[roadbuster]

Hello,

I am in need of a compact, low-cost alternative to the BUF634 for an output stage in a headphone amplifier. I have heard positive things about the EL2001, but I can't find a single distributor for the IC (has it been EOLed?). An output stage constructed from discrete components would be both cheap and high-performance, but it would take-up far too much space. What are my options?

 

[dokebi]

You could try etching your own. Try searching for jung buffer and diamond buffer.

 

[Dreamslacker]

How about the buffer used in the PPA? The HA3-5002. I believe Newark carries them and they're cheaper than the BUF634's.

 

[bg4533]

The EL2001 has been discontinued for a long time now.

The PPA buffers are not pin compatible with the BUF634, but if you are using protoboard or designing your own boards that doesn't matter. The PPA buffers also lack output protection, so could cause a problem if someone pulls the headphone plug out with music playing. A lot of people think the PPA buffers sound better than the BUF634 too.

You could also check Ebay for BUF634s. They sometimes go pretty cheap.

 

[setmenu]

I would certainly recommend the HA 5002 buffer, I use this chip in a pimeta
and my dac/amp combo.
I waaaaay prefer the sound of it to the buf 634!
As for the lack of output protection, it is possible to current limit the device,
I have not tried it myself but the general consensus is that the use of a current limiting resistor would compromise the sound quality.
Other than the above one could use a locking jack or loop the headphone cord around something to create a strain relief and prevent accidental partial
disconnection.
The only downside [for portable use] with the 5002 is it consumes more quiescent current than
the low bandwidth 634 .[in my view the 634 in low bandwidth mode sounds extra horrible anyway]




Setmenu

 

[guzzler]

sijosae is your friend (for Jung buffers at least

)

http://www.headphoneamp.co.kr/bbs/vi...esc=asc&no=124

Original headwize thread

http://headwize.com/ubb/showpage.php...ijosae;buffer;


For anyone here old enough , there's a blast from the past near the bottom of that thread; our old friend MERTON. What a guy....

g

 

[roadbuster]

Quote:

Originally Posted by dokebi You could try etching your own. Try searching for jung buffer and diamond buffer.

Too big

Quote:

Originally Posted by Dreamslacker How about the buffer used in the PPA? The HA3-5002. I believe Newark carries them and they're cheaper than the BUF634's.

Ah, thanks. I was searching for the HA2-5002 earlier, but couldn't find a useful supplier. I was really hoping I could find something carried by Digikey or Mouser, though, because I'm located in Canada and I don't want to deal with a third company (shipping is a killer).

Quote:

Originally Posted by setmenu The only downside [for portable use] with the 5002 is it consumes more quiescent current than the low bandwidth 634 .[in my view the 634 in low bandwidth mode sounds extra horrible anyway]

Well, 0.2W leakage power is a lot more than what I'd like, but the cost savings are important.

Regarding the low-BW mode of the BUF634, it seems odd to me that it would sound worse than the high-BW mode. The high-bandwidth mode has a higher input bias current and a lower input impedance.

Quote:

Originally Posted by guzzler sijosae is your friend (for Jung buffers at least )

Wow, thanks for the link. I scribbled together a buffer made out of discretes last night, but it was about twice the size as the one implemented by Sijosae. The problem is that Sijosae's is still about 3x wider than what I can afford. In fact, I'd much rather go with a 5 pin DDPAK or 8-SOIC than an 8-DIP in order to conserve space.

Thanks for the responses, everyone!

 

[guzzler]

You can rearrange the sijosae buffer to make it vertical mount...

SOIC or DDPAK is going to limit your choices, and you don't have many to start with. As setmenu said, the HA3-5002 is nice, and you can get in SOIC. Other alternative is to use a high current opamp. I've never used one personally, but a lot of people have used the OPA2604 in this role...

edit: meant to add, the sijosae buffer can be shrunk even more, believe it or not, by judicious tombstoning of resistors, and possibly even going SMD...

g

 

[tangent]

Quote:

it seems odd to me that it would sound worse than the high-BW mode.

Higher bandwidth means lower distortion. And in this particular case, higher bandwidth means the buffer is operating deeper into class A.

Some tests have shown that the widest bandwidth setting isn't as good as backing off slightly from that, so there may be something to your intuition.

 

[roadbuster]

Quote:

Originally Posted by guzzler You can rearrange the sijosae buffer to make it vertical mount...

Wouldn't be very mechanically-rigid at the joint if it were vertically-mounted, and that worries me.

Quote:

edit: meant to add, the sijosae buffer can be shrunk even more, believe it or not, by judicious tombstoning of resistors, and possibly even going SMD...

I'll actually think about this. Using SOT-23 transistors and 1206 resistors would make for an interesting design. A quick sketch tells me that it would be at least 14mm x 17mm in size, though. A DIP8 package has a footprint of about 7mm x 9.5mm

Quote:

Originally Posted by tangent Higher bandwidth means lower distortion. And in this particular case, higher bandwidth means the buffer is operating deeper into class A.

What is the cause of this distortion?

 

[guzzler]

You can do it with 90* SIP pins on the buffer that solder directly onto the board. A wee bit of epoxy round the base and it would be pretty stable.

The Glassman buffers were SMD Diamond buffers and took 4 DIP spaces. I'm not sure of the dimensions though. Also, Peranders has an SMD buffer as well

g

 

[tangent]

Quote:

Using SOT-23 transistors and 1206 resistors would make for an interesting design.

Been done. Seek ye out the Glassman buffer. It's about 1 inch square, and about as tight as it can get without removing features.

Quote:

What is the cause of this distortion?

Lots of things; RFI for one. But the cause isn't especially important to this discussion. What's important is that higher bandwidth generally means more available feedback, which means the amp can better quell those distortions.

Look at it this way: a uA741 has enough bandwidth (on paper) for audio, but when was the last time you saw one used for that? Ummm, 1979, maybe?

 

[robzy]

Would it be worth replacing the 634's in my PIMETA with one of these passive(?) versions of a buffer? Such as the jung one guzzler linked?

[edit]: And here is a stupid question, what are the LEDs for on the jung one?

Rob.

 

[tangent]

Quote:

Would it be worth replacing the 634's in my PIMETA with one of these passive(?) versions of a buffer?

If you mean "will you hear a difference", then yes, probably so. Whether it's worth it or not is up to you. It won't be easy to put together, particularly considering how you'll have to bend things to fit into the crammed area the PIMETA board affords.

And no, a buffer is not passive. There is no such thing as a passive buffer. Passives are resistors, capacitors and inductors. Anything with a semiconductor in it is considered active, especially if it has gain. (And a buffer does have gain: current gain.)

Quote:

what are the LEDs for on the jung one?

Use a standard ~2V green one until you know enough to choose a different type.

 

[robzy]

Sorry, i didnt mean passive... i meant discrete. Im still kind of new to all of this.

As for the LEDs, i meant to ask what they did? Are they just a "power-LED" or do they light up to signify something?

As for putting them in a PIMETA, i guess i could stack up some sockets in order to get it high enough for it to fit, but the gains in quality would probably be drained out by the degradation

Rob.