EMI everwhere !

[Nick]

EDIT

The EMI problem outlined in this post turned out to be due to a production problem with the Oscilloscope I bought to look for noise in my music system. More details of what happened with the scope are given in my post on page two ( http://www.phasure.com/index.php?topic=2404.msg27276#msg27276 ) but I wanted to save anyone wasting time reading thought the post.



Regards,

Nick.


I wanted to post on the subject of the effect of environmental EMI noise. I am on the journey of understanding and diagnosing an audible "noise" problem in my system any pointer / experience that anyone can offer would be greatly appreciated (I know there are some experienced engineers amongst us).

The audible sound quality issues manifest themselves as a number of sound quality symptoms. It is possible that some people may also be experiencing them to greater / lesser extent possibly without knowing it. The audiable symptoms of the problem include:

• Faint playback of music when playing at a -144db XX volume setting (this happens whilst playing with my NOS1, Pauls NOS1 and just this weekend with my friend Pete’s Musical Fidelity  NuVista Dac).
• Harshness if the upper frequencies (this is something that may be quite widespread reading some of our posts).
• Volume imbalance between left and right channels (usually about 2 to 5 db differance, estimated subjectively)
• Generally very variable and poor sound quality performance from my system, on occasions sound quality has shone and my NOS was fine playing at Pauls house but 90% of the time my system has refused to perform as well as I know it should (Pauls system set the standard).

Until recently I had worked on the basis that there could be a single rouge bit of equipment in my system causing the problem (eg a part of the PC, the NOS1, Gainclones, speaker cables etc, etc, etc). My approach was therefore to replace the system components one by one until I stumbled on the cause. The problem is that with 60 % of my system changed the problem was still most defiantly there so a new approach was needed. Enter the oscilloscope that I had been promising to buy myself for many years, at last I have gained some “eyes” to look into what is actually happening.

So how did it look with the oscilloscope ? .... not good, that's for sure !

It turns out that on virtually a wet bit of string anywhere in my house I can pick up air bourn interference in the 100Mhz range which happens in short pulses, the pulses have a frequency of about 64khz. This stuff is EVERYWHERE, on the mains, on my speaker cables, in the digital and analogue sections of my NOS1 and my Gainclones (who needs a PC to create noise when you have this much about already) .

When I say wet bit of string mean for instance the 30 cm length of untwisted cable in the picture below which is picking up 0.6us bursts of ~100mhz at an amplitude of 10mv !

The general bursts of ~100mhz do have a pattern of sorts this is a sample of the general profile if the ~100mhz spikes – ouch ! (the vertical scale is 2mv per division). 

And a closer look at one of the larger spikes showing the 100mhz wave form.

So for a few weeks I have been trying to understand where the EMI is coming from and what it is doing to my audio system. So far as the system is concerned since basic Gainclones have no regulation on power supply and no HF rejection on their inputs and outputs the airbourne EMI is certainly causing problems. I think the effect on the Gainclones and the interaction of this effect with my  NOS1 might well be causing the NOS’s word clock signal to become out of sync with the bit clock as 32bit words are processed by the DAC section. I think this may be resulting in LSB word data from one 32bit word entering later adjacent 32bit words at more significant bit locations.  This super jitter however it is happening just ruins music quality. A couple of things it is important to point out are that this defiantly a Gainclone / DAC interaction problem. As I mentioned before, my friends Musical Fidelity DAC (played using SPDIF from a M2Tech) exhibits the problems even more than my NOS1. Whilst using an amplifier with better RF rejection (A Denon midi component as it happens) the problem goes away completely so far as I can tell, the only issue with this solution is that the sound quality of the Denon is way below the potential of the Gainclones.

So where is the RF coming from ??

I checked the house and the immediate area round my village expecting to find an SMPS about to explode somewhere close but nothing so far as I could see. I did a lot of reading about EMI on the net about possible sources and started driving round in my car with the oscilloscope to “look” for the source (fortunately the scope runs from a battery as well as the mains). This evening I drove around half the county trying to work out where the source was. After a couple of hours I ended up parked at the base of a 100m transmitter 20 miles from my house with the amplitudes of the 100mhz bursts 7 X higher than back at the house. I think it is a TV transmitter. The culprit ??

So I have some questions which I need to research now.

• Is it a TV transmitter ?
• Should digital TV signals have this 100mhz modulated pulse structure at about 64Mhz ? is that normal.
• The EMI signal shows as a 5mv pulse on my NOS1 outputs which I know from Peter has a noise spec of about 6uV !!!! That is not good for signal to noise ratio coming from my DAC.
• Are my specific set of Gainclones more susceptible to the interference than they should be  ? I plan to build another balanced set to try this theory (another upgrade in the name of EMI  ).

No resolution as yet for these these points, I might still be looking in entirely the wrong direction but I think there is smoke here so the fire should be close as they say.

If anyone has any experience of this stuff I would love to hear your thoughts, anything that can shorten the path to sorting this out would be fantastic.

From an EMI buzzing Suffolk.

Nick.

Ps I just hope I done have to build a copper clad faraday cage with no windows and power my system from batteries before I can listen to it at its best 

[manisandher]

Hey Nick, excellent post.

I can't really help you other than to say my oscilloscope is coming out tomorrow. I'll see what sort of readings I'm getting at 100MHz and let you know.

I don't have a copper-clad Faraday cage, but do have the 'studio' in my cellar totally lined with 'plasterer's wire mesh' connected to ground (effective enough to take my mobile phone from 3 bars 3G just outside to 0 bars inside). One thing I will do is see how much EMI at 100MHz is eliminated with this approach.

On a last note, for really serious late-night listening, I always disconnect my wifi router and DECT phone. This simple step gives me a sweeter and more musical sound. Have you tried this?

Good luck with your quest.

Mani.

[Nick]

Mani,

Thanks.

I did suspecting DECT and WiFi but could not see anything worrying with the 'scope. I agree that it's going to be a good precaution to turn both off when going for best possible sound.

I was skimming arround on the net just now. A little more information on the transmitter that I ended up parked under. The tower is 300m high not the 100m I had guesstimated !!. A summary of output is as follows.

CURRENT TRANSMISSIONS
 
Digital Television - No transmission
 
Analogue Radio (FM)
 Kiss 105-108 FM:
 106.4
 20kW
   
Digital Radio
 
BBC National (Block 12B)
Digital One (Block 11D)


So the transmitter is for Radio not TV. Kiss FM seems to be verrry close to the frequency I am seeing. Now I need to understand if the signal structure I am seeing is normal for these services. It seems odd that there should be such pulsed variation of the signal intensity.

More time on the net needed....

Happy new year,

Nick.
 
 
 
 
 

[PeterSt]

Nick, man ! To-tal-ly interesting. Yea, well, sad it hurts you all so much.

Heck, I'm dying to hop over just out of interest ...

Djeezzz ...


For others who don't know : Nick has been through all sorts of problems at least me myself an I never were able to understand. What about perceiving sheer (though distorted) music without even any D/A section operable. What has been spend on this all together ? a 200 emails ?
I have said so often : hounted house. UK-familiar.

But with a scope you will hunt down the real ghosts ?

Damn Nick. You must be getting somewhere ...

[PeterSt]

Nick,

Ciska here, our NOS1 builder as we know by now, was shown this topic as some sort of teaser, and she right away comes up with these links :

http://www.dailymail.co.uk/news/article-2231351/Homes-robbed-TV-signal-new-4G-mobile-network-10-000-reconnected.html

http://www.dailymail.co.uk/news/article-2145566/4G-mobile-signal-hit-TV-reception-UK-warns-culture-minister.html

So you might investigate whether 4G is transmitted from this tower.
No time to digest this myself. Dinner time and such.

Peter (well, merely Ciska)

[Nick]

Hey Ciska,

Thanks for the reading links above, I will check out what is being transmitted. The site for the transmitter only lists radio services, but that is not to say that it may not be out of date and other services are not being transmitted.

I looking for information on the DAB transmosion standard just now. There might be another hit on the frequesncy I am seeing.

"Following the revision by the CEPT multilateral meeting which took place in Constanţa, Romania, 02-04 July 2007 the WI95revCO07 Special Arrangement governs T-DAB in the frequency bands T-DAB in the frequency bands 47-68 MHz, 87.5 - 108 MHz and 230-240 MHz."

The thing that is unexpected is that the signal pulses so much with 10s of khz modulation. The stronger pulses are must be circa 10 to 20x greater than the base transmission level. Its these pluses that seem to be the problem. I am still not sure exactly how the DAC gainclone problem happens when excited by these pulses but even with the NOS1 turned OFF and still connected (you would think it would ground the amp inputs) and the mv 100mhz pulses are there on the amp inputs and amplified by the Gainclones. They appear as a very noisy output pulse on the speaker connections in the 10s of mv range as a pulse in the 10s of khz range (the Gainclones cannot pass the 100mHz signal). Not good at all....

Don’t let me keep you from your dinner :-),

Kind regards,

Nick.

[PeterSt]

Nick,

I don't know much about this, did not sort out a thing and did not Google, but I'm afraid we'd have to consider this normal. Consider this as how radio is and we all live with that. So when I'm right ("radio is"), the culprit must be on your side, hence must be solved on your side.

What I would do first is brew myself a very short interlink. Try to use the same material (for cable) as you have now, or otherwise create a normal length you use of the same material as this short cable you are going to make.
Now compare the two at the amp output;

If there's a difference and the short cable shows less of this interference, start thinking about an interlink with sufficient shielding. However, maybe I wouldn't take the effort (but I'd use Triax to start with - well, double shielded coax that is like for sat receivers that should be used) ... and find yourself a sheet of mumetal. Possibly very hard to get in a shop, but the internet will provide it somewhere. Get yourself some money for it because you will need to cover all of your interlink (make a roll of it I'd say). You could first try it on your short test cable.
Remember, only when the shorter cable helps in the first place.

The above shouldn't help a thing because you claim that with the Denon there's no problem. However, this seems inconsistent to me re the 5mV on the NOS1 outputs. So ?

Well, you could try to apply the same test with the USB cable. Does a longer look worse on the NOS1 outputs ?

For your GainClone, think about ANY shorter or longer sticking out (loose) wire, resistor etc. So, you can have a test situation in there and can have cut a resistor at one side. Wrong, because now it's an antenna (look how your wet wire operates).

In the NOS1 there shouldn't be much of this (unnecessary long bare wires etc. and what is there was unavoidable (and still should be).

Although you will have set up a ground scheme that works for you at the moment (or seemed to), experiment with PE connections again. You even might - the other way around - create yourself a mains cable for the NOS1 which has no PE wire in it. But otherwise you can remove the ground connection to the cabinet right near the power inlet. Do notice that PE is only connected to the cabinet and nothing else. Carefully consider whether you implied a connection to something else yourself or thought to use the cabinet for some ground. Remove that again. Not permanently, but to see whether things change.
I can imagine that the cabinet acts as a receptor especially when connected to PE. But when not connected to PE also (and then connect to PE for trials). You may wiggle the USB cable from the input terminal to the interface inside and then for distance to the chassis bottom.

You can try to change angles of your apparatus and try to avoid the waves (which I think won't help at this high frequency).

Although you see it on the mains, do not think right away it comes from the mains; most probably you feed it yourself to it. See above NOS1 - PE example. But also think this may bring you the solution; find that antenna in the house (hence disconnect everything and keep on measuring the mains).

Prove to yourself that it's not your measuring wires picking it up (not sure how to do that; measure something in your car ?)

When you found sources to change, again get that mumetal. But read yourself into how to apply it. Do notice that there's no free lunch and that although the shielding may work, at the sides and holes etc. there will be even more collection. You key is to bend off the radiation in the direction of your liking. Again, read yourself into applying it.


One addition from my own experience :
You may run into the idea of contacting the head of the tower (he will stand on the top of it) with a "what the f*ck you think you are doing here". Well, useless. Once frequencies are above a certain frequency (I have 14KHz on the top of my head) there are no rules for it. But worse : because no rules/regulations exist there will also nobody there to help you. It is not needed (for testing / certificates etc.) and the test equipment won't even be available. You will be on your own (but have Google).

1c
Peter

[Jud]

There are cables I own that are made of mu metal, if I recall correctly.  They are the best sounding ones I have, other than the Mapleshade/Omega Mikro.  They were called Lindsey-Geyer when I bought them decades ago - don't know if they are still available.

If they are not available, I have one set of spares (set of analog interlinks, I think 2m, maybe 1.5) and likely another set will become spare soon.  If you're interested, PM me and I could send them along.

[Jud]

Well, the Lindsay-Geyer cables are still "available," if you consider that adjective to apply to $650 for a 1 meter pair.  I will lend you my spare pair gratis if you'd like, to see if they work in your environment.

[Nick]

Well, the Lindsay-Geyer cables are still "available," if you consider that adjective to apply to $650 for a 1 meter pair.  I will lend you my spare pair gratis if you'd like, to see if they work in your environment.

Jud,

Thanks for the kind offer. 650 dollars per meter per ouch ! Then again I have been looking at the price of mu metal and its not cheap or easy to use, so there could be quite a cost associated with building the cables. They could be very useful.

I have ordered an amerture radio (radio ham) handbook on emi diagnosis and management. I think the problem is going to need a systematic approach.

The radio signal everywhere in the system, speaker cables, amp inputs, power supplies, inputs, throughout my amp and NOS etc. Currently there is only one point where PE and signal earth are bonded and that is back in the PC. The NOS and amp signal grounds are not connected to PE. This means that efforts to shield components has to involve local PE grounding within each component. Fitltering of psu, signal or amp output noise to signal ground currently means the noise has to pass all the way back to back to the PC signal ground to reach a PE (earth) connection. This has the potential to route the noise from the amp and speaker wiring (big aerials) back through the NOS (I think this may be a part of the problem) to the PC. Because of the length of the path to PE ground and the VHF frequency of the noise, the inductance and resistance of such a long path the seems to be making it hard to earth the noise. This is going to tale some working out.

I am guessing the answer may lie in a joint approach of minimising the opportunity for the system to act as an arial (as Peter suggests) and working out how to run noise from shields, psus and signal circuits back to PE / signal ground without spoiling the sound quality. This is going to be a step learning curve I think 

I am keen to do some reading to come up with an approach before I start trying things out so there may be a bit of a delay before I can try the cables if that is ok.

I will post a couple more traces, boy there is lots of noise in the system....

Cheers,

Nick.

[Scroobius]

HI Nick - well I guess that explains a lot I wish you the best of luck it looks as though you have much work ahead. Might be an idea to bring that scope with you next time you come down here and compare yours with what a "normal" environment has.

Best of Luck

Paul

[Nick]

Paul hi,

Thanks, we are thinking the same thing about taking some more normal measurements, it would be really useful. The joys of living < 15 miles from a 1000 foot 20kw transmitter !!

I'm looking forwards to hearing your system, problem is I might end up wanting to move house 

Cheers,

Nick.

[PeterSt]

problem is I might end up wanting to move house 

That was the first suggestion here. But I said "no no, Nick's (hotel-)house is way too nice to leave behind".

So don't you dare do that !
(wimpy )

[Jud]

No problem at all, Nick, I will likely have a second spare pair by the time you want them.

If they would wind up being helpful and you would like to keep them, I am sure we could work out something satisfactory that isn't nearly in the stratospheric range these cables now apparently command.  I hope they do help, but if they don't, then just send them back and all's fine.

[christoffe]

http://www.monoandstereo.com/2011/06/why.html

Interesting to read!