I started out designing a multistage CE amplifier for small signals just to learn how to design these stages with proper biasing as well as impedance matching.

Here is the first part of this experimental design.



When I got that to work properly, I began working on a Class AB power amp, using the available supply of 8v.

I chose 8V. so I can use a 9V. battery to run this circuit.

Transistors Q3 and Q4 below schematic, are designed only to increase Zin to the push pull pair.

Here is the power amp.



Then I combined both circuits together to form the completed small sig. amp.

I tried to use emitter bypass caps. only where neccessary, and no feedback cap. that would hinder high freq. response, without C1 in the circuit the Av. ~= 20 due to signal in being small so as to not overdrive the amp.

Even at this small gain I tested the amp, by putting it through an actual sound test, I placed the input speaker ontop of the speaker of my portable DVD player and turned the volume down and listened and had excellent sound quality.

Then I added the Cap. C1 and feedback R1.

And heard everything very clear and louder also I was avle to pick up all the data handling sounds in my laptop computer, as well as the sounds off the LCD screen the sounds inside my DVD player ect..., to make sure it wasn't oscillations, internal in my circuit I placed the speaker away from everything and heard silence using a 9 volt battery.

This is my experimental first try in designing in a more systematic way trying to keep impedances matched as much as possible as well as not going overboard with base collector feedback capacitors.

here is the final circuit.

The one at the bottom is just a reapeatr I couldn't delete.

 

Just a suggestion, crop that picture. It will be smaller and clearer on the thumbnail, and you can display it full size in your post.

Also, use a different format. JPG is lossy, and tends to add trash. For B&W .gif is much better, and .png is better yet.

How to Display Attachments Full Size

Keep up the good work, I'm following it with interest. My own inclination is that you could eliminate around 2-3 transistors there. What is the gain you're aiming for?

 

Thanks but I can't run paint prog. to good.

Thanks for the link.

 

a few suggestions on the power amp:

1) you should definitely put a small capacitor, like 47p - 220p between Q3's base and collector. it helps make sure that you don't get oscillation.

2) Output stage's bias is basically indeterministic. I would put a small pot in serial with the diodes there to help control bias.

3) think about temperature compensation: you don't have one and if your OPS heats up, you may have a run-away situation here.

4) you may think about bootstrap R3 to increase (open loop) gain.

5) R4 is too big and is limiting the amp on the next swing.

6) I would put the whole thing in a negative feedback loop so you can define its gain via resistors.

7) Re are a little big at 2ohm but that helps here since you don't have a Vbe multiplier.

 

a few suggestions on the power amp:

1) you should definitely put a small capacitor, like 47p - 220p between Q3's base and collector. it helps make sure that you don't get oscillation.

2) Output stage's bias is basically indeterministic. I would put a small pot in serial with the diodes there to help control bias.

3) think about temperature compensation: you don't have one and if your OPS heats up, you may have a run-away situation here.

4) you may think about bootstrap R3 to increase (open loop) gain.

5) R4 is too big and is limiting the amp on the next swing.

6) I would put the whole thing in a negative feedback loop so you can define its gain via resistors.

7) Re are a little big at 2ohm but that helps here since you don't have a Vbe multiplier.

Thanks
I'll look into all that.
It's good to know that it is ok to use caps. between base and collector when needed. as long as I put them in the right places.

That's more to study up on thanks again.

 

if you fix what I suggested, you basically have nad3020, a very well known British power amp.

 

if you fix what I suggested, you basically have nad3020, a very well known British power amp.

Wow
didn't realize it was that close to being a profesional design.
Thanks.

 

and if you use a npn transistor for Q2, you gets the famed JLH1969, which is a transistorized version of the Williamson tube amp, the very first "hi-fi" power amp human beings ever had.

see how easy it is to be good?

you just need to be 70 years earlier,

 

a few suggestions on the power amp:

1) you should definitely put a small capacitor, like 47p - 220p between Q3's base and collector. it helps make sure that you don't get oscillation.

.

IN place of the R1 100K, or in series with it?

 

a few suggestions on the power amp:

1) you should definitely put a small capacitor, like 47p - 220p between Q3's base and collector. it helps make sure that you don't get oscillation.

2) Output stage's bias is basically indeterministic. I would put a small pot in serial with the diodes there to help control bias.

3) think about temperature compensation: you don't have one and if your OPS heats up, you may have a run-away situation here.

4) you may think about bootstrap R3 to increase (open loop) gain.

5) R4 is too big and is limiting the amp on the next swing.

6) I would put the whole thing in a negative feedback loop so you can define its gain via resistors.

7) Re are a little big at 2ohm but that helps here since you don't have a Vbe multiplier.

Just goes to show
Theres a lot of stuff that needs to be learned, even this close to a good design, there's still a whole lot more that needs to be worked at.

Thanks again for all your help..

 

i was talking about the power amp, where Q3 is the 2nd transistor in the middle.

 

Oh ok thanks for clearing that up.

I used autodesignation on circuitmaker when I was drawing this up and it redesignated everything from the first schem to the last.

Thats why the designations arn't consistent.

You said about R4 too high, after this post I'm assumiong the power amp schem. and not the complete circuit scem., right?