1. Register NOW and become part of this fantastic knowledge base forum! This message will go away once you have registered.

Unbalanced output to balanced inputs

Discussion in 'Accessories / Connections' started by achase4u, Dec 30, 2014.

  1. achase4u

    achase4u Active Member

    I am currently trying to read up on this subject. My audio interface accepts either balanced or unbalanced 1/4" jacks into its inputs, and out of its outputs.

    I am going to a remote session where they are using a bass preamp that will generate line level signals out of two unbalanced 1/4" outputs, for stereo operation. I could simply use TS cables and be fine, however - I am curious if using an adapter to take the cable from TS output to TRS input would help the noise issue. Or is that only necessary if my inputs were balanced only?

    I have read that you shouldn't use a balanced cable from an unbalanced output to a balanced input, in Mike Seniors "Recording Secrets" book.
     
  2. Boswell

    Boswell Moderator Distinguished Member

    There are several points here.

    Most balanced inputs are actually what are technically differential inputs, that is, they amplify the arithmetic difference between the + and the - signals that are applied to them, whatever those two signals happen to be (within a specified range). Balanced is just a special case of this where the - input signal is always equal and opposite to the + signal. This means that a standard audio input described as "balanced" will perfectly happily accept an unbalanced signal, as long as whichever input terminal that is not carrying the unbalanced signal is connected to a clean ground. If possible, this ground should be that of the source rather than a ground at the input connector so that the interference rejection properties are still active (see below).

    When it comes to outputs, you should not, in general, use a TS cable to connect a balanced output to an unbalanced input. This is because the - output will be grounded by the sleeve of the TS plug inserted into the TRS output jack socket. The upshot will be that the output stage will be trying to drive the full -ve signal into a short circuit, with the certain result that this will cause large currents to flow in the output drivers, and these could affect the quality of the +ve signal. There is also the possibly of damage to the output stage if it is not designed to cope with a permanent short circuit to ground. If you are wanting to drive an unbalanced input from a balanced output, you should take only the +ve output signal plus ground, leaving the -ve output signal unconnected. This will result in only half the output voltage (a 6dB signal loss), but no stressful circulating currents. There are two exceptions to this rule that apply to higher-end gear: floating transformer output drives and electronically balanced outputs. In these cases, if one output line is grounded, no excess currents flow and the full output voltage is available on the other line.

    The final relevant point is to do with interference rejection. The reason for balanced connections is not only to achieve maximum signal amplitude from a given set of power rails, but also to give rejection of interference. This latter property arises from interference generally affecting the + and - conductors of a balanced feed equally, hence not registering in the voltage difference between the + and - signals. The differential receiver thus rejects (attenuates) the interference while preserving the signal. If you need to feed a balanced input from an unbalanced output, ideally you should use a balanced (TRS) cable, making sure that the -ve signal (ring contact) is grounded at the sending end. This preserves the rejection characteristics of the connection. However, unbalanced output jacks are often not fitted with a ring contact connected to local ground, so the local grounding of the ring contact of a TRS plug is not guaranteed.

    I've probably confused you with all that. To boil it down to the case of your bass amp, because these are line-level signals (presumably) running over a short distance, I would use TS cables. Your interface will be happy with these and it avoids the problem of whether the jacks at the sending end have grounded ring contacts.

    I'm intrigued by a bass pre-amp generating a stereo output - does it have its own built-in effects?
     
  3. Kapt.Krunch

    Kapt.Krunch Well-Known Member

    Last thing first:

    Using a balanced TRS cable from an unbalanced output to a balanced input will cut off half the signal. The balanced input wants to see tip AND ring (+) signals, and sleeve (-) ground. Since the unbalanced end at the preamp output will only be connecting the tip and sleeve, the ring will be floating in the jack unconnected to anything. That means the ring at the balanced end won't get its signal, which is supposed to be 180 degrees opposite phase, and then the circuitry in the unit flips it back into phase, and sums them both together, effectively (kinda) doubling its signal level. The signals run out of phase with each other through the cable to help reject external electrical noise sources. (But...you may already know that!) So, you'll get signal...but it'll be even worse than a straight TS - TS unbalanced connection.

    If it's a relatively short run, an unbalanced TS cable would be fine. But, yeah, a balanced connection would help signal level and noise, especially over a longer cable run.

    Here is something to check out. http://www.rane.com/note110.html

    A transformer-coupled device that converts balanced to unbalanced works fairly well...and as always, how well depends on the quality of the unit. Those Radio Shack ones that are built into a cylinder work...but...are they less noisy over a long run than a simple unbalanced short run would be? The signal needs to be balanced directly after the output of an unbalanced out. To balance it, the phase needs inverted. Can't really do that without some kind of circuitry.

    Lastly..."bass"..."stereo"?

    Why? Are you sure that preamp has two line outs that are actually "stereo"? Does the preamp have effects built in? What preamp is it? Are you sure it's not just a pair of paralleled outputs, or some other usage? Bass is not NORMALLY run in stereo...but it's possible. Just wondering. If it's not actually "stereo", there is no point of connecting both. You'll just have two tracks recorded of the same exact thing, which just eats up recording processor power, hard drive space and makes the hard drive work harder. If it actually IS stereo, and you have a reason to use it for effect...then...OK.

    Kapt.Krunch
     
  4. paulears

    paulears Well-Known Member

    You can, of course, connect the unbalanced circuit to pins 2 and 3 and leave ground unconnected - this works nicely. Even though you do lose half the signal when you short 2 or 3 to 1, it's not normally an issue. Using the pins 2 and 3 version still provided rejection for common mode noise, as it's on both conductors, but this kind of noise is never as common as we think - most noise comes in the form of ground loops with a few volts of AC running through the ground - so losing the ground is useful.
     
  5. achase4u

    achase4u Active Member

    Thank you all for responding. I do understand the cable structure as I make my own with Mogami quad and Neutrik EMC's - it's the input and output hardware itself and how it functions together that eludes me, and you are shedding light on this for me now.

    So - just as a simple overview - a TS cable carries the signal via a conductor, while the other part of the connection is a ground/shield/signal reference?

    Balanced being two conductors that carry an identical mono signal, with a shield as well. One signal side is 180 phase flipped along the path, then returned to "in phase" operation at the input. The reason being, noise which is collected along the cables length will be "in phase" during it's passage through, thus cancelled out when the phase is flipped at the input, leaving the signal and removing the noise.

    However, I'm not sure what a differential system entails.

    "Most balanced inputs are actually what are technically differential inputs, that is, they amplify the arithmetic difference between the + and the - signals that are applied to them, whatever those two signals happen to be (within a specified range)"

    Can you define what part of a balanced cable is the + and -? Is the + the sum of pins 2 and 3(signals) and the difference between that and ground/shield are the differential? This would make sense - because there technically should be no difference between the pin 2 and 3 signals, right?

    "This means that a standard audio input described as "balanced" will perfectly happily accept an unbalanced signal, as long as whichever input terminal that is not carrying the unbalanced signal is connected to a clean ground. If possible, this ground should be that of the source rather than a ground at the input connector so that the interference rejection properties are still active (see below)."

    I imagine it is difficult to tell if the jack itself is connected to a clean ground for the unbalanced signal? I figure if the manual of my Octa Capture says that XLR, Balanced 1/4" and unbalanced 1/4" are acceptable, then it must.

    "When it comes to outputs, you should not, in general, use a TS cable to connect a balanced output to an unbalanced input. This is because the - output will be grounded by the sleeve of the TS plug inserted into the TRS output jack socket. The upshot will be that the output stage will be trying to drive the full -ve signal into a short circuit, with the certain result that this will cause large currents to flow in the output drivers, and these could affect the quality of the +ve signal. There is also the possibly of damage to the output stage if it is not designed to cope with a permanent short circuit to ground. If you are wanting to drive an unbalanced input from a balanced output, you should take only the +ve output signal plus ground, leaving the -ve output signal unconnected. This will result in only half the output voltage (a 6dB signal loss), but no stressful circulating currents. There are two exceptions to this rule that apply to higher-end gear: floating transformer output drives and electronically balanced outputs. In these cases, if one output line is grounded, no excess currents flow and the full output voltage is available on the other line."

    So if I understand correctly, in the case of a balanced output and unbalanced input, you CAN use a TS cable, because one half of the connector is ground(shield) and there is only one other conductor, being the signal/tip. Tip is +ve yes? Or do you mean it is necessary to use a TRS or XLR connection at the output stage because a regular TS jack would ground the -ve signal and cause the current load? That would make sense. So you would need a balanced cable coming from the balanced output, and to snip the -ve(pin 3 XLR or ring, correct?) AFTER the output so that it isn't short circuiting? The current simply flows through the connector and "floats" by not connecting to the other jack at the unbalanced side?


    Ok - I think we are talking a 20 foot run here of line level TS to my inputs.

    Yes indeed - stereo bass had me scratching my head briefly, but it has many effects like chorus etc. The music is a bit punk, a bit Radiohead - so bass FX are possible.

    It's a Behringer Bass V Amp - it has 2 XLR balanced outputs, I have discovered now that I found a manual online.(The gear belongs to the client away from here) I may get long runs of XLR female to TRS to feed my preamp. This would solve the issue. However, I won't have to build the cables and shell out the money if I simply use TS cables from the 1/4" unbalanced outputs.

    "A transformer-coupled device that converts balanced to unbalanced works fairly well...and as always, how well depends on the quality of the unit. Those Radio Shack ones that are built into a cylinder work...but...are they less noisy over a long run than a simple unbalanced short run would be? The signal needs to be balanced directly after the output of an unbalanced out. To balance it, the phase needs inverted. Can't really do that without some kind of circuitry.

    Lastly..."bass"..."stereo"?

    Why? Are you sure that preamp has two line outs that are actually "stereo"? Does the preamp have effects built in? What preamp is it? Are you sure it's not just a pair of paralleled outputs, or some other usage? Bass is not NORMALLY run in stereo...but it's possible. Just wondering. If it's not actually "stereo", there is no point of connecting both. You'll just have two tracks recorded of the same exact thing, which just eats up recording processor power, hard drive space and makes the hard drive work harder. If it actually IS stereo, and you have a reason to use it for effect...then...OK."


    Like the ebtech units? I may have to invest in one of those for such occasions. Thank you. I seem to recall that these units only work one way, meaning that you can connect a balanced output to it, and out comes unbalanced for the times you have an unbalanced input(I don't own any - isn't this rare?)

    But what happens when you want to balance a signal from an unbalanced output? Are there transformer devices for this? AKA an Ebtech hum reducer?

    I only record bass in mono myself, however, this unit has L/R outs both balanced and unbalanced. There are effects galore on it like Chorus etc that are panned. Pretty wild!

    Paulears - Do you have any advice on ground loop eliminating devices? At home, I am on one outlet at least.
     
  6. paulears

    paulears Well-Known Member

    When talking about balanced circuits, it's easy to think about things in terms of polarity, or 180 degrees worth of phase flip, but this isn't really how it works - if you look at any waveform, it's not simply a mirror image, and phase of course, is a time function - while polarity is a different thing. The phase buttons on mixers for years carried the phase title and symbol, but were just polarity reversers!

    We gliby promote balanced systems as having noise reduction properties, and telephones for years ran on unshielded cables, that were balanced, referenced to something else, and the something usually being ground. It's difficult to detail all this in a forum, because the full story is very complex, but easily googleable. The usual explanation, which is the mass market simplified one, is that any interference hits both conductors equally, and because the device they are driving is simply quantifying how far the signal pulls away from 0v (the differential tag often used) any input applied to both cancels out and that is that. Two conductors that differ in capture area won't cancel completely, so the noise will be added.

    In practice, all this noise reduction is only important at low level inputs - microphones typically. The interference vs the wanted signal. In a line level signal case, the ratio between any noise and the wanted signal is much, much greater, and the importance of balanced connections is much less.

    On the ground loop issue, it's critical to understand that not every hum is a ground loop. If we are using the screen of a cable to be a protection from external interference, then it needs to be at the same potential as ground. Depending on the type of power delivered to your premises, your actual ground can be the real ground potential at your location, by a metal rod being buried in the actual ground, and the electrical ground is the same potential as the 'real' ground. However, some types of supply bond the neutral conductor to this real ground where the power enters the building, while others supply the neutral and the live as totally floating sources. If your electrical system is grounded a fair way away, the it's possible for the neutral to be a few volts different from the ground in the outlet. Equally, something in the room can also be connected to the real ground, so you can have multiple versions of 'ground'. This is quite normal.

    The first problem we have is when items are grounded at different places, and then audio cables connect pieces of kit together. At some point, the mains ground at one potential gets connected to a ground of a different potential by the building ground wires, AND the cable screens. This difference circulates in the two paths in a loop, and because there is voltage difference, there will also be current flowing, and because of the mains frequency, it's very easy to generate significant audio levels at 60Hz (or 50Hz). Breaking the loop solves the problem. In a balanced circuit, it's simple - just lift a screen, the loop is broken, current flow stops and the hum vanishes. The screening function still works fine, because one end is still attached to a ground. Lifting the mains ground does the same job at one of the pieces of kit - BUT, especially in the UK with our higher 230V mains power, this is considered both stupid and legally problematic. We used to do it, but now, listen ground wires mean a test fail - as most places now regularly test appliances. In XLR circuits, a simple male-female barely connector with no Pin 1 connection often cures ground loop hum. Worth always trying, as are the switches on things like DI boxes and transformer splits - they do the same thing.

    The usual advice to run everything from a single local outlet simply makes sure that the ground potential is that same on all devices connected.

    One very brief warning about another hum inducing problem. With mains powered equipment that is properly designed the mains ground will also be attached, for sensible safety reasons to the chassis and case, if it has a 3 conductor mains cable. In a two conductor cable, often the figure 8 shaped ones, the metal work is all floating and not connected to ground at all - in the UK, we term it double insulated. If kit has 3 pin XLRs for balanced connection, it's considered bad form to connect Pin 1 to the chassis or external metalwork because if you screw it into a rack, all the metal is joined into one big unit, so any ground currents circulating away can easily be induced to flow in the cable grounds. Most kit therefore has floating electrical/safety ground and signal ground - it just takes one bit of kit, or even a cable where the pin1 connection is joined to the metalwork to start the hum problem - and very often, the bit of kit causing it doesn't actually hum - it's a more sensitive device elsewhere that suffers.

    To sum up - I personally don't really bother with balanced circuits for line level as a preference. If both bits of kit have XLRs, then I'll use an XLR-XLR to connect them up simply for convenience. Making up a balanced to unbalanced cable is a pain, so if both bits of kit will run unbalanced, I'll take that choice. If we're talking about microphones, then everything balanced, and I even have a few Star Quad cables (Google them for the explanation) which I use if I know interference is likely. In practice, they offer some extra protection from outside noise, but realistically, in most cases, normal cables are just as good - Star Quad is just extra protection.

    This is again, pretty simplified in many areas, but it's really not something to worry about. Running a pile of cables through a tester revealed quite a few that were lacking a Pin 1, or had 1 connected to 2, and were still in service - working fine.

    Balanced operation simply needs a circuit - two conductors. Grounding is nice, but not an absolute requirement - try it and see if you want to experiment. I have a few old loudspeaker cables from the 80s where they used 3 pin XLRs, wired with 3 core mains cable, unscreened. I've accidentally used them many times for mics - and noticed, because they can let noise and hum in, but at line level, between mixer and amplifier they work fine!
     
  7. Kapt.Krunch

    Kapt.Krunch Well-Known Member

    Some good info!

    Also, just to answer one of your questions:

    Those work both ways. They are passive, basically transformer-coupled. It doesn't matter which way you use them, except one thing:

    Going from -10 to +4 will result in a slightly small loss in signal strength. Going from +4 to -10 won't. The small loss can easlily be made up at the source output, or, possibly, the destination input, or both.

    Kapt.Krunch
     
  8. Boswell

    Boswell Moderator Distinguished Member

    There's a danger of over-thinking things here. I would try simple lead interconnects before investing in isolating transformers or other methods of connection.

    What I suggest you do is carry out a simple test using a multimeter on the bass amp with its power switched off. Plug a TRS-TRS jack lead into one of its outputs and at the other end of the lead measure the resistance on the plug between the ring contact and the sleeve (ground). If this is either near zero or is a few hundred Ohms, then this indicates that the bass amp output jack is fitted with a ring contact and it is connected to ground, maybe through a low-value resistor to balance up the impedances. It means a TRS lead is OK to use between the bass amp and your interface. If, on the other hand, you see an open circuit or a resistance greater than a couple of KOhm, then it's likely that the output jack does not have a ring contact and you should use TS leads for the interconnection. Twist the plug around in the jack socket a little while making the measurements to check that the contacts are solid and not momentary.
     
  9. achase4u

    achase4u Active Member

    I really appreciate all this info, guys. I am trying to soak it all up. It's allot to take in for a fella who has very little electronics background. Some of the terms lose me.

    In a slightly ironic twist of fate, I think I will use the SPDIF out on the bass preamp, to avoid having to do A/D twice on the way in. However, there is no info on the units bit depth internally - I know I can simply set to 44.1khz to match my daws sample rate, but I suppose it will just record it at 24 bit wav depth?
     
  10. Boswell

    Boswell Moderator Distinguished Member

    The wordlength of S/PDIF data is inherently 24 bits, but it's up to the sending end how many of these bits contain real data. If you connect an S/PDIF data stream to the Octa-Capture coax (RCA) input, the sampling clock for all the Octa-Capture input and output is switched to be taken from the S/PDIF incoming data. You have no choice over what rate your DAW uses.

    As an important result of this, it means that the quality of all the analog input and output sampling in your system would be determined by the quality of the clock in your bass pre-amp. You might want to think carefully about that before electing to use the S/PDIF input.
     
  11. achase4u

    achase4u Active Member

    Ok - I understand. I would be ok with the 44.1khz rate and 24 bit. However, I'm not sure I trust the clock. I will probably settle for doing A/D/A/D in that path. Thanks!
     

Share This Page