Balanced inputs vs rca

[DrWho]

So, would cables like THESE be useful or not?

They'd be useful for the unbalanced approach. The only problem I have with those specific cables is the extended metal sheeth...they don't do their tension protection very well and like to unravel. I suppose some versions are built better than others, but I prefer to go with the "guitar style" rubber jackets.

But at $10 a pop these are a great bargain and if you're never moving things around I suppose the reliability isn't as huge of a concern.

[sivadselim]

but could they help eliminate a hum?

[DrWho]

no

most hums are caused by ground loop issues, which are just a headache

to correct. If you have noise problems I would first suggest getting

one of those fancy power strips....especially if you have cable TV (and

don't forget to wire the cable signal through it too). In the studio I

revamped 2 years ago we were able to drop the noise floor by about 40dB

by moving to power conditioners! (including both hums and hiss in the

system).

[Rivendell61]

Or.....for a start, you could use the wire configuration I suggested :-)

It should help.

Doc, did you get a chance to ask?

Mark

[DrWho]

lol, "your wire configuration" is the exact same thing...what am I missing here???

no i haven't....I've been really swamped because we have a week off for

thanksgiving so tests and all that are crazy this week. And then I'll

be off campus and then when I get back it'll be crunch time for finals.

But it's definetly on my to do list (I'm gonna have to find a prof that

is familiar with the topic too).

[Rivendell61]

lol is right....the one I suggested back on page 1. With the cold lead and shield tied together at the RCA output......

The ones I see in the link look like just adapter equivalent cables.

Did you check out the Whitlock links I posted? They will explain it.

Mark

[DrWho]

Yes I did.

The pictures weren't too great, but the circuits are electrically

identical. It doesn't matter where you choose to connect the ground with

pin 1, they're both going to have the same electrical potential

(technically they're on the same "node").

[NOSValves]

Craig,

I suspect some of the 'pro' oriented places have slightly better prices (Redco.com, Haveinc.com, etc) but Blue Jeans prices seem good to me too--and they have a much easier web site to navigate (for me anyway). Also the customer service is excellent. I ordered several sets of interconnects last summer. Ordered each late in afternoon--always had an email within two hours that cables were assembled and mailed Priority from Seattle--and had them in two days in Virginia. Really sold me.

Mark

Mark,

I ordered the night you mentioned www.bluejeanscable.com and recieved yesterday! I was just ordering regular RCA terminated IC's and let me tell you the quality of build is astonishing for the price! 3' pair with belden 1505 or 1694a with Canare ends that will work fine on vintage amps for $25 add $3 extra for Techflex choose your color to boot! What a deal thanks for mentioning them! I will refer all my customers there. Hell it makes putting youy IC's together yourself a waste of time.

Craig

[Rivendell61]

Hey Craig,

Great--After all the free email advice I've gotten off you......I'm glad I finally gave some back!

Doc, there is a great (full page, in color!) pic of the "pseudo-balanced" circuit on p.82 of the AES link I posted.

I think we are kind of talking in circles here.

In any event, the shield-tied twisted pair, unbal. to bal., version will reduce noise/hum ignored by the simple adapter version. As Whitlock describes--by allowing "power line leakage to flow in the shield"....and...."allows the balanced input to sense the signal *at the unbalanced output* and reject common-mode noise".

Mark

[DrWho]

ooo nice picture...how did I miss that before? You're gonna have to

explain to me how the "LO" and "SH" in that picture are carrying a

different voltage...or is that not the point? Does the LO and SH in

"3-C" somehow pickup interference differently than the single SH in

"2-C"? I would think the interference would be the same. I don't think

his "Relative CMMR Performance" chart on the next page is entirely

accurate either based on pesonal experiences....but that said I'm not

about to argue with Whitlock, or especially something published by the

AES [] I just don't understand the "why" is all...

[DrWho]

Ok, so I tried one professor and she had no clue about the subject....but I she gave me the name of another instructor who I can meet up with in a little over an hour...Just so you know, I am delaying my trip home for this....so you better be right []

[Marvel]

Mark and Shawn,

I understand about not needing a signal on both legs to get the CMRR,

but I'm not sure I get the 30db NR in the unbalanced to balanced just

by having the twisted pair (like a low Z mic cable) with the rca end

tied to the shield. I'm thinking you could tie the one line to the

shield on the XLR end and have it be the same, even if electrically

they are.

Nice Powerpoint presentation even if it does sell the Jensen products.

Bruce

[DrWho]

bah, the second professor already went home for the holidays! Oh well, looks like I've got a week to wait.

Btw, I was told that the only possible explanation had to do with the

"non-ideal" behavior of wire. So basically the difference in resistance

and capacitance would be the only possible explanation....

[Rivendell61]

DrWho--I'm no help with understanding the underlying mechanism....

If it helps....I can tell you I first learned of this method from the guy who designed my current amp [he was Chief Engineer at Philips Digital Labs] who insists that this is the best way to connect unbalanced to balanced. He used the term "Kelvin Sensing" saying: "The most important thing to do is use the differential [balanced] inputs to extract the output signal from your [unbalanced] preamp referencing your preamps ground. Outside audio this is known as Kelvin Sensing." Then he points to Whitlocks wiring description.

Benchmark Media (who mainly deal with broadcast usage) also describe what I think is the same thing in an Application Note, calling it "Forward Referencing" saying: "unbalanced outputs must be connected [to balanced input] so the signal is forward referenced. This is accomplished by connecting the non-inverting input to the output terminal, and the inverting input to the 'ground' (signal reference) terminal of the unbalanced output. Now, any difference of potential that still remains between the two pieces of equipment will be ignored by the input as a common mode signal.

Mark

[PrestonTom]

Mark and Shawn,

I understand about not needing a signal on both legs to get the CMRR, but I'm not sure I get the 30db NR in the unbalanced to balanced just by having the twisted pair (like a low Z mic cable) with the rca end tied to the shield. I'm thinking you could tie the one line to the shield on the XLR end and have it be the same, even if electrically they are.

Nice Powerpoint presentation even if it does sell the Jensen products.

Bruce

I agree with Bruce / Marvel (who has been quite correct throughout this entire thread).

If the noise is correlated between the two channels, then the inversion/addition will sum the signals and cancel the noise. The signal-to-noise ratio will be be quite large.

If the noise is uncorrelated between the two channels, the signal will sum (+6db) and the noise will sum (+3dB since it is uncorrelated). In this latter case the improvement will only be +3 dB in terms of signal-to-noise ratio. This latter case, would be unusual and would indicate the limiting noise is perhaps in the electronics and not pickup along the line.

There is no magic here.This is the exact same strategy is also used in transmission of TTL pulses and has been around for sometime.

Good Luck,

-Tom

[DrWho]

If the noise is uncorrelated between the two channels, the signal will sum (+6db) and the noise will sum (+3dB since it is uncorrelated). In this latter case the improvement will only be +3 dB in terms of signal-to-noise ratio.

Actually, If you were to send the signal down both pins then they are both -3dB along the cable, so when summed you get no boost in volume at the reciever. Any uncorrelated noise ends up being +0dB at the reciever as well (so any uncorrelated voltages picked up on the cable = the voltages going out). In other words, there is no noise rejection whatsoever for any uncorrelated interference.

The end volume of the output must be the same regardless if you send the signal down both pins or down just one (the summed total MUST equal the same output of a single pin carrying the signal). Also, it would be quite annoying if your signal got louder simply by plugging more devices into the signal path - just imagine the headache in the studios where you could have 10 devices in series with the signal...at +6dB between every device that would be +60dB at the end! Thankfully that is not the case because that would require a noisefloor of at least -80dB on every piece of equipment.

[Rivendell61]

Marvel,

The 30dB reduction is not fixed, it is just the amount derived for the example Whitlock provides in what he thinks represents an average consumer set-up. The actual amount would vary depending on several things (as it does in any balanced connection)--but is fundamentally based on the ratio of the balanced input impedance of the reciever to the output impedance of the unbalanced source--which is why I told Coytee the effectiveness of his set-up would be maximized using the low impedance setting on the Peach.

As far as why the shield is connected at the unbalanced end.... Benchmark points out there is advantage to tying the shield at the 'send' end rather than the 'recieve' end in: "reduction in coupling of high-frequency noise to the audio signal". I'm not sure if that is why Whitlock does it.

Mark

[DrWho]

As far as why the shield is connected at the unbalanced end.... Benchmark points out there is advantage to tying the shield at the 'send' end rather than the 'recieve' end in: "reduction in coupling of high-frequency noise to the audio signal". I'm not sure if that is why Whitlock does it.

Well all the graphs show the reduction occuring in the HF...which is actually kinda odd because phase reversals tend to work better at lower frequencies (because they don't have to be as perfectly lined up). I've got a high frequency noise in the channel going to my sub amp that has been corrected using a transformer to get rid of the hiss, and since I have so many spare XLR cables I'm going to have to give this a try (the adapter cable I bought does the ground connection on the recieving end).

Assuming it does make a difference, then wouldn't you get half the attenuation if you connected the shield in the middle of the cable? That will be another experiment I try....In fact, that would be really easy to do because I just use my prebuilt cable and then plug it into an XLR cable to carry it the rest of the way.

[PrestonTom]

DrWho,

Re: the +6 dB gain.

You were a bit quick to reply on that one. No matter if the inputs are attenuated. My analysis (not mine actually since this is standard engineering) is in terms of signal to nose ratio. Obviously one would not want the signal to have a net gain when going thrugh a balanced cable. I was just clarifying that you MUST have the inversion/summation to get a gain in signal to noise ratio. My examples were meant to be helpful in that this exactly what you would expect if you simulated this with a signal generator and 1 or 2 noise generators and looked at the output on a scope/analyzer

In terms of gain in S/N ratio, the two extremes are the cases where the noise is correlated or when it is not correlated.

This is standard stuff and I was confused by why others were confused.

Good Luck,

-Tom

[Rivendell61]

DrWho,

I think you are jumping too fast re the shield.

Remember that noise reduction is unrelated to signal symmetry/phase.

It is only due to identical impedance on both conductors. And follows the explanation above, that in the pseudo-balanced wiring scheme, the balanced input senses the signal at the unbalanced output and rejects the common mode noise.

The question of how essential it is to have the shield connected at the unbalanced end to achieve this is all I was responding to. I don't know. The point Benchmark made was indicating a 'possible' advantage regarding HF noise. But that was not related to the overall noise reduction cited by Whitlock.

Mark