Schiit Mjolnir 3: power output and what it will drive

The Schiit Mjolnir 3 is a desktop amplifier at $1199. It delivers 1.0 W into 32Ω and 500 mW into 300Ω, which ranks it 84 of 280 for raw power among the sources in this database.

Measured at 100 dB SPL it has comfortable headroom on 560 of the 567 dynamic and planar headphones catalogued here, enough but not generous on 6 and not enough power for 1. Hybrid tube/solid-state. Beautiful tonality with slam.

The manufacturer publishes 1.0 W into 32Ω, 500 mW into 300Ω. Everything else in the table above is worked out from those figures, and this table says which is which.

The balanced output is rated separately: 2.0 W into 32Ω through the XLR4 jack.

Derivation holds output voltage steady between and beyond the published loads rather than output power, because that is what an amplifier actually does. The full method is written up here.

The noise figure above is converted, not read. Schiit does not print microvolts for this one. It prints a signal-to-noise ratio together with the voltage that ratio is referenced to, in the same block of the same table: “SNR: >98db, A weighted, referenced to 2V RMS”. From those two numbers the noise follows by one division — 2.0 V ÷ 10^(98/20) — which is where the 25.18 µV comes from, at the single-ended setting, A weighted. The ratio is printed as a bound rather than a measurement, so the figure is a ceiling: the noise is at most that, and may be lower. This is the same situation as a sensitivity printed per volt instead of per milliwatt: published, but in a convention nobody converts.

Source: schiit.com.

Ordered from easiest to hardest to drive. Headroom is measured at 100 dB SPL, which is louder than almost anyone listens — a small negative number means the source still plays loud, just without reserve for peaks.

These figures come from each manufacturer's own sensitivity rating. A handful of planars — the HiFiMAN Susvara and the Abyss AB-1266 in particular — are widely reported to need far more than their published sensitivity implies, so treat a bare pass on those as optimistic rather than settled.

At 25.18 µV of output noise, here is where that noise lands in the five most revealing in-ears in the database.

This one hisses. With the most sensitive in-ears in the database its own noise lands around 49 dB SPL, which you will hear between tracks and in quiet passages.

At 0.13Ω output impedance this source satisfies the one-eighth rule with every headphone in the database, down to the 3Ω in-ears. Nothing here will have its frequency response shaped by the source.

The rule is a rule of thumb, not physics: a source impedance below one eighth of the headphone's keeps the interaction under about 1 dB. The checker works it out for any pair.

Of the 280 sources here, 15 do everything the Schiit Mjolnir 3 does as a headphone output: reach 100 dB on every headphone it reaches, stay inside the one-eighth rule wherever it does and stay under the audible-hiss threshold wherever it does. The cheapest of them is the Schiit Magni Unity at $119, against this one’s $1199 — $1080 less.

This counts a headphone output and nothing else. Inputs, converters, microphone preamps, balanced outputs, the volume control, the case and the warranty are all outside it, and any of them can be the reason to buy the dearer box. What the count does say is that the three numbers a maker can be held to do not separate them.

Worked out in advance against its closest rivals: vs Gustard H20 · vs Gustard H26 · vs Schiit Asgard 3 · vs Schiit Jotunheim 2 · vs Schiit Lyr 5 · vs Schiit Lyr+ · vs Schiit Magni Heresy · vs Schiit Magni Piety · vs Schiit Magni Unity · vs Schiit Magni+ · vs Schiit Midgard · vs Schiit Ragnarok 2 · vs Schiit Valhalla 2 · vs Schiit Valhalla 3 · vs Schiit Vali 2+ · vs Schiit Vali 3 · vs SMSL VMV P2 MK2 · vs SPL Phonitor se · vs Violectric V222. Each of those pages lists the headphones that tell the two apart.

Or put any two side by side in the source comparison tool, or run your own headphones through the amplifier matcher. Or see every Schiit source in the database, and how much of what this site says about them came from Schiit’s own documents.