Opamp Power Amp - Will it Work?

[jnmar]

Hi

I´ve been reading this forum for the last few months, and i fell ready to build something. I like to design my own schematics so i´m going to start with something simple.

I simulated the design and with works fine but i don´t no how it will work in real life. So any comments are welcome.


The first opamp does the voltage gain and and the feedback. Opamp 2 and 3 generates the bias to drive the mje15031/30. The output is biased to class A. The output power is around 16 W/8Ohm that i believe is more than enough.


Thanks
João

[darkfenriz]

The hard part is to make good, thermal tracking and adjustable voltage sources V3 and V4. Otherwise it won't work good.

[Original Burnedfingers]

I believe it can be make simpler yet. Check out what Altec Lansing did in their 9446 power amp. EV made a nearly identical model the 3200. They both use a 5532 op amp to drive the output stage.

I guess I should mention that both amps haven't been made for a few years now. There are a lot still in service.

[AndrewT]

Hi,
the opa134 outputs have a resistor to Vrail to pull the output stage to single ended ClassA.
But, one is pulled to V+ and the other two are pulled to V- ,
will this work? Are the 134 output devices quasi or complementary?

Do V3 and V4 put a fixed 1.8V bias across the driver bases?
If so then the output stage is going to be seeing a lot of current. and the drivers only have ~6mA of collector current. where will the output device base current come from when Iout=>2Apk

[janneman]

Originally posted by darkfenriz
The hard part is to make good, thermal tracking and adjustable voltage sources V3 and V4. Otherwise it won't work good.

Indeed. Also note that the two driver opamps have a gain of two so the bias is also amplified by two, giving more than 3 amps quiescent current. Isn't that a bit too high?

Jan Didden

[jnmar]

About thermal track i´m thinking of doing it with the diodes in the output bjt's.


The quiscient current it's only 1.1 A at 15 V gives 16 W not much.

The U2 and U3 are adders with voltage gain = 1. So i´m only biasing the drivers with 0.9 V and 100 uA.

They only shif (+0.9 U2 and -0.9 U3) the signal from U1.

I think that pull up or pull down in the output of the opamp will both put it in class A.

Thanks for the comments.

I know that some say that opamp will degrade sound, but will these kind of opamp (0,00008 % THD) destroy much the quality of sound?

[janneman]

Originally posted by jnmar
About thermal track i´m thinking of doing it with the diodes in the output bjt's.
The quiscient current it's only 1.1 A at 15 V gives 16 W not much.
[snip]

Sure? I see about 1.1V across the the 0.33 ohms Re's in your diagram and that's > 3 amps.

Edit: No, you're right! The bias is halved and then amplified x 2 still 0.9V gives about an amp. My bad!

Jan Didden

[Gigapod]

Originally posted by jnmar
...

I know that some say that opamp will degrade sound, but will these kind of opamp (0,00008 % THD) destroy much the quality of sound?

Depends on the load they are driving.
I believe AndrewT questions above are related to that.

[jnmar]

Made some changes remove V3 and V4, bias generator and replace them with a zener to give me around 5V. Then regulate the bias for each output via R52 and R56. Maybe a 7805 will give better results but i can´t find the spice model for it.

What kind of Thermal Tracking do you suggest?


Thanks for the help.

[bikehorn]

from what i've read, resistors alone don't make very good current sources (re the opamp in class A). at the expense of some complexity, consider a JFET current source? maybe that would help matters.

neat page on biasing opamps into class A:

http://tangentsoft.net/audio/opamp-bias.html