Chip buffers

[Atilla]

I know a lot of different buffers have been discussed in quite a lot of those forums, but I've been wondering if anyone has experience with the latest and the greatest that is out. What good, chip buffers are there and still available nowadays? The latest mentions for chips that I've seen talk about the LME49600. How does that compare to other commonly used buffers, compared to people's experiences with it? For example, compared to BUF634.

How about trough-hole buffers, for all those SMD-weary? (I'm sure I can solder a TO-263 allright, but just wondering)

They all have a catch somewhere though, the LME49600 is hard for me to find, the BUF634 has definitely worse specs and costs quite a lot of money...

Is anything else worth the try?

[peranders]

LME49600 seems to be a slightly better version of BUF634 and you can choose the package according to your own preferences. Since the buffer is a big part it's rather easy to solder even in SMD verision. If you want to solder the one in the picture, just burn for some time and two soldering irons will of course help but it's not necessary.

[Atilla]

Actually, with this type of package, I've done trough-hole work quite a few times, with the help of some leg-bending. It's not the best solution of course, but it's worked for some small prototypes I've done.

Your headphone amp is quite impressive piece of work, by the way

[peranders]

Thanks!

[theAnonymous1]

Atilla,

I've built the heaphone amplifier in the LME49600 datasheet and I'm very pleased with it.

Let me know if you are interested in any of these, I have some I can spare if you have trouble sourcing them locally.

[Atilla]

Mr Anonymous, you've got mail.

So, there's no other buffer IC that people have used that is popular? I know there's noting too much to ask from such a device so the choices are limited. I guess except for those units people would go for discrete diamond and Sijosae buffers...

[zener_diode]

The LME49600 is actually quite excellent as a buffer. I measured the THD+N
of it and there is no difference between it and going strait into my ADC.
THD+N at a level of around 1.7v RMS is < .0006% (the limit of my DAC/ADC)
with a load of 5k ohms. Also to note, the higher order harmonics didn't seem to
change, which means that the output transistors on the chip are very closely
matched and don't seam to be adding any crossover distortion. Same spectral
content at about 5mv RMS both with and without the buffer. The noise floor
was also about the same. All this without any negative feedback!

[Andrea:]

Originally Posted by zener_diode

The LME49600 is actually quite excellent as a buffer. I measured the THD+N
of it and there is no difference between it and going strait into my ADC.
THD+N at a level of around 1.7v RMS is < .0006% (the limit of my DAC/ADC)
with a load of 5k ohms. Also to note, the higher order harmonics didn't seem to
change, which means that the output transistors on the chip are very closely
matched and don't seam to be adding any crossover distortion. Same spectral
content at about 5mv RMS both with and without the buffer. The noise floor
was also about the same. All this without any negative feedback!

Did you use it open-loop?

[Andrea:]

Originally Posted by peranders

LME49600 seems to be a slightly better version of BUF634 and you can choose the package according to your own preferences. Since the buffer is a big part it's rather easy to solder even in SMD verision. If you want to solder the one in the picture, just burn for some time and two soldering irons will of course help but it's not necessary.

Made with a Macintosh? Does that make it prouder?

[zener_diode]

Originally Posted by Andrea:

Did you use it open-loop?

Yes, open loop (that's why I said "without any negative feedback").

I also measured the IMD and there was no difference with or without the
buffer, but my setup seems to have high IMD, so maybe there is a tiny
difference that I could not see.