Usefullness of a external quality power (linear PSU or Battery) for USB Pci card (SOTM, PPA,...) for use with an Asynchronous, completely 5V decoupled DAC ?

[malarz]

Hi everybody,

 

I am questioning myself of usefullness of a external quality power (linear PSU or Battery) for USB Pci card (SOTM, PPA,...) for use with an Asynchronous, completely 5V decoupled DAC ?

 

Thanks for your opinions.

 

Have a nice day.

[malarz]

Hi,

In fact my question is as follows:

The USB audio cards PCI (SOTM, PPA, ...) are equiped with an extra external power supply input for quality power (linear PSU or Battery) instead the primary power from PC, is this external power designed ONLY for supplying quality 5V for devices plugged into this card (and which need this 5V) OR it improves ALSO the performance of this USB audio card (output data signal for DAC , ...) ?

I am asking that, because I have an Asynchronous DAC that don't need the 5V power and is reclocking itself the incoming signal, so maybe (for me) buying the expensive extra external power supply for my SOTM Tx-USBexp card is wastefulness ?

 

Thanks for your opinions.

 

Nice day everybody.

 

[Piopio]

PPA USB needs external power to deliver the best quality. I tried to power it from MOBO (even powered by linear supply) but the sound didn't differ much from the ordinary USB port.

 

Cheers

[malarz]

Thanks a lot.

 

I suppose it' should be the same for my SOTM Tx-USBexp card ?

[k6davis]

PPA USB needs external power to deliver the best quality. I tried to power it from MOBO (even powered by linear supply) but the sound didn't differ much from the ordinary USB port.

 

Cheers

I've got a PPA USB card and a TeraDak power supply to use with it, but I'm lacking a power cable to connect the two. Can anyone point me toward a USB to Molex power cable?

 

Thanks.

[Michael-Elijah Audio]

[Michael-Elijah Audio]

I know this is for SSD power , but wiring setup is the same for 4 pin Molex- cut yellow wire and one black wire from Molex plug - leaving the Red wire and one Black wire, solder to USB plug as shown

[Harpy]

Hopefully I am not missing anything, but if your DAC doesn't need 5v why feed it any? The Elijah Audio Isolate CU USB cable carries just the Data and Ground.

[mrtnsko]

Hopefully I am not missing anything, but if your DAC doesn't need 5v why feed it any? The Elijah Audio Isolate CU USB cable carries just the Data and Ground.

 

I think the chipset (on the card needs) 5v.

[Harpy]

My bad. I am not that familiar with internal cards. I use a battery on my external convertor, but am using upgraded clocks and a low noise voltage regulator.

[panhead]

USB to Molex:

 

USB to Molex Fan Adaptor Cable 100cm | eBay

[Pepsican]

Hi,

In fact my question is as follows:

The USB audio cards PCI (SOTM, PPA, ...) are equiped with an extra external power supply input for quality power (linear PSU or Battery) instead the primary power from PC, is this external power designed ONLY for supplying quality 5V for devices plugged into this card (and which need this 5V) OR it improves ALSO the performance of this USB audio card (output data signal for DAC , ...) ?

I am asking that, because I have an Asynchronous DAC that don't need the 5V power and is reclocking itself the incoming signal, so maybe (for me) buying the expensive extra external power supply for my SOTM Tx-USBexp card is wastefulness ?

 

Thanks for your opinions.

 

Nice day everybody.

 

 

Even if you have a linear psu, having a separate, clean supply for the PCIe card matters. The main reasons are:

• switch mode psus on the motherboard itself still introduce ripple and rf noise
  
• SATA drives and HDDs will also introduce noise whenever they spin up
  
• The mobo itself is not insulated or protected and hence is susceptible to outside influences
  

 

Good that your DAC is asynchronous and good that it is insulated. But there will still be some current leaking across.

 

In conclusion, it will make a difference although in your case less than when the USB connection is not insulated.

[malarz]

Thanks a lot Pepsican

[barrows]

Belt and suspenders in this case. the more one reduces noise the better. The USB interface in my DAC totally isolates the clocks, re-clock circuit, and I2S output from the USB input, but I still run an isolated, separately regulated power supply feed from my LiFePO4 battery pack to the SOtM USB card. It is worth it.

[sandyk]

EVERYTHING MATTERS

 

+1

[aps]

A related question: What is the power consumption of the PPA USB Card? Is it 5V/1AMP so 5W?

[firedog]

Note in Chris' original article about the CAPS Zuma, he wrote that he was surprised that running both the USB card and the MB off 2 different outputs on the battery supply, (acc'd to needed wattage for each) sounded noticeably better than just the MB running that way.

[barrows]

Note in Chris' original article about the CAPS Zuma, he wrote that he was surprised that running both the USB card and the MB off 2 different outputs on the battery supply, (acc'd to needed wattage for each) sounded noticeably better than just the MB running that way.

 

This makes total sense, as those two different outputs likely provide some degree of isolation between the two feeds. If one just connects a single supply to both the MoBo and the USB card, the noise from the MoBo (actually the modulation of the power supply caused by the high speed switching components on the MoBo) will feed back and be shared to the USB card via the power supply wiring, no matter what type of supply it is.

 

How much power a USB requires will vary, as some DACs will draw up to 500 mA from the USB port, and some DACs do not use USB power at all. I would suggest it best to make sure the USB card power supply be capable of passing at least 1.5A to be not he safe side and have some headroom.

[sandyk]

This makes total sense, as those two different outputs likely provide some degree of isolation between the two feeds. If one just connects a single supply to both the MoBo and the USB card, the noise from the MoBo (actually the modulation of the power supply caused by the high speed switching components on the MoBo) will feed back and be shared to the USB card via the power supply wiring, no matter what type of supply it is.

Surprisingly, even then it doesn't give the same degree of isolation as 2 separate PSUs. I found this when I decided to put two +5V linear PSU PCBs with JLH Add-ons in the same case , but using separate transformer secondary windings. They didn't sound as quite as good as 2 separate PSUs.

 

Alex

[Hifidelit]

It depends on the card design and what you are connecting to. If we take the JCAT usb card then the design is as follows (from Marcin):

The JCAT board has a linear regulator which can provide additional filtering to the NEC chip. However without 5V connected, the 3.3V from computer will be applied directly to the NEC chip without much filtering. Even without a linear 5V power supply it might be better to connect 5V from computer. To sum up: even for self-powered USB devices, the JCAT USB card will benefit from high quality 5V PSU. 3.3 is not used to power the NEC chip if 5v is applied.

 

Nige200 asked:

So you can tape over the 3.3v pins on the jcat card and supply via molex to prevent noise contamination from the motherboard to the clock and usb chip?

Marcin replied:

The 3.3V is still used for 1.05V regulator for the NEC digital core. We wanted to use only 5V but it was not possible due to power dissipation.

http://jplay.eu/forum/jcat/jcat-battery-psu/

 

Pearse (sligolad) mentioned this tweak on the usb card thread:

 

The one benefit of the cheap ebay riser cables is that if you have a 3.3v linear supply you can break into the wiring on the riser cable and cut the 3 off 3.3v lines and supply these direct off the 3.3v linear supply.

I found it a nice improvement on the JCat USB card to feed it directly off the Teddy Pardo 3.3v Linear supply and have no 3.3v feed off the Mobo with all the junk on the Mobo supply even with all linears to the ATX supply.

All my grounds are on a switch so the JCat card is powered only when I switch in the grounds at power up and all works fine with a cleaner sound supplying the 3.3v for the JCat off a linear supply.

 

http://jplay.eu/forum/jcat/jcat-usb-card/page-10/

 

 

[barrows]

Surprisingly, even then it doesn't give the same degree of isolation as 2 separate PSUs. I found this when I decided to put two +5V linear PSU PCBs with JLH Add-ons in the same case , but using separate transformer secondary windings. They didn't sound as quite as good as 2 separate PSUs.

 

Alex

 

Agreed, and further isolation can be had with a fully separate transformer as it will not be subject to capacitive coupling of the windings. And even further by separate power cables going to the opposite phase feed from the power box, and even more isolation by having a separate sub panel... and on...

But, at some point diminishing returns are going to kick in. I love using battery power for my server, as every AC based approach which I have tried still allowed plenty of MoBo caused noise to get back out on the AC line and then couple into the rest of the audio system. But, two separate battery packs are too much hassle for me to deal with, hence I feed both the MoBo and the USB from the same pack. My solution is to use very short wiring from the battery pack to a bank of Silimc IIs, and then to a discrete voltage regulator which provides 8 VDC to my SOtM card, the output of the LiFePO4 batteries is very low impedance, and the regulator is very good, and the Silimic IIs are good at damping the high frequency noise from the MoBo-hence: not total, but very good isolation from the MoBo noise. I sure cannot detect any noise on the output of the dedicated regulator to the SOtM card, and I powered it from a separate battery supply once, and could detect no sonic difference, so I am happy that my approach is a good as I can detect in my system.

[Superdad]

...every AC based approach which I have tried still allowed plenty of MoBo caused noise to get back out on the AC line and then couple into the rest of the audio system.

 

Perhaps you should try an LPS that was designed to not let crap get back out to the mains. I happen to know of such a device…

[barrows]

Perhaps you should try an LPS that was designed to not let crap get back out to the mains. I happen to know of such a device…

 

Hahahaha! No need actually, batteries keep me entirely sure that no noise gets back to the mains, and they are onboard my server so no extra boxes. I hate having too many boxes and wires in my rack. But, of course, my solution is DIY, and not available to folks who might be less inclined to experiment and build. For those folks a choke input filtered LPS might be just what they are looking for.