Test showing EMI/RFI affect on XLR cables

[BadAudioAdvice]

Sounds Speeds YouTube Channel just released what IMO seems like a fair test of cable performance in rejecting EMI/RFI.

It certainly showed some differences.

I wonder if a setup like this could be added to the ASR cable testing suite?

 

[Chrispy]

Damn he's got an annoying voice, couldn't handle but a few seconds. What apparatus did he use for testing?

 

[Jim Shaw]

Damn he's got an annoying voice, couldn't handle but a few seconds. What apparatus did he use for testing?

Some voices inspire incredulity. This is one. I lasted maybe to sentence 3.
.....
A few years ago, I was buying an AVR from Best Buy. The sales guy introduced himself as the store manager. I asked why the receiver was competitively priced, but his stock of HDMI cables seemed outrageously expensive.

"We gotta make money somehow... we earn as much on two or three cables as we do on that $150 receiver."
I was only surprised by his candidness and bought the cables from Monoprice.

 

[Blumlein 88]

The guy clearly has sped up his voice. While not a pleasant voice, it is far less annoying if you'll open the video settings on youtube and reduce playback speed to .75 of normal. However, that makes his tedious video last several more minutes. He also does not pronounce Monoprice correctly.

 

[Blumlein 88]

I skipped around so hopefully didn't miss anything important.

Firstly he drops cables behind his computer desk and loopbacks to his Sound Devices interface. The Amazon cable showed some noise, and the Cable matters just a little more than other cables, but less than the Amazon. I wouldn't use either for a microphone lead, but probably not an issue for line level.

He then used an EMF meter to find the noisiest spot around his desk. Retested the same way and this time Cable matters was worse and Amazon not far behind the good cables. Since those were the two picking up some noise I would have lightly twisted them and tested again.

The Amazon looks just like the cheapest garbage cables I've seen. I've opened some of those up when they gave noise issues. The plus and minus leads are not twisted very much at all and they have very porous minimal shielding or simply use an also lightly twisted ground lead without shielding. The Amazon cable doesn't mention shielding and the Cable Matters say copper which I bet is some loose weave with maybe 50% coverage. Moral of the story, people will make garbage if they can sell it cheap and still make a buck. Go just very slightly upstream in price and you have nothing to worry about with XLR cables.

 

[Jim Shaw]

I will not buy any cable, at any price, that doesn't carry out the empties and empty the ashtrays first thing in the morning.
Harrumph.

 

[fpitas]

If I was worried about noise pickup in XLR cables I'd get the Blue Jeans cables made with Canare Star Quad: https://www.bluejeanscable.com/store/balancedaudio/index.htm

Having said that., I am happily wired with a lot of cheapy Amazon XLRs that work fine, but my house is pretty far away from everything.

 

[Blumlein 88]

If I was worried about noise pickup in XLR cables I'd get the Blue Jeans cables made with Canare Star Quad: https://www.bluejeanscable.com/store/balancedaudio/index.htm

Having said that., I am happily wired with a lot of cheapy Amazon XLRs that work fine, but my house is pretty far away from everything.

Yeah that is the other thing about a test like this. You don't have to move the cable very far at all from such sources of noise and it drops well below any level to matter. The worst case in this test was over a known hotspot near some harddrives and a power supply. And the noise was a bit lower than -90 db down. You likely wouldn't have heard it and moving it even 3 inches would drop it considerably.

 

[atmasphere]

He didn't mention how he was reading the noise.

XLR cables don't block noise!

If your equipment supports AES48, the receiving end of the cable will not amplify noise that is common to pins 2 and 3, depending on the Common Mode Rejection Ratio present in that receiver. So how well the cable is shielded doesn't seem like the thing to measure.

 

[chris719]

There’s so many things wrong with this “test” that it’s not worth the time to break it down. Also, any English-as-a-first language speaker that pronounces Monoprice like that is a monster .

 

[fpitas]

I'll mention a big problem with testing bargain cables like that: you have no guarantee or expectation they will remain the same design or construction. So, even if the test is perfect, your data is ephemeral.

 

[Blumlein 88]

He didn't mention how he was reading the noise.

XLR cables don't block noise!

If your equipment supports AES48, the receiving end of the cable will not amplify noise that is common to pins 2 and 3, depending on the Common Mode Rejection Ratio present in that receiver. So how well the cable is shielded doesn't seem like the thing to measure.

Yes he did. He was using a 150 ohm dummy load plugged directly into the xlr input jack of his audio interface to read the base noise level. He then would connect that load to the input of his interface with the cable. So any noise difference would presumably be from the cable picking up the noise. He also used an EMF meter to get an idea what sort of electromagnetic fields were in the area in microteslas. So in general he was reading levels of -112 or -113 db. Any reading higher with a cable connecting would be from the cable under test picking up noise from the environment in addition to the thermal noise of the 150 ohm dummy load and noise inherent to his interface input.

So one cable appeared not to be picking up anything above what his interface input had, and another had a little something, but not even enough to alter the result 1 db. Cable Matters and Amazon cables picked up enough to make a few dbs or even 10 db differences.

 

[DonH56]

I have not seen the video, and probably won't, but there seem to be a couple of things conflated here... EMI/RFI shielding and common-mode noise rejection are two different (albeit often coupled) things.

1. How well cables block EMI/RFI depends upon the shield coverage, connector terminations/construction, and interference frequency. Better shielding (e.g. braid + foil) with better connectors to a proper isolated shield (AC) ground will be more resistance than loose (braided, or no) shield with large gaps in the shield at the connectors.
2. A differential system will reject common-mode noise, which ideally XLR provides, but it depends heavily upon implementation. "Balanced" does not mean "differential" -- it can be a range from passive impedance balancing through quasi-differential circuits to fully-differential Rx/Tx using active or passive (transformer) devices. But even then common-mode rejection only works within the circuit's bandwidth to reject it; CMRR decreases as frequency increases. An amp with 120+ dB CMRR at power-line frequencies may only have a few dB at RF frequencies.

There's a lot more to it but I think those are the two basic discussion points for noise rejection...

Actually measuring shield effectiveness is a PITA I hope to not need to do again... For me it involved a screen room or anechoic chamber and fancy set of equipment to measure field strength at the cable as we swept frequency and power, e.g. from an antenna pointed at the cable driven by a high-power (W's to 10's of kW depending upon the frequency) sinusoidal or noise source.

FWIWFM - Don

 

[Axo1989]

The guy clearly has sped up his voice. While not a pleasant voice, it is far less annoying if you'll open the video settings on youtube and reduce playback speed to .75 of normal. However, that makes his tedious video last several more minutes. He also does not pronounce Monoprice correctly.

One of those cases where we really don't want pitch correction. Now, do I listen to see how he says "monoprice" ...

 

[restorer-john]

Actually measuring shield effectiveness is a PITA I hope to not need to do again... For me it involved a screen room or anechoic chamber and fancy set of equipment to measure field strength at the cable as we swept frequency and power, e.g. from an antenna pointed at the cable driven by a high-power (W's to 10's of kW depending upon the frequency) sinusoidal or noise source.

Shouldn't have stood in front of that magnetron waveguide for so long Don...

 

[Blumlein 88]

One of those cases where we really don't want pitch correction. Now, do I listen to see how he says "monoprice" ...

Mono-pris is how he says it.

 

[DonH56]

Shouldn't have stood in front of that magnetron waveguide for so long Don...

...And definitely should have taken the chocolate bars out of my pocket.

Maybe that's what happened to all my hair!

 

[atmasphere]

Yes he did. He was using a 150 ohm dummy load plugged directly into the xlr input jack of his audio interface to read the base noise level. He then would connect that load to the input of his interface with the cable. So any noise difference would presumably be from the cable picking up the noise. He also used an EMF meter to get an idea what sort of electromagnetic fields were in the area in microteslas. So in general he was reading levels of -112 or -113 db. Any reading higher with a cable connecting would be from the cable under test picking up noise from the environment in addition to the thermal noise of the 150 ohm dummy load and noise inherent to his interface input.

So one cable appeared not to be picking up anything above what his interface input had, and another had a little something, but not even enough to alter the result 1 db. Cable Matters and Amazon cables picked up enough to make a few dbs or even 10 db differences.

The interface with the cable is the thing. If the noise is common mode its supposed to be rejected by the CMRR of the input the cable is driving. That's the part I was talking about. The noise present on the cable otherwise simply isn't a thing- the CMRR of the input is.

Balanced operation can be seen as an exotic cable technology where the source and the receiver are doing all the heavy lifting- all the cable has to do is be built correctly- a twisted pair running inside a shield. If the source and receiver are properly built the noise in the cable doesn't amount to a pile of beans.

 

[Blumlein 88]

The interface with the cable is the thing. If the noise is common mode its supposed to be rejected by the CMRR of the input the cable is driving. That's the part I was talking about. The noise present on the cable otherwise simply isn't a thing- the CMRR of the input is.

Balanced operation can be seen as an exotic cable technology where the source and the receiver are doing all the heavy lifting- all the cable has to do is be built correctly- a twisted pair running inside a shield. If the source and receiver are properly built the noise in the cable doesn't amount to a pile of beans.

Read what DonH56 posted. Then realize with actual devices some cable shielding and construction is not well done. Which can result in various levels of noise intrusion. There are some very poorly made cables out there. Passive impedance balanced circuits are common even in rather expensive gear.

 

[atmasphere]

Read what DonH56 posted. Then realize with actual devices some cable shielding and construction is not well done. Which can result in various levels of noise intrusion. There are some very poorly made cables out there. Passive impedance balanced circuits are common even in rather expensive gear.

If the equipment you are using has a high CMRR then the shielding really won't amount to much. In fact in some systems it can be absent and actually get lower noise. The reason the balanced line system works isn't because of the shielding- the equipment on either end of the cable is what does the heavy lifting. So there can be quite a lot of variance with the cable and it will be of no consequence either audible or measurable.

We were the first company worldwide to offer balanced line amps and preamps for home stereo use. Since we started we've seen a lot of other companies jump into the fray but I've noticed that most of them (in high end audio) are either ignorant of or choose to ignore the balanced line standards (AES48) and accompanying low impedance advantages. As a result the performance of balanced line connections with equipment like that is compromised. Ground loops, noise, all the problems associated with single-ended connections are there as well. A few years ago one popular tube amplifier by a major manufacturer in high end actually had a CMRR value of less than 20dB... When you have CMRR values that low, you're asking for trouble. Then shielding and the like will make a difference. But if the equipment is designed competently it won't.