[udailey]

I am amassing components to build this amp to power speakers. Great read. Great effort. Do you still use/enjoy this amp? Any further thoughts on its sound?
Uriah

 

[wakibaki]

You're going to build my design? Or you're going to build a similar multi-5532 design? Either way, I'll do anything to help that I can.

I have lots of headphone amplifiers that I have designed. I don't listen to this one very frequently, but I remember the sound as highly satisfactory with the phones I was using at the time. The first thing I do with an amp when I build it (other than checking for fatal errors) is listen to it, kind of obviously, but critically, for problems. Hiss, crackle, distortion, flatness. Is the gain appropriate? Do the channels match? Is there any noise when the controls are operated? Are pops at switch-on and -off suppressed? When I've done that, though, I don't analyse the sound any further, other than notice problems if they should arise. I'm not really seeking some holy grail, I'm more seeking to create a finished product or range of features, but with essentially transparent performance. This amp let me test some digital stuff, it was quirky, and it wasn't expensive to put together. The feature here is that both channels can be adjusted in half-dB steps. Of course it can be built fully analog.

In common with other of my designs, I have numerous schematic variations. If I have a new idea, I rename the schematic rather than discarding. The nature of CAD is that if the schematic is correct, then the connectivity on the board is generally (although not absolutely always) correct. This enables me to evolve layouts. This does mean, however, that component values may sometimes not be shown correctly, as I always have the option of changing a value as I place the component.

I do have some recent tweaks I would regard as desirable to include.

Keep us posted with your progress or let me know if I can do anything.

Fred

 

[udailey]

Hi Fred. My aim is to build a similar design that will power speakers as well as headphones. I think I will have selectable gain or at least a headphone/speaker gain switch. The gain stage will likely be on a separate board so I can swap gain stages and treat the buffers as only a generic output stage. Like the F4. Im itching to try a transformer input with buffered secondaries to drive the output boards. In memory of my good buddy Mark I will also try a tube gain stage.
I have purchased the opamps and output isolating resistors. Already have .33uf caps and other required caps. I need to check out prototype qty pcb prices tonight to decide how big each output board will be and if its 2 or 4 layer.
Thats all so far. If you'd share your further tweaks I certainly am interested.
Oh, probably just an unregulated supply with a few voltage suppression devices to protect. Self's new amp book goes on and on about sticking to unregulated if you have good psrr.
Uriah

 

[DingoSmuggler]

I don't really see the point of paralleling regular dip opamps for speaker use, other than novelty factor, unless you're only going for a couple of watts of power output. That said, novelty is a valid pursuit in the world of DIY.

 

[udailey]

It IS mostly novelty. I have 1775 NE5532. I can mive some serious speakers. Noise is cancelled every time more opamps are paralleled. It should be easy to get noise to -120db and THD of an opamp is better than a mosfet amp.
Uriah

 

[wakibaki]

If you join Elector (free) you can download a .PDF of Doug Self's design, or you could in the past. Or I have a copy, since they gave it away, I don't think there's any issue about passing it on.

 

[wakibaki]

Here's a 2-channel * 16-opamp unity gain buffer:
 

 
It's 100*100mm, so U$25+shipping for 10 from seeedstudio.com

 

[AudioCats]

I don't see  the need for using such a big circuit board (which probably cost a lot more than the 5532).
Why not just use high quality 8 pin socket on a perf board, and stack the IC's, like how they do the TDA1543 DAC's. You can even "roll" different IC stacks. Imagine 25 OPA627 per stack...... very expensive...

 

[udailey]

I know you are talking to Wakibaki but I'll respond for myself anyway. The opamps I bought, NE5532, are already bought and for the price I paid I would never have gone DIP8. These are SOIC-8 and the output resistors, which are Vishay wirewound WSC-1-1.3-1R86, are 2515 SMD resistors, so 2 surface mount items and the caps will be through hole. The SMD SOIC-8 are easier to solder than dip sockets (when you use lots of good flux and a chisel tip iron). The way I am using the capacitors I will be able to NOT populate any of the caps until I test the board. So I will populate all opamps, then apply power to the empty through holes where the caps will eventually be. If the opamp is good the little headphone speaker I will have attached will play some music and I will know all is well with that particular opamp. 
Off track from your question I guess so far but back to it.. I definitely dont want to stack. If an opamp goes bad later I've got problems diagnosing which one it is in a stack. Boards look cool and with some work put forth on the couch with the computer I can forgo a LOT of wire to wire work. I want the pcbs to take up almost all the room in the chassis and I want them large to dissipate heat whereas a bunch of small boards would need to be stacked and generate heat. Thats another argument against stacking. DACs may be a bit of a different story but these opamps are going to be working hard at times. As to expense its just not much of an issue when I can get boards from a place like this:
http://imall.iteadstudio.com/open-pcb/pcb-prototyping.html
Super inexpensive and so far I like the boards I have ordered from them in the past. 
Uriah
 
EDIT: Now that I think about it more I wont be able to manage my capacitor trick so I am thinking maybe the output rails will be a solid strand of silver wire soldered to surface mount pads on the output of the resistors. This way I can test each opamp before soldering in the output rails by touching the speaker leads to the output landing pads, then I can put in the rail. If ever one fails its easy to remove some of the wire to see when the problem disappears, indicating which row of opamps is the culprit.

 

[udailey]

A bit of what I am planning so far. This would of course be multiplied several times.
Well, Picasa and Microsoft both are giving me issues sharing the photo so I guess I'll post a link. I always prefer to upload but Microsoft thinks I ought to upload to them, then share from there. No..
https://plus.google.com/photos/107238671950065164765/albums/6053062025973392625?authkey=COf3j5vRyZiO_AE
 
EDIT Wow the sure give you a long time to edit a post. I reconfigured the layout to something that would actually work. I had previously had a serious mistake in it. Now it makes more sense to me. Also gave surface mount pads in the film cap position. These are on both top and bottom layer so I may add whatever caps I have available.