Simple SMPS test

I thought I’d share here prompted by an article I was reading and my own experiences. SMPS are often incorrectly maligned, a switch mode powersupply can be as noise free if not better as an analogue transformer power supply if built and designed appropriately. But how do you know if you don’t have an EM/RF detector connected to a spectrum analyser? I know one can use one’s ‘ears’, however I am not sure how reliable that is unless powersupply quality is very poor given expectation bias and other variables. So a simple check is get a LW (MW or SW would work but not as ideal), and place right by the SMPS powersupply, and then try it’s mains cable and then it’s down stream low voltage cable ... turn on radio and tune around away from a radio station... can you hear any buzzing or rasping or other strange doors... you might also hear a rough tone that is slightly unsteady (this may be the switcher’s oscillator) ... any way if you hear such a noise that then stops when you disconnect the powersupply then you have found RF / EM noise radiating from the device... ideally you want no noise, but importantly you don’t want noise on the mains cable or down stream cable...  Also if you hear such a noise irrespective of the SMPS you have other sources of RF/EM interference, that where strong should be investigated... this all affects our audio.

This an ideal test for EM/RF emissions and you can use for Ethernet and other electrical and digital  equipment... and even some so called low noise regulated transformer power supplies , Ethernet/usb  bridges etc....

Just a thought to share for the obsessive types, and perhaps helps prevents chasing rainbows.......

Original Post

Getting hands on LW radio might be stretching it a bit Simon,   but I agree the test idea should be possible,  LW is 150 to 375 kHz & SMPS's work in the 50 KHz & 2 MHz region.     However my little Fluke scope tells me my fully ferreted SMPS's are down to nothing to worry about.

Yeah, mike you are listening out for the interactions (intermodulation) with the radio IFs as well... yes SMPS work on many many many different frequencies depending on design... and of course the harmonics from poor switching and filtering is often where the issues lie rather than the switcher oscillator. The radio test really works well, and LW is ideal as it more sensitive  to EM interference... and also remember not all interference is coupled through wires, so in which case ferrites will have no benefit at all.

I have used this technique with local farmers to identify leaky insulators on electric fences for cattle.. these can be notorious RFI generators affecting MC cartridges etc, as well identifying rampant powerline adapters in my village.

I have also used this technique to isolate noisy Ethernet switches from the likes of Netgear etc... and you can also hear low quality carrier clocks as they tend to warble rather than be a steady tone... one of the reasons in the early days I advocated the Cisco 2960 devices for audio setups.

This technique I discovered very recently is also advocated as a low cost / low skill method of RFI identification by the Radio Society of Great Britain... so I thought I’d share here... you don’t always need expensive test equipment...... and currently I am relatively free of RFI.. and yes I do use some quality SMPS.. and cheap consumer ones from the likes of Apple, Netgear, Noname etc.. I keep well away from my audio equipment.

Simon-in-Suffolk posted:

...and currently I am relatively free of RFI.. and yes I do use some quality SMPS.. and cheap consumer ones from the likes of Apple, Netgear, Noname etc.. I keep well away from my audio equipment.

Simon, I’d love to know which are the quality SMPS’s you’re using/ would suggest.  I have a Netgear NAS as well as an Apple TV 4K. 

Hi, near to my audio equipment  I’m using a Hugo bundled SMPS, seems behaving well, and I also use the SMPS in my Cisco Catalyst switches...

My very recent LG OLED TV seems well behaved .. certainly better than my previous Panasonic Plasma..

I  also use a specialist SMPS by Yaesu for some of my sensitive radIo equipment, not really related to audio, but shows very low noise SMPS are possible.. 

Thanks Simon. I will check out Yaesu.

If I could please get some other wisdom from you : I’ve just sprung for a TP psu for my Unitiserve. I think it’s making quite a difference so I was wondering about getting one for my NAS drive. Possible issue is the nearest spec available is (same as the TP Unitiserve PS) a 12V, 4.5 A whilst the Netgear is spec’d at 12V, 5.0A. Do you think :

a) it is near enough to make no difference

b) inadvisable as too far from the right power ( I note that the original Naim US supply is rated 4.2A and the TP is 4.5A, but seems to be fine)

c) a linear PS of any kind won’t improve the sound from the NAS 

cheers, Kevin

I discovered this week that the much praised iFi 5v PS on my RPi is a significant sonic degradation in my setup. There is also an audible buzz via the loudspeakers when the iFi is used. A random PS works better, unfortunately.

What I need to do is to check out if the buzz is created by the iFi, or that some protection circuit is not present on the iFi and that it is just passing thru the distortion. But I'm afraid its the iFi which needs to go back.

kevin J Carden posted:

.....the nearest spec available is ..... 12V, 4.5 A whilst the Netgear is spec’d at 12V, 5.0A.....

That sounds dodgy to me - I don't think I'd chance an underpowered PSU on a NAS. I would either ask TP if he can supply one to the correct spec, which he probably can, or look elsewhere.....or try Simon's LW radio test, and be happy with what you've got!?

Hi Kevin - unless your NAS is physically close to your audio equipment and you believe its introducing electrical noise onto your mains, then I would leave your NAS power supply alone... I agree with Chris above NAS power supplies need to be correctly specified for reliability and robustness. You don't want a PSU failure to destroy your disks...  BTW Yaesu make radio equipment and don't produce independent power-supplies as far as I am aware..

Simon

I understand Mean Well and Channel Well Technology make good quality SMPS's and hence use them myself.

Output voltage and current rating are, obviuosly, important parameters but, when looking at the specification sheet, what else will help us make a good choice?

Dave

Thanks Simon,

I took an old AM transistor radio and dialed between stations. I found the worst offender by far was the Hugo psu followed by the laptop psu. From now on I will just unplug the power supplies when listening to music. 

FWIW, I also tried approaching the radio near the HCDR and it was almost silent but that was not the case when putting the radio over another 'hcdr wannabe' that is very noisy!

dave4jazz posted:

 

Output voltage and current rating are, obviuosly, important parameters but, when looking at the specification sheet, what else will help us make a good choice?

i think the radio test as performed by Cat354 above is the acid test - make sure you can send units back if not happy .....

I'm not sure I'm convinced.   If the SMPS case is emitting a radio signal,  whats to say that same noise is being transmitted into the DC & AC cables.      I get a MW signal off my SMPS cases out to a few inches but I get nothing from the cables once away from that area.

As I said above - I believe there are three areas - upstream mains, downstream low voltage and the apparatus itself. 

The radio in this mode is detecting near field EM radiation as opposed to what we generally perceive as a radio signal which is more far field EM and as such the EM radiation is quite different between the two

Near field EM can especially couple into nearby wires and enclosures... its clearly good if the connected cables are not radiating - but i usually found a noisy PSU will also radiate from its wires to some extent through coupling - but clearly careful design can mitigate this. Anyway if the radio test is all quiet on the leads - (try coiling around the radio) then you are good.

I also knocked up a little receiver coil for my oscilloscope a few months back to look at this - I wanted to see the shape of the noise... thats when i found how relatively poor my Apple power supplies were.

A few years back I could actually see the noise of Plasma TV / Plasma TV power supply and its harmonics from a house about 150 feet away... I could tell what they were watching by the noise pattern!!! - thanks goodness they got a new TV and all is quiet

Mike-B posted:

Getting hands on LW radio might be stretching it a bit Simon,   but I agree the test idea should be possible,  LW is 150 to 375 kHz & SMPS's work in the 50 KHz & 2 MHz region.     However my little Fluke scope tells me my fully ferreted SMPS's are down to nothing to worry about.

Which Fluke scope are you using ? 

HI Simon     Re:  Anyway if the radio test is all quiet on the leads - (try coiling around the radio) then you are good.

That's the way I see it,   I am confident I have my SMPS noise isolated & to some extent part suppressed,     that is apart from any radio emissions from the cases themselves.     All my audio related SMPS's (plus phone) are on a single power strip that is fed from a UPS with its C&D mode choke/X&Y caps & an isolation transformer,  that locks up the SMPS's on the AC side.        The DC side has an iFi iPower ultra low noise SMPS on the Cisco switch,  the others are OEM SMPS but all have LF (150kHz - 2MHz) ferrites with numbers of wire turns around each.  OK OK I know ferrite is only good for suppressing & are not total zappers,  but this is as far as I'm prepared to go on the DC side.

Keeping it all in perspective,  these days everything has a SMPS,  TV's, & all its add on boxes, radio's, phones, all things computer & the various chargers that go with them, fridge freezers, cooking ranges & microwaves,  & so much else is controlled in some way with switching & chopping,   the way I see it's only radio ham & OCD audiophiles that are aware of SMPS problems

Thanks for the suggestion, Simon, this prompted me to get an old portable radio out of the loft and wander around the house with it. The results were quite an eye-opener.

The cupboard that houses my NAS, switch and router gave out a steady hum, although I haven't yet isolated any individual culprits yet. I have a couple of Airport Express, and one was almost silent, but the other (which is temporarily connecting up my NDX pending a new cable installation) was surprisingly loud. iPhone chargers and cordless phones hummed away quite audibly. Some of the worst culprits were light bulbs. I still have a couple of CFLs running, and these were particularly loud. More annoyingly, though, the LEDs I have replaced them with were quite noisy. Some of these are GU10, and I bought Philips ones rather than cheapo unbranded, thinking this might be a safer bet, but they are really quite loud.

Quite how much of this noise pollutes the HiFi gear is another matter, of course. I've installed dedicated mains, and my LAN has optical cables, so I would hope that I have some isolation from this mush, but I might have to investigate a little further now. 

 

A few years back I could actually see the noise of Plasma TV / Plasma TV power supply and its harmonics from a house about 150 feet away... I could tell what they were watching by the noise pattern!!! - thanks goodness they got a new TV and all is quiet

Hi Simon

This statement has me a bit concerned,my Panasonic plasma is mounted about three feet from my Fraim stack,What,if anything could this be doing to my audio equipment?

Eloise posted:
james n posted:

Do you watch TV and listen to music at the same time ?

Only during Proms season... well and Glastonbury… when Later with Jools Holland is on… and for some Radio 2 concerts.

:-)

Yes ok - I do the same but for serious listening it's from local sources and with the TV off so don't worry about my Plasma (on the wall 2m or so from my kit) being an issue.

james n posted:
No quarter posted:
 

This statement has me a bit concerned,my Panasonic plasma is mounted about three feet from my Fraim stack,What,if anything could this be doing to my audio equipment?

Do you watch TV and listen to music at the same time ?

Sometimes I mute the TV and listen to music,if a sporting event is on.I can not tell any difference with it on or off...maybe it is just paranoia.

Hi NQ, if you can hear no difference then fine... I could hear a slight robbing of micro detail when on, a typical sign of RFI. I could live with it.. but for serious listening the plasma went off. (Physically off as opposed to stand by).

HI Chrissu, I guess with modern living there is going to be some EMC pollution. it’s about being proportional, and keeping it away if you can from your audio equipment.. especially if more high end.

As far as fibre Ethernet, you probably know my view.. if done properly with quality SFPs and care and consideration given to minimum lengths then it’s fine.. but as Naim don’t currently support fibre it’s questionable what is being achieved other than creating a new random noise profile that may or may not sound preferable other than mitigating bad common mode noise from a cheap or errant network component. I suspect  almost certainly it will be introducing clock noise to the Ethernet link... of course you might like the effect of this added digital noise on the analogue signal... the effects of digital noise can be strange.

Put your radio across your media converters and their powersupply leads ... if almost silent the chances are you won’t be adding common mode noise, however you might hear the effects of poor clock stability possibly as a faint unsteady tones on certain frequencies... if the other noise is quiet enough.. this is bad in my opinion as it can cause Ethernet physical layer clock intermodulation ... I believe I found this first when troubleshooting noise from a Netgear switch that was breaking into my Naim tuner.

One of the  biggest noise culprits from networks I believe is from heavy broadcast traffic on your subnet....

I've been wandering round the house with the LW radio again this afternoon, feeling like a bit of a weirdo, but nobody else is in! Some further observations re. sources of noise:

I still have a few fibre media converters on my network, and these do make a bit of noise. Not the worst offenders, but not silent. A couple of them run on iFi iPower PSUs, which were a little noisy, despite the manufacturer's claims (if they're that good, why do they need a ferrite on the DC lead?)

I have a Cisco 2960 connected to my NDX, and this was probably the only electrical device in the house that seems to be absolutely dead silent. I guess that's a good sign. Curiously, I have a Cisco 300 series switch at the other end, and this is very noisy, so I might have to investigate it a bit further.

As well as iPhone chargers being quite loud, my iPhone is noisy when active, with no lead connected. Maybe miniaturisation makes noise harder to mitigate, as my MacBook isn't as bad.

Light bulbs - some of these are awful. I've recently put in a number if 'filament' type LED bulbs, and these are very quiet, with the exception of a single Tesco own brand one, which sounds like a tractor. As do the few remaining compact fluorescents I still have. Time to ditch those, I never liked them anyway!

Does any of this noise actually do any harm to sound quality? Hard to say, but I'm going to try and eliminate some of the worst offenders, anyway.

 

 

Chrissu, interesting post and your point about ifi power supplies is spot on.. especially about the need for ferrite chokes 

Also interesting about the switches, and your point illustrates perfectly what I have said before... just because a switch says ‘Cisco’ on the front, it doesn’t make it ideal for our uses.. it’s the type and model that is more relevant... and also your show SMPS can be designed to produce almost no noise at all as you found out.... SMPS’s are  not automatically a bad thing... the 2960 uses a quality SMPS.

As far as affecting SQ... well it creates noise on the mains or connected cabling... this noise can pollute our audio systems.. whether you hear that pollution is going to vary from person to person and system to system I guess...

Simon-in-Suffolk posted:

As far as affecting SQ... well it creates noise on the mains or connected cabling... this noise can pollute our audio systems.. whether you hear that pollution is going to vary from person to person and system to system I guess...

I'm curious as to whether the cumulative effect of all these noisy devices adds up - in which case eliminating any devices that are a bit noisy might help. Or is it more the case that a single, very noisy device might be powerful enough to reach the audio gear, and only that is worth eliminating?

Likes (2)
Paul Quigley ieMr Frog
×
×
×
×