Is it good to make the buffer of a hybrid headphone amplifier with LM1875?
Aug 2, 2008 at 10:55 PM Thread Starter Post #1 of 5

yzriver

Head-Fier
Joined
Jun 20, 2008
Posts
52
Likes
0
Hi, all, Is it good to make the buffer of a hybrid headphone amplifier with LM1875?

I have built a SOHA with an OPA2134 as buffer. I admire you guys the developments of the discrete buffers as JISBOS, SOHA and so on. However I wonder whether I can simplify my life by using audio power amplifier chip such as LM1875. It is supposed to provide enough power to play high dynamic sound.
I draft the circuit as following. The input can be the tube section of SOHAII or STACKII. In the power amplifier section, I can change the Vcc to adjust the output power. For example 10V Vcc can provide 4W to 8Ohm load; I can play the gain by changing R2/R1; I can add a R0 as attenuator for my headphone. I have a AKG501, which is 120Ohm.

LM1875poweramp.jpg



What do you think? Is the LM1875 a good audio chip, with a low distortion, and low noise?

Thank you for your reading.
 
Aug 4, 2008 at 12:59 PM Post #2 of 5
the minimum gain for that IC amp to be stable is 10 according to its data sheet. This is a little high for a line level source and you certainly wouldn't want any preamp before it unless you were amplifying the output of a record player. If you were to use this IC it would be best to use it on its own. You have neglected to label the non inverting / inverting terminals on your schematic. IC power amps do however make serviceable headphone amplifiers; until recently I used such an amp with resistors on the output.

*people may argue with this but with modern headphones its general best to have as low output impedance as possible and so your attenuation should be a duel tracked log pot on the input to the power amp. To see how to do this examine tangents cmoy schematics. Remember to chose sensible values of R1/R2.
 
Aug 4, 2008 at 1:15 PM Post #4 of 5
As a buffer: no. As the current amp stage: well, also no.

As kipman mentioned, its not going to like 1x gain very much at all, which rules it out as a 'buffer'. For the current amp stage after a tube, the package will end up with too much voltage gain after the tube and its own gain. I was also under the impression that that series of chips were actually kinda noisy, sure they're quiet with speakers but headphones end up quite sensitive.

You might be able to pull off an EHHA style amp, where you strap the negative feedback of the system to a differential amp at the start. That way the chip should still be able to operate with 10x gain but the system would be lower. That ends up being a fair bit of negative feedback I would think though...
 
Aug 4, 2008 at 7:07 PM Post #5 of 5
The LM1875/3875/3886 also like to put out a bit of DC. Not enough to harm speakers, but probably more than I'd be comfortable with on headphones.

If you poke around on DIYaudio.com, they do have a thread on how to throw away some of the gain with different feedback topologies, while still keeping the chip stable.
 

Users who are viewing this thread

Back
Top