Recent Posts



Optimised USB isolation explained

The correct place to isolate the digital audio signals & gain the maximum sonic benefit, is on the incoming USB signals & NOT on the I2S signals coming out of the USB receiver - the horse has bolted at this stage - as whatever noise/disturbances are riding on the USB signal lines, causes distortion to be baked into the I2S signals.

The few USB audio DACs that are advertised as being 'galvanically isolated' are actually isolating these I2S signals but this has no effect on the distortion already baked into the I2S signals by the USB receiver.

This optimal approach to USB isolation is the exact approach taken in all my ISO devices...... read on.

It’s been on the wish list of many who have tried USB isolation &/or USB reclocking devices & found it has audibly improved their USB audio systems - to eliminate the 3 or more boxes & interconnecting cables needed & combine it all into a single device which incorporates the same functionality. The ISO (isolated) range of Ciunas devices provides this & more.

The concepts:

The Ciunas ISO-HUB provides an all-in-one combination of true isolation of USB 2.0 high speed signals along with the reformation of the USB signal waveform itself. It also handles USB 3.0 signals & cleanses/reforms these signals before output while also isolating the USB ground.

Combining this technology inside a USB to SPDIF converter or USB DAC is the next logical step & the new Ciunas products ISO-SPDIF & ISO-DAC are exactly this - a combination of the acclaimed Ciúnas converter/DAC with the above ISO-HUB technology.

Finally the new Ciunas ISO-PS is an off-grid, isolated, ultra low noise, fast & stable power source that delivers two isolated voltages 3.3V and isolated 6.6V outputs. It can be charged from the 5V of a USB port or an external higher current 5V supply.

A stable, ultra low noise power supply is the foundation for all devices to OPTIMALLY improve signal waveform or signal integrity. In these devices this low noise power is provided by very powerful but very safe, batteries which are user replaceable.

In the Ciunas devices, through various techniques, direct ultra low noise, clean power is provided from these sources without any intervening voltage regs. I have found that using voltage regulator (even the best ones) on the battery output, negatively affects the quality of battery power output. I believe that most digital audio devices require large currents in bursts & it’s important that the power system is capable of delivering these current demands without generating noise. Batteries are capable of this large dynamic current delivery - the batteries I use have 120 Amps instantaneous current capability. The result of achieving this are large sonic benefits in this area of digital audio.


The Issues:

It seems that there are at least two factors at play - leakage current mitigation & stable, fast, ultra-low noise power (particularly low noise at very low frequencies).

Current leakage is an inherent phenomena in all mains powered equipment - the reason why medical grade electrical devices exist is a recognition of this fact & keep leakage currents within the limits of medical standards - usually < 100 uA for patient safety considerations.

Why are leakage current issues being focussed on now when they have always existed in our playback systems? The rise in use of SMPS power is at the root of the problem. These tend to have higher leakage currents than linear PSes. Another factor are the RFI filters needed in order to pass pass certification in devices handling high speed signals & these filters increase current leakage to ground/chassis. SMPSes which have a typical switching frequency of 40-80KHz tend to actively produce a strong leakage current at this frequency & its harmonics. Linear supplies are a passive window which pass 50-60Hz mains frequency & its harmonics.

Leakage currents are not the same as the 50/60Hz ground loop noise heard as hum or buzzes from speakers. This ground loop noise occurs when two or more grounds in the audio system are at different voltage levels & a current flows between these points. Leakage currents arise between the AC & DC of both linear & SMPS but are usually higher in SMPS. A small current is “leaked” onto the DC & often flow in signal cables in order to get back to ground or neutral.

Leakage currents is not noise identified as a separate audible entity but rather noise which is interwoven with the music signal. How this sounds is variable because it depends on the leakage characteristics of the power supplies through which the leakage currents are flowing. In a lot of cases it is perceived as a dull, flat uninteresting sound lacking dynamics, life & vitality.

Off-grid power supplies, such as batteries are one step towards addressing these leakage currents by simply eliminating one of the leakage loop pathways. USB isolators block another leakage current pathway - through the actual signal lines, not just the ground or 5V power line. These two approaches, implemented in digital audio devices, prevent leakage currents from entering the sensitive analogue areas of these devices.

Another current finding in digital audio is that a well-formed digital waveform seems to result in improved sound quality. Why, is not currently known or revealed by current measurements. To achieve optimally formed waveforms requires a low noise, stable & fast power supply as the stable reference needed for generating such clean waveforms.

In my experiments with USB isolators, USB signal cleaners on the outputs always improved the sound. Why? The reason is likely reducing the jitter which all isolators create on their outputs.

My experiments with reclockers/regenerators showed an obviously audible improvement when powered by battery or similar clean, off-grid PS even when these devices use highly regarded, low noise voltage regulators, LT3042 or TPSTA4700. An even greater improvement was heard when these internal voltage regulator were bypassed & direct battery power applied in certain critical areas such as clock & USB hub chip. I have posted instructions for performing this minor modification on an existing USB reclocker here & those who have tried it have reported the large audible improvements. The importance of this power supply approach cannot be underestimated. I believe that this unmistakably audible improvement in sonics is to do with the very low noise of these batteries particularly at very low frequencies. If you look at all the ultralow noise regulators datasheets, including the two above, you will see that noise plots stop at 10Hz - they don’t show anything below this frequency & the noise is shown rising from from 100Hz to 10Hz. Extrapolating this slope to < 1Hz will show significantly rising noise levels for these regulators. Ultra low noise regulator datasheets that show noise measurements below 10Hz (ADM7150) show noise of ~30 to 50mV @0.1Hz under light load of 10mA current.

LiFePO4 batteries used in my devices don’t have this issue, staying at ultra-low noise levels all the way down to DC when supplying much higher currents.

The result of optimising all of these approaches - off-grid clean, fast, ultra low-noise & ultra stable power supply; USB isolation; & USB reclocking, results in a noticeable improvement & much more realistic sound - a deeper, more 3D soundstage where each element is more naturally portrayed - a portrayal which has an unforced flow & naturalness to it.