Hmm. I hadn't thought about the USB port possibly underpowering the GO, but it should be supplying 500 mA at 5V as specced.

I should clarify that I'm using the GOV2 in low gain into the Leck when testing that way.

Edit: computer volume maxed when used as a DAC, of course.

Volume is voltage, for a particular pair of headphones. As long as the amps stays within their respective power ranges, the one with more voltage available when the volume control is at max will play louder. Basically, the amp with the most gain without clipping will play the loudest.

Power (Watts) has to do with how much current an amp can deliver at a particular output voltage. Drive a 100 Ohm headphone at 1 Volt and you're putting out about .01 Watts. Drop the impedance of the cans to 1 Ohm and you'll have to put out 1 Watt to drive them. If the amp can put out 1 Volt easily, but can't drive it at 1 Watt---or one amp of current, in this example---the output devices will overheat and blow up. So the output power capability of an amp is essentially a rating by the manufacturer of when the amp is operating in its maximum safe output current into a particular resistive load without blowing up. Most amps will put out more power than they're rated for for a short period of time---before they blow up.  That's why you sometimes see power ratings for for both instantaneous and continuous safe power output.

I'm not sure. The Leckerton is supposed to do 90 mW into 100 ohms which is about 3Vrms while the GOV2 is supposed to do about 4Vrms. The GO depends on what the computer USB port can supply though.

I did some measurements on the GO1k some time ago, and I think it was able to do ~500 mW limited perhaps by USB port voltage swing and current:
http://www.changstar.com/index.php/topic,1678.0.html

Haven't measured the GOV2 yet though. But it seem to give plenty of juice in high gain to my HD600. Note that volume control is fully in the computer side now.

Nice. I saw these measurement first time here. Nicely done. ;-)

GO V2's full voltage swing on balanced out is 4.2V... it is a little bit higher than Geek Out 1000.  Also, due to the full balanced structure, we have TWO dedicated headphone amp inside for two channels. which let them work under easier environment, which is always beneficial.

I believe that. Here is what I got unbalanced:

Into 33 ohms -> 0.65 Vrms (low gain) / 2.18 Vrms (high gain)
Into 300 ohms -> 0.65 Vrms (low gain) / 2.18 Vrms (high gain)

I didn't measured balanced but it would give about double that so somewhere around 4.4 Vrms which is close to the claimed 4.2 Vrms if things don't current starve (depending on the load).

--- Power ---

This translates to the following power numbers:

Into 33 ohms high gain -> 144 mW (unbalanced) / 576 mW (balanced - assuming the USB port can give 123 mA)
Into 300 ohms high gain -> 16 mW (unbalanced) / 63 mW (balanced)

--- Current ---

Into 33 ohms I'm getting 144 mW so the current draw is likely 66 mA which is close to the limit of USB standard. The 1000 mW number subject to 4.2 Vrms limitations means that the Geek can support 238mA into 17.64 ohms as long as the USB port can supply that amount of current.

--- Bottom Line ---

Volume as far as the Geek V2 is concerned will indeed be a function of:

1) Voltage (as Tyll said) which is 2.18 Vrms unbalanced and very likely 4.2 Vrms balanced.
2) Whether the cans are driven balanced or unbalanced.

It is possible that the Leckerton will play louder if using the unbalanced out of the V2, but things may even out if using the balance output.

As it is I get more powar that what I get from my 2i2 and is more practical for listening. Higher rates for playback and DSD are a welcome bonus.