Transistors

[Nowhere-Man]

(3) Beta is pretty much similar to amplification factor in valves.

Sorry, no it's not! Amplification factor in valves is a well-defined characteristic relating to the effectiveness of the anode and the grid at controlling anode current. Beta is a highly variable characteristic comparing collector current to base current.

It would be possible to define amplification factor for a transistor (Early effect would play a part), but not beta for a valve.

I didn't say amplification factor was identical. It's just the closest you can get to grid/base, collector/anode. I believe this was already discussed a few pages back (current vs voltage amplification). Therefore, I won't go back over old "ground" or pick any "holes"

I sometimes think it would have been better to hand-pick my vintage transistor radios as there are many very easy designs. My Bush, for example, has OC transistors point soldered to a chassis and pretty much fits the same model where valves were used. It's dated 1959 and still works fine. If I recall, it may not even have a ferrite aerial.

One point we haven't debated is something I was looking up while walking my dog last night. This is changing the base bias of a GE transistor and fitting a Silicon. This would make life easier for me as I got kind of tired chasing around after GE transistors and being directed towards E-bay. Let's take the example I posted earlier in the thread of a 3 volt supply through a voltage divider. Let's assume we have a PNP GE transistor that needs 0.3 volt to "kick in". We located a suitable silicon equivalent but that requires 0.6 volt. Thus, by my reckoning, substitution without thought would mean the base of a silicon transistor staying idle because only 0.3 volt is reaching the base. What about altering the voltage divider? Or is there some known loophole to enable the base bias to be adjusted?

[Terry]

My Bush, for example, has OC transistors point soldered to a chassis and pretty much fits the same model where valves were used. It's dated 1959 and still works fine. If I recall, it may not even have a ferrite aerial.

*panto-mode* Oh yes it does! */panto-mode*

It is also quite a substantial size compared with many others of the same period and is unusual in that both MW and LW coils are connected in parallel for MW, increasing the sensitivity.

If, as I suspect from your description, it is a TR82, it is more closely related to its mains/battery valve predecessor - the MB60 - than you might realise.

Take a look at the back. Notice the flat oval indentation at one end? In the MB60 this is where there was a hole to accept the mains socket on the detachable mains lead. Now look at the inside of the back. See those raised circles moulded into the plastic? This is where circular sorbo-rubber pads were glued. They had a central hole so that they fitted over the sealing pips of the valves, keeping them firmly in their bases when the back was screwed on.

Actually, the positions shown are for the first version of the MB60 which was later redesigned to take different batteries with a different chassis layout, thus meaning that all the valve positions were different. The sorbo pads were still fitted - obviously located by a template - although they were no longer located in the moulded circles!

Apart from the red rexine on the central part of the MB60 body and the hole punched in the back, the dial,knobs and entire cabinet are identical.

That mains socket area came in very useful on the TR82 as we received numerous requests to fit an output to feed a tape recorder. The obvious place to put the sockets was in that oval depression!

[Mark Hennessy]

This is changing the base bias of a GE transistor and fitting a Silicon. This would make life easier for me as I got kind of tired chasing around after GE transistors and being directed towards E-bay. Let's take the example I posted earlier in the thread of a 3 volt supply through a voltage divider. Let's assume we have a PNP GE transistor that needs 0.3 volt to "kick in". We located a suitable silicon equivalent but that requires 0.6 volt. Thus, by my reckoning, substitution without thought would mean the base of a silicon transistor staying idle because only 0.3 volt is reaching the base. What about altering the voltage divider? Or is there some known loophole to enable the base bias to be adjusted?

A lot of the time, there's nothing to change. Many circuits designed for germanium work just fine with silicon replacements.

If we'd stuck with the tutorial a couple of months back, we would have got to that. In short, the circuit with 4 resistors is known as "fully stabilised bias", meaning that the DC conditions don't change radically. To prove it, just build it. Or, re-run the exercise, but substitute 0.3 for 0.6V.

Just like designing for a particular value of current gain (Beta, or Hfe) is bad practice, it's also bad practice to require a particular value of VBE. Some circuits do - and might require matched or selected transistors - but you're not likely to encounter those in a transistor radio.

Remember; 0.3 and 0.6 are just simplified assumptions - in reality VBE depends on base current. Just like the voltage dropped across a diode depends on current...

There are places where things are more critical, but in a typical transistor radio, we're only talking about the output stage where you might expect to re-bias.

[Kalee20]

That would be really instructive! Yes, changing the Ge for Si should require a bit of bias adjustment, but like Mark says, it probably won't matter.

The notes from a few pages ago, recommended 'wasting' about a volt across the emitter resistor, and then if you add the base-emitter voltage (0.3V? Although I'd probably allow 0.2V) you need 1.3V on the base. If you swap the germanium device for a silicon type, without changing anything, it'll gobble up 0.6V of this 1.3V before anything happens, thus you'll only have 0.7V across the emitter resistor. So the silicon device will pass 30% less current, but that may still be enough to do its job.

You could return the current to the original design by reducing the emitter resistor, or altering the base resistors, and seeing if it makes any difference. In an amplifier stage which is coupled to the next, by capacitors or transformers, you won't upset anything by doing this, and it can be fun and instructive.

[Nowhere-Man]

The very old engineers in 1940's used to put a variable resistor in the grid. F Camm used this method in the very early days of TRF. Anyway, for now, I will just not overly worry about the bias. A bit later I may reboot that schematic of my VEF to see what comments are made. I still need to repair the RF stage but the circuit was a bit ahead of me. I am hoping when I return to it I'll have gained a bit more insight into how transistors work and the currents involved.

[Nowhere-Man]

A quick glance at the VEF schematic last night indicated the output stage is push pull. That makes sense as battery life is always a factor. Last season with that radio I got as far as determining the audio stage is OK but the RF stage has an issue. I went to a lot of trouble to try and get suitable, spare GE transistors and found Ebay has a near monopoly on GE sales. These transistors simply pull out of the sockets so can be tested but I wanted a few spares first in case the legs fracture. At this point the boat season kicked in and radios put to the side.
Push pull for me with valves is very clear. Chas Miller noted the early battery valve receivers used it to save battery power.
Surely anyone who manufactured retro GE transistors would have a good company. Even in Russia the P Top Hats are hard to get. There is a different sound too as my Bush 1969 tranny sounds really good.

[Mark Hennessy]

Sorry, but the idea of "germanium sound" has rather more to do with the circuit topologies employed at the time rather than the devices themselves. And in a complete radio, it's the loudspeaker/cabinet combination (plus any EQ included) that dominates the overall sound. There are threads on UKVRRR with my measurements to demonstrate that.

I really don't understand the romanticism around germanium (Ge, not GE). They worked amazingly well, considering, but the world has moved on. I only use them for sympathetic restorations, but would have no hesitation in fitting silicon if original germanium were not available.

[Crackle]

Oh thanks Mark, I thought MW man was fitting "General Electric" transistors, I couldn't work out why he wanted to be so particular with an old Vega/Vef radio.

Mike

[Nowhere-Man]

I can't comment as to whether the circuit or the GE transistors create the difference. My ear though is pretty sensitive to pitch and acoustics. My feeling is just based on sound is audio quality has regressed. With decline in pop and rock I think consumers are less fussy over audio quality so sound systems are mass produced. I do know George Martin did lots of experimentation using just the ear and from experience recording I know why. A recording will sound different from player to player so often tests would include audio representation on just a cheap tranny. I imagine Lyle Mays would have a lot to add as he repaired synths and used them. The vintage Moog synths still cost a bomb today and first came out in the late sixties.

[Nowhere-Man]

I browsed online generally about GE transistors. Most audio engineers agree with Mark that silicon far outperforms germanium. Gain and higher frequency range were quoted. So really demand for germanium is rooted in the music sphere, especially pedals and fuzz. Musicians can be fussy over audio detail - there is still a certain awe over the Hammond tone wheel rock organ. At any rate I am forced to agree the GE transistors were less advanced although I loved the way they were brightly coloured in yellows and reds, as well as the larger size.
Someone commented germanium transistors were now in the hands of the snake oil vendors and fly-by-nighters which is how I started to feel. The GE transistors I now have came from my own car-booting at little cost but are sold online up to 15 quid! I have a fair few of the dreaded AF117's, the mention of which causes people to make the sign of the cross and retreat.