My good friend Mr. @Cardri and I have been chatting about this whole Matrix and external LPS scenario. He actually reached out to Arthur with some questions and learned that the the Matrix actually using a step down stage on the external power supply section of the Matrix that reduces any voltage fed to the Matrix down to 5V. So, that being said, it makes no sense to feed the Matrix anything beyond 5V and this very well could be the reason that the external port creates more measurable noise than the USB 5V buss.
This also makes me believe the Matrix was primarily designed to be fed power by the 5V USB buss and the external addition was a secondary design decision. I am sticking with USB 5V bus with data only USB cable and AQVOX LPS as I detailed earlier. However, I do have an SBooster 5V LPS so I will be doing a proper A/B test between the AQVOX USB fed LPS vs the SBooster 5V connected via the barrel plug.
I am in the process of reaching out to Sbooster about their 5-6V LPS and a connector for the USB-B and also their Vbus isolation connector to see if they just lift the 5V or filter or touch the shield (GND). My approach is to use the Sbooster USB-B and Vbus isolation and AQVOX 50cm great USB cable… this is fun and unfolding… I think I will move my (2) T5s to the AS4000 and get rid of the PSA Bridge II… BOOHA! Jeff will help w/ SQ and the A/B on the external Sbooster to Matrix and or the AQVOX LPS to USB-B connector and “data only” USB cable… thanks Jeff…
They are using an LDO (Low Drop Out) DC/DC for the external PS input. I think they designed the internal PS first using the USB 5V cause they have no wall wart. Then, down the road they stuck on the external PS and tied it into the design with a voltage sense that turns off the USB 5V when external is sensed. I would have used a pi filter (isolation) for the external input so when I asked I just got we use an LDO DC/DC… assume range is 5-9V and the lower the voltage the higher the current drain.
Yes, definitely want to separate the power from the data bus on the USB cable as best you can. For me, I am using a Mad Scientist Heritical “Data Only” USB cable the the AQVOX LPS which feeds 5V power to the USB port NOT the external power port on the Matrix.
And I just noticed that the price went up from 98.00 Euro to 119.00 Euro in the last two weeks! Still a great price at 119.00 Euro if it performs as advertised.
@JeffofArabica - yeah, going the “data only” versus the Sbooster Vbus isolation could be tested. I am waiting for an answer on what they did in the Vbus isolation connector. Did they just lift the 5V or filter, or anything else. I assume they just lifted the 5V. Then if that is the case, using a really good AQVOX 25-50cm USB cable may work great.
@JeffofArabica - looks like there is filtering on the external PS input (see caps, diode, and some small LDO external resistors) with the LDO DC/DC following. Please pay attention to the Sbooster SQ on the external when you hook it up. Not sure the audiophile you sighted that did the A/B (external vs. USB-B) had a good LPS used for external PS. Don’t have schematics so I don’t know how the USB 5V is routed; either to LDO DC/DC or straight onto the Vbus after filtering for the Xilinx Spartan FPGA and other components. The plot thickens… Thanks @Jaime for the Matrix PWB, just gained more insight (one pic) than asking Arthur at Matrix (distributor) any questions.
Not sure when the EXT/INT PS switching is done wrt the routing of the supplies. Looks like the switch is to the left of the cap in line with the USB-B filtering… schematics would help. I assume at this point the external PS goes through an LDO then meets up with the USB PS and switched when the ext PS is detected.
It’s the voltage step down that I suspect. They designed this thing to run off 5V. So it could be that if it was fed only 5V from an external LPS, the added noise would be minimized or eliminated. But again, it is clear they designed this thing around 5V USB power and forces the external LPS to comply with the voltage drop step.
@JeffofArabica - the DC/DC has an input DC range - 5V to 9V… the lower the voltage in, the higher the current draw. I assume the LDO DC/DC is for the external PS and the USB is just filtered… looks like it was designed for the USB power and external PS was an afterthought…
@JeffofArabica - Feedback from Sbooster is not to use the Vbus isolator, they have a filter in the shield and it is not intended for that Matrix application. Also, Sbooster does not have a USB-B connector that isolates the 5V… so the short answer is… follow Jeff… AQVOX USB-B LPS w/ “data only” USB cable… The only stone left is Sbooster into external PS connector w/ “data only” USB cable bypassing the AQVOX USB-B connector vs. AQVOX LPS w/ USB-B and “data only” cable. That is an SQ follow up Jeff will tackle when he A/Bs the Sbooster against the AQVOX… Is it lunch yet…
In my system improvement in the order of greatest magnitude – 1 being the greatest…
Uptone LPS 1.2
WW SS HDMI Cable
Audio Sensibility .3m Silver DC Cable
The point of all this for me was an alternative to USB. The difference between this combo and straight USB to the DS is night and day. It also stomps pretty hard on the Toslink options I tried.
@JeffofArabica - I believe the external PS input uses a DC/DC convertor (6V-9V input range) and the USB-B supply (called internal) uses an LDO (Low Drop Out) linear regulator for the internal power. An LDO linear regulator will usually have mV drop (in/out); however, the DC/DC convertor is usually switched. This is what Sbooster and I have been discussing this morning. So, if I do this approach, I’m following you on the AQVOX USB-B LPS and “data only” cable.
If the assumptions are correct, that is the reason the AQVOX USB-B LPS had better noise measurements than the external PS input. Counter-intuitive (internal better than external PS input) to me; however, if the Matrix is set up that way… I can see it… so, there is really no internal PS; no wall wart with the Matrix. Just powered over USB or an external PS that disconnects the USB supply when sensed.
From Matrix Website - yes, verified…
Multi-level LDO - The power supply units of X-SPDIF 2 are constructed by multi-level, multi-group LDOs and solid capacitors, they can filter out the superimposed noise from power supply effectively. All the circuit powered by low noise or ultra-low noise LDO, which avoided the hi-frequency switch noise caused by DC/DC circuit. You also can obtain high quality power from USB bus.
So, the USB-B probably avoids the DC/DC switch convertors and uses just the LDO regulators to distribute the 5V that feeds the circuitry. This seems to be the reason the external PS is noisier (according to measurements)…
Edit[1] - As @cudfoo pointed out, the external PS input is ranged from 6V to 9V, not 5V to 9V as I incorrectly indicated above. Changed it to the correct range. My apologies.
Hi - So you’re saying that the info in their manual is incorrect:
" Power supply specifications USB bus power supply: 5V/500mA External power supply: DC 6~9V，≥800mA"
If 5V should be used why would they recommend 6-9 and not 5V?..
" With the external power supply port, convenient to upgrade the external linear power supply. The device will auto-disconnect the USB bus power supply after connected the external power supply, and can be connected to the devices which do not support power output, such as mobile phone"
@cudfoo - not at all… just how the supply routes the options… A DC/DC switch mode convertor has an input range… the higher the voltage, the less the current draw…
