Ferrum WANDLA GoldenSound Edition Measurements & Discussion 2024
There have been a number of updates to the WANDLA GoldenSound Edition, and the measured behavior is slightly different from the original.
Ferrum WANDLA GoldenSound Edition Measurements & Discussion
DISCLAIMER – THIS IS NOT A REVIEW
This is not a review, but rather an effort to be open and honest about the WANDLA GoldenSound Edition’s objective performance and the changes made to it.
I get a portion of the proceeds from the sale of the WANDLA GoldenSound Edition, which is a joint venture between Ferrum and me. I will not be making this product because reviewing it would put me in a blatant conflict of interest due to the direct cash incentive I am receiving.
I will present the measurement data without offering any subjective interpretation, except in cases where I believe it is necessary to clarify something for the reader. As a result, this post will be slightly more sparse than usual.
Before I had the chance to collaborate on this product with Ferrum, you can read my opinions on the original Wandla here if you’re interested in a review of the Wandla:
Do not buy this product just because of this post or my opinions on it; instead, wait for reviews and feedback from actual customers before making a purchase.
Design and I/O
By keeping the same external appearance and form factor as Ferrum’s OOR, ERCO, and HYPSOS units, the WANDLA GoldenSound Edition allows consumers who own various items in the lineup to still create a matched stack. The unit’s splash screen and user interface retain the features that distinguish the GoldenSound Edition.
Owners of both the OOR and the HYPSOS were curious about the possibility of powering both vessels from the same HYPSOS power source, or if they would have to decide between the two. In response, Ferrum introduced a splitter unit that enables the HYPSOS to supply power to two devices at once: an OOR and a WANDLA.
With its matte black chassis, corten steel square accents, and lighted Ferrum logo, this device follows in the same vein as other Ferrum offerings. However, unlike the HYPSOS, it has a much larger display that doubles as a touch screen. This will make it easier to use all of the settings and capabilities that the WANDLA offers. You can adjust the level and other features without taking your hands off the device thanks to the metal remote that comes with it.
Optical SPDIF, coaxial SPDIF, AES, I2S, HDMI ARC, and RCA analog inputs round out the WANDLA’s seven input possibilities.
The WANDLA may double as an analog preamplifier and DAC since its analog RCA inputs bypass an ADC and go straight to the outputs via an all analog signal channel that also features analog volume control. Those who own a vinyl turntable will like this feature, since it allows them to use the WANDLA as the main controller for their whole system. It should be mentioned that the unbalanced signal that is inputted into the analog input is instantly balanced within the device, and it stays balanced all the way through, including the output stage and the volume control.
Although the DAC does not have balanced analog inputs, it does have balanced outputs.
Internals
When inspecting the WANDLA’s inner workings, the central daughterboard, SERCE, may be the first object that catches the eye. Because the WANDLA does not use an XMOS chip, this all-in-one audio control board serves as the device’s central processing unit (CPU), processing all digital audio inputs (including USB), and handling all controls and outputs to the display. It also provides flexibility for higher performance digital audio processing than what is available on a standard DAC chip.
Drawing on this potential, Ferrum collaborated with Jussi Laako, creator of the high performance upsampling program ‘HQPlayer,’ to create a suite of oversampling filters for the WANDLA that enhance reconstruction performance. These new ‘HQ’ filters offer a filter response that completely attenuates by 22.05 kHz, yet has no significant early rolloff below 20 kHz. If you prefer the more traditional ESS filters or would like to hear how they stack up against the HQ ones, you can find those in the menu as well.
Regarding the DAC itself, Ferrum’s IV stage design is to the top right and left, and the ESS9038 chip is visible in the lower right, surrounded by components for voltage regulation and power filtering, as well as the clock. I was informed that they tried a few different IV stage designs that had great measurement findings but didn’t sound well enough in their opinion. This IV (current to voltage conversion) stage was designed using both measurements and listening as validation. They decided on the design you see here because they thought it had the best sound after going through several iterations and redesigns.
The output stage itself and two MUSES laser trimmed resistor attenuation integrated circuits (ICs) that enable analog volume control are visible above this.
The user has the choice of selecting digital volume control, in which case all attenuation is done in the digital domain by the SERCE processor, or in the analog domain, where the DAC converts the digital signal at full level, and then attenuation is done by the resistor ladder ICs.
Some enthusiasts feel that less digital manipulation is always better, but some would argue that digital domain attenuation can allow for objectively more accurate results than analog in some instances. In any case, the user has the choice available on the WANDLA to use whichever they prefer.
Additionally, in the WANDLA GoldenSound Edition, a new hardware voltage divider circuit has been added, which allows the user to bring down the output voltage of the WANDLA from just under 10V, to just under 4V, thereby ensuring compatibility with any amplifier, without needing to add the MUSES chips into the signal path or apply any digital volume control. This new feature can be used in conjunction with either the analog or digital volume control as well if desired.
GoldenSound Edition Changes and Features
For the WANDLA GoldenSound Edition, I worked with Ferrum to develop six new features and changes that I personally feel provide an enhancement to the user’s listening experience:
– Elevated digital headroom: Many modern musical tracks are produced in a way that leaves very little headroom in the recording, often referred to as the “Loudness Wars”, This can cause clipping and additional distortion during playback on most DACs due to the oversampling filter being unable to accurately reconstruct strong transients. The WANDLA has its own digital processing headroom, but GoldenSound Edition has elevated headroom to accommodate it to its additional DSP functions, which are described below, and therefore, unlike the vast majority of available DACs, the WANDLA GoldenSound Edition is highly immune to intersample overs even in bypass mode at full output level, allowing you to play back any music free from additional distortion.
– Spatial Enhancement: The Ferrum SERCE module’s power is harnessed by the WANDLA GoldenSound Edition to create a unique spatial enhancement that expands the soundstage and improves the clarity of separation between musical elements while avoiding the problems associated with other methods. To provide an improvement specific to each use-case, there are two modes available: Headphone and Speaker. This is accomplished without adding extra reverb, distortion, or coloration to the frequency response.
Tube Mode: The WANDLA GoldenSound Edition has a tube mode that purposefully amplifies the distortion contributed by a tube amplifier to the even-order harmonic distortion generated by the DAC. Without really altering frequency response, this produces a richer, warmer sound without any of the other problems associated with tube-based devices, including increased noise.
Impact Plus: For those that prefer a little extra low-end in their music, the WANDLA GoldenSound Edition has an Impact + option. This function, however, uses a specially designed two-band EQ in place of a standard bass shelf, which not only raises bass but also adds punch and dynamic impact.
No MQA: The WANDLA GoldenSound Edition does not employ any MQA processing; instead, all content is preserved as unprocessed PCM, with the exception of sound tuning options that you may choose to activate. The SERCE module of the WANDLA is heavily utilized by all of the DSP features of the WANDLA GoldenSound Edition. Consequently, MQA decoder is not included in the GoldenSound Edition to guarantee flawless operation.
Hardware Voltage Adjustment: The WANDLA GoldenSound Edition has an output voltage adjustment feature that lets the user switch between an output voltage of slightly less than 10 Vrms and slightly less than 4 Vrms. This is made possible by a specific hardware voltage divider in the DAC’s output stage, which enables the user to choose a lower output voltage and guarantee compatibility with all amplifiers without requiring the use of analog volume control components or digital volume control in the signal path. With board revision v4.0, this hardware voltage divider will be accessible on WANDLA standard edition and WANDLA GoldenSound Edition machines.
While they will support this capability, revisions prior to v4.0 will do so in a somewhat different manner. Instead, the current analog volume control chips will be used in these updates.
Measurements
Measurements Include:
1 THD plus N (SINAD)
2 THD+N in Relation to Frequency
3 Low Level Signal Output Upsampling/Reconstruction Filter
4 Linearity of Jitter Thd against Output Level
5 Resistance of Output
6 Volume Correspondence
7 Intermodulation distortion caused by noise (IMD)
8 Multitone Crosstalk
Additional test data and measurements are included in the complete report.
Test Setup
1 Analyzer for Audio Precision APx555 B-Series
2 The Furman SPR-16-Ei WANDLA, which was warmed up for six hours before testing, provides regulated 230V power to the measurement setup and the equipment being tested.
3 U nless otherwise indicated, WANDLA is in “bypass” mode.
4 Used as a stock power supply for all tests
5 Unless otherwise noted, USB input and balanced output are utilized.
6 A 100kΩ load is used for all tests unless otherwise noted.
7 The Wandla GoldenSound Edition, board revision V4.1, firmware V1.4.0 beta, is used for the tests.
8 The entire reports contain specific analyzer/filter configurations for every measurement.
CH2 (Red) = Right, CH1 (Blue) = Left
Full Measurement Reports
Results
IMD SMPTE: -105dB
Noise Level RMS (20-20khz): 7.99uVrms
DC Offset: 45uV active, 102uVidle
Susceptible to intersample overs: No
Output Impedance: 44Ω Balanced, 22Ω Single-Ended
THD+N / SINAD
Note that SINAD is a measure that incorporates a few of performance factors while leaving out a large number of others, so it should not be used to assess a product’s total performance. When assessing a device’s total objective performance, it should be considered one component among many, and the kind or structure of distortion can frequently be more significant than the distortion or noise level per se.
1khz Sine 10V (Bypass) Balanced Output:
1khz Sine 10V (Bypass) Balanced Output – Tube Mode Enabled:
The DAC’s 2nd order harmonic distortion is greatly increased in tube mode, while other harmonics are barely modified. In order to offer further personalization, I intend to include choices in the future that offer distortion profiles that closely resemble particular tube models.
Note: This functionality is accomplished by purposefully misconfiguring the DAC itself such that it causes additional distortion but only in the intended way. It is NOT performed via DSP in the SERCE module.
Impact+ Mode
The “Impact+” setting boosts the bass frequencies but adds another sense of “punch” or dynamics/impact in addition to “more bass” by using a multiband adjustment instead of a traditional bass shelf. The outcome of the modification is comparable to that obtained by adding an air gap to some headphone models, such as the Abyss 1266, without sacrificing subbass.
THD+N vs Frequency (96khz capture bandwidth)
(The analyzer employed a bandwidth of 96 kHz. Since the test setup is different, do not directly compare this to standalone audible band THD+N measurements.)
Low Level Signal Output
To look for any odd behavior, this test runs a very low level signal through the device.
Reconstruction Filter (Nyquist/Oversampling Filter)
HQ Apod filter fully attenuates by 22.05khz, whilst not rolling off under 20khz
Jitter
The performance of Jitter is practically flawless. It would be best to use USB alone with the WANDLA. It does not require pairing with an additional DDC.
THD vs Output Level
The WANDLA’s distortion profile is continuously proportional to output level.
Linearity
This test uses the gadget to play a tone at 1 kHz while gradually lowering the level. The degree to which the device’s actual output closely resembles its intended output is measured by its linearity. There is a 0.3dB nonlinearity, which would appear as a 0.3dB rise on the graph, if the level was stepped down by 20dB while the device’s output was 19.7dB quieter.
This test uses the gadget to play a tone at 1 kHz while gradually lowering the level. The degree to which the device’s actual output closely resembles its intended output is measured by its linearity. There is a 0.3dB nonlinearity, which would appear as a 0.3dB rise on the graph, if the level was stepped down by 20dB while the device’s output was 19.7dB quieter.
Bandpass filtered, this is merely gauging the 1 kHz tone’s accuracy.
No bandpass filter, this factors in noise into the result
Output Impedance
XLR (Balanced): 44 Ω
RCA (Single-Ended): 22 Ω
Volume Matching vs Attenuation
This test uses the volume knob to change the amplifier’s level and displays the dB difference in level between the two channels.
Certain amplifiers will exhibit channel imbalance, especially at lower volumes. This test determines how closely an amplifier’s two channels match as the level is lowered.
The WANDLA has two modes of volume control: analog volume control with a resistor attenuation chip that has been laser-trimmed, or digital volume control. With the analog volume control, channel matching was always within 0.05dB, and with the digital volume control, it was flawless.
Noise
No switching noise or other sources of noise observed upto 1Mhz.
IMD
Crosstalk
Multitone
