Power BJT strange functioning - am I doing something wrong?

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

Hi all,

 

I obtained two NPN BJT power transistors from DigiKey. They are by ST Micro and the part number is BUT30V. These are 100 amp, 250 watt parts in the "ISOTOP" (SOT-227) package.

 

They are intended to be used as pass transistors for a 13.8 volt, 30 ampere linear power supply regulator. I attached each one to it's own heatsink and used Indium foil as the thermal interface material, along with two heavy springs to maintain a constant pressure on the heatsink during thermal expansion and contraction. Before wiring them in, I intended to run around 20 volts DC at up to 10 amperes through them to insure the TIM was properly conducting the heat to the heatsink. To do this, I setup a power supply set to 20 volts with the current limit set (initially) to 3 amperes, intending to go up to 10 amperes eventually. The base was driven via a 10 ohm, 1 watt resistor connected to a 1.5 to 20 volt adjustable power supply, starting out at 1.5 volts. (See the test setup image below).

 

Anyway, when I turned on the 20 volt "big" supply, I had a collector current of around 3 amperes (caused by the 1.5 volts through the 10 ohm resistor). I slowly turned up the base drive and the collector current increased as I expected. At around 7 amperes collector current, the transistor collector voltage went down to around 7 volts and the collector current at around 7 amperes (it "jumped" to that value). I turned off the "big" supply and removed the base drive (the base was now open - free floating). Turned the big supply back on and got the same 7 volts, 7 ampere load from the transistor! It was conducting with no base drive. Then I tried connecting the base directly to the emitter and no change - still 7 amps collector current.

 

Thinking that somehow I blew this monster of a transistor, I tried the second one. Same thing, it started out with full collector voltage of 20V and zero amperes. Upon connecting the base drive, I got the same low collector current as the first one. Upon ramping up, this transistor suddenly snapped into the same "conducting with no base drive" mode as the first one. Connected the base to the emitter - same thing - no change. And the second transistor had virtually the same collector voltage and current as the first one!

 

Then I tried applying a NEGATIVE bias relative to the emitter of -5 volts through the 10 ohm resistor. No effect, no change.

 

To see if I was maybe doing something wrong, I stuck an old 2N3055 I had lying around on a heatsink and connected it the same way. IT worked. I could control the collector current, the collector current dropped to zero when removing the base drive, and it got hot as I expected it to. NOTE that the monster BUT30V transistors didn't even have time to get hot (or warm). However, connecting them to the "big" power supply and just letting them draw the 7 amperes with a 7 volt collector voltage makes them heat up, along with the heatsink (at least the indium foil works!).

 

Thinking possibly they were oscillating, I hooked the silly-scope to the collector. Nothing. Pure DC.

 

Here's a link to the PDF datasheet:  BUT30V Datasheet

 

These transistors are supposedly "high current bipolar modules" intended for "motor control, smps, ups and welding equipment", yet both seem to have died under conditions that a little TO-220 part could handle. What the heck? Did I do something wrong?

 

Thank goodness I did this test, because if these went into their "semi-shorted" mode in the power supply, it would have nuked my Kenwood TS-590SG HF radio (a $1700 to $1800 radio!).

 

I would appreciate it if someone could take a look at the data sheet and tell me if I did something wrong, because I'm not seeing it.

 

(click thumb for full res)

JPEG Image

 

Thanks!

 

Attachment(s): 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

Last Edited: Thu. Jul 23, 2020 - 12:50 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

BJTs have positive thermal feedback. That is, as  temperature rises, the base current increases, which increases the collector current, and so forth. Potential  Death Spiral. The expectation MAY have been that negative feedback is present, to limit such behavior.

 

Jim

 

 

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

 

 

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

Looking at the "Safe Operating Areas" chart 20V at 7A is outside the DC region.
That '250W' rating is an "Absolute Maximum Rating"

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

ka7ehk wrote:

BJTs have positive thermal feedback. That is, as  temperature rises, the base current increases, which increases the collector current, and so forth. Potential  Death Spiral. The expectation MAY have been that negative feedback is present, to limit such behavior.

 

Jim

 

 

Yes I know. But temperature wasn't a factor. The transistors didn't even have time to warm up. And, the power supply driving the collector-emitter circuit has a current limit function. Anyway, it's not like the transistor was getting hot and the collector current ran away. It never got a chance to warm up, and if this did happen, the PSU current limiter would have stopped it. As a point of information, when these were to be used in the regulator circuit, there would be two in parallel, with emitter resistors to balance the load between the two, so I'm fully aware of the NTC characteristics of a silicon BJT.

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

Last Edited: Thu. Jul 23, 2020 - 05:00 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mikech wrote:
Looking at the "Safe Operating Areas" chart 20V at 7A is outside the DC region. That '250W' rating is an "Absolute Maximum Rating"

 

I think you are misreading the "Safe operating areas" graph on page 3. "Safe" is usually more than 1/2 "max"....

 

A TO-220 BJT could handle 20V/7A easily (if kept from overheating). The "big" transistors never even got NEAR 20 amps Ic before they failed (or even got warm). The 7 amp value was what the Ic was after the transistors failed (i.e. the collector current with NO base current).

 

A test with a 2N3055 and a 2N3771, however, handled the same test conditions just fine (test time limited to a few minutes to prevent overheating of course).

 

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

Last Edited: Thu. Jul 23, 2020 - 04:55 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Looking at the "Safe Operating Areas" chart I see with 20V on the collector the current for DC conditions (I.e. not pulsed) is only 2A.

 

You've blown these transistors up.

 

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


Krupski wrote:
They are intended to be used as pass transistors for a 13.8 volt, 30 ampere linear power supply regulator

 

For an example of a 13.8v 35 amp power supply, you can't find a better one then Astron brand, the schematics are readily available online, here is one example.

These things are almost indestructible:

 

Astron 35m

 

Jim

 

 

(Possum Lodge oath) Quando omni flunkus, moritati.

"I thought growing old would take longer"

 

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

But in that supply the 25A continuous load current is shared between 4 transistors; each of which have only (18-13.7) = 4.3 Vce. That's a world of difference to the OPs 20V circuit.

 

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

Krupski wrote:

mikech wrote:
Looking at the "Safe Operating Areas" chart 20V at 7A is outside the DC region. That '250W' rating is an "Absolute Maximum Rating"

 

I think you are misreading the "Safe operating areas" graph on page 3. "Safe" is usually more than 1/2 "max"....

 

A TO-220 BJT could handle 20V/7A easily (if kept from overheating). The "big" transistors never even got NEAR 20 amps Ic before they failed (or even got warm). The 7 amp value was what the Ic was after the transistors failed (i.e. the collector current with NO base current).

 

A test with a 2N3055 and a 2N3771, however, handled the same test conditions just fine (test time limited to a few minutes to prevent overheating of course).

 

Mikech is correct.  Read your graph, and keep in mind the axes are logarithmic, not linear.  20V/7A is not MUCH outside the SOA (safe operating area / secondary breakdown) but it is.  There is no such thing as "... Safe is usually more than 1/2..."

 

And your transistors have blown into short-circuits, which is what BJTs do when run into secondary breakdown.  S.

 

PS - A 2N3055 is a TO-3 can device, not a TO-220.  So's a 2N3771.  The chips themselves inside those cans are much larger.  Read their spec sheets -

 

https://media.digikey.com/pdf/Da...'s/2N3055-ssi.pdf

 

Something mangled that URL, so here it is with spaces in the HTTP.  Lurv the 'preview' button!

ht tps://media.digikey.com/pdf/Data%20Sheets/Solid%20State%20INC%20PDF's/2N3055-ssi.pdf

 

https://www.onsemi.com/pub/Colla...

 

 

Edited to fix 7V to 20V.  S.

 

Last Edited: Thu. Jul 23, 2020 - 07:07 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Just for grinsies laugh , I had a go at this - making LINEAR graphs of Safe Operating Area for your BUT30V and your 2N3055.  I think this makes clear why your super-duper new transistor isn't quite as good as your lumpy old ones - at the point of interest.  Hope y'all can read this screenshot.  It's not real complicated.

 

Some (most?  All??) of the data points were sorta estimated from the spec sheet graphs.  Charming how their specifications get a bit strange when you try to use both of them together!  Error correction is welcomed!  S.

 

Two Transistors in Safe Operating Area comparison

 

Edited to change the picture to one I thought more useful, compact, and informative.  So there.  S.

Last Edited: Thu. Jul 23, 2020 - 08:55 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The transistors didn't even have time to warm up.

The transistors might disagree...their little die is inside handling your 7A*20V =140Watts....so they may be the hot point ...with their heating slooowly spreading.  But before then, their SOA rating plummets, because at 50 deg C die temp it's down to 80%, and at 80C, it's down to 50% SOA.   Try a mosfet, you may be happier, they have a positive temp coeffecient...if overheated, their conductivity goes slightly down. 

 

What are the Vin & Vout for your desired supply?  Note the transistor you have is rated for 200V*...so there is a penalty in other parameters (maybe soa) to get high voltage rating...why not uses a 60V part?

 

* this is especially true in mosfet, the 200V device Rds goes up a lot compared to 30V or 60v parts.

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: 2

From memory (>20 years ago, so probably wrong), I seem to recall something about current crowding is a problem with large BJT. Once an area of the silicon heats up, all the current goes to that spot, and it runs away in an instance (so fast the heatsink remains cold?).

 

http://ecee.colorado.edu/~bart/book/book/chapter5/ch5_4.htm#5_4_4

 

That might be why it can only do small currents above 10V. I think there is a lot more about this in old books than is on the mind virus repository that some call the internet.

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

N.Winterbottom wrote:

Looking at the "Safe Operating Areas" chart I see with 20V on the collector the current for DC conditions (I.e. not pulsed) is only 2A.

 

You've blown these transistors up.

 

 

I don't want to doubt you because I really don't KNOW, but it SEEMS ludicrous to me that a big monster of an SOT-227 transistor rated for such strenuous duty can be blown with Vceo=20V and a puny 2 amps Ic.

 

To me it seems akin to saying "I've got this heavy duty tow chain rated for 1000 pounds but I tried to pull a 10 pound load and it snapped". In my mind, "it does not compute".

 

I must REALLY be missing something.

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

N.Winterbottom wrote:

But in that supply the 25A continuous load current is shared between 4 transistors; each of which have only (18-13.7) = 4.3 Vce. That's a world of difference to the OPs 20V circuit.

 

 

I can see you really aren't "getting" what I'm trying to do. Here is a schematic of an older linear power supply that I built (and it works great). The one I'm working on now is a very similar design, but with only two "big" transistors instead of four TO-3 types:

 

(click thumb for larger)

JPEG image

 

The "20 volt thing" was ONLY a test... to make the new transistor dissipate power and warm up, so I could feel the transistor body AND the heatsink to determine if the indium foil I was using as thermal interface material was working as I hoped.

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

ron_sutherland wrote:

From memory (>20 years ago, so probably wrong), I seem to recall something about current crowding is a problem with large BJT. Once an area of the silicon heats up, all the current goes to that spot, and it runs away in an instance (so fast the heatsink remains cold?).

 

http://ecee.colorado.edu/~bart/book/book/chapter5/ch5_4.htm#5_4_4

 

That might be why it can only do small currents above 10V. I think there is a lot more about this in old books than is on the mind virus repository that some call the internet.

 

AH!!! I was reading around the web trying to stumble upon some reason why I'm having this problem. I also ran across the topic of current crowding and can visualize how it can happen. Maybe THIS is the problem and I simply blew up the transistors... sad

 

Isn't this just a "large die MOSFET" problem, though?

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

ron_sutherland wrote:

From memory (>20 years ago, so probably wrong), I seem to recall something about current crowding is a problem with large BJT. Once an area of the silicon heats up, all the current goes to that spot, and it runs away in an instance (so fast the heatsink remains cold?).

 

Roughly speaking, yes.  You get a localized meltdown.

 

Krupski wrote:

To me it seems akin to saying "I've got this heavy duty tow chain rated for 1000 pounds but I tried to pull a 10 pound load and it snapped". In my mind, "it does not compute".

 

I must REALLY be missing something.

 

Imagine you hooked up your tow chain to a jetliner taking off, and left a lot of slack in it - Then it suddenly snaps tight at 400mph.  F = ma, you know - and yes, a 10lb weight might well break it.  Your new transistors certainly have a lot more strength (current, voltage) but they're also lot more rigid - won't stretch at all - and so snap where a slightly springier (read: old) transistor might just yank it into motion.

 

That help any?  smiley  S.

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

avrcandies wrote:

The transistors didn't even have time to warm up.

The transistors might disagree...their little die is inside handling your 7A*20V =140Watts....so they may be the hot point ...with their heating slooowly spreading.  But before then, their SOA rating plummets, because at 50 deg C die temp it's down to 80%, and at 80C, it's down to 50% SOA.   Try a mosfet, you may be happier, they have a positive temp coeffecient...if overheated, their conductivity goes slightly down. 

 

What are the Vin & Vout for your desired supply?  Note the transistor you have is rated for 200V*...so there is a penalty in other parameters (maybe soa) to get high voltage rating...why not uses a 60V part?

 

* this is especially true in mosfet, the 200V device Rds goes up a lot compared to 30V or 60v parts.

 

Please see this post: LINK for the answer to your question.

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

Krupski wrote:
Isn't this just a "large die MOSFET" problem, though?

 

No, MOSFET do not current crowd, they have other problems though.

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

PS - A 2N3055 is a TO-3 can device, not a TO-220.  So's a 2N3771.  The chips themselves inside those cans are much larger.  Read their spec sheets -

 

Yes I know that. I mentioned TO-220 to make the point that "even a relatively small transistor should be able to handle my test".

 

I wasn't saying that a 2N3055 or 2N3771 was a TO-220 package part.......

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

There has also been a LOT of development work in BJTs since the heyday of the 2N3055, but as the graphs show, it has been towards high currents at low voltages.  Your other post has a few things to say about voltage-constrained systems and why BJTs might be better there - which is why they're going for that market. 

 

Most real high-power stuff, though, is FETs and IGBTs, on the grounds that if you're throwing kilowatts around you can afford to pop up with a slightly higher drive voltage somewhere.  Same with high voltage - on modern BJTs that's mostly an afterthought, because everyone wants to use a MOSFET there.

 

Takeaway lesson here?  READ THE SPEC SHEETS.  wink

 

IMHO, anyhow.  Have fun!  S.

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

ron_sutherland wrote:

Krupski wrote:
Isn't this just a "large die MOSFET" problem, though?

 

No, MOSFET do not current crowd, they have other problems though.

 

So the current crowding thing is a BJT problem???

 

How is it that in my 45+ years of EE I've never heard of this problem until yesterday/today?

 

(Then again, I DID plot the grid/plate characteristics of a 6C4 triode by hand, on paper in one of my labs...) We didn't have MOSFETS back then and most transistors were the crappy PNP germanium type. And we had dinosaurs. Really.

 

Have you ever smelled a burning selenium rectifier?

Ever smelled a popped aluminum electrolytic capacitor?

Do you know the pleasant smell of a warm tube ("valve" for you guys on the other side) radio?

Ever put little permanent magnets near a TV horizontal output tube to eliminate the Barkhausen oscillations causing vertical bars on the left side of the CRT?

 

"What a drag it is getting old".

 

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

Scroungre wrote:

There has also been a LOT of development work in BJTs since the heyday of the 2N3055, but as the graphs show, it has been towards high currents at low voltages.  Your other post has a few things to say about voltage-constrained systems and why BJTs might be better there - which is why they're going for that market. 

 

Most real high-power stuff, though, is FETs and IGBTs, on the grounds that if you're throwing kilowatts around you can afford to pop up with a slightly higher drive voltage somewhere.  Same with high voltage - on modern BJTs that's mostly an afterthought, because everyone wants to use a MOSFET there.

 

Takeaway lesson here?  READ THE SPEC SHEETS.  wink

 

IMHO, anyhow.  Have fun!  S.

 

Yup. RTFM is always good advice!

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

Scroungre wrote:

Just for grinsies laugh , I had a go at this - making LINEAR graphs of Safe Operating Area for your BUT30V and your 2N3055.  I think this makes clear why your super-duper new transistor isn't quite as good as your lumpy old ones - at the point of interest.  Hope y'all can read this screenshot.  It's not real complicated.

 

Some (most?  All??) of the data points were sorta estimated from the spec sheet graphs.  Charming how their specifications get a bit strange when you try to use both of them together!  Error correction is welcomed!  S.

 

Two Transistors in Safe Operating Area comparison

 

Edited to change the picture to one I thought more useful, compact, and informative.  So there.  S.

 

Very interesting... especially how going from log to linear shows it so much more clearly.

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

Scroungre wrote:

 

 

Imagine you hooked up your tow chain to a jetliner taking off, and left a lot of slack in it - Then it suddenly snaps tight at 400mph.  F = ma, you know - and yes, a 10lb weight might well break it.  Your new transistors certainly have a lot more strength (current, voltage) but they're also lot more rigid - won't stretch at all - and so snap where a slightly springier (read: old) transistor might just yank it into motion.

 

That help any?  smiley  S.

 

Yes it does. I didn't think there was any big difference between different BJTs other than Hfe, Vceo max, Ic max, Ft, etc... 

 

To see if I'm really seeing it properly, it seems to me that a BJT with a SMALL die would be less susceptible to current crowding and runaway because the die isn't big enough to have "slightly different personalities in different areas". Or, a hot spot tends to conduct it's heat to the rest of the (small) die, therefore large temperature differences get "absorbed", sort of like "thermal capacitance". Is that an accurate idea?

 

 

Gentlemen may prefer Blondes, but Real Men prefer Redheads!

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

My first job after college was at this outfit.

 

https://www.apexanalog.com/

 

They used a lot of power transistors; sadly, I don't recall those things very well; Scrounge's graph may have jogged a connection that was nearly atrophied.

 

At work, I had a CRT that had a button I could push to operate the degaussing coil; it got a lot of use, for example: tools, floppy disk people accidentally left on my desk.

 

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

ron_sutherland wrote:
...  I seem to recall something about current crowding is a problem with large BJT.
ON Semi low VCE-sat transistors spread the current across the die; some of these have been second-sourced by Diodes.

Diodes acquired Zetex (known for high gain power transistors some of which are "large" die so somewhat low noise)

 

ON Semiconductor

Technology [page 1]

 

via Technical Documentation | NPN Transistors - PNP Transistors (ON Semiconductor, White Papers)

https://www.diodes.com/products/discrete/bipolar-transistors/transistor-bjt-master-table/#Part%20Number:%22DSS%22

 

"Dare to be naïve." - Buckminster Fuller

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

Compared to a few other power transistors I looked up, the BUT30V has a pretty horrific SOA curve.  At 20V long-term, it drops like a rock.  Looks like you got the "lucky" pick.

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

Krupski wrote:
Do you know the pleasant smell of a warm tube ("valve" for you guys on the other side) radio?
and vacuum electron device by the professor one annoys wink

Tubes: Luckily, They’re Still Around | Electronic Design

by Lou Frenzel

JUN 08, 2020

As Noah said to his family after dinner, “Those unicorn steaks were excellent!”

...

 

By the Book

...

These are vacuum electron devices (VEDs).

...

 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Thu. Jul 23, 2020 - 11:33 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

"Dare to be naïve." - Buckminster Fuller

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

Krupski wrote:

Have you ever smelled a burning selenium rectifier?

Ever smelled a popped aluminum electrolytic capacitor?

Do you know the pleasant smell of a warm tube ("valve" for you guys on the other side) radio?

Ever put little permanent magnets near a TV horizontal output tube to eliminate the Barkhausen oscillations causing vertical bars on the left side of the CRT?

 

"What a drag it is getting old".

Yes, yes, yes, and no!

 

  #10  -  Lately, I've noticed people my age are so much older than me.

  #14  -  I thought growing old would take longer.

 

 

 

(Possum Lodge oath) Quando omni flunkus, moritati.

"I thought growing old would take longer"