I certainly agree with Gary's response - the Passport's 12VDC output will allow the MG3 only 10Wpc into 8-Ohm (vs. 24VDC, which allows the MG3 to deliver 32Wpc into 8-Ohm).
But I'd like to offer some warnings about my 24VDC battery-powered solution for the MG3.
I am currently using this 3600 mAh, 6-cell LiPo to power the MG3 at 25.2V (fully charged) down to 20.4V (when I put it back on the charger):
And I recently purchased this larger, 5300 mAh, 6-cell, LiPo (having the same length as the one above), on sale at the manufacturer's web site: http://www.gensaceusa.com/98p-30c-5300-6s1p.html
These things are somewhat dangerous, however, as this 30C-rated, 5300 mAh version can supply 159 Amps continuously, without overheating (peak current is much higher still).
Their discharge must be monitored during use, to prevent then from dropping below 3.0V per cell - which can damage the battery and even cause fires that emit toxic soot and smoke, even if they don't burn down your house. Charging requires some know-how, too, using a "balanced charger" like the Thunder AC6, to ensure that individual cells are not charged to voltages higher than 4.2V, which again, can cause damage and increase the possibility of a fire.
In other words, I don't recommend this solution unless you're "into batteries" and can enjoy the ritual of being disciplined in safe handling.
@Chris J - I see this again and again in various articles I've found online: the recommendation to use a 3:1 ratio (where the series resistor has three times the value of the parallel resistor) when designing a 2-resistor network.
Can you please tell me what happens sonically, with dynamic headphones especially (as opposed to planar magnetic), when the ratio is increased or decreased?
As previously posted, the resistor network that Jan Plummer (of TBI Audio) built for my MG3 amp uses two resistors, but their values have a ratio of 5:1.
I just used 3:1 as an example.
With low numbers like this, your amp should sound good with just about any headphone. But it's really just trail and error. I'm not surprised "the box" sounds good with the HD600, but is disappointing with the LSDs, the power/current/voltage ratios are quite different.
The numbers you show look just as valid, but:
If R1 = 10 Ohm and R2 = 2 Ohms you would only get more attenuation (approx -15.6 dB), i.e less hiss, and also more likely to move you to a "nicer" part of your volume control range.
I say nicer because we often have to run the volume control way down around 8:30 or 9:30 because the whole chain may have too much gain for headphones when we use speaker amps.
I used to have a Yamaha CR-2020 receiver with an output impedance of approx. 120 Ohms (or was it 220?).
Anyway, I had to press the -20 dB button on the receiver to get away from using the very bottom range of the volume control.
Without the -20 dB button pressed in the channel tracking could be a bit dodgy. Irritating!
With the -20 dB button engaged the volume control was more in the range of 9:00 to 11:00 o'clock. Very useful.
If I was doing this I would do it this way:
1. try driving phones directly
what do we hear? too much hiss? do we need to operate the volume control at the bottom end?
2. try values like R1 = R2
still too much hiss? still problems with volume control?
3. try this: R1 is three times higher than R2
still too much hiss? bad volume control travel?
4. try R1 is 5 times higher than R2
5. if still too much hiss, try R1 is 9 times higher than R2.
this is approx. -20 dB of attenuation.
Really what it comes down to is, subjectively, where is the best compromise between low hiss, good volume control travel and hopefully we still have enough power delivered to phones.
As you have found out, sometimes you just can't find a good compromise.
Jan must have done a few calculations, but if he doesn't actually have a pair of LCD-2s, it's hard to nail down good real world values for those specific phones.
Maybe try a lower ratio?
A friend of mine drives LCD-2 straight off the output of a small DECware vacuum tube power amp, I think it's about 10 Watts into 8 Ohms?