difference of 1st stage

[clovi_kate]

hi..

what would be the best thing to choose for these two types of 1st stage amplifiers?..i have some samples of different amplifiers especially on its first stage..is it nice to use 2driver transistors at 1st stage or the others that use 1 transistor for audio signal..
what would be the best type when building large amplifiers with regards to 1st and 2nd stage?..


regardsss,
Reign..

[tomahauk]

first example is very not good for hi power amp
second is beter but still not good

in both examples the SR is poor

this will be good for hi power amp

[CBS240]

Hi

Cascode of the input stage is good because it places a finite value on Vce and 'linearizes" the input better. However, I like to reference the bases of the cascode transistors to the CCS on the emitters of the input pair.(resistor to juntion of R8 & R9) and loose the Zener. I don't like Zeners. IMHO, J-fet CCS biasing small metal film resistor || with a small cap, is less noisy. Also a base resistor on Q12 in case Q12 were to saturate, it would not affect the bias of Q7 as much.

[clovi_kate]

any specification of that design?..and also for the value of those individuals..what would the max voltage can this be?..

[tomahauk]

The c300 up close
This 300 Watts/RMS amplifier is meant for those who are not only looking for higher power, but superior performance as well. In order to achieve this, the c300 features some advanced techniques that are absent in its' smaller counterpart.

Additions in 1st Gain Stage

Cascodes
Right at the very first gain stage, cascodes (Q5,6) are adopted. They serve to improve the high frequency performance of the c300. These cascodes are biased to approximately midpoint between 0V and +V by zener diode D1 (33V).

Current Mirrors
The first stage also contains current mirror Q3,4. As the name implies, the mirror forces equal current in the LTP (long tail pair). It is known for its' active loading and high gain properties.

Emitter Degeneration Resistors
Slew rate of the input differential is improved by resistors R6,7,8,9,10. In the absence of matched transistors, preset R10, is used for trimming DC to a minimum at the output of the amplifier.

Buffering the 2nd Stage

The VAS mod
The 2nd stage is direct coupled to the differential via a darlington Q8. This effectively buffers Q10, the main transistor that is amplifying the voltage from loading the preceding stage. Q10 is biased into class A by constant current source Q12. Capacitor C9 sets the dominant pole in Miller compensation.

Thermal Tracking

The remaining parts of the circuit is conventional. Vbe multiplier Q11, adjust the bias for the output transistors which is in full complementary EF configuration. Q11 must be thermally coupled to the main power heatsink for proper thermal tracking. VI Limiting network consists of Q13,14, R25~30 and D3,4. This network is optional, and can be omitted if desired.

Biasing of output transistors
All THD readings were done with outputs biased to 20mV across 0.39 ohms emitter resistor. This works out to approximately 55mA per output transistor in idling state.

[danville]

I agree with Tomahauk that neither amplifier is very good although I would guess that 1 is the better.

It seems that this might be a good teaching moment.

You can learn a bit from the amplifier Tomahauk submitted. He explains some of its attributes.

Lets start with Amp 2. This is a common style of dual diff pair front end. Advocates of this design like the symmetry.

The first horrible mistake is that R12 and R13 are not current sources. They are very poor approximations. You can improve this amplifier by replacing this section with a transistor current source it will take one or two transistors and make a big difference. Q5 and Q6 are supposed to be mirrors. It this circuit they are voltage to current converters (but not goiod ones) Power supply rejection of this circuit is poor. The bias circuit could be improved with an additional transistor as well.

Both amplifiers will suffer from slew rate issues due to miller capacitance of each diff pair. The cascode used in Tomahauks example is used to linearize the input and improve the bandwidth.

Amp 1 looks like a very simified version of Tomahauls example. Take a look at the difference and see how that amplifier addresses the problems outliners above.

You would have seen a lot of these kind of amplifiers in the 1970s. I know, I designed amplifiers professionally inj this period. They were not good then. Certainlty better than Quasi comp of the day (which I still believe is a horrible way to design an amp, but I digress). Maybe they won't break glass at fifty feet but they will not be very impressive. Certainly the bass will be mushy.

Using better parts will not make these amps good.

I have seen a lot of good designs that have been posted on this forum. I would take a look at lots of them, especially the complicated ones, and try to understand why the designer through more transistors at the problem.

A very simple concept is to look at a circuit and ask, could I make this section better. If the answer is yes, usually the amplifier will perform better with the changes. For example, replacing an R that approximates a current source with a more ideal current source will probably improve the amplifier if no other changes are made.

I started out by modifying existing designs. Its amazing how much you learn by making before and after comparisons and studying other peoples circuits . Later on, if you are a good designer you might come up with more novel solutions.

Al Clark