Oh no! There is DC on my output!

[drZoidberg]

Hi guys!

I'm currently building a power amplifier as a school project.
I have built the amp and it sounds great!

The only problem is that there is about -200 mV of DC on the output..
I don't have a lot of experience with amplifiers, but i kinda know how all the stages work.
I don't know how much DC is acceptable on the output, and how much 200 mV will damage my speakers.

I'm trying to figure out what stage affects the DC level on the output. I suspect that the differential amp stage has something to do with the DC, but I'm not sure.

I'll try to measure some more to find out if I have done something wrong..

Also, when I tested it without C12 it oscillated at 1MHz.. But that went away as soon as I put in C12.

Oh, and here's the circuit:
http://dl.dropbox.com/u/12774933/Amplifier.png

Any help would really be appreciated.
Thanks!

[tecman]

Try new xsistors at Q1 and Q2.  I suspect that they are not close enough in characteristics.

paul

[Zero999]

Try adjusting the 2k pot to trim the current as near zero as you can.

I suspect 25mA of DC won't do much damage to the speakers.

[drZoidberg]

Oh, yeah! That makes sense.

And thanks for the fast reply

But is there any way to test if they are close in specs?

Is it better to do that is the circuit, or should I desolder and test is somehow?

[drZoidberg]

Oh! And one more thing:

When I tried to adjust the pot to GET crossover, and THEN adjust it up to a decent level, I couldn't get any crossover.
It didn't seem to go that low.. Is that normal? I would expect that I would be able to go lower..

[Zero999]

Oh yes, sorry, the pot is to eliminate crossover distortion, not change the DC bias.

What do you mean you couldn't get any crossover? You don't want crossover distortion, it's bad.

[drZoidberg]

Yeah. But I expected the "sweet spot" for setting the pot to be somewhere near the middle.
So therefore I wanted to turn the pot all the way down, and then turn it back up until the crossover went away.
But it would not go that low..
But still, I don't know if that would help any (?). Or if that suggests that there is something else wrong (maybe wrong value resistors ?).

[mzacharias]

I work on a lot of DC coupled amps, and it is common not to be able to adjust til a notch appears. Simply adjust for the specified current through the output device (usually measured across an emitter resistor) generally in the 7.5 to 15 millivolt range. 200mv DC offset is not damaging but it is unacceptable. Less than 10mv would be pretty ideal but certainly it should be less than 50mv.

Mark Z.

[blackdog]

Hi drZoidberg,

With a small modification to your schematic you can resolver your offset problem, place a 100 Ohm potmeter in the emittors of q1 and q2.
The wiper is connected to the collector of q3.
See schematic: www.bramcam.nl/apex/drZoidberg.jpg


Testing without your speakers!!!

Bram

[Zero999]

With a small modification to your schematic you can resolver your offset problem, place a 100 Ohm potmeter in the emittors of q1 and q2.
The wiper is connected to the collector of q3.
See schematic: www.bramcam.nl/apex/drZoidberg.jpg

Yes, you're right, that will do the job. I was struggling to think of a solution. It's years since I did this kind of work.

[drZoidberg]

Hi guys!
Thanks for the suggestions!

If I understand the circuit u gave me correctly the point is to get the same current through both transistors ?
And that will offset the DC?

I'll definitely try that. But maybe I can just change the resistor R1 to adjust the current in Q1? So I don't have to etch another PCB ;p

I'll give you an update on the results :->

Thanks!

[blackdog]

Hi drZoidberg

The only good solution is the potentiometer in the emitters of Q1 and Q2.
You can make the amplifier more stable with temperature when Q1 and Q2 coupled Thermic, see the picture how to do it...
The resistence in the collector of Q1 is not your problem...
It should not be to difficult to place neatly a additional potentiometer on the circuitboard.
A good working amplifier is moor important than a perfect looking circuitboard

www.bramcam.nl/apex/drZoidberg-2.jpg

You mentioned the release of C22 and the amplifier went in to oscillation, thats normal, the gain is dropping to 1x
and the amplifier is not compansated for a gain of 1!

My 2 cent ;-)

Kind regarts
Bram

[drZoidberg]

Thanks a lot Bram!

I will try putting a potentiometer there tomorrow. That looks like a good solution, at least in simulation. But I guess I'll find out for sure tomorrow

Meanwhile, I tried to change the value of R1 at the base of Q1, and that also seemed to work. But I had to almost double the value, and that doesn't look right to me..
So I'll probably go for the potentiometer.

I will also thermicly couple the transistors.

Also I have another problem. The output clips on the low side. It should clip at over 30 V amplitude, but on the low side it clips at about -17 volts.
I'm giving it a 1kHz signal and it only clips when I attach an 8 Ohm load resistor.

First I thought that the problem perhaps had something to do with the protection circuit (Q7 and the voltage divider R17/R15), but that should also happen with no load, right?

Can the problem be that Q4 is not getting the current it needs to pull the output low? (since it swings perfectly +/- 30V when no load is present)

It also did some funky stuff when clipping at higher frequencies (like 10kHz). I'll upload some scope photos tomorrow.

Thanks again
drZoidberg (\/)(\/)

[blackdog]

Hi drZoidberg 

Try this...
Disconect D3 and D4 en test you amp.
Note that you now have no current limit for you power transistors...
Is the amp still clipping a-symetrical?
If the amplifier is now behaves normally check the component ??values of R14,15, 16, 17, 18, 19, 21, 23, 24 and 25

Regarts,
Bram

[drZoidberg]

OK

I have put the potentiometer in place, and I changed R1 to 33k Ohm.
Now the DC is sorted.

And you were right about it being a diode. It D6 was turned around by mistake..
So now it doesn't clip.

Thanks a bunch.

But like I said, there is some funky stuff going on with the signal when I test with higher frequencies.

Here's a photo (the blue is the input (for all the photos), and the yellow is the output):


This only happens when I test at high output amplitudes, and is not affected by adjusting the bias for the output transistors.

Another thing I don't really understand is the voltage on the base of Q4. The waveform looks distorted even when the output amplitude is low and there is no distortion on the output.

When there is  no distortion on the output it looks like this;


And when the distortion occurs it look like this:


Another thing I noticed about the base voltage of Q4 is that it follows the AC ripple on the negative power rail, but I don't know if that affects the circuit at all..
Here is the base voltage on Q4, zoomed out to show the low frequency ripple:


The other two problems I kind of understood, but with this one I really have no clue.. :/

Thanks for the help

Zoidberg

[qno]

Maybe check the current gain on the power stage.
The current gain of the BD911 is only 10.
I wonder if the drive capacity of the BD140 is enough.