Engineers Needed How do you reduce THD in audio, how to get more gain, and more

[ski]

Greetings!

I posted a few weeks ago about the preamp for the tape head I'm working on. I solved for all of my resistors and capacitor to get the right frequency response, and below is the theoretical frequency response, followed by what I actually get:
http://www.littleboyinc.com/uploader/uploads/graph3.JPG
http://www.littleboyinc.com/uploader/uploads/graph1.JPG

Here is the schematic of the circuit:

http://www.littleboyinc.com/uploader/uploads/preamp2.JPG

And the preamp WITH the Class AB power amp. I need 400mA at 1kHz (no DC) at the output to drive the 8 Ohm speaker so I can get at least 1 Watt, but I'm in the microamps range... crapola :(

http://www.littleboyinc.com/uploader/uploads/ampbig.JPG

Questions (if you can help me with even one, I will be very grateful!)

1. I'm only getting 30 dB relative difference on my preamp at 50Hz and 2kHz. I have tried different values for R1, R2, R3, and C1. Is there something I can add to get better gain, or other specific values I should try?

2. I need less than 0.1% THD at 1Khz for the preamp. I'm getting upwards or 2% and 3%! Yikes. What are some ways to reduce THD in the preamp. Also, I'm getting 2000% THD at the load (end of the power amp), what type of feedback should I try to add to reduce this?

3. Simple one -- I need to split the 13.6V car battery into +6.8 and -6.8. I did that in the circuit above, but when I connect it to the class AB power amp, the top voltage line drops from +6.8 to +1.5, and the bottom line goes from -6.8 to -12.1. I'd like to run this off of just the car battery, so is there any way I can rig this so I can hook up the voltage lines to the power amp without them changing? You can see I just put separate DC sources in for now.

Thanks... I know this is long, but hopefully I have described everything okay! :)

 

[Averagesupernova]

What sort of load are you trying to drive with the 324?

 

[ski]

It's a standard 8 Ohm speaker... I added a coupling capacitor in line with the speaker so no DC current would go through it.

Anyone want to take a stab?

 

[Averagesupernova]

There is no way that 324 is going to be able to drive a speaker. Really all that the power supply splitter does is set a bias point for the input. Consider it 'virtual ground'.

 

[ski]

I changed the LM324 to a uA741 on the preamp (first opamp) to reduce the THD in the preamp.

Should I try a uA741 for the second opamp?

 

[ski]

This is a VERY elementary question, but I'll ask anyway to make sure:

If you have voltages VDC and VAC (a sine wave offset by a DC voltage), how can you get rid of the DC offset to make the sine wave (AC) centered about zero?

 

[louie]

You can try connecting a 10 kohm pot between the offset null pins (1 and 5) of the op amp. The wiper arm of the pot. connects to the negative supply voltage. Then adjust the pot. to zero the offset. It should be noted that this only cancels the output offset at one particular operating point. With temperature changes or simply over time, the circuit might need to be readjusted.

 

[Averagesupernova]

Take a closer look at the data sheets for the op amps you are using. It takes a different op amp than a 324 or 741 to be able to drive a speaker. The 741 won't do it either. I'll commen more later. Sorry, kinda busy.

 

[Averagesupernova]

Ok. My apologies. I didn’t realize that you were driving the speaker with a transistor output. I failed to look at the second schematic. But, I can give you some advice on how to make the circuit work better and simplify it all at the same time. I also have questions for you.

In the schematic you have drawn 3 batteries. Are you actually using 3 separate batteries or isolated power supplies?

In the output stage you have the emitters connected directly to each other. Bad design move. You need to use a low value resistor on each emitter and have the resistors tie together and feed the output. This prevents a thermal runaway condition and may help your crappy output.

Pin 5 of U2B is floating. Bad design move. This pin should be tied to ground through a resistor to prevent the voltage on it from floating around. Since it is connected only through a capacitor its DC voltage can float anywhere it wants which in turn causes the output to directly follow it.

You have the source showing a 0 VDC output. Is this really the case? Is it wired directly to the tape head? If so, you don’t need to use the capacitor C4. For testing purposes you may want it, but when wired to the tape head it’s DC voltage should be zero. However, keep the capacitor C2. Its job is to prevent the DC voltage gain from being the same value as the AC voltage gain. You may think that since the input is 0 VDC it shouldn’t matter. But since no op-amp is perfect there is a slight input offset and the C2 keeps that offset from being amplified. Now, let’s take a step back to pin 5 of U2B. I mentioned it should not be left floating. Since we can assume that the output of the first stage has VERY little DC offset on the output why can’t we just drive the next stage (pin 5) directly without the cap? The truth is that you can. Go ahead and eliminate it. Actually, try it both ways. Experimentation is how we learn.

Now for your THD and gain problems. Where are you measuring the THD at? And what are the gains throughout the whole thing? I think that you should split the circuit up and test preamp with the output stage totally disconnected. Even from the power supplies. I am considering the second part of the op-amp part of the output stage. Reducing THD from 10% to 9% is easy. It’s an easy place to gain 1%. But reducing it from 2% to 1% is another story. There are MANY more subtle things that can keep that THD up there where you don’t want it. Any noise on the power supply can cause it. I don’t see any bypass capacitors on the power supplies. These are a must. I’ll leave it at this for now. There are a few other things I want to address but let’s get a few things nailed down first. It shouldn’t be a hard project for you to get off the ground with a guiding hand.

 

[Averagesupernova]

* B U M P ! *