headphone dist. amp design

Hi,

This is my first audio design and I was wondering if someone here could look it over before it is commited to fibreglass. Its a pretty simple design, I just wanted to check I hadn't missed anything obvious.

The design is for a simple headphone amp with individual volume controls. The 'amp' modules just dasy chain off the balanced line input module, so you can scale it up to however many outputs you want.

Also, are there any particular guidelines when laying out PCBs for audio? The layout below is just a quickie done without any particular layout considerations, would this layout be ok for this project? (I'm not going for amazing quality here, its destined to be for cueing)

Cheers,
Fergus Noble.


http://www.angusnoble.com/fergus/input.gif
http://www.angusnoble.com/fergus/inputbrd.gif

http://www.angusnoble.com/fergus/amp.gif
http://www.angusnoble.com/fergus/ampbrd.gif

Looks fine (quick look only) except that I would spend a 100 nF ceramic or film bypass cap between +15 V and -15 V for every opamp.

Samuel

It doesn't look good. Feedback resitors to noninverting input?

Regards,
Milan

If there is any d.c. offset in the source you might want to a.c. couple at least into the level adjust potentiometers on the headphone amps. If you anticipate long cable runs between the input diff amp and the individual headphone boards you might want to provide for series resistors in the outputs of the diff amp to avoid instability with the reactive loading of the long cables.

If you are confident there will be little out-of-band interference then you are probably fine with the design as shown. But for flexibility you might want to add locations for additional bandwidth reduction with small C's across R5,6,11,12 on the diff amp, and possibly across R3,5 on the headphone amps.

The polarities of the inputs are reversed on the headphone amp schematic. EDIT I see moamps spotted that as well.

You might want to put locations for small R's in series with the outputs of the headphone amps.

EDIT2: PS, welcome to the forum.

Wow, thanks for the speedy response!

Oops! that was a silly mistake with the op-amp inputs!

I have put 100R in series with the outputs of the line inputs. What value would you suggest for the caps you suggested across thoes resistors?

** Edited out stupidity **

Obviously the PCBs are wrong now, but any tips on layout for audio or is it non critical?

That input amp "can" be overloaded. Balanced outputs these days are two opamps fed +/-15V, can output 24V peak-peak each or 48V p-p across the balanced jack. That's +27dBu, which is insane, but insanity can happen. The input opamp can't output +27dBu, so can't take such a hot signal at unity gain. A safer plan is to take 2:1 loss in the input stage (and any make-up after the knob).

A constant-voltage headphone output will be MUCH louder in typical 32 ohm phones than in 300 ohm phones. There is no universal solution, because power-match is a curve, and because headphones of the same impedance can vary 20dB in efficiency. However, I believe you will really want a series resistor to the headphones. 7Vrms behind 27 ohms will drive about any headphone to near its power rating (an odd coincidence) and give more even output for various impedances. Considering your 8V or 10V capability, I'd figure 33 ohms series resistor.

This decreases headphone damping. I have not yet seen a "32 ohm" phone with enough impedance variation to matter. I have seen "300 ohm" phones with 600 ohm peaks, and 33 ohms will damp that well.

I have not looked at the power capability of that chip. If it has the guts to drive 33 ohms to 6V RMS, then the 33 ohms should be 1 Watt to reduce smoke when shorted headphones happen and someone cranks-up trying to get sound.

I was going to ask if it made sense to have a bunch of little boards, or one big 16-output board that you could populate as needed. Then I realized you probably want to put each volume knob (and associated amp) AT each musician.

http://www.paia.com/hdasch.pdf

I have used this and it can be put on a small board. May offer some tips on
the input variation.

It doesn't look good.

Click to expand...

Thanks for correcting me--my look was obviously to quick.

What value would you suggest for the caps you suggested across thoes resistors?

Click to expand...

You are talking about R5/R6/R11/R12 on the input and R3/R6 on the output amp as suggested by Brad? I'd throw in a 47 pF each, preferably 5% or better for best CMRR (though this is likely uncritcal in your application).

BTW, I'd replace C1/C2 on the input board with two 1 nF caps each going to chassis--this provides much better RFI protection.

Check the OPA551 for a beefier output amp.

Samuel

josan:

I dont see what advantage your design has? They seem almost identical to me, only without the balanced inputs! Thanks though it is reassuring you came up with the same solution.

PRR:

thanks for the info, I will put in some series resistance. I don't ever foresee the input being driven that hot but I will put in some zeners on the input as a crude limiter. Is this an acceptable solution?

Thanks for all the other suggestions, I will post the updated schematic shortly.

Cheers,
Fergus

PS. All the boards were going to be mounted in the same box, I just made it modular so I could make one output board first to test, then scale up when I am happy it works.

Also I thought I can use the same PCB to add a headphone output to other projects easily or make small boxes with only one or two channels.

> I will put in some zeners on the input as a crude limiter. Is this an acceptable solution?

Same difference. Some gear can output 18Vrms cleanly, this input won't take that without clipping.

I'm just being paranoid. I've had to take absurd inputs and give good output. If you run your studio right, you won't put yourself in this pickle.