[yzriver]

Hi all:
I need more eduacation from you.
I just measured the Vin Vout of this headphone amp. I don't like the performance of it. I have a 300Ohm resistor as a load at the output. I captured the curve of voltage drop on this resistor. Please look at the curve. Input is 1KHz sin wave.
The output voltage saturates at +-5V, as the red line indicates. What cuases this problem? Is it from the OPA saturation? I have a powerful DC source, 24V 3A, so I don't think it is the problem of DC supply.

The other problem is that it is pretty slow. It takes 10 nS swich from -5V to 5V. Is it the slew rate 1V/1uS? This number dispoints me. The C2kIII have 70V/us.

Look at the circuit. Is it a hopeless circuit? Is it the negative feedback cause this problem? How can I tune it to swing higher than +-5V, like +-10V.

Thank you for you reading. I appreciate your reply.

 

[tangent]

Quote:

Originally Posted by yzriver /img/forum/go_quote.gif The output voltage saturates at +-5V, as the red line indicates. What cuases this problem?

What do you get with no load on the output? I don't see that the peaks can be any higher than 9 V from ground, due to the +/-3 V output-to-rail limit of the OPA2134.

Quote:

Is it from the OPA saturation?

No, it's buffered.

Quote:

I have a powerful DC source, 24V 3A, so I don't think it is the problem of DC supply.

Indeed. You're only asking for about 40 mA with what you show.

That's a very curious number: the op-amp itself is capable of this, so if the circuit isn't capable of more with the buffer, why have the buffer?

Quote:

The other problem is that it is pretty slow. It takes 10 nS swich from -5V to 5V. Is it the slew rate 1V/1uS? This number dispoints me.

It's probably the huge capacitive loading from having that electrolytic on the output. You can only charge and discharge the thing so fast. Microfarads require microamps to charge quickly.

This might get back to the clipping problem, too.

 

[amb]

The low-pass RC filter at the input and the compensation cap in the feedback loop might be excessively large, causing the slow square wave response. You don't show any parts values so I don't know what their effects are. Since you have a scope, you can try smaller cap values to speed things up.

 

[yzriver]

I really your experienced guys' time on my questions. Sooner or later I will make a CKK or tube hybrid headphome amp, before that I would like to get more understand on the audio circuit.

If the Vout limited by the output swing of OPA2134, which is about +-2.5V, I can replace it with NE5532 that is capable of +-13V. The following curve is measured from the amp with NE5532. I also put the Vout of open and Vout on 300Ohm. The Vout on 300Ohm is about half of the Vout on open because there is a about 220Ohm resistor on the output. Is the clip from the input of the OPA, since the clipping starts at Vin about 0.6V, or from the negative feedback?

the rising speed is nearly the same as with OPA2134.

Following is the circuit with components’ value. I measured the resistors on board. The 63pF caps are those with F530 labeled. Is it correct? I don’t know the type of the diodes, so I just pick up any one I know.

I also attached the Orcad project. I am new to Orcad, there are several problems. For example I don’t know how to link the BD139.lib to *.olb. Could any Orcad expert please fixed it? The load resistor should be 1M/300Ohm.

Thanks a lot.

 

[amb]

Are you sure that C6 is 47uF? That could be the problem right there...

 

[yzriver]

sorry, C6 is 0.1uF,63V.

 

[Uncle Bob]

I've been following this thread with interest because I've got an OBH-11 as well and have been thinking about doing some mods.

I've a question though. There are three components on the board, labelled LK1/2/3, which I assume are inductors? I wonder what they're intended to do in the circuit? I notice that you've ignored them on your circuit diagram so far (good work btw, gives me a headache trying to figure these things out from the pcb)

 

[yzriver]

LK1/2/3 are small resistors, 1 or 2 Ohm. They are used for jumpping over the thrace.

 

[tangent]

The OBH-11 board is single-sided, so these are used as jumpers, where a dual-sided board would just take a via to the other side. They're nominally 0 ohms. The prefix might stand for "LinK".

 

[Uncle Bob]

Ok, thanks guys

live and learn....

 

[Uncle Bob]

I decided I'd try an op-amp upgrade to my OBH-11 - I happened to have an AD8397arz on a browndog adaptor which I could try. I wasn't sure whether it would work but it certainly seems to...

I say 'seems to' because although I've been using it for over a week and it sounds a lot better than the NE5532 it's replaced, it does seem to run very hot.

Does this indicate that it might be oscillating or that there's too much voltage being supplied? Or is this normal for this class of op-amp?

 

[Pars]

The AD8397 will run a bit warm. It is also a cranky part, so I would measure the current draw of the part or the amp and see if it looks like it is in line with datasheet values. If not, I'd suspect its oscillating...

 

[yzriver]

if you heard of a good sound, the opa is working well. However, the AD8397 is not suit in this amplifier well. The rating power supply of it is 24V, and the voltage at V+ is a little higher than 24V in my Creek OBH-11. So your AD8397 is working under over bias, although the AD8397 can suvive under 26V.
The AD8397 has a high current capability, which could heat it up.

 

[Uncle Bob]

Well I finally got round to measuring the current draw of the amp with the following results;

no chip installed - 6.4mA
standard NE5532 installed - 34.9mA rising to 35.4mA after about 5min
AD8397 installed - 47.4mA rising to 48mA after about 5 mins

both tests ran with nothing connected to either the inputs or output, Vcc measured at Pin 8 is 23.5V

From the spec sheet, the Icc of the NE5532 @Vcc=15V is 8mA typical / 11mA Max while the AD8397 @Vs=24V is 22mA typical / 30mA max

So an extra 13mA of current draw would appear to be in line with the specs of the chips, indicating normal operations - correct?

 

[tangent]

Quote:

Originally Posted by Uncle Bob /img/forum/go_quote.gif So an extra 13mA of current draw would appear to be in line with the specs of the chips, indicating normal operations - correct?

Half correct. The fact that it isn't drawing lots of extra power means it probably isn't oscillating wildly. It could still be misbehaving. The fact that it sounds good mitigates against that possibility; I'm just being picky.

Most likely, that extra heat is just due to the higher Iq of this chip.