My project consists of 6 goals @700Hz:
Rin = 250kOhm
Rout = 150Ohm
Rload = 150Ohm
Gain = 30
Voltage Swing = 6Vpp
Vpowersupply = 12V


Design is up to me and my partner.

We have gone through about 7 different designs. Attached is our latest.
Our main problem is the final stage. We can't get an output impedance of 150Ohms, while driving a 150Ohm load, without screwing up our signal.

We calculated that we need to feed a 0.2Vpp(0.1Vp) signal into our amplifier (with a gain of 30), and be able to get a clean output of 6Vpp(3Vp) without clipping. So when we simulate, we usually do 1mV, 10mV and 100mV all at 700Hz.



Can anyone help with the final stage? We have tried just about everything, Common Collector/Emitter/Base. We thought of using another MOSFET at the final stage also, haven't tried it. We just can't seem to get it... spent ~2 weeks already.


Attached is our design and Vin versus Vout waveform (From the points on the schematic) Final stage is not part of the simulation, due to ... it not working.

Thanks (if anyone can help via IM, please do so at ydoesgodmockmeso on aim/yahoo-im)

 

I got a question about your circuit. How does stage 3 connect to stage 4? Looks like you're trying to use a common emitter output stage... but there's no input to it that i can see.

I'd suggest using an emitter follower as the output stage. The tricky part (as you've noticed) will be balancing the output resistance and the voltage swing. Try choosing the emitter resistor on the output stage to be around 1k or so so that the voltage swing will be satisfied. You'll also want your biasing resistors to be pretty large so that the output resistance (parallel with these resistors) will stay small. Also, with these values large, your gain will not be sacrificed.

Now the trick to keeping the overall output resistance low is to realize that the output resistance is largely dependent on the collector resistor of the previous stage. So you may have to adjust your biasing of that stage so that you can lower that resistor whilst still having plenty of gain. Which... since you're using two gain stages... i would think gain isn't a problem.

Thats gain V/V, right? not dB?

Hopefully this will give you some ideas...

 

The final stage wasnt connected because it wasnt working properly.

The thing about using an emitter follower(C.C.), or anything other than a C.E., is the output resistance of the amplfiler needs to be 150Ohms. W/ a CC the output resistance depends on too many factors that we cant totally control(beta for example). Whereas w/ a CE or a CB, we can make Rout equal to 150Ohms....or so we thought.

 

The reason i suggested using a CC is because its a buffer... generally high input resistance and low output resistance.

Ohh ok. Well if you select your resistors properly (i.e. the emitter resistor should be sufficiently large with respect to the parallel of R39 and R40... if i remember right, i usually use Re/alpha = 10 * (R39 || R40)/beta), then the dependence of the circuit on variations of bata will be negligible. Thats one of the main reasons for using a four resistor bias network.

So in the end, the output resistance of the CC amplifier is MOSTLY dependent on the thevinen resistance of the previous stage (which is dominated by the collector resistor of the previous stage and terminal output resistance of the previous stage's transistor).

Since terminal output resistance is Vt/Ic, if you raise your Q-point current of your second to last stage (lowering terminal output resistance), then you'll lower your total output resistance (because terminal output resistance in parallel with the collector resistor is close to the terminal output resistance (i.e. low) because the collector resistor is much larger than the output resistance. Also, for the same reason, Re || Rth/beta (last stage) approaches Rth. Rth being approximately the terminal output resistance of your transistor of the second to last stage).

Wowza. That was a mouthful haha. Hopefully its clear what i'm driving at though.

The last thing you have to do is to make sure that the emitter resistor of the output stage is sufficiently large so that Vce and Ic are appropriate for the amount of voltage swing you need.

Hopefully all this discussion is helpful to you... hehe. Good luck!

 

The final stage wasnt connected because it wasnt working properly.

The thing about using an emitter follower(C.C.), or anything other than a C.E., is the output resistance of the amplfiler needs to be 150Ohms. W/ a CC the output resistance depends on too many factors that we cant totally control(beta for example). Whereas w/ a CE or a CB, we can make Rout equal to 150Ohms....or so we thought.

If the Rout has to be exactly 150 Ω, and not just something less than 150 Ω, then use an emitter follower to get something much lower than 150 Ω, and add a resistor in series with your output to bring the final Rout up to 150 Ω.

 

I agree with both Guitarguy and the Electrician. Use a follower at the output.

But I think you need only 3 stages. Your first stage is good and satisfies the 250k input resistance requirement.

For stage 3, use a follower -- FET or BJT -- and if the output resistance needs to be exactly 150 ohms, do as TheElectrician said, just put a resistor in series between the emitter and the load resistor.

Stage 2 is for voltage gain. Here is the hardest part of your design. You have to take the output resistance of stage 1 into consideration because you will lose gain due to this. Call the input resistance of stage 2 Rin and the output resistance of stage 1 Rs. Then the voltage at the base of stage 2 in Vin*(Rin/(Rs+Rin)).

Also the input resistance of stage 3 will attenuate the gain since Rin of stage 3 is in parallel with the output resistance of stage 2.

Here is another point you need to consider. Your input signal is 100mV peak. This presents a problem for CE stage with the emitter resistor totally bypassed. 10mV peak at the input is the most you can do in this case, otherwise you get nonlinear distortion.

This means you need some unbypassed emitter resistance. In this case use 220 ohms of unbypassed resistance at the emitter. Bias the base at about 3 volts so that your emitter is at 2.4 volts. Ideally, bias the collector such that it is at least 5.4 volts in order to get the signal swing you need (2.4 + 3 = 5.4). But practically, your last stage will attenuate this by a factor of .75 or .8 which means you'll lose a volt of gain so bias Vc at 6.4 volts for a Vp-p greater than that of your output stage. Vc can be larger than 6.4 volts but no more than 12-4=8 volts. (This is for the necessary swing from the supply to the collector).

Finally, remember the gain of stage 2 is approximately -Gm*(Rc(2)//Rin(3)). Also, if you need a gain of exactly 30 put a 1k pot in place of the 330 ohm resistor and find the exact value experimentally.