Local 5V 1A power supply for Logitech Duet

[DSK]

A friend has bought a new Duet (he is currently having digital mods done to it) and would like to buy a better power supply for it. He doesn't wish to pay hundreds for an "Audiophile" one, just an off the shelf unit.

Anyone know of (or have tried) any locally available linear regulated 5V 1A power supplies? Presumably the connector will need to be changed.

Thanks guys.

Edited February 16, 2009 by DSK

[GryphonGuy]

I bought this one from Jaycar for my Squeezebox and the difference in sound was audible and significantly better. I still use it today after about 1 year and still pleasing results.

http://jaycar.com.au/productView.asp?ID=MP3086&CATID=27&form=CAT&SUBCATID=381

[kye]

If you're able to find one I would suggest that you go for one that doesn't have a regulated output.. anyone that has looked at the output of a voltage regulator will know there is all kinds of HF hash on it, which is the last thing you want being pumped into a device that is trying to do high end digital processing..

if you're able, i'd suggest building a linear power supply and using good quality bypass caps - or perhaps ask one of the sponsors if they can make one up for you if you're not up to the DIY yourself... you can certainly get a kickass linear powersupply for the cost of that jaycar unit mentioned above.

OR.. if you can't go this route and have to use one with a regulator i'd make an adaptor cable that has a bunch of bypass filtering caps that try to get rid of the HF that the regulator will be pumping out.

[DSK]

Thanks guys. No, he doesn't have DIY skills and doesn't want a large variable voltage output supply, just a small fixed voltage supply that the can hook up and forget. Something without dials that his kids could turn up to 30V and cook his new Duet.

[A9X]

If you're able to find one I would suggest that you go for one that doesn't have a regulated output.. anyone that has looked at the output of a voltage regulator will know there is all kinds of HF hash on it, which is the last thing you want being pumped into a device that is trying to do high end digital processing..

That is not true. It is not difficult or expensive to make a linear regulated supply with very low broadband noise.

[Zaphod Beeblebrox]

If you're able to find one I would suggest that you go for one that doesn't have a regulated output.. anyone that has looked at the output of a voltage regulator will know there is all kinds of HF hash on it, which is the last thing you want being pumped into a device that is trying to do high end digital processing..

Incorrect. I suggest you examine the data sheets for any of the popular regulators. They will do the job just fine, as long as appropriate attention is paid to bypass caps. And, like any decent power supply, bypass caps should be placed as close as possible to the equipment. Long supply cables are a disaster waiting to occur. It is, for this reason, that power supplies on the PCB are the best approach.

if you're able, i'd suggest building a linear power supply and using good quality bypass caps - or perhaps ask one of the sponsors if they can make one up for you if you're not up to the DIY yourself... you can certainly get a kickass linear powersupply for the cost of that jaycar unit mentioned above.

Why re-invent the wheel? Good quality power supplies are easy and cheap to buy and use. Back in the bad old days, I would have sold my mother into slavery to get hold of an LM317T. There was no such animal. They offer phenomenal performance, they're easy to use and adjustable. You can buy them for a Buck now.

OR.. if you can't go this route and have to use one with a regulator i'd make an adaptor cable that has a bunch of bypass filtering caps that try to get rid of the HF that the regulator will be pumping out.

Regulators do not "pump out HF".

Edited February 15, 2009 by Zaphod Beeblebrox
Brain fart

[seano]

I believe that Hugh at AKSA has a power supply for the Duet (and the plain old SB). It won't be as cheap as the Jaycar version but then it probably isn't the same inside...

http://www.aksaonline.com/products/products_Holo3.html

Nope.......not as cheap.....$690. But since it is essentially similar to the Ultimate Nirvana PS supplied by Bolder Cables.....not a bad buy if you like that sort of thing.

[joz]

Would one of these external power supply upgrades show any impovement if using an external Dac or just the onboard ones??

[tricka]

I think Joz we are back to a discussion of jitter and how well your external dac deals with it.

[Guest]

Check out John Swenson mods for jitter reduction if your serious about using one of these as a transport . Do a google search .

Hugh makes an excellent produced very well made and finished a couple of people who have tried it never gave it back .

Cheers

[Decky]

I fully agree with Zaphod - you can choose to go along the "super-regulator" path and, do not get me wrong, there are some merits in that approach. However, using a standard linear regulators like LM7805 or 317T is very effective as well. Good capacitors, fast diodes and good PCB with a nice thick ground plane are all very important and the cost is minimal compared to the super-regulators.

[Zaphod Beeblebrox]

I fully agree with Zaphod - you can choose to go along the "super-regulator" path and, do not get me wrong, there are some merits in that approach. However, using a standard linear regulators like LM7805 or 317T is very effective as well. Good capacitors, fast diodes and good PCB with a nice thick ground plane are all very important and the cost is minimal compared to the super-regulators.

Just to clarify, for the pedants. The order of 'goodness' for regulators is this (worst to best):

7805T

LM340-5 (Marginally superior to the 78##) Usually considered as interchangeable with the 78## series.

LM317T A significantly superior regulator to the above. Slightly more complex to implement.

LM723 The best of the lot. A PITA to implement, as significant support components are required. One of the earliest IC regulators and still one of the best.

I use LM317T and LM337T for pretty much everything, as they are easy to use, very versatile and offer stunningly good performance.

[tricka]

Zaph

can you post examples of the type of psu you are taking about. Sorry for being ignorant.

Cheers

tricka

ps I am also interested (actually more interested) in a 12 V one.

[Zaphod Beeblebrox]

Zaph

can you post examples of the type of psu you are taking about. Sorry for being ignorant.

Cheers

tricka

ps I am also interested (actually more interested) in a 12 V one.

What is the application?

[tricka]

weeeell...I have walwart's for both my SB3 (5v ) and KRK Ergo (12v 1.25A) and was toying with improving the psu of each. The KRK is my DAC as well.

Cheers

[markm1111]

I have been having thoughts about improving the power supplies for my SB3 (5v) and Duet (9v), as long as it was in a cost effective manner. Don't know much about the Bolder ones, or Aksa, but a couple I came across from Channel Islands Audio are around USD250 before shipping. Something with noticeable performance improvement aroud the $100 to $150 mark would be perfect.

[Zaphod Beeblebrox]

weeeell...I have walwart's for both my SB3 (5v ) and KRK Ergo (12v 1.25A) and was toying with improving the psu of each. The KRK is my DAC as well.

Cheers

The LM317T (with suitable pass devices, for more current), with appropriate heat sinks, should do the job quite nicely. Go to the National Semiconductor site and research the application notes for your particular needs.

Like all such things, you should keep DC cables as short as possible and use plenty of decoupling at the device.

[DSK]

Oops ... just realised that I had 5V 1A instead of 9V 1A in the thread title (Was thinking of the SB PS instead of the Duet). Have fixed the title now.

Yes, I know about Hugh's PS for the Duet and have home auditioned a prototype of his SB version ages ago. But my friend who has bought the Duet does not want to spend this type of money. He is looiking to spend less than A$200, less than $100 if possible.

[kye]

That is not true. It is not difficult or expensive to make a linear regulated supply with very low broadband noise.

I didn't say it was difficult, but I doubt very much that it is commonly done.. especially by people that sell pre-manufactured PSU units with the little DC jack connector on them!

Incorrect. I suggest you examine the data sheets for any of the popular regulators. They will do the job just fine, as long as appropriate attention is paid to bypass caps. And, like any decent power supply, bypass caps should be placed as close as possible to the equipment. Long supply cables are a disaster waiting to occur. It is, for this reason, that power supplies on the PCB are the best approach.

Why re-invent the wheel? Good quality power supplies are easy and cheap to buy and use. Back in the bad old days, I would have sold my mother into slavery to get hold of an LM317T. There was no such animal. They offer phenomenal performance, they're easy to use and adjustable. You can buy them for a Buck now.

Regulators do not "pump out HF".

Have you actually looked at the output of one of these on a CRO?

I got given a Rakk DAC unit and the power supply module that the manuf also makes and in order to see how good it was I hooked up the CRO to the output of it and it was awful - wideband hash in the 10mV+ range, which for those that like to talk in dB is about 30dB down from the 12V rail it was supplying.

The effects on the output would be complex being that it's supplying a digital circuit, but if that were supplying an analogue device such as a DAC chip, i'd be very unhappy about it.

By contrast I have 250V HT rails in my valve preamp that have less than 1mV of noise on them with full signal going through the preamp. That's 81dB down if I did my maths right..

I don't know what kind of powersupply optimisation you guys are prepared to live with, but it's a little less than the optimisation that I have in my $5k setup. I very much doubt that by ignoring 10mV of PSU noise that your comments are useful across the board, considering that there are people on this board with $250k setups..

and if we have seen something consistently happen it's that someone says "that's not important" and someone else with different equipment says "it's audible in my setup and i've done X Y and Z to eliminate it".

[Zaphod Beeblebrox]

I didn't say it was difficult, but I doubt very much that it is commonly done.. especially by people that sell pre-manufactured PSU units with the little DC jack connector on them!

Have you actually looked at the output of one of these on a CRO?

Only properly designed ones, yes. Clean as a whistle. No HF, no ripple, nothing at all. Nada, zilch, zip. The LM317T/LM337T are the ones I use the most and any noise produced is beyond the capacity of my test equipment to resolve (150MHz CRO, Sound Technology 1700B distortion analyser, PC-based analyser, etc). All one needs to do is to follow the manufacturer's suggested applications and the result will be excellent. For even better performance, one can use the venerable 723 regulator, but it is a PITA to implement.

Here are the performance data cited by the manufacturer (and easily acheived in real-world conditions):

http://www.national.com/ds/LM/LM317.pdf

And:

http://www.national.com/ds/LM/LM723JAN.pdf

I got given a Rakk DAC unit and the power supply module that the manuf also makes and in order to see how good it was I hooked up the CRO to the output of it and it was awful - wideband hash in the 10mV+ range, which for those that like to talk in dB is about 30dB down from the 12V rail it was supplying.

The effects on the output would be complex being that it's supplying a digital circuit, but if that were supplying an analogue device such as a DAC chip, i'd be very unhappy about it.

Incompetent design is not an adequate reason to damn LM317T regulators. Of course, the 78## style regulators are somewhat poorer than those, but their performance (when properly implemented) is far in excess of the figures you've cited.

By contrast I have 250V HT rails in my valve preamp that have less than 1mV of noise on them with full signal going through the preamp. That's 81dB down if I did my maths right..

I don't know what kind of powersupply optimisation you guys are prepared to live with, but it's a little less than the optimisation that I have in my $5k setup. I very much doubt that by ignoring 10mV of PSU noise that your comments are useful across the board, considering that there are people on this board with $250k setups..

and if we have seen something consistently happen it's that someone says "that's not important" and someone else with different equipment says "it's audible in my setup and i've done X Y and Z to eliminate it".

I accept that you may have experienced some badly designed products, but that is no reason to damn modern regulators. Many designers simply can't read application data nor can they design for real world conditions. 10mV of noise from an LM317T is laughbly bad. The manufacturer should be shown, in no uncertain terms, the error of his ways. In hyper critical applications I would use a capacitance multiplier/pre-regulator ahead of the 3 terminal regulator. I would design for much better than 10mV (p-p) noise and ripple, BEFORE even getting to the regulator.

[kye]

Only properly designed ones, yes. Clean as a whistle. No HF, no ripple, nothing at all. Nada, zilch, zip. The LM317T/LM337T are the ones I use the most and any noise produced is beyond the capacity of my test equipment to resolve (150MHz CRO, Sound Technology 1700B distortion analyser, PC-based analyser, etc). All one needs to do is to follow the manufacturer's suggested applications and the result will be excellent. For even better performance, one can use the venerable 723 regulator, but it is a PITA to implement.

Here are the performance data cited by the manufacturer (and easily acheived in real-world conditions):

http://www.national.com/ds/LM/LM317.pdf

And:

http://www.national.com/ds/LM/LM723JAN.pdf

Incompetent design is not an adequate reason to damn LM317T regulators. Of course, the 78## style regulators are somewhat poorer than those, but their performance (when properly implemented) is far in excess of the figures you've cited.

I accept that you may have experienced some badly designed products, but that is no reason to damn modern regulators. Many designers simply can't read application data nor can they design for real world conditions. 10mV of noise from an LM317T is laughbly bad. The manufacturer should be shown, in no uncertain terms, the error of his ways. In hyper critical applications I would use a capacitance multiplier/pre-regulator ahead of the 3 terminal regulator. I would design for much better than 10mV (p-p) noise and ripple, BEFORE even getting to the regulator.

I am glad you agree that 10mV of noise is bad! I guess my question now would be that if an "audiophile" company can produce such bad results with one, then what is the hope of something made in china to a price?

My original statement still stands, even if now it's because I am commenting on the likelihood of getting a good result instead of the possibility of being able to get one. A linear power supply will not have these issues... it may have 50Hz ripple and a higher output impedance (when designed to a price) but it won't be actively *making* noise like a badly implemented regulator..

regardless, a power supply designed by someone who cares about such things would be better than one from an electronics box shipper such as altronics or jaycar.

yet another reason to have one designed by a local hifi mob

[markm1111]

yet another reason to have one designed by a local hifi mob

Any volunteers in the OP price bracket??

[kye]

OP.........?

[markm1111]

OP = original poster

Sorry, should have said DSK

[DSK]

...

regardless, a power supply designed by someone who cares about such things would be better than one from an electronics box shipper such as altronics or jaycar.

yet another reason to have one designed by a local hifi mob

Agreed, but not an option in the price bracket my mate is looking at.