transitor: the base pin.

[c4757p]

So my statement remains : no base current is no chance of collector current .

Repeat after me: "correlation does not imply causation".

Correlation does not imply causation.
Correlation does not imply causation.

Bloody hell...

[baljemmett]

Here is the core of the problem of this thread (post#1):
„When applying a current to the base pin to allow the current to flow from the collector to the emmiter(or vice versa) where does the current from the base pin go?...I haven't crossed any text that explains this clearly.“

And this was my answer:
However, you always should know what you are doing and, thus, realize that this is a model only. The physical reality is that the BJT is, of course, a voltage-controlled device (Ic=f(Vbe)). This is a proven fact.

… which, you might notice, has nothing whatsoever to do with the actual question asked.  You took a question which could have been - and in fact was - answered within a couple of posts and turned it into yet another tedious rehash of a discussion that has been had many times already, and now you're spitting the dummy because you don't like the behaviour of some of the local argumentative bores.  Congratulations.

[free_electron]

So my statement remains : no base current is no chance of collector current .

Repeat after me: "correlation does not imply causation".

Correlation does not imply causation.
Correlation does not imply causation.

Bloody hell...

i throw a frozen pea at a jar of peas and nothing happens
i throw accelerate the same frozen pea supersonic at the jar and it makes hole allowing other pea's to roll out.

i apply a field to a base emitter of 0.1 volt. nothing happens.
i apply a field sufficiently large to get an electron to flow from emitter to base. i have now created the hole for an electron to go from emittor to collector.
you can sit there until doomsday  : as long as that initial pea does not make the inital hole you have no flow of pea's

without that initial base current , no collector current . yes you need a field to get electrons to flow. fortunately for us we only need about half a volt. if the force required to get electrons in valence band was a million volts it would have been bloody hard to make bipolar transistors.

correlation indeed does not imply causation.

case in point : base resistor 0 ohms. i apply 0.1 volt and nothing happens. i apply 0.5 volts and get collector current. i apply 0.6 volts and i get a larger collector current. i plot this out as a curve.
Vbe vs Ic. cause- effect corellation. great.

now. put a teraohm resistor in that base path. apply same 0.5 volts. see how much current you get in that collector. oops... your carefully plotted vbe /ic curve just went to snot.

you will undoubtedly argue that the 0.5 volts stands over the teraohm resistor... fine. Ohms law says that for that voltage to be there there must be a current flowing through it.
Do two things :

1) figure out where that electron that creates a voltage across the base resistor came from.
2) figure out why  , the drift speed of that electron ( which is what resistors do ) has an impact on the flow in the collector. according to you it was purely field induced...

[G0HZU]

i apply a field to a base emitter of 0.1 volt. nothing happens.

An electric field is defined as an electric force per unit charge. eg it has units of newtons per coulomb   (or volts per metre)

[T3sl4co1l]

So my statement remains : no base current is no chance of collector current .

Repeat after me: "correlation does not imply causation".

Correlation does not imply causation.
Correlation does not imply causation.

Bloody hell...

Now now.  Back in the days of germanium transistors, negative base bias was occasionally required.  And yes, I mean negative in the sense of turning off, not just a technicality because germanium transistors were almost exclusively PNP.

i apply a field to a base emitter of 0.1 volt. nothing happens.

Current still flows, even if you can't sense it; for 1pA = IS (typical within orders of magnitude for a 2N3904), one should expect Ic ~= 2nA at Vbe = 0.1V.

It is true that hFE is severely reduced at low Ic (due to recombination), but nothing causes Ic to reach zero unless Vce also goes to zero.

Leakage current stops at Tabs = 0 K, but so do the dopants, so you can't have a functional non-leaky BJT.

Tim

[G0HZU]

In my day the recommended learning method for stuff like this was to listen to an experienced lecturer at college/university and to also read the (recommended) literature on the subject. This was a fairly safe and reliable way to learn.

Two pages back we had both of these sources of knowledge on this very thread. Now they are gone...

Today we also have the internet and the casual and curious googler can gain access to all kinds of subject matter in a few seconds. It is a truly remarkable place to share knowledge. However, the risk here is that the subject matter they find (and believe) may be controversial or at the very least unreliable and I'm not sure it is the safest method of learning or passing on information to beginners.

Trying to debate and learn stuff on a hobby forum like EEVblog is also fraught with risk because the experts are massively outnumbered and not easy to spot amongst the pretenders or 'google professors' who are very keen to use their keyboard and show off.

A little bit of knowledge can be dangerous

[edavid]

Now now.  Back in the days of germanium transistors, negative base bias was occasionally required.  And yes, I mean negative in the sense of turning off, not just a technicality because germanium transistors were almost exclusively PNP.

Now too - it's also true for silicon, at a high enough temperature.

[T3sl4co1l]

The democracy of information is at once both liberating and stifling.

[T3sl4co1l]

Now now.  Back in the days of germanium transistors, negative base bias was occasionally required.  And yes, I mean negative in the sense of turning off, not just a technicality because germanium transistors were almost exclusively PNP.

Now too - it's also true for silicon, at a high enough temperature.

Indeed, or at a low enough current one would suppose (Iceo vs. Icbo should be different by approx. hFE times, for whatever hFE is at that current).

Tim

[free_electron]

Of course there is thermally induced current.
That leakage is caused because temperature accelerates electrons. If they start movi g you get current.

It's the same thing with photodiodes or transistors. Expose the base region (actually the base collector region, the emitter is not sensitive) to light and every photon will knock of an electron creating an electron hole pair. The recombination creates a base current and off we go.

Every transistor is light sensitive. Every junction is light snesitive for that matter. Photons produce free electrons that start moving. Moving electrons is charge displacement . charge displacement is current.  How much more clear does it need to be made ?

[T3sl4co1l]

I'm not sure offhand how sensitive (or not) the emitter is, actually; I know it's not usually targeted (for obvious structural reasons).  I would think it would be just as effective, as long as the light is absorbed within a few diffusion lengths of the B-E junction.  Photodiodes are made with a PIN structure to enhance the drift / depletion zone width, making absorption more likely to yield useful photocurrent at the terminals.

But be careful what you say; earlier you said "no chance of current", which I took to mean, NO current period, zero.  Yet you just said "of course there is thermally induced current", which means nonzero current.  As with the majority of this thread, it's a perfectly legitimate approximation to make in many circuits, but there are many others which will fail if this factor is forgotten.  And if you go on wording such statements as absolute fact, when they are actually highly conditional approximations, you will find many more problems, both technical and interpersonal, in your life!

Tim

[free_electron]

Photodiodes are pin co struction indeed. (Most of them, the ones that are not are photovoltaic cells)
That is the reason you need to reverse bias those things. The intrinsic layer creates a larger depletion zone. If i remeber ot right it is the depletion zone that absorbs the photons to get a current going. That intrinsic lump of semiconductor makes a 'thicker' junction thus it becomes more sensitive.

Same goes forregular pin diodes used for switching rf. You bias those things in reverse to turn em off. The thick junction caused by the intrinsic material means that the plates of the parasitic capacitor are very far away . Instead of a few tens of atoms in a regular p-n recombination zone, they are now the width of the intrinsic material. That means these diodes have extremely low capacitance so no rf energy will leak through.  They are susceptible to charge injection.

In the harddisk preamps we use an array of pin diodes to switch gain and bandwidth of the amplifier (switching resistors and caps essentially) i once  had a problem whereby after a few seconds the gain would suddenly drop. If i turned off a certain block in the chip this would not happen. We speculated there was leakage somewhere. I pushed the setup i to the probe station , turned off the lights in the probe room and targeted the laser, at very low setting, to the bank of pin diodes. After a few minutes iof prodding with my 'light probe' i could reproduce the effect. I had found the diode that produced that exact drop. Turning on the illumination of the microscope i looked around at adjacent components and noticed that there was no via ring around a control structure. Turn off the light again and aim at the control structure. By modulating the laser i tensity i could create charge accumulation and let it dissipate again . This caused the pin diode to react and show the gain behavior we saw on bench.
It turned out that, under certain operating modes (these are bicmos products) the base of a bipolar transistor went floating as the cmos driver tristated when the control block was turned off.
That foating base picked up stray electrons and became conductive enough to toggle the pin diode.
So we put a few gigaohms of resistance between the base and ground and the problem was solved. (We actually made the bottom drive mos 'leaky')

[T3sl4co1l]

Well, quantum efficiency doesn't change much with bias (having the N/P encroaching on the intrinsic section doesn't really prevent charges from finding their way, there's still a built-in potential).  Photocurrent is generally written as independent of bias.  This seems to be reasonable, even in the forward direction (PV panels have a squarish V-I curve; basically the only reason current drops is because it's shunted internally by forward I = Is*exp(V/Vth) current).  It does do *wonders* for capacitance though!

Heh, nice story.  Lesson learned: don't leave nodes floating at undefined voltages!  They'll inevitably do something unexpected or undesirable.  Hope it wasn't a gotcha, like, from a bad appnote... those are just mean.

Surprised light did much around ICs, unless there was, like, glass body diodes or something.  Are PIN diodes usually in glass format?  Were those..?

Think I measured this before, a 1N914 is so-and-so, I think around 1nA at room temperature and ambient light levels; rising to 2-4nA under the influence of my (admittedly pretty intense) LED flashlight at point blank range.  An effect not attributable to temp rise, because it doesn't vary over time (a valid concern, because I can physically feel the warmth from the LED when turning it on and off near a sensitive region, like my lips).

On the other hand, I've got a circuit with 2N3904/6 collectors driving a capacitor; when both are fully off, leakage is in the ~pA range, for a voltage drift in the 10mV/min range.  At least... at room temperature it is.

If you need a really good diode, I've heard a plain old 2N3904 is at least as good as those so-called picoamp diodes (PAD-1 etc.?).  Go figure, putting the tiniest bit of effort into making something other than a diode and you get a wonderful diode...

Tim

[c4757p]

Surprised light did much around ICs, unless there was, like, glass body diodes or something.  Are PIN diodes usually in glass format?  Were those..?

I get the feeling that these were all in the IC - perhaps with a transparent top for just this purpose?

[SeanB]

Under the right circumstances silicon junctions emit light, as they are basically a poor LED. They thus can emit light inside the package which, will affect the other sensitive junctions nearby in some cases.

As well most packages these days are thin enough that light can penetrate, though this is likely to be very little in typical use. Might be an issue if you have the package in direct sunlight out in space, but otherwise hardly an issue.

However Vince was likely using a development chip, made in a ceramic package and likely without a lid, so that they could probe it electrically during development. In final use it likely was in a QFP chip on a board, but there it was a LCC carrier. Having an onboard photodiode is used in some secure chips to defeat probing after decapping, if current flows through the photodiode the chip will kill itself.

Glass encapsulated diodes do however make very good if poor quality photodiodes. This is a problem in high impedance use where you might find high leakage or noise if the case is opened.

[T3sl4co1l]

Under the right circumstances silicon junctions emit light, as they are basically a poor LED. They thus can emit light inside the package which, will affect the other sensitive junctions nearby in some cases.

Mmm, it's not like an LED (carrier recombination) -- actually, I don't remember what the mechanism is!  It's an avalanche phenomenon, as far as I know, on the rare occasions when this actually occurs.

In a dark room, a 2N3055 with the top cut off is just visible with ~100mA E-B (it'll drop about 6V and start to heat up..).  It's kind of a yellowish green.  I wonder if it's a continuum spectrum or if it has peaks.

One of the more esoteric bits of semiconductor lore, indeed; supposedly, Pease had a riddle, whereby, in a circuit with absolutely no oscillation and no negative voltage to start with -- just resistors and transistors -- a negative voltage, small and feeble, but negative nonetheless (or if you prefer: a voltage beyond either supply rail -- same thing, upside down), can be produced.

Tim

[SeanB]

Try your 2N3055 using a phone camera which is sensitive to IR light..........

[dannyf]

Not even 3055: it is true for most power amps. And if you google, you will find those pictures of glowing emitters.

It is a well known phenomenon.

[free_electron]

yes this was all inside the same chip. this was a prototype for a new harddisk preamp we were designing. remember that in my previous life ( up till 3 months ago ) i spend 23 years mucking around in silicon ... i have SEEN electrons flow.. ( e-beam prober . you could measure electron density with that thing. discarge an internal cap through a node and it would show you how many electrons were now in the touched area. you say the number decay over time. )
I'm nowhere near the levelof the peopl that actually design the circuits or design the transistors , but i can find my way around a chip and debug the darn things. gimme a probe station ,a few picoprobes, a green / uv laser and a darkroom and i'll prod around in there like the best. i've had setups where is nipped off gainstages ,b rought out the inputs, outputs and control lines and measured a buried diff amp with programmable gain. 7 or 8  needles in contact. requireing drilling through multiple metal layers, cutting passivation , metal , poly. i've done it all. i do understand the principles how electrons travel , how to create effective dams , know they have a tendency to roll where we don't wan t them. i can spot a floating well a mile away ... i can show you where the latchup occurs by smearing liquid cristal on the bare die , exposing it to polarised light and pulsing the chip in latchup so i create a hotspot. liquid cristal changes its spin under temperature. so the polarized light would not bounce back up when cold , but come back when hot. by carefully lettingthe chip go in latchup and cut the current before thermal runaway let the whole thing heat upyou would get a flashing beacon of light showing you where exactly the parasite sat that caused latchup. and then you could go sniff out the well that was not contacted.

the chip was in standard black epoxy body and had the problem that after a few seconds of operation it's gain would suddenly drop. this coincided with the turning off of the configuration block. turn off tha particular section and you could time it with a stopwatch...

so i decapitated the chip shoved it under a probestation , made the room dark and started prodding with a 'light probe' in the form of the laserbeam under the microscope ( the probestations have a resolution down to 200 nanometer so i can see the structures clearly )

there was metal over the structures but a few blast of the laser drilled a hole down to the junction area. turning of the microscope illuminator but leaving the laser illuminator running gave me a spot . by playing with the laser aperture array i can create a rectangle. setting the rectangle slightly larger than the size of the intrinsic area ( i had a sun workstation next to me with the chip layout so i knew what i was looking at ) i could toggle the diodes on and off trying to find out which one was toggling ( i could not turn the control block on because the control block powered solved the problem. the problem only starte a few seconds after the control block went in standby. so i used light to toggle the diodes. once i foudn the one that caused exaclty that drop in gain : microscope light on , look at image , look at layout on the workstation : click the node and up comes the schematic. follow the drive patway for that diode and find it hits a bipolar transistor.. the base drive was in standby. so you have a floating base at that point. that stuff is so high impedant ( teraohms ) in combination with the base capacitance that any stray electron will tickle the transistor in conduction.

so we could do mulitple things :
- eliminate the stray electrons by building a wall of ground (substrate) via's
- add a pull down resistor

all of which would require layout changes to make room for that stuff to be placed.. that meant full maskset . at a million dollar a pop , 12 masks .. ain't gonna happen.

so we altered only one mask to alter the doping of the lower mos so he became 'leaky' essentially a mos with a resistor path in parallel. you can't really turn the mos off completely. that fixed the problem.

this was an oversight of the people designing the control block (pure digital , generated from verilog and plonked down) and the people doing the analog block. when the circuit was connected the analog guys did not know the digital block would go in standby and become tri-stated. the digital guys. well those only know about ones and zeroes . if you show them a resistor they panic ...

so the transistor had a base series resistor , but no pull down ...
not a true problem a , no current means no transistor effect. except in this case there was a leakage current. if that had been trapped by substrate contacts no problem. in this case the electrons went into the base and the damn thing turned on... it took a while to reach the correct biassing but once there wer eneough going in all hell broke loose.

[T3sl4co1l]

Cool stuff!

[atferrari]

Not sure if cool...but close to sci fi for me! 

Wondering what somebody would say to those that commited the "mistake"? Is it? Can you actually blame them? 

[free_electron]

Not sure if cool...but close to sci fi for me! 

Wondering what somebody would say to those that commited the "mistake"? Is it? Can you actually blame them? 

no. it is an oversight that wasn't even caught during design review. shit happens. deal with it. that is life. only when playing with 'refined sand' mistakes become very expensive ... but that is life

[SeanB]

Even Intel has them..... Look at the errata and you will see functions that are marked as "do not use" with various steps and such.

[atferrari]

Not sure if cool...but close to sci fi for me! 

Wondering what somebody would say to those that commited the "mistake"? Is it? Can you actually blame them? 

no. it is an oversight that wasn't even caught during design review. shit happens. deal with it. that is life. only when playing with 'refined sand' mistakes become very expensive ... but that is life

I see.

BTW, how a mask comes to cost 1 million?  What is the essential reason for it? Aren't they just implementing what somebody else (you in this case) has designed already? Shouldn't their work be somewhat an ordered sequence of well defined steps? Sorry but you see I know actually nothing of all that.

[atferrari]

Even Intel has them..... Look at the errata and you will see functions that are marked as "do not use" with various steps and such.

At my level, when I print the datasheet of a new micro I intend to use, I look first for the last errata. Learnt how convenient it is, the hard way.