Oh yes indeed, and in glorious stereo
This is how its done:
Welcome to the (virtual) listening room.
I have National Semiconductors Audio Handbook from 1977.
There are suggested circuits in this under the heading Boosted Power Amplifiers which shows two examples one of which broadly resembles the circuit drawn up by Ian Hegglin.
This is based around the "old" LM378 dual op.amp rated at 2W capable of 100 m.a. output currents.
The main difference between this and Ian's illustration is the placement of the 15 Ohm equivalent series resistor to the co-joined base junctions of the power devices
In National's circuit this is a 5 Ohm resistor in parallel strung between the bases of the output transistors and the emitters. In this way the LM378 will drive the load directly via the 5 Ohms.
Should the voltage drop across this fall below the threshold of conduction of about 0.53 Volts, the LM378 will also add feed as in Ian's illustration via the output transistor base emitter junctions.
One would have to consider that the first circuit shown in post 6 used germanium transistors and these have less abrupt turn on characteristics than silicon transistors.
The question is how do silicon transistors fit into the mold of the 1962 design and why was there no further development. I think that would have been explored at the time.
The question is how do silicon transistors fit into the mold of the 1962 design and why was there no further development. I think that would have been explored at the time.
Sorry. These are not terms.
When I say a transistor is nearly, fully closed, I mean, there is some collector emitter current, which, may be insignificant. Fully closed means, theoretically, 0 current, practically, just the leakage.
When I say a transistor is nearly, fully open, I mean, there is some collector emitter voltage, which, may be insignificant. Fully open means, theoretically, the collector emitter voltage is 0, practically, Ucesat.
I do not know. The only thing I know is 20KHz is a very low frequency for almost anything not mechanical. I say this, not because of your post ( as I said : I don't know ), but, because, there are some people in this forum who really think 20KHz is a high frequency. Strange. 20KHz is nothing even for modern tubes.
LM378 is, probably, very powerful. TI have some extremely powerful IC amplifiers ( 40W ), probably, still, extremely expensive.
Yes. And this is why I have made another thread here to ask for Germanium transistors.
Although I have been able to find something, I continue to ask for your help. In case you know Germanium transistor numbers and names, please, be kind to post them there. This will be extremely helpful.
These devices are a rarity now, I did a search "buy germanium transistors" which gave some results however npn complements for the predominant range of pnn types gave no obvious npn matches. l have Towers International Transistor Selector issue of 1977 which includes germanium devices.
Germanium types in the 2n range have part numbers below 711 while others have prefixes A as the first of two letters. 44 years on to source these would involve searching surplus stock suppliers for out of production items or offerings on ebay.
Below a Vb of about 0.53 there will be no IC current and thus a dead conduction zone arises.
Above that an IC to Vb plot will curve up to Vb of about 0.7V and tend straighten out above that level - linear meaning straight and not curved.
In the first circuit the Integrated circuit is feeding the output transistors through non linear base emitter junctions along the switching process.
Although NPN and PNP transistors are nominally complements, there are still some differences.
If Silicon is doped with Phosphorus to create an electron rich N zone and Aluminium is used to dope an electron deficient P zone, the respective levels of dopant for each will have to differ as it would using these materials in reverse to create a complementary transistor.
This is to do with the energy bands in an atom where electrons are bound tight so to speak or loose. You can check the atomic structures of the materials mentioned.
It is also worth mention that emitter diodes have a resistance which is non-linear. So the voltage drop across that will be non-linear. The resistance is quite small but still a nuisance for that reason.
Adding an emitter resistor in series is helpful because the voltage drop across the combination will be largely due to the added resistor. To compensate the output stage then needs some means of biasing the output transistor bases.
Last edited:
Thanks. Thank you for the number range.
Germanium transistors are still sold on eBay, AliExpress and everywhere. Some of them are old stock. Looks like, NTE still makes them and say these are for replacement.
There is a huge availability of Soviet Germanium transistors. I do not know whether they still make them, yet, they are excellent.
Please, note, one of the functions of the amplifier is to ensure, theoretically, perfect linearity. Also, the transistors are in buffer configuration, which, means only the variation of Ube ( Ice ) ( and Ube ( Ibe ) ) will introduce non linearity, which nonlinearity, will be compensated by the amplifier.
In some cases, Ube has to be >= 0.6V for NPN power transistors ( 2N3055 measured ) and >= 0.62V for PNP power transistors ( MJ2955 measured ). All at 25C.
In regards to Rbe, when between the amplifier and the output, this is supposed to make the amplifier regulate at very tiny voltages. However, thereafter, the amplifier still have to jump. The trick is to make the resistor so low, so the amplifier can regulate to, say, 0.7V. This is not possible with normal amplifiers, low resistance speakers and high voltage amplitudes.
Except from the power consumed from the output of the amplifier, the higher the consumed current, the lower the maximum voltage of the amplifier.
LM7171 and LM6171 lose around 1.5V at very low current, around 2V at 15mA and much more. Most likely, a maximum current between 15mA to 20mA can be considered.
Also, Ube increases with high currents, hopefully, the control transistor will not see too much current of a Darlington must be used for the second Sziklai.
Thank you, Sir! This is immensely helpful. I will check all of them ( your first post and your second post ) and reply in the other thread, called Germanium Transistors! Help!
Mistake : replace " jump " with a " steep climb ".
The best way to understand the base emitter resistor is to assume a perfect amplifier as a buffer, then, a resistor ( which will act as the base emitter resistor ), then, the feedback connection, then, the speaker WITHOUT any transistors.
The resistor and the speaker act as a voltage divider. In case the transfer coefficient of the voltage divider is, say, 10, then, the when the amplifier reaches the maximum voltage, of, say, 15V, the output will be at 1.5V. Thus, the amplifier will be able to regulate between - 1.5V and 1.5V.
The transistors are supposed to react before the amplifier reaches the maximum voltage and the system gets to work normally.
The best configuration of the transistors may be Sziklai with one primary transistor and one or many secondary transistors. When the secondary transistors are more than one, they can be in Darlington and the system will be a Sziklai Darlington configuration.
Therefore, although in steeper climb, the base emitter resistor helps.
However, this is not the problem. The problem is different : the machine gun noise ( I made up this " term " ). This is a banging noise, one or two shots per second.
This may be only in the setup I have and not a global problem.
The problem does not depend on the transistor and the amplifiers, because, I have tried a few. I have tried base emitter resistor as far as I remember. The problem persists. Most likely, the problem is caused by oscillations.
This problem does not appear to be present when ALL of the transistors are always open and never closed with the same setup.
To get rid of the problem, an extremely low ESR, NTE, Ceramic, 1uF, 100V capacitor was used. This capacitor, also, acts as a start up capacitor for the transistors.
I have discussed these things in other threads I have published here.
With Germanium transistors as primary, the leakage may keep the secondary transistors open, yet, the primary, Germanium transistors will be closed around 0V. However, Germanium transistors have low Ube, as low as 0.2V nominal ( at 25C ) and seem to have similar temperature coefficient as Silicon.
Ube of 0.2 will make the amplifier work in low signal mode around 0V, which, will make the switch immensely fast.
Anyway, yes, Germanium transistors are amazing.