NJM4556A mentions driving a 150ohm load to +/-10.5V peaks, should sound plenty loud enough (360mWRMS, ear splitting on any normal sensitivity 'phones).
Thanks for the tip! The problem is that I don't like the way NJM quotes their specs---kinda in a "cloakish" manner that makes them hard to compare---NO current noise specs, for example. The 70mA drive capability is attractive, however. The 4556 show about a degree of magnitude worse THD than a 4562, but it's so low that I wonder whether it's audible.
They are pin compatible, aren't they - just try it? If you build a "CMoy"-style amplifier it's fairly easy to try different opamps and evaluate the sound quality 🙂
If you're ultra-concerned about THD and still need high current then you're better off looking at ADSL driver ICs (TPA6120, AD815). But then you have to abandon DIL.
The OPA1612 has monster specs for everything except current drive---looks to be about ~ 45mA. A bit light for headphones maybe.
Perhaps you're giving too much credence to the THD curves in the NJM4556A datasheet. They're obviously not real-world measurements. How do I know this? Because the output swing cannot reach 10VRMS on 15V supplies with a 200ohm load (70mA) - at best the output voltage swing would be 13V peak to the +ve rail.
You might run into power dissipation trouble with the DIL8 and also have to make this circuit survive shorting to ground. One huge advantage of some of the ADSL drivers is that they do have thermal protection
Yeah, I see your point. They quote a maximum voltage swing of +/- 13.5 volts into 2KΩ with a +/- 15 volt supply in their specs. So the 10.5 volts rms into 150Ω must be with the maximum supply voltage of +/- 18 volts. They are definitely playing specsmanship games with their datasheet; which makes you not want to trust anything they say!!
That's 150mW in 150 Ohms.
That ought to be plenty for anybody.
If 32 Ohm phones are around, it is 31mW, which is often ample but not super-power.
Often the real concern is how gain falls-off with real low impedance, hurting NFB.
NJM4556A is known to be an excellent and simple solution for headphone driving in many applications. No, NJM does not give great specs, they rely on designers actually *trying* their parts and liking them.
15 years ago I bought DIP-8 OPA2227 specifically for its 45mA current (as high as it got then?), and I have a couple left to play with...
But that's peak power, not effective / RMS. For a sine, peak power is twice as large as RMS power, which explains the discrepancy.
If you want to learn some more about how the NJM4556A behaves in practice, read up on the O2 amplifier design. It seems to be quite a well-behaved chip, as you would expect from something that first appeared in about 1984. Peak current output is about 80 mA. And yes, JRC/NJR datasheets suffer from "'80s datasheet disease".
If you want to learn some more about how the NJM4556A behaves in practice, read up on the O2 amplifier design. It seems to be quite a well-behaved chip, as you would expect from something that first appeared in about 1984. Peak current output is about 80 mA. And yes, JRC/NJR datasheets suffer from "'80s datasheet disease".
Last edited:
For conventional reasons, we quote "audio power" on Sine wave.
This is arbitrary. Any amplifier can be driven to near Square Wave. Actual speech/music signals are much less demanding than Sine. But that's the convention.
And what sgrossklass said about peak to RMS.
This is arbitrary. Any amplifier can be driven to near Square Wave. Actual speech/music signals are much less demanding than Sine. But that's the convention.
And what sgrossklass said about peak to RMS.
- Status
- Not open for further replies.