ВЭФ 12 (Astrad VEF 12 Project) - including a transistor tutorial

[Mark Hennessy]

This might be controversial, but...

If you want to learn about how transistors work, forget about old text books. They are next to useless in my experience (happy to be pointed towards an exception). They were written by people who were too stuck in the valve era and didn't really understand how to apply them (fair enough, in a way), and spend far too much time banging on about physics. None of that helps you to understand what's happening in a radio.

I'm not dismissing the physics out of hand, but I've never understood how that helps beginners - I suppose the academic argument is to create a bridge between something you already know and something new, but why assume that a budding electronics engineer is already a highly qualified physicist? But while I find the physics is really useful for understanding some of the more obscure stuff that comes up later (e.g., Early effect), I'd really advise you to forget about holes and electrons, and focus on what the transistor actually does for the time being. Come back to the physics later if you like, but it's not essential, even for quite advanced circuit designs.

There are lots of modern text books and web sites that give good explanations. Start by downloading this: http://wouterjan.deds.nl/The%20Art%20of%...20Hill.pdf (it's about 27MB, but worth it!)

Fundamentally, transistors appear to work in two ways:

1. The current control method. A small base current controls a larger current in the collector. Look up "transistor man" on page 64 of that PDF.

2. The voltage control method. How it really works. The voltage between base and emitter accurately determines the collector current.

Sadly, many people believe that the current control method is the way transistors actually work, and you will encounter many arguments on forums to that effect. Many text books also assert that. Sure, when you apply a voltage to the base-emitter junction, a current flows, and that current is roughly proportional to collector current. The ratio between the two currents is called Beta, or hfe. But Beta (or hfe) is a very poorly defined parameter that varies between device sample, temperature, collector current, planetary alignment, etc, etc. Whereas the theory that drops out of the voltage control approach is dependable and precise - much more believable and useful.

In practice, you need to think about both when analysing and designing circuits. But you treat the current control method with the caution it deserves, and avoid designing circuits that require a specific value.

The PDF covers a lot of ground that can be skipped at first, but goes on to discuss the voltage-control method on page 79. It introduces the Ebers-Moll equation, and gets to the point that the transistor is a transconductance amplifier (voltage in, current out - just like a valve). For a transistor, the transconductance - gm - is simply 40 times the collector current. Knowing that, it's really easy to understand (and design) transistor amplifiers.

Another really good book is here: https://books.google.co.uk/books/about/A...edir_esc=y

A lot of our electronic fundamentals training here at work was based on this book. I'm not aware of a PDF download, but the Google preview gives you a flavour.

I appreciate that this post will take some digesting, but if you want to fix transistor radios, you need to start at the beginning (but skip the physics!). I'd recommend doing some experiments using a breadboard, but whatever works...

[Mark Hennessy]

Just another point. It is hopefully "stating the obvious", as it applies equally to valve circuits, but just in case...

DC conditions first. Then add signal.

In other words, you must always bias your valve or transistor before it can amplifier the way you want. The AC signal basically "wobbles" the circuit either side of the quiescent DC condition.

The signal at the base-emitter junction of a transistor consists of a DC offset and and AC signal. Learning to mentally separate these two components so you can consider each separately is probably the biggest favour you can do for yourself.

That DC offset is likely to be close to 0.2V for a germanium transistor (0.6V for silicon). Give or take...

If there is an emitter resistance, it'll have a DC offset on it. More on that later if needed.

[ppppenguin]

Mark, really excellent approach to learning about transistors. Quite right about not needing to understand the physics first.

Minor caveat: Mark is quite right about the voltage control aspect of transistors but this is really only true for common emitter configuaration and, with slightly different calculations, for common collector. Not true for common base. But then a valve in grounded grid is a current controlled device too.

What some of you may not know is that Mark is a professional educator, teaching BBC engineers about aspects of their craft. He's very good at giving lucid and relvant explanations.

[Nowhere-Man]

Mark, I will get back to your other points gradually. It's just kind of ironic (and interesting) you addressed something that was on my mind for some weeks. That is, I discovered I don't have an engineering approach but a strong theoretical one. And I agree without the actual experience of messing with radios, you don't learn how to do repairs. As I discovered that I literally had to force myself to get the soldering iron out and a test meter. This has helped. I am going to continue to try and develop these engineering skills even if I work more slowly. Meantime the bookish side still seems where I feel more relaxed as somehow I love messing with sums and ratios and comparisons. My girlfriend calls it "numbers" and laughs I have OCD. The biggest weakness at the moment is my lack of a diagnostic method as well as tools such as Signal Generator. Yes, I've got my hands dirty with radio work but am a bit weak at systematically tracing faults. However, I am more aware of it now. My hope is to try and balance the theoretical stuff with forced project work and I keep reminding myself I must get some decent glasses to see better.

[Nowhere-Man]

Just another point. It is hopefully "stating the obvious", as it applies equally to valve circuits, but just in case...

DC conditions first. Then add signal.

In other words, you must always bias your valve or transistor before it can amplifier the way you want. The AC signal basically "wobbles" the circuit either side of the quiescent DC condition.

The signal at the base-emitter junction of a transistor consists of a DC offset and and AC signal. Learning to mentally separate these two components so you can consider each separately is probably the biggest favour you can do for yourself.

That DC offset is likely to be close to 0.2V for a germanium transistor (0.6V for silicon). Give or take...

If there is an emitter resistance, it'll have a DC offset on it. More on that later if needed.

"The AC signal basically "wobbles" the circuit either side of the quiescent DC condition."

I sort of had some issues lately with signal voltage and DC bias. Here is where I got curious: In a valve radio, you will know the signal that hits the antenna is voltage-wise very weak. Millivolts. So, the first stage is to amplify the signal. In the very old TRF's of course the first valve would be the HF amplification. The modern sets just use the Mixer Triode Hexode. At any rate, to get to the point we wind up with an IF that passes through the IF transformer as, say, 455 KHz. Normally the IF transformer is shown as primary winding between HT and the anode of the Mixer (a hefty voltage). The secondary will connect between the detector grid and either AVG (or is it AVC) or chassis. Here is the point: The detector grid is DC biased at less than zero volts. Thus you only need very small amplitudes of AC signal to drive the grid into more positive values. In AARL, the example used is a grid at - 5 and a signal input of 2 volts+. We're told the grid will swing to - 3 on positive waves or - 7 on the negative amplitudes. For every positive swing, current is going to increase. What had me foxed is a very simple observation. If the actual IF amplitudes are at high voltage, how do we wind up with a signal on the grid at just, say, 2 volts AC? My only guess is that the sheer speed of the alternating cycle (1000's per second) are too fast to literally push the grid over the zero bias. So maybe it's not a good idea to try and apply graphs too literally? Of course, then we have grid leak which also balances out the bias. Also, possibly the capacitors may attenuate the signal. Even so, no textbook seems to have addressed this and seems to assume it's too obvious to highlight. However, it bugged me. So, your remarks about not mixing up DC and AC remind me of that. What I do know is there's a voltage drop across the diode load resistance that responds to the grid swings as current wobbles about. So, we eventually get the audio decoded. Now with transistors, the principle I imagine is the same but the holes and so forth, plus the smaller voltages required are no doubt a science in itself.
Agreed, though, none of this is needed in repair work.  

[pwdrive]

There won't be a high signal voltage at the anode of the mixer, only a high DC voltage, by the time the signal reaches the detector the signal will be large enough for further AF amplification by two AF amplification stages at most to give sufficient volume from the loudspeaker once the signal is detected but not large enough at the detector to cause overloading of the detector if the receiver is designed properly.

As Mark said, there's the valves DC conditions then there is the dynamics of the AC signal pertaining to the valve.

A weak signal at the antenna would be considered as microvolts or part thereof, not millivolts.

Lawrence.

[Mark Hennessy]

NM,

Your journey will be much easier and quicker if you start with audio frequencies first. Keep it simple.

I prefer to not draw too many analogies to valves. First of all, there are too many types, secondly, the more I learn about valves, the less I understand. Let's regard valves as basically OT for this thread.

Attached is an image that I have shamelessly stolen from the internet, and added some component values, etc.

Your mission is to work out the DC conditions of this circuit. In other words, what are the voltages at the base, emitter and collector, with respect to ground?

You are allowed to ask for hints and help. This is an "open-book" test - you are allowed to use the internet and text books to help. But you must explain your thinking to demonstrate understanding. And as much as possible, stay focussed on this task.

No one else is allowed to chime in with the answers. This is about helping NM. I - unusually for me - reserve the right to edit or hide posts that don't help the narrative.

Best of luck. It's not too difficult, I promise.

[Nowhere-Man]

I'll see what I can do. At the moment I'm viewing it on my Blackberry but it's blurred. I saved and zoomed but still not clear to see. So I'll print it at the library. There won't be any problem on a library PC.
Today I hopefully am ordering transistors from Russia. St Petersburg where I lived. Possibly I may have some to spare. I wanted the original USSR types. I know there are alternatives but we shall see.

[Mark Hennessy]

Clicking on the attachment or this link should give you a clearer view: http://golbornevintageradio.co.uk/forum/...?aid=15790

[Nowhere-Man]

That's fine. I can see the voltage divider and the rest of it quite clearly. I promise to get back to you on it and see if I understand it. Any questions I will ask.