AB Amplifier with clasics transistors

[Thevenin]

Hi!,

I'm new here!

I was designed this amplifier some time ago, but I never want to build it because I don't know if it really works, and I haven't a symmetric power supply yet.

Here you can see some graph and the schematic of it.

By the moment it hasn't protection of any type

Do you think that it can work properly?

Thanks in Advance.

Gain 30 dB
Power supply is +-25V (simetric)

Input Signal: 200mVpp 1Khz.


Amplififier



Frecuency response is between 15 and 31 Khz
Frecuencia



Analogic


Output Power
Vpp=16V about 8 Ohms =32Wpp, 32/2sqrt(2)=22.6Wrms


Fourier


I attach a Proteus simulation.

[lineup]

for a supply +-25V
you can not count for more than like 15-16 Vrms output.
So for full output 1 Vrms in, you need only like 15-20 voltage gain.
If you reduce gain to like 15-20 you will most probably get better quality out

We should not use more Voltage gain, than we need.
Because it is when we let transistors do lots of gain we get distortion.
For a follower transistor, gain<=1.0, we get the lowest dist.
The more the gain, the more the stress onto devices, the higher distortion.

Otherwise it is a nice amplifier project you have there.
Easy to find & buy those transistors anywhere.
Well done!

Regards Lineup

[Segran]

The Zobel network on the output looks kind of strange to me...

Shouldn't you have L1 in (and 3-8 ohms in parallell with L1) in series with the speaker and a 0,1-0.68uF capacitor where L1 is in series with R13 to ground?

[jerluwoo]

I would find another transistor for the vas, tip31's just don't sound good here. They make fine drivers but they're cob is way to high for a good voltage stage. I've used them alot since I can get them cheap locally, you will notice the treble doesn't sound good compared to a lower cob device.

Have to agree with the above post, the zobel is incorrect.

[wg_ski]

Couple of suggestions....

Increase current in Q7,Q2 (decrease R1, R12). These are a little low, i.e, Q2 should be about 6mA, one or two in Q7. As-is it would work, but as you crank up frequency they'll current starve and have more distortion than they could.

Split R12 and make a bootstrap.

The "zobel" on the output looks more like a "load" to me - but 8mH is a bit high. More like 100-500 uH is more realistic for a load.

[wg_ski]

Originally posted by jerluwoo
I would find another transistor for the vas, tip31's just don't sound good here. They make fine drivers but they're cob is way to high for a good voltage stage. I've used them alot since I can get them cheap locally, you will notice the treble doesn't sound good compared to a lower cob device.

Have to agree with the above post, the zobel is incorrect.

If you're trying to find a cheap high performance NPN VAS, go for a video output transistor (old-school CRT tube screen driver). You may be able to scavenge them from dead TVs for nothin'.

[Beto]

I think the wrong place at the L1 and R13 use Zobel filter using the capacitor in series with 100nf resistor of 10 ohms.

[Thevenin]

Hello everybody

and lot of thanks for the answers! :-)

wg_ski is right, L1 and R13 simulates the load, and I never know what inductance value a speaker has. Where do you obtain this kind of data, because I think that I get the value from a forum some think like this...
I think that I read some data specifications of speaker, but never saw it.

I forgot to include the Zobel Network and the damping resistor as everybody says, I don't know why, the design is old, and maybe is one of many of my attemps.

In High Power Construction Manual, Chapter Seven Pag.210 Says in a picture:

Zobel Network
===========
(0.1uF Serie 8.2Ohms) Parallell with Speaker

Damped Inductor
============
1uH+10Ohms Series Speaker

Is is right, isn't it?

[Thevenin]

Originally posted by lineup
for a supply +-25V
you can not count for more than like 15-16 Vrms output.
So for full output 1 Vrms in, you need only like 15-20 voltage gain.
If you reduce gain to like 15-20 you will most probably get better quality out

We should not use more Voltage gain, than we need.
Because it is when we let transistors do lots of gain we get distortion.
For a follower transistor, gain<=1.0, we get the lowest dist.
The more the gain, the more the stress onto devices, the higher distortion.

Otherwise it is a nice amplifier project you have there.
Easy to find & buy those transistors anywhere.
Well done!

Regards Lineup

Great, you're right. I try to redesign decreasing the gain vor the VAS,

Originally posted by jerluwoo
I would find another transistor for the vas, tip31's just don't sound good here. They make fine drivers but they're cob is way to high for a good voltage stage. I've used them alot since I can get them cheap locally, you will notice the treble doesn't sound good compared to a lower cob device.

I can't understand, What do you mean with "COB"? I search it in Google but

[lineup]

Cob = output capacitance

This is a value that some say is very important for VAS and driver transistors.
You can find it in datasheets.

Here is where I download datasheets:
http://www.datasheetarchive.com/

TIP31 datasheet
Other common driver transistors often used are MJE15030/31 and BD139/140
You can compare COB for these and TIP31/32