Have I gone mad?

[king.oslo]

Hello there,

I thought this was correct: "NPN transistor, small base-emitter current enables large collector-emitter-current."? I thought so, until now. I am very confused:

I found two transistor in my parts bin with with C368 marking. I assume they are BC368 1A 20V NPN transistors.

My PSU has 1 ampere current limit. I connected 2K2 resistor between 5VDC and collector, and emitter straight to GND. When I applied 5VDC (via my Fluke 87V in current ampere mode) to the base, and my PSU goes straight to current limiting (1amp) which my multimeter confirm is flowing in through the base. Consequently, the transistors get very hot. This is the case for both my C368-marked transistors.

This surprised me so much. How come a large current plunge through the base to the emitter, when I have 2K2 resistor between 5VDC and collector.

This makes sense to me. Is my knowledge of NPN transistors to blame, or is there something else wrong?

Thank you for your time.

Kind regards,
Marius

[IanB]

You have fallen into the classic error of logic:

"All dogs have four legs. The creature I see in front of me has four legs. Therefore the creature in front of me is a dog."

Do you see the problem?

[IanB]

The above assumes you really have a BC368 transistor and that you have identified the pins correctly.

But also note there is no guarantee a component marked C368 is actually a BC368 transistor, nor that you have correctly identified which legs are B, C and E.

[grumpydoc]

Nothing in your set-up limits the base current. At the end of the day the base-emmitter junction is just a diode and you more-or-less created a short circuit from the +VE terminal on your lab power supply, via the B-E junction to ground. No wonder it current limited, I'm surprised the transistor survived (did it?).

Putting a 2.2k resistor in the collector circuit will limit the collector current, not the base current.

[IanB]

You wrote: "small base-emitter current enables large collector-emitter-current".

Maybe so.

But THIS DOES NOT MEAN that restricting the collector-emitter current will limit the base current.

[king.oslo]

IanB: You're saying the transistor isn't what I think it is. What is it then? I found the base using the diode measurement on my multimeter. That is in compliance with the datasheet. But I have wired these in all ways thinkable, and I still get results which makes no sense to me.

grumpydoc: Limit base-emitter current? Never heard that is necessary. Is this the case? I have a setup in LTSpice right now with the same setup which only draws a few microamps through the base.

Kind regards,
Marius

[IanB]

I'm not saying the transistor isn't what you think it is. I'm saying it may not be.

But what I'm really saying is what grumpydoc also says.

Just because a small base current will permit a large collector current, it does not mean the reverse is true. A small collector current does not force a smaller base current. Just because all dogs have four legs, it does not mean that all four legged creatures are dogs.

You should have a current limiting resistor on the base of a transistor.

[king.oslo]

  Thanks M

[Simon]

The Base emitter path is a diode, and as with a diode it will carry as much current as it is given once the voltage drop is exceeded (0.7V). So you limit the base current and the collector current will follow the base current.

Don't believe everything a simulation tells you, the key word being: simulation, not real life that is much more complex.

[grumpydoc]

Limit base-emitter current? Never heard that is necessary. Is this the case? I have a setup in LTSpice right now with the same setup which only draws a few microamps through the base.

Can't comment on LTSpice - don't have a copy to play with but surprised if it wouldn't model the high current correctly.

Yes, in the set-up you describe you need a base current limiting resistor. The B-E junction acts as a diode with a forward voltage drop of 0.7V, so the voltage across whatever resistance is offered by your Fluke current sense will be 4.3V - chances are this is a few tenths of an ohm. Ohms law will give you the current that would flow if your PSU's current limiting didn't kick in first.

As both IanB and I said - limiting the collector current does not limit base current. Transistor current gain does not work backwards.

[king.oslo]

How about this then,

Say I put a 10K resistor between VCC and base. Base-emitter is a short circuit (minus diode drop), which means that I will always have .65V at the base, which means that the NPN transistor never can switch on (the V_EMITTER is always V_BASE - 0.6V.

This cannot be right?

[c4757p]

I just ran your setup in LTspice with the similar 2N2222 and I see 400mA into the base - already huge. Thermal runaway will make this much larger in real life. I think you have something set up wrong.

As for your recent question, yes, it can switch on just fine, because the base-emitter voltage (or current, depending on the model) determines the collector current. V_BE is not exactly 0.65V, it varies with applied current. The more current you apply through the limiting resistor, the higher the voltage goes, and the more current the transistor will allow to pass into the collector.

Just like a diode, there is no fixed voltage drop, voltage varies with current. (Or current varies with voltage.) But if you plot a graph of current vs. voltage, there is a sharp corner around that 0.65V, and the current drawn shoots up fast. You could in theory apply the proper voltage directly to the base without a resistor, but it would be something in the 0.65-0.9V range. (With a 2N2222, LTspice gives me a range of roughly just about 0mA to 300mA for that voltage range.) In real life you need the resistor to protect against variation in this I/V curve (it varies from transistor to transistor, but also varies significantly with temperature), and it also helps you hit that tiny voltage more accurately.

Try it in LTspice. Apply a voltage {VB} directly to the base (include the curly braces), then add a directive (Edit -> SPICE directive) ".step param VB 0.65 0.9 0.05" (make sure to include the dot at the beginning, do not include quotes), and run an operating point analysis. Check the collector current. Try the same thing with a steady voltage, but add a varying resistor (you can vary resistance the same way as voltage - .step param PARAM_NAME LOW HIGH STEP_AMOUNT")

[king.oslo]

I just ran your setup in LTspice with the similar 2N2222 and I see 400mA into the base - already huge. Thermal runaway will make this much larger in real life. I think you have something set up wrong.

As for your recent question, yes, it can switch on just fine, because the base-emitter voltage (or current, depending on the model) determines the collector current. V_BE is not exactly 0.65V, it varies with applied current. The more current you apply through the limiting resistor, the higher the voltage goes, and the more current the transistor will allow to pass into the collector.

Just like a diode, there is no fixed voltage drop, voltage varies with current. (Or current varies with voltage.) But if you plot a graph of current vs. voltage, there is a sharp corner around that 0.65V, and the current drawn shoots up fast. You could in theory apply the proper voltage directly to the base without a resistor, but it would be something in the 0.65-0.9V range. (With a 2N2222, LTspice gives me a range of roughly just about 0mA to 300mA for that voltage range.) In real life you need the resistor to protect against variation in this I/V curve (it varies from transistor to transistor, but also varies significantly with temperature), and it also helps you hit that tiny voltage more accurately.

Try it in LTspice. Apply a voltage {VB} directly to the base (include the curly braces), then add a directive (Edit -> SPICE directive) ".step param VB 0.65 0.9 0.05" (make sure to include the dot at the beginning, do not include quotes), and run an operating point analysis. Check the collector current. Try the same thing with a steady voltage, but add a varying resistor (you can vary resistance the same way as voltage - .step param PARAM_NAME LOW HIGH STEP_AMOUNT")

Thanks!     M

[Shuggsy]

I have a setup in LTSpice right now with the same setup which only draws a few microamps through the base.

You aren't alone... even some of the greats have had similar issues at times... no real replacement for experience and real-world testing! SPICE can be very useful as long as you understand the limitations. Everything is a model!

See pic below for Bob Pease (analog guru from National Insturments, sadly passed about a year and a half ago) tossing his computer off the top of National Semiconductor's building back in the day.

RIP RAP!

[ftransform]

lmfao!~!!!!!

[Simon]

is that a commodore 64 ? or did he just hurl the monitor and keyboard ?

[Ed.Kloonk]

You have fallen into the classic error of logic:

"All dogs have four legs. The creature I see in front of me has four legs. Therefore the creature in front of me is a dog."

Do you see the problem?

My dog has no nose.

[Shuggsy]

is that a commodore 64 ? or did he just hurl the monitor and keyboard ?

My guess is the whole PC. I don't think Bob Pease would have been satisfied with just smashing a monitor and keyboard...

[grumpydoc]

How about this then,

Say I put a 10K resistor between VCC and base. Base-emitter is a short circuit (minus diode drop), which means that I will always have .65V at the base, which means that the NPN transistor never can switch on (the VEMITTER is always VBASE - 0.6V.

This cannot be right?

If you connect the emitter to ground and the base to 5V via a 10K resistor then the BE junction will start to conduct. There will be a forward voltage drop of approx 0.7V.

So, the emitter is at 0V (you grounded it, remember), the base at +0.7V the transistor is on and the base current is (by ohms law) (5-0.7)/10,000 = 430µA.

The gain is somewhere between 85 and 375, lets guess a typical value of 200 so up to 200x430µA collector current is possible or 86mA - if we stick to your 2.2K collector resistor then that is what is going to limit the collector current. The transistor will be saturated - we can't get more collector current if we increase base current from this point. The datasheet says the collector-emitter saturation voltage is 0.5V so with the emitter at 0V (grounded, remember) the collector will be at 0.5V and the collector current will be (5-0.5)/2200 which is a shade over 2mA

Make sense?

[king.oslo]

How about this then,

Say I put a 10K resistor between VCC and base. Base-emitter is a short circuit (minus diode drop), which means that I will always have .65V at the base, which means that the NPN transistor never can switch on (the VEMITTER is always VBASE - 0.6V.

This cannot be right?

If you connect the emitter to ground and the base to 5V via a 10K resistor then the BE junction will start to conduct. There will be a forward voltage drop of approx 0.7V.

So, the emitter is at 0V (you grounded it, remember), the base at +0.7V the transistor is on and the base current is (by ohms law) (5-0.7)/10,000 = 430µA.

The gain is somewhere between 85 and 375, lets guess a typical value of 200 so up to 200x430µA collector current is possible or 86mA - if we stick to your 2.2K collector resistor then that is what is going to limit the collector current. The transistor will be saturated - we can't get more collector current if we increase base current from this point. The datasheet says the collector-emitter saturation voltage is 0.5V so with the emitter at 0V (grounded, remember) the collector will be at 0.5V and the collector current will be (5-0.5)/2200 which is a shade over 2mA

Make sense?

Yes it makes sense. Thank you.M

[vk6zgo]

You have fallen into the classic error of logic:

"All dogs have four legs. The creature I see in front of me has four legs. Therefore the creature in front of me is a dog."

Do you see the problem?

My dog has no nose.

My God! How does he smell?

[vk6zgo]

You have fallen into the classic error of logic:

"All dogs have four legs. The creature I see in front of me has four legs. Therefore the creature in front of me is a dog."

Do you see the problem?

I would suggest that this creature is a 2SC368!

[PA0PBZ]

You have fallen into the classic error of logic:

"All dogs have four legs. The creature I see in front of me has four legs. Therefore the creature in front of me is a dog."

Do you see the problem?

My dog has no nose.

My God! How does he smell?

He smells bad, but fortunately he doesn't know 

[T4P]

Usually BC parts are marked as BCXXX and Sanken transistors are marked as CXXX which are actually 2SCXXX