I havent had much experience with audio circuits, so i thought id jump in and have a go. Its much tougher than i though it'd be and i have sooo many questions, be glad of the help!

I have a speaker that is 8ohm (0.3W) - using v^2/R - will that give me the Maximum RMS voltage i can apply to the speaker? This must be AC coupled to the speaker so that no DC passes to it, because that heats up the coil and may damage the speaker right?

Ive read, from the way the speaker works, that its just a coil moving in the magnetic field of an armature magnet, so is this 8ohm, just the DC resistance of the coil, or.. the AC reactance of the coil? If thats the case, i thought this inductor (because its a coil?) is a function of frequency how can it be fixed at 8ohm? - i read a few posts that suggest that this changes at resonance or something!

Lots of the circuits ive seen on the net have mentioned alot about Class A and Class AB operations. So from what ive read, you can tell a circuit is Class A because the output transistors are on all of the time, and Class AB if one transistor is one during the positive of the cycle, and the other is on during the negative - as in a push pull output stage - is this right?

Im trying to simulate amplifier circuits, but im not sure what to use as the input, im not sure how to model 'sound'. I understand its a mixture of frequencies 20hz - 20khz, and will have a maximum voltage depending on the type of input. Can i just use a sinewave input at these frequencies to simulate my circuits, or do i need to take anything else into account?

I know if you pull too much current out of a battery its voltage drops, so how do i get around the problem of supplying power to an output speaker rated at 300mw, can a 9v battery do that? Is there a minimum power i can deliver from a battery and i should choose lowered powered speakers to make sure the rail doesnt drop - or use big capacitors across them to hold them tight?

Thanks for any help or points in the right directions.

Megamox

 

yes .. the coil shld be ac coupled , if DC component is there , the coil movement will be unbalanced and cause distortion.
I have heard that most of the spkr has ~10% efficency , ie 90% power is wasted as heat !!..

the DC resistance of the coil will be less than that. the coil resistance is measuread (AFAIK) at 400Hz ( if i remember correct)

well what u said abt class AB has a slight mistake(practicaly), ie it is a class B , for an amp to be a class AB , a sight forward bias is also needed for the push pull o/p stage (remember abt the 0.7 vlt drop at base), else waveform will be chopped at bottom.

yes u can use sine wave
a power amp needs burst of power , so use a high value capacitor as buffer

 

What kind of Whats? RMS continuous Watts, "maximum" (which is simply double the RMS number) whats or "music-power" momentary whats?
For how long a duration?
At what frequencies?
At how much distortion?
Speakers have a mechanical power rating (the coil hits the end of the magnet structure) and an electrical heating rating.

Correct. DC also causes the coil to be offset in the magnetic gap causing awful distortion (but some guitar players like it).

The impedance rating of a speaker is at its lowest, just above its resonant frequency, where the low inductance doesn't raise it much like happens at high audio frequencies where it could be 20 ohms. At resonance its impedance could be 40 ohms or more. Its DC resistance is probably 6 or 7 ohms.

Correct.

You need to use many sinewave frequencies. You also need to combine two frequencies to measure intermodulation distortion. You need to use many levels to see an amplifier's behaviour at clipping and crossover distortion at low level. Frequently squarewaves are used to see if an amplifier is approaching instability which causes ringing on the transitions of the wave. Frequently an amplifier is tested with 2uF across the load to simulate phase-shift caused by speaker resonances or a crossover network.

An audio amplifier isn't used at max output all the time unless you like to play a siren. When it clips the peaks slightly at max output, its average power is probably 1/10th with most music or speech. A little 9V alkaline battery should last a long time (go to www.energizer.com to see) with a drain of only 30mW (maybe only 8mA including idle current). At max output of 300mW a little 9V alkaline battery is fine to provide momentary 33mA. A battery's internal resistance increases as it runs down so a big (1000uF) capacitor across it helps to keep its voltage from flucuating too much.

__________________
Uncle $crooge

 

A battery's internal resistance increases as it runs down so a big (1000uF) capacitor across it helps to keep its voltage from flucuating too much.

... although the average dc output will continue to diminish. It's only the voltage variance caused by the very short-term load changes that will be helped by the big cap. The battery's still going to "run down" or "go flat" depending upon which side of the pond you're on.

Dean

__________________
Dean Huster, Electronics Curmudgeon
Contributing Editor emeritus, "Q & A", of the former "Poptronics" magazine (formerly "Popular Electronics" and "Electronics Now" magazines).

R.I.P.

 

I've tried Energizer's new disposable lithium AA cells. Its too bad they don't make them in the 9V size. They keep going and goin' and goin' and goin' and goin' and ... :lol:

__________________
Uncle $crooge

 

Thanks for the help guys. The speaker i have (i think its from an old broken portable radio) just has '8ohm 0.3W' imprinted on the back of it, thats all i know about it. I just didnt want to put too much power into it, so i wanted to know what the maximum voltage i could feed into it was. So im assuming if i feed in no more than 2v peak (~1.5VRMS), this works out at about 0.25W RMS power. Ive drawn out a design to see if i can do something basic like amplify the signal from an electric mic (20mV) and output it to a speaker, using a 9v battery.

The output transistors are power transistors, rated at 1A. The others are BC547B, gains of about 300. The resistors on the emitter's are there to stop thermal runaway, i wasnt sure what value to use, as ive seen many different values in other designs, so i put in 8ohm, to try and make output impedance = input impedance of the speaker, and have maximum power transfer.

Originally i calculated to 470k to be 5M, to bias the collector at mid voltage and get maximum swing, but this gave a non symmetric output. I think the combination of the 100n coupling capacitor and this resistor might have been acting like a filter. I didnt want to reduce the 100n because i wanted to couple the low frequencies from the input, so i reduced the resistor down to 470k, which seemed to work.

With an input of 20mV, the peak voltage across the speaker is about 1V.
Would this be a good circuit to try, or can it be done much simpler.

Thanks in advance.
Megamox

Attached Images

speakeramp.jpg (26.4 KB, 465 views)

 

You don't need 1A power transistors in a simple amplifier producing max power of a puny 250mW. The BC547B and BC557B are rated at 500mA and will be fine by themselves.

High power amps use emitter resistors but your have the diodes to stop thermal runaway.
Amplifiers don't match impedances for max power transfer. They have an extremely low output impedance for low voltage loss and good damping of the speaker.

Where are you going to find a transistor with a gain of exactly 300?
What happens to its gain when the temperature changes?
What happens to its operating current when the battery voltage runs down?
How much distortion does a transistor have without negative feedback?

I think your amplifier has too many transistors and needs DC and AC negative feedback. Here is a much simpler amplifier that works very well, but needs a separate preamp transistor to operate from a microphone:

Attached Images

simple_amp_847.png (14.5 KB, 452 views)

__________________
Uncle $crooge

 

Thanks audioguru! I figured that id need a rating of 1A because at the highest output, the emitter would be at 7.6V across the 8ohm = 950ma? When i said the gain's were 300, i meant on average, id like the circuit to be indepedant of small hfe differences in transistors. Btw what does the bootstrap part of the circuit you added do? It seems like a capacitor on the emitter will reduce the AC impedance of the emitter and have larger swings of current through the transistor - at least thats what happens when a cap is added on the emitter of other common emitter amplifier's ive seen.

Megamox

 

As Audioguru says, you DON'T match impedances in transistor amplifiers, it's a really BAD idea. However, I disagree with Audioguru about not requiring the emitter resistors at all, I would always fit them to help prevent thermal runaway, which the diodes on their own wouldn't do.

Personally I don't like the two diode method of bias compensation, and with no resistors it's highly likely that the amp will destroy itself - unless the diodes are perfectly matched to the transistors, and extremely well thermally coupled to them. I far prefer to use a Vbe multiplier, so you can adjust the current, yet still get the same thermal compensation.

Bootstrapping (or some other constant current source) is vital if you want to get any sort of power out of the amplifier - the base current for the top transistor is provided by the resistor from the positive rail, as the base swings positive the voltage across the resistor gets less, so eventually it can't provide enough base current to the transistor. This causes low power and horrible distortion, with good negative peaks and poor positive ones.

BTW, your design would suffer from the same problem on negative peaks as well, so it's not very practical at all!.

Unless the design is for a specific school project?, I would suggest using a small IC amp - far easier than a discrete design.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

No, the speaker is capacitor-coupled to the amp's output which rests at about 4.5V with a brand new alkaline 9V battery. If the transistors don't have any loss (they do) then 4.5V peak into 8 ohms is only 562mA.

Design circuits with DC negative feedback so they work with transistors that have just about any current gain.

The emitter-followers have a very low output impedance. They drive the bootstrap cap with about the same signal as their input:
1) the junction of the 1k resistors actually swings above the supply voltage. Therefore the NPN transistor has low loss.
2) The lower red 1k resistor has the same signal at both its ends resulting in a constant current through it. The constant current load for Q3 is very linear reducing Q3's distortion dramatically, and since it is a very high impedance then the voltage gain of Q3 is very high.

Here's a sim like I've seen on my 'scope many times of a transistor like Q3 without bootstrapping and without AC negative feedback. Its distortion is awful and is so high that you can't measure its voltage gain.

Attached Images

sim_transistor_distortion_730.png (34.6 KB, 438 views)

__________________
Uncle $crooge

 

Thanks guys, ive got alot to think about. Hopefully when i get the basics down i can try those IC chips like the LM386. I didnt really think about feedback or bootstrapping but i can see the benefits, ive been looking at those distorted waveforms all day wondering why they were happening. I didnt know the technical term for 'why are my waveforms more pointy at the bottom!'

Just wondering, how do you work out the input impedance of these designs?

Megamox

 

In the simple audio amplifier circuit I posted, the input impedance is difficult to calculate because the inverting circuit with negative feedback to its input isn't "perfect" like an opamp. It is a little more than its 27k input resistor.

__________________
Uncle $crooge

 

Just a quick question about the power ratings on the output PUSH-PULL transistors. Maximum current is drawn when they're around 4.5v and that current is 562ma. So the power in the trans is (9-4.5v)*(562ma) = 2.5W.
The 2n3904 is rated at 500mw, wont that be a problem?

 

Transistors have a very low current gain near their max collector current. The 2N3904 and 2N3906 have a max collector curent rating of 200mA, but their current gain is a minimum of only 30 at 100mA. They don't have enough input current and don't saturate well at high current so the voltage across the 8 ohm speaker will be much less than 4.5V peak.

The two series'd 1k resistors feeding current to the upper transistor have about 1.9V across each one which calculates to only 1.9mA. With a current gain of 30, the collector current will be only 57mA. If the current gain is 100, the collector current will be 190mA if the transistors saturate enough.

You don't operate audio amplifiers continuously at max output. The average power for music and voice is about 1/10th the max. These little transistors won't heat much in this simple amp circuit.

__________________
Uncle $crooge

 

Just had a thought, if we added another 2n3904 in parallel with the PUSH 2n3904 and fed the base from the same input, the collector currents would be divided equally between both of them, so each one would have half as much current and so a higher hfe. Would this work and push more power into the speaker at peak?