Current flow in semiconductors.

[MajorWest]

Split from this thread:-

https://www.vintage-radio.net/forum/...d.php?t=108226


Hi, folks.

I found something interesting about that book. Did you know it only covers BJT transistors with emphasis on PNP? I didn't know MOSFETS must have come later on. This makes sense as I have a vintage BEREC radio dated 1960 and the transistors will be BJT's PNP type. Maybe 0C49 or something like that and black colour.

Anyway I made an attempt to read through segments and you find the bulk of information still refers to valves as Transistors seem not yet to have damped enthusiasm for the valve. I find the book not to not be as detailed as I like but it's definitely useful. It gives two different connections for PNP and NPN. For PNP I seem to recall (correct me if I'm wrong) that the emitter goes to B+ and collector to neg. Whereas NPN is B+ to collector.

I think this book for me is a goldmine even though I figure it could be better.

[Phil G4SPZ]

Conventional current flows in the direction of the arrow, which is always the emitter. In the PNP device, the arrow points towards the base, and vice-versa for NPN. The most usual 'common emitter' amplifier has the emitter connected to signal ground, usually drawn on a circuit diagram as the solid line at the bottom. The earliest circuits used PNP transistors, and therefore the B+ or battery positive connection was made to what would be colloquially known as 'earth' which is a bit counter-intuitive, and contrasted with valve circuit diagrams where B+ or HT+ were always shown at the top. When NPN transistors were developed, the battery potentials were reversed and the convention of positive at the top and negative at the bottom was restored.

The Americans have always drawn their schematics in ways that can appear peculiar to British eyes. Learning solely from an American text book may therefore cause confusion later when reading British circuits, where terms such as Vcc will be encountered rather than B+ and so on.

Oh yes, MOSFETs were developed much later, after the JFET and the IGFET, I guess sometime in the late 60s or early 70s. The bipolar transistor was first developed as early as 1948, I believe.

[MajorWest]

I always sort of dislike the term "conventional current" although I know this term continues as normal today and even heard students were urged not to use books that prioritise actual electron flow. For me, it just makes it easier to adhere to electron flow process but I'm now aware "holes" flow within the transistor too. It's just when I look at a valve, say, a triode, I see electron clouds around a cathode, heading to the anode (with B+ connected) and on to the battery + but either hindered or forwarded by the grid mid valve.
NPN BJT's were the first ones I looked at. It's a bit of a headache when it's explained. The way I simplify it now is to just consider positive bias on the base of an NPN as providing holes to break down a virtual cell junction (forward bias) so electrons can flow from the emitter and out of the collector. The base I suppose is a bit like a grid. In the book I described in my thread the collector is called the anode in its diagram of NPN. I also try to bear in mind you have various PD's (emitter/base, base/collector, emitter/collector.
PNP is a bit new to me as most examples I looked at were more modern NPN.
The American sets you refer to I think are different as you say. I hard a lot about these All American Five sets. I especially like the Zenith sets and think they are the cream of vintage.

[MajorWest]

Originally Posted by Phil G4SPZ

Conventional current flows in the direction of the arrow, which is always the emitter. In the PNP device, the arrow points towards the base, and vice-versa for NPN. The most usual 'common emitter' amplifier has the emitter connected to signal ground, usually drawn on a circuit diagram as the solid line at the bottom.

Not sure I follow everything you said, probably because "current" has different meanings for different people. The same happens all the time in books. That is, bipolar junction transistors as the name clarifies has conventional + (holes) flow and then - electron flow.
The common emitter you refer to I think is the most widely known. I have a diagram with me now and, as you said, there is an emitter to ground connection. There is also the load resistor between HT+ and the collector. So for me this is pretty similar to our familiar triode valve, except semiconductors do function slightly differently. And where I get confused often is we could either talk about the flow of holes with arrows going one way, or the flow of electrons. Again, as you stated, American books may have the arrows indicating the flow of holes. I try to be careful here and very often scratch my head trying to discover what current a writer refers to.
In closing, I suppose the vintage radio enthusiast only needs to deal with the BJT's. The book I have dedicates most of its pages to valves. Maybe they also figured valves were still the better option in terms of audio quality.

[P.Pilcher]

In the late 60's my University lecturers were describing the rate of drift of the parameters of IGFETS - insulated gate field effect transistors as dependent on the rate at which shovelfuls if coal were delivered to the power station boilers! In those days ICs were made using normal junction transistors and because of the failure rate of such devices a yield of as low as 10% of working devices from a wafer was considered acceptable.
In the very early 1970's it was discovered that the IGFET drift was due to a continuous chemical reaction between the insulating layer of silicon dioxide and the evaporated aluminium interconnecting layer. Once this was stopped by sputtering on an additional layer between the silicon dioxide and the aluminium, IGFETS became drift free and the yield of such "surface" transistors became very close to 100%. This opened the door to LSI and the CPU chips containing over a million switching transistors which we are familair with today. I never did find out what this additional layer was!

P.P.

[MajorWest]

Some of it loses me as I only recently delved into semiconductors. I do seem to recall, though, that metal rectifiers worked via depletion zones and PN junction.
I think I did find ways to make sense of PNP or NPN junctions. The way I see it is when we talk of forward bias, it's a bit like dismantling a virtual cell (usually 0.6 volts). By connecting a battery up so battery polarity counters the virtual cell, the cell is broken down like a seaside wall that crumbles.
Funnily enough, I have one diagram where the base of a transistor has a bias battery and then emitter and collector connect to the main H.T. battery. So, I wondered if early transistor radios had 2 batteries, one for bias. Or was this just a feed from a voltage divider?
My problem is I do too much theory and not enough time spent actually fixing a set. I'm one of those types who stays up late at night going through books and diagrams with a mug of coffee but have hardly soiled my hands yet with an actual repair job!

[Phil G4SPZ]

Perhaps you're thinking too deeply about device physics. Fault-finding and repair is more about current flow, bias and voltage drops rather than hole or electron flow.

[paulsherwin]

Solid state electronics is very difficult for non physicists to understand as a good grasp of quantum mechanics and electron theory is needed. This is in contrast to valve operation where it's possible to envisage electrons boiling off a hot cathode and being attracted to a charged anode - this is only a metaphor and isn't what's really going on at the subatomic level, but it's simple enough to be grasped by most people.

The equivalent simplified explanation for semiconductor devices has always been electrons and 'holes' moving about, but this is very counterintuitive and not a good approximation of what's actually going on.

As Phil says, you don't have to understand the subatomic physics to work with electronic devices. (Just as well in my case.)

[G8HQP Dave]

Trying to analyse a circuit while sticking rigidly to electronic current or 'conventional' current can lead to confusion, especially when you then turn to an identical circuit using devices of opposite polarity.

Better to think in terms of signal currents and bias currents. So the base bias current of a BJT always flows in to the base, the collector bias current always flows into the collector, and their sum always flows out of the emitter. Then the signal current flows into either base or emitter, coming out amplified at either emitter or collector - depending on the circuit.

Alternatively, think in terms of bias and signal voltages. Some circuits are better analysed in this way.

[MajorWest]

Originally Posted by paulsherwin

Solid state electronics is very difficult for non physicists to understand as a good grasp of quantum mechanics and electron theory is needed.

I agree with you so far as physics per se is concerned. Take, AC current flow. How the heck can it "flow" if polarity is reversing at 50 HZ per second? I was once very puzzled and winded up on a physics site where there was an attempt to explain it.
However, I'd like to say that it's not a bad idea to delve into the dreaded holes and electrons because I found it does fall into place.
One thing you will notice is, in a circuit, they often talk about drops in P.D. between 2 points. For example, positive pulse is delivered to the base of a BJT and voltage between collector/emitter of the NPN BJT drops. Why? Well, if electrons flow via the base and exit the collector, said collector is a bit less positive, simply because the more electrons you get at a point, the more negative the potential. That, of course, depends on two points of a PD.
Now here is something: Chas Miller was saying in his book that with valves you often get a leaky grid feed cap. Should a cap leak + voltage onto the grid of an output valve, current goes wild. It can even burn out an output transformer. As we all know well enough, + bias on a grid can dramatically increase current (electrons). So, Chas explains if you short out the suspected cap to chassis, and if you then see an increase in voltage at the anode (via an AVO), suspect a leaky cap to the grid. Why does positive voltage increase suddenly? Well, we removed the leaky pos bias on the grid and this slowed electron flow to the anode.
I accept that, yes, this can give you a ****** good headache. And it gets worse with semiconductors. Also, to be fair, it has little impact on ones' ability to fix a radio. Also agreed to try and picture all these ****** holes jumping around is enough to tip you over the edge.

[MajorWest]

Originally Posted by G8HQP Dave

Trying to analyse a circuit while sticking rigidly to electronic current or 'conventional' current can lead to confusion, especially when you then turn to an identical circuit using devices of opposite polarity

I think where last I left it, was with NPN BJT's you have the base emitter junction PD. This is usually 0.6 volt. The N emitter material has lost electrons to the P type Base so is positive at the junction (vacant holes). The P type base gained electrons at its junction so is negative at the junction. It is a virtual cell of 0.6 volts between layers. If you bypass at a higher voltage to overcome the 0.6 potential hill, clearly something will happen.
That is the base emitter junction. I know we all know this but I'm putting it in terms I tend to understand best.
There is also a base collector P.D. virtual cell of 0.6 volt.
Finally the emitter collector P.D. That's a wee bit like cathode and anode in a valve (well almost).
Here is the kicker. If you bias the emitter base P.D. so battery connections oppose the virtual cell, the virtual cell is broken. Electrons are free to flow to the base/collector junction where the emitter collector p.d. takes over. Thus current flows from battery cathode, to emitter, through the bass (as gate)out of collector and back to anode of battery.
The big deal is that it's the base current that amplifies the emitter/collector current.
P.S. it's easy to confuse these things so don't take my word as gospel on any post but I try to get it as accurate as able.

[G8HQP Dave]

Originally Posted by MajorWest

The big deal is that it's the base current that amplifies the emitter/collector current.

That is not really how it works. Bipolar junctions transistors are actually transconductance devices: base-emitter voltage sets emitter-collector current (the base current is just an unwanted complication). However, the simpler 'current amplifier' model serves to get people started in electronics, as it is understandable by people who don't know about things like exponential functions.

(In theory at least) I understand how transistors work, but I don't think about that when designing a circuit. I just treat them as three-legged devices with certain properties and with inputs and outputs determined by the circuit.

[kalee20]

I agree with G8HQP Dave - it is the base-emitter voltage sets the base-emitter current. Then, if you are lucky, most of this current carries on to the collector. What doesn't, flows out of the base.

Regarding current flow, I'm one of the people who does think in terms of electron flow, when considering direction. Swapping to PNP transistors and P-channel FETs etc is a bit of a challenge in the very few occasions when the -ve supply rail is drawn at the top.

What does faze me completely is the Hall effect in p-type semiconductors - the polarity can ONLY be explained by considering positive charge carriers. Thinking of the current flow as electrons moving in the other direction just does not work. Maybe there is a way forward for the PNP valve after all!

[MajorWest]

Originally Posted by G8HQP Dave

That is not really how it works. Bipolar junctions transistors are actually transconductance devices: base-emitter voltage sets emitter-collector current (the base current is just an unwanted complication). However, the simpler 'current amplifier' model serves to get people started in electronics, as it is understandable by people who don't know about things like exponential functions

To really understand how it works you need physics. So, I try to grasp it in terms I understand.
With the BJT, to my mind, the base takes the role of a valve grid. There are differences as it's bipolar. The current also with a BJT isn't that minute and can be measured in milliamps. With a MOSFET, however, current at the gate is miniscule and voltage creates an N Channel between source and drain.
However, the really simplest way I know of looking at it is, forward bias simply dismantles the P.D. of a PN junction. This is 0.6 volt. They call it a potential hill, which I find a bit weird. Reverse bias just makes the virtual cell stronger. Above all current flow, will alter a polarity status quo.
In physics, I know it's a lot deeper than that.

[MajorWest]

Originally Posted by kalee20

Regarding current flow, I'm one of the people who does think in terms of electron flow, when considering direction. Swapping to PNP transistors and P-channel FETs etc is a bit of a challenge in the very few occasions when the -ve supply railis drawn at the top

I agree. I find PNP confusing too. Especially in view of the fact some transistors have a common collector or common base.

While we're at it, AVC bias derived from negative smoothing in rarer valve sets is another grinder to contend with.

To be honest, as I'm into vintage radio I don't really have to bother with MOSFETs or JFETs. It makes more sense to grapple with the bipolar transistors as these were quite popular in the early sixties. I do actually have some PNP 0C49 transistors on a vintage set. On the lid it says, "This radio was constructed using the latest technology of transistors."

As you say, though, with PNP it's like backwards for me. I think I pretty much have an idea with NPN.

[Herald1360]

Never heard of an OC49. What's the set?

[Station X]

Originally Posted by kalee20

Thinking of the current flow as electrons moving in the other direction just does not work.

I remember our lecturer telling us to think of a row of draughts with one draught missing (a hole). If you keep moving draughts to say the right to fill the hole, the hole keeps moving to the left.

[kalee20]

It doesn't work - yes, granted that moving the draughts to the right allows the vacant hole to move to the left. But when you put a magnetic field across the thing, the moving draughts would get deflected to give a small transverse voltage of one polarity (same as you'd get in a copper wire). And instead, the p-type semiconductor gives a Hall voltage of the opposite polarity.

[Radio Wrangler]

The draught board analogy, or those sliding tiles puzzles or cars in a car-park are also interesting because they demonstrate the much higher mobility of electrons compared to holes.

There aren't any equally-mobile wandering positive charge carriers, and so there is an essential asymmetry in what we can do in semiconductor physics.

Being able to make reasonably matched complementary transistors is quite an achievement, or it's an indication of how far we yet are from absolute physical limitations - whichever way you look at it!

David

[MajorWest]

Originally Posted by Station X

I remember our lecturer telling us to think of a row of draughts with one draught missing (a hole). If you keep moving draughts to say the right to fill the hole, the hole keeps moving to the left.

I finally made a start on the book that started this thread going (Amateur Radio). It made the point that electrons in a valve are free flowing. Just a cloud of electrons in a tube that are attracted to the anode.
As a footnote, I also found errors in the said book (a grid bias battery at 5 volts bias showing the wrong polarity battery connection respect to the grid and filament).
Anyway, I was wondering what is to be made of that. Do you think valves are actually better than semiconductors in terms of efficiency and given the fact current flows in a vacuum?
I do sometimes draw diagrams of base, emitter and collector NPN to work out the flow of voids and direction of electrons (which are not free electrons here).
Usually I start right at the beginning with the virtual cell. P material will automatically attract electrons from N material by virtue of the doped silicon. Thus N material becomes positive at its junction and P material negative at its junction. The limit to this is a depletion zone of 0.6 volt. Whereas bias on a grid is another matter and seems to work more like a deflective, charged shield.
Anyway that 0.6 potential hill is where I start working out what is going on. Clearly reverse bias will simply increase the voltage of the potential status quo of 0.6 but forward bias is where we get more creative in circuit design.