

I also once designed a 450V linear regulator that used an 800V rated TO-220 MOSFET as the pass element. It's still working, but this is probably just luck! As that application note says, MOSFETs have got worse at linear operation over the years as their switched mode performance has got better. The device I used was the BFC62 which is now discontinued, and it was specified to run in linear mode with 200V across it.

One thing that always surprises me, since I learned about these limitations, is commercial audio amp designs that use vertical MOSFETs in the linear region. A**** made a few high-powered amps for bass guitar that used parallel banks of IRFP250 and IRFP9250s, and they show up in repair shops a lot with blown MOSFETs.

You can buy lateral MOSFETs that are fully specified for linear use, and really do share current in parallel. But they're pretty expensive and don't come in voltage ratings above 250V.

...

Tue Jan 22 2008, 03:32PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

Really all that is necesary to make the mosfets share a load gracefully is to add a large amout of resistance in series with the drain, that will swamp the difference in resistance in the fets due to temperature, etc.... From what I understand that was the plan anyway, so all that you need to do is split the resistor bank so taht each fet gets its on resistor.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob