TPA6120A2 best headphone amp in the world?
Jun 8, 2004 at 9:53 PM Post #31 of 110
Hmmm......... I may have to build this
biggrin.gif
 
Jun 9, 2004 at 4:52 AM Post #32 of 110
Here are pics of it on a 16 pin DIP socket and a penny heat sink soldered to it. It sounds quite good with a dual voltage +14v and -14v supply. Drives my Grado headphones with very good bass. The penny does get very warm so a binder clip keeps it cool.
 
Jun 9, 2004 at 5:26 AM Post #34 of 110
I'm amazed it how much power it can put out and how simple it was to put together. The 700 mA output current spec is what caught my eye and it certainly has more bass than my double buffered BUF634 X12 amp. Only my X16 (gilmore) amp has more bass.
 
Jun 9, 2004 at 3:46 PM Post #37 of 110
i thought ur Gilmore V2 amp in DIP sockets was amazing but now you bring smd parts into the equasion, and with a penny as a heatsink! Now that is awesome
cool.gif
 
Jun 9, 2004 at 4:34 PM Post #38 of 110
To try to find weaknesses of this part, beyond the noise spec that may not be the best it could be, I notice that the performance at low frequencies is not as good as at mid and high frequencies...


JF
 
Jun 10, 2004 at 1:41 AM Post #39 of 110
Mr. Gilmore, would you mind sharing the resistor values you used with your build? I think I am going to try a dead bug arrangement with this sucker and sure as hell do not feel like swapping resistors out. Thanks!
 
Jun 11, 2004 at 11:38 PM Post #42 of 110
Quote:

Originally Posted by kevin gilmore
board picture
http://gilmore.chem.northwestern.edu/micro.pdf
Yes that is real size .9 x .9 inch.

You solder the chip thru the backside of the board then solder
on a piece of copper that is .3 inch wide and about 1 inch long.



I think that PDF's scaling is a bit off. I printed it out and there's no way you could solder the chip on that -- the pad spacing is off by a few mils, and the board is for a chip package .193in wide vs. the TPA's .292in...
 
Jun 12, 2004 at 12:50 PM Post #44 of 110
Correct, wrong package. Working on fixing that now. Glad i did
not have boards made yet. This whole heatsinking thing is a pain.
Nice thing however is that the chip is symetrical. So with a little
work you can solder it in upside down. Then you can attach a real
heatsink.

fixed file with same name uploaded. This is the top copper and silk only.

On my printer the vertical size is exact and the horizontal is ever
so slightly small.
 
Jun 12, 2004 at 1:06 PM Post #45 of 110
Quote:

Originally Posted by kevin gilmore
Nice thing however is that the chip is symetrical. So with a little work you can solder it in upside down. Then you can attach a real
heatsink.



That's exactly what I was thinking. Going to pick up some random cheapo soic chips localally to practice bending the leads with some tiny needle nose pliers. I am somewhat worried about the mechanical strain this places on them. Mostly the output pin. Besides making it weaker, doesn't it also add a little resistance at the bend? With regular through hole stuff, when I need to bend it, I always use the shaft of a small screw driver to get a nice rounded bend, thus minimizing the strain. I am kinda suprise this chip is not available in any other type of package.

Started working on a PCB design myself. But I think I want to use two of these chips for balanced output. One chip per channel. One side inverting, the other non-inverting. Should do the trick nicely. Though this doesn't apply to my intended implementation, the voltage rails could be a lot lower this way too. Downside is you now have two protection output resistors in series with the load, thus slightly lower damping factor. I use HD650s, so with their 300 ohms, it shouldn't be a big deal adding 2x 10 ohms.
 

Users who are viewing this thread

Back
Top