Use an EE8 Switch and had it powered by a Farad Super3 LPS with Upgrade 1 cable. This had a Chord ShawLine Mains cable and similarly, I thought âWow" when changed from an iFI iPower2 SMPS.
After a while, I felt that something wasnât quite right. I then put back the original stock SMPS that came with the EE8.
Another âWow" moment, this was much better, the music was more emotive, engaging and just more toe-tapping.
Just because a power supply is based on SMPS doesnât make it bad by default. A well designed SMPS can sound amazingâŠ.just that well designed SMPSs typically arenât cheap
Not all LPSs sound great. The really well designed ones typically do
Then thereâs running in. In total, during my 5 years of system development I probably listened to power supply running effects at least 10 times. The effect was always the same. It would sound fine for a few days, then it would become unbalanced as different parts of the spectrum changed. Bass became overemphasised, then better, same for treble. The system would develop sibilance, which would gradually disappear. With Mundorf capacitors, running in periods with units switched on 24 hours per day but properly working 3-5 hours per day took about 3 months. SMPSs didnât seem to need much time at all.
During running in I often found myself distracted, reading album meta data and the like and not particularly enjoying the music. Once run in the system became 110% involving.
I did consider whether it hadnât run in properly. Iâm very familiar with the effects of running in. My Chord M6 mains block was purchased just before Christmas and it still seems to be gradually improving nearly 3 months later!
However, although I only gave it a few days this time, the previous time I gave it a few months and it showed no signs of getting better.
So I donât know. Maybe it needs even longer?
Plus Iâm not convinced itâs a very good LPS. Plixir, which I have on my Melco to excellent effect, actually make special LPSâs for network components like switches and routers. Iâm very tempted to try, but I really canât afford to risk wasting another ÂŁ600. Plus the BT router needs a DC plug that is all but impossible to source as it is an unusual size. My MCRU LPS used a lead/plug assebly cut off from a BT SMPS as it was the only way to get a plug to fit.
From Farad3 manual related to burn in. In Farad3 there are big cap reservoars that need long period of burn in. Actually it behaves more like a battery on the DC output.
Breaking in
Like all devices in audio, the Farad power supply and cables will show break in behavior the first 500 hours on AC voltage. The character and quality of this supply will show immediately, but will improve during this period. Sound will become more tight, more at ease and fluent. Always keep this in mind when evaluating the power supply during this period.
A Silent Angel Forester F1 LPS was a worthwhile upgrade on my EE8 switch, even though Chord Co. advise against this, preferring their SMPS. The F1 has the added benefit of being able to drive another 5v device, if required. Upgrading the power cable to the LPS and the power feed connection from F1 to EE8 also improved the EE8.
I have Ifi iPower supplies on my fibre box, router and Netgear switch (feeding all network connections in my house) which are away in a cupboard. Adding each brought apparent improvements and all add into an overall result with blacker backgrounds and better detail in how music is presented, so worth experimenting.
I have been mulling over adding a multi output LPS in place of the Ifiâs but the comments above about sound degrading over time with LPSs is interesting and Iâve seen similar comments in other forums.
Just by way of information and interest, Plixir, regarding their LPS optimised for network switches and routers, say this:
Particular attention is done to isolate the component noise from polluting the AC mains as well as filtering optimisation for high speed digital application.
Also, Network Acoustics, who make a high end switch, supply this with what they descibe as a âhybridâ PS. They say this:
This unique hybrid design delivers ultra clean power normally only associated with the best Linear supplies combined with the essential dynamics only a switched-mode can provide.
The Chord Company also say that they went with a SMPS for their switches as no LPS was fast enough.
All this seems to me to point to the fact that SMPSâs are better suited to âhigh speedâ digital applications like network switches and routers. On the other hand we have companies like Plixir who make LPSâs for that specific purpose. So do Melco with their switch.
So who knows. Itâs all very confusing. A little knowledge is a dangerous thing perhaps?
Iâm minded to stick with the stock BT SMPS with my router. Same for my EE8 switch.
Or it might be that a quality LPS for say 500GBP would totally shift their target group. After all a minority spend such money on a LPS for a switch. Even a switch for 500GBP is for the majority snake oil.
You have to have in mind that a manufacturer of a product have many other interests that need to work except product performance. So what is THEIR truth might not be my truth.
There are hundreds of crappy low prices LPS on ebay and other places that does not even fulfil EU regulations and safety standards that people put in their system. I believe many that get disappointed try to find short cuts in that an LPS is an LPS. Thatâs like saying an amp is an amp. Knowing what 555PSDR do to the digital side of say a 272 or NDS then you will understand that SMPS is there mostly due to low price, small size, small weight. Itâs good enough for most people. Some are better and some are worse like with all products on the market.
Iâve also reverted to the EE8âs stock SMPS after testing it with a couple of LPSs (Uptone JS-2 & a cheapo Chinese model).
For me, both LPSâs added a degree of weightiness to the sound, but at the expense of PRaT. As others have reported, thereâs an initial âwowâ response, followed by a sense of something missingâŠ
Iâd also echo @GraemeHâs suggestion to have the EE8 plugged into a different outlet than your Naim boxes, if possible.
Thatâs for sure, all the network infrastructure should ideally use different line of power supply which doesnât intersect with the components so theyâre connected with the ethernet cable only. Otherwise thereâs a loop via ground conductor of the power cables.
No doubt thatâs true. But then one would expect them to say something along the lines of âthe supplied PS is perfectly adequate to achieve excellent performance but replacing it with a more expensive LPS may bring further improvements stillâ. What would they have to gain by covering this up?
In fact Iâve discussed this very thing with them on the 'phone. I was told that their R&D department had experimented with various LPSâs and found nothing to really shout about.
The other key thing you should add in my opinion, is that devices need to typically have a matching PSU in terms of DC line noise unless they are designed to use a generalised PSU. Specifically the PSU should have matching HF decoupling in the power lead to the powered device - and this helps present a low impedance to any stray HF/RF voltages that are passing up through the DC power lead from the power device. These may even come and go over time. This can become rather important for digital devices with clocks in etc. Without this you might be adding more RF noise and breaking your products EM compliance when using a non matched PSU.
If in doubt I would solder 100nF capacitors across the DC lines within the PSU and in the powered device and keep the DC lines as short as possible.
I think the LPS/SMPS is largely immaterial if quality PSUs are used, the key thing in my opinion is the noise decoupling and if regulated, the regulator response. This can be really rather important as otherwise the regulator can add performance robbing noise into the DC lines.
Naim focussed on this with their own regulator designs that produced a response that matched their devices to be powered.
An example of a noise bandpass for a linear regulator.
The intricacies of the technical side are largely beyond me Simon but I agree completely.
Iâm now inclined to the view that itâs best to stick with the PS supplied by the component manufacturer - LPS or SMPS, unless the manufacturer specifies the possible use of something better. An example of this would be Melco, who supplied a medical grade SMPS with my N100 but also state that upgrading to a Plixir LPS will bring improvements, so thatâs what I did, and indeed it does.
Supplies are closely matched to the products they are intended to be used with it seems. Itâs a lot more than the correct voltage and adequate current delivery, as you say. Replacing them with something âbetterâ is not so simple as it may first appear.