Headphone amplifier using SJ2520

[Electroma]

Hello everybody,

Usually I found all I want, looking in the different forums, with all contributors in DIYaudio, there is a lot of information and help around.

Now to the problem.

I'm trying to recycle a desktop speakers (creative SBS340), and I found that board have a couple of SJ2025, and a small F4558 chip. What I want to do is, use these part to build a nice headphone amplifier (portable), that works with 9V (for the portability factor), with volume, bass and treble control (and if its possible gain).

I will be really thankful for any schematic base in the use of the SJ2025 chip. The datasheet that I found is in chinese, and google translator does not do a good job with it (SJ2025 datasheet pdf, SJ - SJ2025).

thanks for your help.

 

[sgrossklass]

F4558 ought to be a basic dual audio opamp, nothing too crazy in terms of GBW or low noise or output driving but usable. (DIP-8 duals all have the same pinout.) Don't think I recognize that manufacturer logo, does anyone else? (Chinese domestic?) Anyway, that one should be useful in a basic inverting Baxandall tone control. (Not sure about impedances yet, I'd consider going 1/10 on the Rs and *10 on the Cs (MF Rs and film caps please). And it'll probably be a single-supply circuit, so you'll have to supply virtual ground. 10k/10k + 100µ should do.)

The output of that one could be run into the volume pot (10-20k typ).

SJ2025 appears to be a Chinese TEA2025 variant, not sure what the difference is.

Anyway, the TEA2025 datasheet sez,

The purpose of Rf is to reduce the gain. It is recommended to not reduce it under 36 dB.

36 dB, or a gain of about 60, generally is way more than you need or want in a headphone amp. It would inflate even the best-case input-referred noise level of 1.5 µV to 90 µV. That's enough to become bothering on any headphones of about 105-110 dB SPL / 1 Vrms up (which is only medium sensitivity). What kind of cans would you hope to be driving with it?

What you could do: Set up a low-impedance external feedback network (let's say 1k OUT to FEED, 100R to ground) and determine stability with various loads experimentally. (Try changing those 0.15µFs into a Zobel network with R~=3-10R. You can afford a few ohms of output series resistance. Headphone cables may be at 1+ nF, so 10 nF better work without any issues.) I suppose you don't have a scope though?

Plan B would involve using some other kind of amplification device. Got anything else? TDA2822Ms tend to be quite well-behaved here, down to at least a gain of 10. Or do you have some other decent audio opamp, a few diodes and some complementary somewhat-more-beefy-than-small-signal transistors?

 

[Electroma]

SJ2520

Thanks for the light in my problem,

I get a Sennheiser HD598, and the idea was recycle was is there, specially the amp chips, but buy required components is not a problem.

I like the Modified Baxandall Tone Controls, and i think could be perfect for a headphone amp, but I will be looking around.

any circuits for the main amp?.

Thanks for your sugestions.

 

[sgrossklass]

Main amp: Well, what kind of parts can you obtain?

HD598s obviously are 50 ohm nominal, and they appreciate an output impedance of less than 10 ohms. You could drive that with a NJM4556 à la cMoy, or employ a weaker but lower-noise part (e.g. NJM2068) plus a class AB output stage using low/medium power transistors (something like BC337/327, maybe types with more constant beta though; I'd suggest Re = 10R, Iq ~= 2.2 mA).

Would you like to keep it to a single 9 V battery or are 2 of them OK as well? Then one could use a proper split power supply, otherwise you'd have to mess around with virtual grounds and capacitor coupled outputs and feedback.