The function of the shield in balanced systems is to extend the chassis ground to the piece of equipment at the other end of the cable, as well as to basically encase the entire cable in a "stretched and flexible" chassis. It does not carry signal current at all. So, what you did is not technically correct, but most equipment is so good that it can still function in less than ideal circumstances. Your cable would probably not work if it were exposed to high levels of RFI or EMI.
Exactly! Which also means that the two signal levels on the pin 2 and pin 3 wire do not have to be equal in amplitude, that has nothing to do with a balanced connection! Full signal on pin 2 and zero signal on pin 3 works equally well as 1/2 signal on each pin 2 and pin 3. Only a few people realize that.
Jan
Andrew read the posts above on the requirement that the signal conductors are symmetrical to themselves and to 'the rest of the world'. Does your trick guarantee that?
jan
All of the noise immunity of a balanced connection happens at the receiving end, by way of its "common mode rejection". To the extent that interfering noise is the same on both signal lines, it's rejected by the receiver.
So the whole trick is to try to keep interfering noise *the same* on both signal lines. Note that this is completely unrelated to balanced *signal* levels.
All good fortune,
Chris
So the whole trick is to try to keep interfering noise *the same* on both signal lines. Note that this is completely unrelated to balanced *signal* levels.
All good fortune,
Chris
While it's won't work with phantom powered mics, you can use one Cat5 cable for 4 balanced interconnects.
You'll notice that for every four microphone cables, there is a fifth cable in the middle. And what is that cable? A Category 5e or 6 cable! Those four microphone signals are carried by the four pairs inside the data cables. These little boxes are called InstaSnake and made by a company called ETS in Fremont, California. (Energy Transformation Systems, Inc. :: Contact Us). These are clever boxes. If you hook them up with UTP, unshielded twisted pair Category 5e or 6, they work just fine but there is no ground connection. For self-powered or dynamic microphones this is just fine. Of course, the noise rejection comes from the fact these are twisted pairs that are differentially-driven balanced lines, with the best balance (CMRR) ever created in a twisted pair. And the best data cable for rejecting noise, among all the data cables, is the one with the "deepest" CMRR.
The above from a Belden Blog.
"The Strange World of Cat 5e and Cat 6"
Posted by: Steve Lampen on October 08, 2013
The Strange World of Cat 5e and Cat 6
You'll notice that for every four microphone cables, there is a fifth cable in the middle. And what is that cable? A Category 5e or 6 cable! Those four microphone signals are carried by the four pairs inside the data cables. These little boxes are called InstaSnake and made by a company called ETS in Fremont, California. (Energy Transformation Systems, Inc. :: Contact Us). These are clever boxes. If you hook them up with UTP, unshielded twisted pair Category 5e or 6, they work just fine but there is no ground connection. For self-powered or dynamic microphones this is just fine. Of course, the noise rejection comes from the fact these are twisted pairs that are differentially-driven balanced lines, with the best balance (CMRR) ever created in a twisted pair. And the best data cable for rejecting noise, among all the data cables, is the one with the "deepest" CMRR.
The above from a Belden Blog.
"The Strange World of Cat 5e and Cat 6"
Posted by: Steve Lampen on October 08, 2013
The Strange World of Cat 5e and Cat 6
I am asking.
The 4 pairs inside the insulating cover of the CAT5 are each balanced pairs.
If they can operate out to over 100MHz as a balanced interconnect when in close proximity to three other pairs, then why should the same pair at <1MHz not be able to work effectively as a balanced interconnect? Again that is a question.
a balanced interconnect does not need a shield. That is NOT a pre-requisite.
The shield attenuates interference AND couples the Chassis to Chassis so that the enclosures are extended to include the cable "inside" the enclosure.
As noted above (in post17) the balanced pair are taken from the screen termination at the Chassis to the circuit by the two wire pair ONLY.
Yes may well be working good. I don't know these cables, so cannot comment on them.
Be aware that 'working' on a digital signal can, in some respects, be easier, more forgiving to not being exactly symmetrical, than with analog.
With digital you can always 'spruce up' the signal in the receiver, there is no need for exact reproduction of the signal. Only the data bits need to be recovered. I am pretty sure that these systems are quite forgiving to crosstalk and interference.
With analog, you need exact reproduction (subtraction or summing). So it may not be the same requirements, but I have no experience here.
Jan
Digital balanced signals (LVDS) is becoming more common, just doing a PCIe interface layout where all the signals are balanced pairs (0.050" skew max between lines).
The problem with CAT5 is that each core has a different twist per metre to minimise crosstalk between the pairs, so you could only use one pair, if you doubled up pairs the noise would not cancel out due to the mismatched twist.
Jan, I was not differentiating between balanced cables and balanced signals (differential mode transmission) yesterday, I quite often have to design a balanced connection on boards for sensitive single ended signalling.
The problem with CAT5 is that each core has a different twist per metre to minimise crosstalk between the pairs, so you could only use one pair, if you doubled up pairs the noise would not cancel out due to the mismatched twist.
Jan, I was not differentiating between balanced cables and balanced signals (differential mode transmission) yesterday, I quite often have to design a balanced connection on boards for sensitive single ended signalling.
Why does the noise affecting one pair not cancel?
If combined with the noise canceling on the other pair, why would the two sets of noise canceled signals not add up to a noise canceled total?
If the cores have a nominal 110ohms characteristic impedance, would a paralleled pair have a characteristic impedance of 55ohms?
The noise affecting one pair does cancel, it is the twists that keep the noise common mode.
using two cores and twisting together, I am pretty sure because of the different twists between the twisted pairs would cause some unbalance causing some noise to be differential, but I may be wrong. I like to keep things simple with balanced and only use two cores.
The 100 ohms differential impedance is at 100MHz usually, if you twist two pairs, I believe that this figure would not change significantly, again I may be wrong. Paralleled pairs would have no advantage as it is the twisting that provides the noise reduction. Where JN when you need him
using two cores and twisting together, I am pretty sure because of the different twists between the twisted pairs would cause some unbalance causing some noise to be differential, but I may be wrong. I like to keep things simple with balanced and only use two cores.
The 100 ohms differential impedance is at 100MHz usually, if you twist two pairs, I believe that this figure would not change significantly, again I may be wrong. Paralleled pairs would have no advantage as it is the twisting that provides the noise reduction. Where JN when you need him
Just wondering that myself, here the page,
Canare Corp.: Star Quad Series: Star Quad Microphone Cable(L-4E6S / L-4E5C)
just reading up.
As I've said with balanced I like to keep things simple and tend to either buy my balanced cables or use a Belkin 2 core cable when I have made them (and that was many years ago).
Canare Corp.: Star Quad Series: Star Quad Microphone Cable(L-4E6S / L-4E5C)
just reading up.
As I've said with balanced I like to keep things simple and tend to either buy my balanced cables or use a Belkin 2 core cable when I have made them (and that was many years ago).
Yes I understood that. I've also been using differential signaling with a single-ended signal.
Jan
post28 link includes this titbit:
extrapolating from that link gives a user the choice of
4pair, no ground
3pair, 2wire ground
2pair, 4wire ground
1pair, 6wire ground.
The article from the Belden employee suggests (strongly) that 4 microphone cables can be transferred into a single CAT5e and not lose any of the CMRR of the 4 balanced pairs.
I'm tempted to suggest a further option for around the "house" distribution:
2pair for audio,
1pair for dual polarity DC,
1 pair for 2wire ground.
Thoughts, please, for suitability for a domestic distribution using the installed CAT5e cables (~53 sets of 4 pair to all the rooms and loft)
extrapolating from that link gives a user the choice of
4pair, no ground
3pair, 2wire ground
2pair, 4wire ground
1pair, 6wire ground.
The article from the Belden employee suggests (strongly) that 4 microphone cables can be transferred into a single CAT5e and not lose any of the CMRR of the 4 balanced pairs.
I'm tempted to suggest a further option for around the "house" distribution:
2pair for audio,
1pair for dual polarity DC,
1 pair for 2wire ground.
Thoughts, please, for suitability for a domestic distribution using the installed CAT5e cables (~53 sets of 4 pair to all the rooms and loft)
Last edited:
Cat 5 is actually entirely usable for carting balanced line level audio around and is used semi routinely to do just this (A very popular trick when temporarily rigging a building for broadcast), it is also good for phones, data and even video service.
Screening is actually far less important then twisting, at least as far as low to medium frequency pickup is concerned.
On the receving end the important figure of merit is the common mode impedance, as this determines the systems sensitivity to minor imbalances in the cable and source output impedance.
The way to think of a differential signalling system is as a wheatstone bridge with the noise being the common mode driving voltage and the signal being introduced by a voltage source in one leg. Seen in these terms it becomes clear that raising the common mode impedance lowers the recevers sensitivity to cable imbalance.
Now sometimes (Particularly if running between buildings) you may need transformers to remove excessively large common mode voltages (Or potentially large common mode voltages in the event of a distribution network fault), but usually within a building it just works.
Screening is actually far less important then twisting, at least as far as low to medium frequency pickup is concerned.
On the receving end the important figure of merit is the common mode impedance, as this determines the systems sensitivity to minor imbalances in the cable and source output impedance.
The way to think of a differential signalling system is as a wheatstone bridge with the noise being the common mode driving voltage and the signal being introduced by a voltage source in one leg. Seen in these terms it becomes clear that raising the common mode impedance lowers the recevers sensitivity to cable imbalance.
Now sometimes (Particularly if running between buildings) you may need transformers to remove excessively large common mode voltages (Or potentially large common mode voltages in the event of a distribution network fault), but usually within a building it just works.
Stephen Lampen of Belden Cable has written two books of interest to us,
"Wire, Cable, and Fiber Optics for Video & Audio Engineers"
&
"Audio/Video Cable Installer's Pocket Guide"
He also has some 70 blog pages and he does presentations at audio conventions.
Somewhere in all the above, he talk's about how each pair in a Cat5/6 cable has a different twist ratio. But at intervals all the pairs swap ratios.
"Wire, Cable, and Fiber Optics for Video & Audio Engineers"
&
"Audio/Video Cable Installer's Pocket Guide"
He also has some 70 blog pages and he does presentations at audio conventions.
Somewhere in all the above, he talk's about how each pair in a Cat5/6 cable has a different twist ratio. But at intervals all the pairs swap ratios.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.