In other words, as mentioned in my 1st comment, yes im being silly and i need to apply a bias voltage to the alongside my ac signal in order to satisfy my base cutoff current and set the qpoint in the center of the operating range.
apologies if some of my questions are a bit stupid, i'm literally doing this for the first time using nothing but a textbook, its quite easy to confuse yourself in such a fashion....

 

Can you show as your amp diagram?

 

No offence intended but just trying to put this in its proper context.
How old are you? What is your level of education? What math, physics, electricity and electronics background do you have?

Designing an audio amplifier is a introductory electronics course at college or university.

That is not to say you need this to build an audio power amp. You can accomplish this by following a readymade design and putting the components together.

If you want to understand how the circuit works and how to design one on your own that is a different matter. In that case you have to study basic semiconductor and electronics theory.

Reading and studying the tutorials here on AAC would be one such approach.

 

Here is part of a table I prepared many years ago for audio amps.



It allows you to take many useful early decisions about an amp design, including the power supply.

Your need to decide whether you are going for a single rail or dual rail supply as early as possible.

Either way the rail to rail (0 to V+ or V- to V+) needs to be about 2 to 4 volts larger than the peak to peak voltage across the speaker.

Note this is only 20 to 22 volts.

With a 30 volts supply you can approach 30 watts.

Also you should look at the current figures in the table. Your power supply will have to supply these currents (2.25 amps for 10 watts)

In general the formula for a power amp is

\(P = \frac{{V_{rail}^2}}{{8{R_{spea\ker }}}}\)

Which allows you to calculate the power for various rail voltages and speaker impedences.

The choice of output transistors will determine how much headroom you need over the minimum peak to peak.

You mention input transistors, but have not said what circuit arrangement you are following. This is often called the circuit topology.

We need to decide this soon to have something to discuss.

 

The whole volume control is not in series with the input transistor. It is connected from the input jack to ground. When turned up for maximum gain, none of the volume control is in series with the input transistor. The impedance of the signal source is the current limiting factor. Still, input amplifiers are designed to use almost zero current from the signal. While you might fancy an all bjt amplifier, I don't think that is a good reason to refuse to use the best parts for the job. A jfet input is almost magical about reducing white noise (compared to a bjt input) and it has incredible input impedance. If you find designing with jfets difficult, a TL071 op-amp meets these needs, too.

 

Hi everyone,
Thanks for the feedback. i can assure you no offence was taken, i need all the help i can get. If anything i get annoyed at myself for making mistakes/asking questions i know to be stupid, due to lack of experience.
I studied acoustic engineering for a year several years ago and have now started on a HNC in electronic engineering.
I have some understanding of the basic principles behind transistors, however i have never put this into use before and have only really dealt with single transistor circuits.
I'm attempting to build a transistor based amp for my project, however it is probably above the level of detail we study transistors to, so I'm having to wing it and research etc to find out more.
Whilst I'm aware that i can just get a schematic, and that there are probably better ways of implementing my ideas and better components for the job, the general idea was to gain a bit of understanding into the use of transistors within amplifier design, the amp itself isn't expected to be anything spectacular.
To be honest I'm in over my head but I'm determined to get something working if it kills me. ^^

i haven't got a circuit diagram with me, but the basic idea is something along the lines of this.

its going to be very very basic. just to get an idea of design really.

I have a max supply voltage of 30V however was actually looking at dropping this to 20 odd as this should be more than adiquate.

The idea is -Q1+2, input pair, minimal gain (unity or maybe 2)
-Q3 VAS, most/all of my 20x voltage gain
-Q4+5 output stage, supplying current for the load.

I've been researching but am struggling with getting small gains at the input pair.
A random example.... (not any of the values i've used, I'm out of town at the moment and haven't got them on me)
Calculating gain.

(C1 = R3, C2 = R4 from the circuit above)

gm = 1 mA / 25 mV = 0.04 A/V.

single-ended gain:

Therefore gain can be altered by adjusting the value of R3 + R4 in the circuit above.
but for unity gain there values would be very small (100 ohms region)
Is this not going to give massive currents at the collector?
what can i change to alter gain without affecting biasing etc?

many thanks in advance,
Will

 

that image didn't work so heres a link
http://upload.wikimedia.org/wikipedia/commons/f/fd/Amplifier_Circuit_Small.svg

 

Well the input stage gain is
Av1 = (R3||RinQ3)/(2*re)
Where:
RinQ3 = (Hfe + 1) *26mV/IcQ3
re = 26mV/IcQ1
Also do you know the difference between open loop gain vs closed loop gain ??
Also do you know that if you design the circuit properly the voltage gain will be equal to R8/R7 ?? And for single supply you need a capacitor in series with R7.
And with 20V power supply you do not get 10W at the output. You need Vpeak = √(2 * 4Ω * 10W) ≈ 9V this give as 18V peak to peak. But saturation voltage for this circuit will be at least 5V or even more.

 

that image didn't work so heres a link
http://upload.wikimedia.org/wikipedia/commons/f/fd/Amplifier_Circuit_Small.svg

Here is the image:

 

You really ought to check out some of Rod Elliot's power amp projects:

http://sound.westhost.com/projects-0.htm#pwr

 

Also look at the Sinclair Z-30 audio amp:

http://forum.allaboutcircuits.com/threads/30w-2n3055-audio-amplifier-circuit-question.30721/

 

I understand the difference but haven't really considered its implications in my design so far, its all been trying to get to grips with the gains.
i was under the impression that if i aim for an open loop gain of around 20 and then make my feedback resistors give around a 20th of this to the base at Q2 i should get something approximately equal to the gain I'm after?

so i can calculate Rinq3, set re to give the current i want at the emitters and then vary R3 to get the gain i need?

are you saying I'm going to have problems with transistor saturation at values way below my desired output voltage?

 

How about something like this:

or this:

 

I think I'm naming a specific here. Gm is not in the first stage gain equation, right after you make the negative feedback loop with R8 and R7. Excessive gain that you are not using is what reduces distortion. Let the input pair run free and quench the excess with R8 and R7.

Yeah, I know that's not the technical way to say it. I do this because math intensive study doesn't reveal the practical side to beginners. I've taught many a BSEE practical applications that they didn't learn in school or derive just from doing the math.

 

so your saying i set the gain for my input pair higher (without feedback) then the feedback attenuates this to the levels i want (which i can set by varying R7 and R8) whilst simultaneously reducing my distortion.

I would love to just grab a schematic like this and make a better quality amp, however i believe i need to show calculations for resistor values etc. I've said i'm using a long tail pair in my spec however i may make some revisions to this if things cary on as slowly as they are currently.

 

I understand the difference but haven't really considered its implications in my design so far, its all been trying to get to grips with the gains.
i was under the impression that if i aim for an open loop gain of around 20 and then make my feedback resistors give around a 20th of this to the base at Q2 i should get something approximately equal to the gain I'm after?

No wrong . Why you want Aol -open loop gain equal 20V/V ? And in real life we want the open loop gain to be as large as we can get.
And then we can set Acl - closed loop gain at 20V/V.
Acl = Aol/(1 + Aol *B)
where
B = R7/(R8 + R7)
So only if Aol >> 1/B the closed loop gain approaches 1/B
And this is why in real live we don't care about firs stage gain. Because in this type of a amplifier Aol is mainly set by Q3 voltage gain.

so i can calculate Rinq3, set re to give the current i want at the emitters and then vary R3 to get the gain i need?

Yes, but this will not be easy, because the voltage droop across R3 is almost constant equal to VbeQ3 = 0.66V, so IcQ1 ≈ VbeQ3/R3 (I ignore Q3 base current)

are you saying I'm going to have problems with transistor saturation at values way below my desired output voltage?

Yes, exactly

 

If the base of Q1 receives a positive going signal, the feedback loop will raise the base of Q2 to almost the same voltage. (If the base of Q2 was raised to exactly the same voltage, there would be zero gain in the differential stage.) The lack of raising the base of Q2 to exactly the same as base of Q1 is what allows some gain in that stage. The gain will be exactly enough to drive the following stages to a high enough voltage gain to drive R7 and R8. Thus, the bases and emitters wag up and down on the oscilloscope screen, but the voltage across R3 changes very little.

ps, If this home-style version doesn't suit you, just say so.

 

thank you, so I'm getting the Aol Acl idea, actually makes my design easier, as i can just aim for good gains at each stage.
why will i have/how can i deal with these saturation issues?

 

Sorry. I'm not math enough to see where you have saturation problems, but I can say: Look at base of Q3. It's one Vbe off the supply rail. If R3 only has to support 6 or 7 tenths of a volt, what's saturating?

 

this may be another completely idiotic statement, but can i not just stick a reasonable size cap before q3 and then bias it using a potential divider?