Anyone ever used LM4780 for headphones?

[KaffiMann]

Short version:
Anyone ever used LM4780 strictly for headphones?
The LM4780 would be more than strong enough to drive several sets at once. From the spec sheet it says supply range 20-80V (v+ and v- 20-80V, not +-20-80V), seems like it could be driven by as low as 2x12V batteries.

Only have 600 ohm cans in-house now. So high-ish voltage is needed anyway.

Have some LM4780 laying about, also have a nice brand new unused case, input selectors, alps pot, nice black aluminum knobs, and a 15v+15v 50VA transformer. If this option is realistic it will only set me back the cost of a decent switchcraft or possibly neutrik connector.

Long version:
Looking at various headphone amps (kits or finished) and/or some lowish end external sound cards with powered mic in. Some seemingly decent kits about for 100$ up including p&p+import tax and such. If I go the sound card with powered mic in, the only half decent option seems to be the Lexicon IO|22, most of the other stuff seems to lack proper headphone out (judging purely from spec sheets). I find it hard to justify cost of headphone amp alone to more than 100$ in parts, shipment and taxes, especially since the Lexicon IO|22 has many other features I desire, no building required, costs MSRP $359.95 and is available here for that price. Locally available storebought amps are 150$ up and do not seem good enough. Seeing as there are several small caches of parts floating around my place, this would be a good opportunity to justify some of it's existence and at the same time save some money.

[Redshift187]

How are you going to deal with the high minimum gain of the LM4780?

[KaffiMann]

...
What am I missing?

I get a rough estimate of 375mVA/channel at 600 ohms with a supply voltage consisting of +/-15V rails. Granted, this is more than I will ever need, so taking it down a notch +/-12V rails give me about 240mVA/channel, just above maximum rating of my headphones, which is 200mW. Adding more headphones will just spread out the power consumption, the power delivered to each set will remain about the same level.

What is wrong with that? Not like I will have to listen to them with the volume turned up to full tilt. That is what my situation is now, full tilt, still audible distortion and not loud enough.

Anyway, the initial thought was to possibly use 15 to 20 v rails, adding a switch that engages an attenuator circuit at input and sets headphones or speakers out, so it is possible to use some small computer speakers with it on occasion. Setting 1: attenuator on input and headphones out, setting 2: no extra attenuator and speakers out. But 12V rails should still net me enough power for speakers, and it will remove the need of an extra attenuator + switch.

So I will grant you the point, rail voltage reduced to 12V.

Any other thoughts?

[Redshift187]

The minimum stable gain of the LM4780 is 10x. A normal headphone amp has a gain of between 1 and 4.

The power per set of headphones is irrelevant here. You can plug in 1 pair or 10 pairs and the LM4780 will deliver the same power to all of them. Don't use the power supply voltage to determine power. Use source output voltage (max) * gain to determine output voltage. Then P = V * V / R.

Example: 1V source output, 10x gain. Gives 10V output. 10*10/600 = 166mW.

10V RMS = 28V p-p. So you would need a power supply of 28V + 6V because it can't go rail-to-rail, giving 34V or +/-17V for a regular 1V RMS input signal.

[Redshift187]

Using your 15V+15V transformer:

30V - 6V = 24V p-p output limit
24V p-p = 8.6V RMS output limit
8.6*8.6/600 = 123mW output power limit

[KaffiMann]

Originally Posted by Redshift187

The minimum stable gain of the LM4780 is 10x. A normal headphone amp has a gain of between 1 and 4.

The power per set of headphones is irrelevant here. You can plug in 1 pair or 10 pairs and the LM4780 will deliver the same power to all of them. Don't use the power supply voltage to determine power. Use source output voltage (max) * gain to determine output voltage. Then P = V * V / R.

Example: 1V source output, 10x gain. Gives 10V output. 10*10/600 = 166mW.

10V RMS = 28V p-p. So you would need a power supply of 28V + 6V because it can't go rail-to-rail, giving 34V or +/-17V for a regular 1V RMS input signal.

Did not notice that. From the spec sheet: "The LM4780 is designed to be stable when operated at a closed-loop gain of 10 or greater". Think it's a bit silly to think of gain in steady increments though, with music you have no control over the average input voltage, maybe some control over maximum voltage. This will have to be attenuated at input anyway, so usually much lower than the 1V you mention.

The part about how many headphones one can connect is irrelevant, I agree, just like my post said, it was just an interesting thought, because when I think amplification it is completely new for me to not worry about other limitations when connecting lots of loudspeakers to the same power amp. Probably common sense to not worry about headphone power, but still a new thought for me.

Anyway. I have a 15v+15v transformer, so that is ac. The suggestions regarding different rail voltages would have to be with regulation. Rectified and with some caps the rail voltage will probably be just over or around 20V pending on smoothing.

BTW wondering about your RMS numbers in the last post, how do you calculate RMS to be 8,6Vrms with 24V p-p? Calculating with +/-12V get me to 8,48Vrms? Difference is negligible but just wondering why?

No matter, doing it reasonably close to your way:

40V - 6V = 34V p-p limit?
34V p-p = 12Vrms
12*12/600=240mW

Edit:
Sorry about the long delay in replying, been busy...