The Schiit Yggy MIB uses the TI chips that actually recalibrate themselves depending on temperature. Temperature affects the resistance of the ladder network.
Stavro,
Temperature-dependent DAC linearity is hardly a new engineering challenge. Back in the 1990s, some telecom equipment already used temperature-indexed calibration lookup tables (LUTs) stored in ROM, automatically selecting correction coefficients as temperature changed, without requiring a new calibration cycle.
Those familiar with resistor-ladder DACs used in 1980s polyphonic synthesizers will also remember that resistor matching and thermal drift were well-known engineering concerns.
Of course, there's a substantial difference between those older discrete resistor networks and the monolithic 20-bit R-2R architecture of the TI DAC11001B used in the Schiit Yggdrasil More Is Better (MIB). Integrated resistor matching and thermal tracking offer significant advantages.
According to TI's DAC11001B datasheet (section 7.3.5), the chip includes a temperature-dependent INL/DNL calibration function. However, this function must be enabled and explicitly triggered by the host controller. The chip doesn't autonomously initiate recalibration when temperature changes.
Schiit's original documentation for the Yggdrasil MIB confirms the use of four DAC11001B converters, but doesn't describe whether or how this calibration feature is implemented. The MIB is also absent from Schiit's current Yggdrasil lineup, which now lists the Less Is More and Singular versions.
There's another interesting aspect to this. Schiit previously recommended leaving the original Yggdrasil powered on for best performance, specifically because its precision DACs benefited from thermal stabilization. That recommendation concerned an earlier converter architecture, but it raises a relevant question about the MIB.
Does Schiit trigger the DAC11001B's recalibration at startup, after thermal stabilization, periodically, or in response to temperature changes? Or does the implementation primarily rely on maintaining a stable operating temperature?
And if temperature-dependent linearity is significant enough to justify calibration, could the converter's measured performance vary depending on whether it was just switched on or had been operating for several hours?
A calibration performed at one temperature may no longer provide the same correction after the converter warms up, unless another calibration cycle is initiated.
This also raises a practical question for listening comparisons: are we comparing DACs under equivalent thermal operating conditions, or could their performance vary with warm-up time and power cycling?
Recalibration is a legitimate engineering solution to a known limitation of precision resistor-ladder converters. Its presence alone isn't evidence of superior conversion performance.
Ultimately, the meaningful comparison is the actual linearity, distortion, and thermal stability achieved by the complete DAC, and whether those differences are large enough to produce an audible improvement over other well-designed architectures.
A sophisticated solution to an old engineering problem isn't necessarily an advance in audio reproduction.
References:
Texas Instruments DAC11001B datasheet, section 7.3.5:
https://www.ti.com/lit/ds/symlink/dac11001b.pdf