Balanced or Single Ended (specific case)

[DiminishingReturnsOfficer]

Hello, I was hoping to attract one of the gurus here and intise him/her into offering me some insight into what I know is a belabored topic. I'm not speaking here in general terms, and I'm aware that the answer, as to which is best, depends on the type of circuitry employed inside the components. In other words; if a componet is not balanced internally, there is little to no advantage (and possibly a disadvantage) to using XLR cables.

I have a Bel Canto DAC3, and will soon have a Bel Canto Pre1. The manual for the DAC strongly recommends using the balanced outputs because of a small amount of DC offset on the single ended out's. The Pre1 internally converts all signals to single ended. They are only truly balanced at the I/O. Would I be able to get rid of the DC offest by using the DAC3's XLR output even though the preamp is not balanced internally? Unfortunately, I have not had much luck getting Bel Canto to reply to my e-mail. I've left several messages on their phones but haven't gotten a single reply. I don't mean to bad mouth Bel Canto, it is possible that I'm not "going through proper channels" (?)

 

[barrows]

balanced to the preamp. This is fairly technical, and it really should be answered directly by bel canto, as ultimately whether or not balanced will be an advantage will depend on how bel canto's PRE 1 converts balanced signal to single ended. The DAC is truly balanced, and you will get the advantage of common mode noise reduction (and less noise pickup along the interconnect) as long as bel canto actually converts the balanced signal to single ended in the PRE 1 (as opposed to just taking the positive half of the balanced signal and shunting the negative half to ground).

I would wait to decide until hearing from bel canto's tech support.

 

[DiminishingReturnsOfficer]

You were actually one of the "gurus" that I had in mind. You've given me solid advise elsewhere on CA and I appreciate it. I've got an e-mail in to Matthew at Bel Canto's tech support and I will await hearing from him. The only information I have regarding the actual conversion to single ended is they say that its done in order to preserve a lower noise floor. I'm hoping there's a good quality Balun in there...

 

[PeterSt]

I'm afraid I have nothing much useful to add, but merely wonder ...

 

I think you (the manual) said that the DC offset is at the SE outs.

 

Would I be able to get rid of the DC offest by using the DAC3's XLR output

 

This isn't "getting rid of" as such, but just avoiding it. Well, if true at all, because I think the story is strange (hey, better use balanced out because our SE sucks).

 

Balanced (differential) as a technical means won't get rid of DC as such. It merely will *create* it, when not done right. So, on this part I don't see the relation with solving anything (no matter how much you may tend to read it like that). It's just the avoiding of a poor other means (the SE outputs).

 

Now the other way around : this "poor other means" as I just mentioned it, is not poor within itself at all. That is, again related to the DC it isn't. However, to me this smells like a poor base design which ... can be solved by means of differential appliances (which "balanced" is). So, now think of additional circuitry which nicely fits in the differential circuitry part, where the former explicitly attacks DC offset.

 

It becomes more complicated when we (ok, I) denote (imply) the whole DAC as poor, only because of some DC offset, of which we don't even know how much it is. The DAC is not poor at all, period.

So, having said this all ...

 

I would use the SE outputs, just to avoid that circuitry.

Also, it seems to me that the DAC is not inherently setup differentially. So, you'd be a kind of "converting". Not a vey big deal, but everything matters.

 

Otoh, balanced cabling sure matters also for the ever so short run. But now it depends on your evironment (pickup etc.).

 

Ok, sorry for the brainstorm (I didn't know the answer myself when I started typing), but the conclusions could be these :

 

- Balanced cables are always better.

- Creating balanced output when it is inherently not there will degenerate a little (usually a row of opamps).

- DC Offset isn't that bad, if it's only a little. I always say 20mV, but I guess 50mV is still fine.

 

Notice that DC offset is also about the distance when talking about two channels. So, one channel -15 and the other +20 totals to 35 for phase anomalies.

One channel -30 and the other also -30 implies no phase anomalies.

 

The DAC is truly balanced

 

This is not what I got from it, but if that is the case I guess it is not all that difficult to re-interpret the above.

But then I really don't see how SE would have DC while balanced has not.

 

Anyway, I hope this helped a litte.

Peter

 

[barrows]

I am 99% sure that all three of the bel canto DACs are true balanced designs from the output of the DAC chip to XLR connectors.

 

[DiminishingReturnsOfficer]

Peter, I do thank you for your comments, however, I'm not sure quite what to make of them. First of all, when I said that I wanted to "get rid" of the DC offset; this is exactly what I meant.

"This isn't "getting rid of" as such, but just avoiding it"

My understanding is that a balanced line has 2 legs with polarity inverted 180*, and a ground established at the mid point. Any signal component having the same polarity on both legs (such as DC) would be canceled through common mode rejection. "Canceling" something is the same as "getting rid of" it for my purposes.

I have 2 excerpts from the manuals provided by Bel Canto:

"We strongly recommend that you use the Balanced XLR output to drive your e.One amplifier. Because of the DC coupled nature of the DAC3 there is a small residual DC offset on the outputs. This is typically less than 10mV (0.01 Volts)."

This is for the DAC3, and the following is for the Pre1:

"The Pre1 output stage uses the latest integrated

technology to combine multiple amplifiers and

Laser trimmed precision resistors in a true

balanced output stage. This balanced stage

mimics the operation of a transformer, providing

ideal line-level audio drive into either a balanced

or single-ended load. This operation eliminates

ground loop issues with power amplifiers when

used with either the RCA outputs or the XLR.

This assures maximum signal to noise and

extremely low distortion with the shortest audio

signal path."

I am not using an e.One amplifier, and the comments above regarding the preamp are all relative to the OUTPUT. They don't say anything about the input section, so I am assuming that they are handled in the same manner. If the input stage also mimics a transformer, I belive that the (0.01V) DC offset will be canceled. The DAC is internally balanced throughout, but I read a review that said balanced inputs are converted to single ended. As Barrows pointed out; if they convert by merely dropping the inverse phase leg, there would be no common mode rejection. I have called Bel Canto and left messages and I've got an e-mail in to Matthew, their Tech. Services guy, but I have yet to hear back from them.

 

[PeterSt]

... for the outlay. But I didn't want to anticipate on that because it looks negative. But I guess now I will ...

 

Your latter alinea will be true within itself, so no comment on that (and really no idea what to do with it).

 

But about "cancelling out" as such ... A balanced connection cancelles out common mode noise, but no DC. As I said earlier : it will merely create it (because of inaccurate resistors and things; also ref. the laser trimmed resistors from the Pre1 story :-).

 

The misconception (or what about mis guidance) is this one :

 

This balanced stage mimics the operation of a transformer

 

Although this comes from the Pre1 (remember, your quote - I just don't know), this is just marketing stuff, told to someone who's supposed to (half !) know. A (double !)transformer indeed will create a ground midpoint. But (mimicing) electronics do not at all. As said, they merely create the problem because of not well balanced resistance, soldering joints and the such.

 

I assumy you take the sory from the Pre1 to project on the DAC3; I wouldn't do that (unless you really know it).

 

What remains is what I said before : that 10mV on the SE outputs really isn't bad. Just use it, unless you think balanced works out for the better for other reasons (I can't judge).

 

But what also remains is why SE would have DC offset while balanced (in a true differential setup !!) has not. This doesn't fit.

 

Please notice : my ideas about this.

Peter

 

 

 

 

[Panelhead]

I never used balanced connections even when available. Could never see where it would help with short cables.

Most gear is single ended inside anyway. But after actually trying balanced it has sounded better in every case. Even with an opamp is used to do the balanced to single ended conversion.

I would not have believed it was so. Just shows it pays to try all the options. A great cable is the Sommer Carbokab. It is expensive for a pro cable. But reasonable for audio cable. Sounds better than Canare, Sommer, and Mogami.

 

 

George

 

[1audio]

I don't see why a DC offset that is present at the unbalanced output is not present at the balanced output. An offset in the output comes either from an issue with the DAC chips, an imbalance and leakage issue with the opamps that constitute the balanced output and then the balanced to unbalanced output conversion (another opamp).

 

How much offset is causing this concern? Balanced cables are no special magic and differential signals are just as likely to have problems. The benefit is higher net level (usually by 6 dB) and some reduction of common mode noise pickup. Nothing about balanced will remove DC.

 

[DiminishingReturnsOfficer]

Ah, another of the "gurus" that I was hoping to hear from. I'm glad to hear from you, thanks. Weren't you also with Spectral for some time? Anyway, here's a quote from Bel Canto regarding the amount and reason for the DC offset, "We strongly recommend that you use the Balanced XLR output to drive your e.One amplifier. Because of the DC coupled nature of the DAC3 there is a small residual DC offset on the outputs. This is typically less than 10mV (0.01 Volts)." I'm assuming that the DC is bleeding over from the power supply. I don't know for certain that the DC is not present on the balanced outputs or not; what I had (erroneously) assumed is that it was there, but would cancel on the ground. I was under the impression that noise was either common mode or transverse mode and I assumed (again, erroneously) that DC would be common mode which would cancel (sum to zero). From what I'm hearing here, DC is not common mode and will not be canceled out. I would like to mention that I'm quoting directly from the Owner's Manual for the DAC3. They are talking about using the DAC as a preamp to drive one of their eOne amps directly. I am just guessing that the DC would be there regardless of what component its hooked to.

 

[PeterSt]

what I had (erroneously) assumed is that it was there, but would cancel on the ground.

 

FYI : DC Offset can't be cancelled (out) with anything. It can only be corrected by means of "processing" once it is there. Of course the real solution is not to have it in in the first place.

 

For your fun (or misery), you can also wonder what the DC Offset is between the two signal pins on a balanced output. So, here again, this is - or works the opposite of what you might have thought in the beginning; here too DC Offset can exist and this is a BAD one. Why ? because the DC difference on both signals (pins) will cancel out MUSIC now. Ok, the signal itself. More nasty things will happen, because a differential setup is a self correcting one. But now it will be correcting the signal, and there go your dynamics. On one channel only perhaps.

 

There is so much more on this little subject alone ...

 

Regards,

Peter

 

[PeterSt]

This is just for people to think over. To think over how bad our audio actually is, and it is so because nobody knows what's going on, let alone how to solve it.

 

Go Google. Google for DC Offset. Try to learn a few things about it. You won't. But at this attempt of learning a few things, there's one thing you will notice - and will be shocked about, if it's only looked at for its proper merits ...

 

You will automatically run into the thousands of forums/threads about how the recording industry deals with it. To this regard you will find *NO* single real solution to the problem. What problem ? well, that there's DC Offset in the recordings. The exact same DC Offset we are all experiencing through our equipment without realizing it (but you can measure it easily). So, the recording industry easily shows that the problem is there always, but with the notice that the larger setups (like a 100 devices interconnected) incur for it somewhat more.

 

All the Digital Audio Workstation software contain DC Offset correction functions. So, this is how the DC Offset is removed from the recording itself.

 

Please notice : without further context many will claim "BS" on this subject, because many audio processing functions create DC Offset just the same (could be a limiter of some kind). So, the software does, and let's assume this is unavoidable (what about destroying music ...). But during your Googling, you will also run into expressions like "but I didn't do a thing ! and still I have DC Offset !!". That's what I'm talking about.

 

The recording engineers know. It takes out unnecessary headroom. So they correct it software wise (an easy and fairly (to sound) harmless job).

 

Nobody is attacking the root cause ... It's all taken for granted because of not understanding.

 

Peter

 

[1audio]

The DC offset is a pretty normal situation. Its the accumulated offset errors from the DAC chip and the opamps. Getting the offset very low (less than 1 mV) usually requires adjustments, additional circuitry or coupling caps. I still don't see why the balanced output is different, however it may be that the balanced input of the amp is AC coupled or has a servo (electronic equivalent of a cap).

 

[DiminishingReturnsOfficer]

I would be remiss not to mention that I heard back from Bel Canto with regard to my question; could I get rid of the DC offset by using balanced cables? "You could and can, but the extra circuitry in the PRE1 that sums the differential (XLR) information to single ended (RCA) may actually sound worse than just using the single ended RCA outputs from the DAC3. Best advice is to try both and see what you prefer." OK, so along with contributions from many of you, I think I have my answer.

 

[barrows]

you received a response, Matt is good guy. Sounds like the pre 1 uses an opamp to sum balanced to single ended, so their advice is correct, both balanced cable to preamp, and single ended cable to preamp will have a downside and an upside.

Balanced: downside, and extra active component in the signal path. upside, common mode rejection which could result in a lower noise floor (this will be system dependent).

Single Ended: downside, might result in a higher noise floor. upside, one less active component in the signal path.

I agree that you will need to try both to determine which will be best in your system (always good advice, BTW, kudos to bel canto for telling the truth).

 

[Miska]

I don't see why a DC offset that is present at the unbalanced output is not present at the balanced output.

 

It may be output offset of the op-amp used to convert differential signal to single-ended. Some op-amps have higher and some have lower offset errors, single ones usually have pins for tuning pot for the offset (NE5534 as an example).

 

Another possibility is that there's common (same side) offset on both differential rails and single-ended output is done by just taking one of the rails against ground, this would propagate DC offset through the single-ended while removing it for the balanced.

 

Those offsets could happen due to virtual grounds.

 

Nothing about balanced will remove DC.

 

If both rails have same offset, on the same side it will cancel out when converted to single-ended, since the output is only the difference between the differential rails and the common part (common DC) is removed.

 

(for example both have +10mV -> +10 - +10 = 0 mV, if there's +10mV and -10mV -> +10 - -10 = 20 mV, if -10 mV and +10 mV -> -20 mV)

 

Naturally this should also disappear on the internal conversion as well.

 

 

[DiminishingReturnsOfficer]

Thanks to Miska and again to Barrows, not to forget Matthew of Bel Canto. It appears that I have my answer. Now, I'm going to have to get a pair of balanced interconnects. Money is tight right now, so I probably can't afford to use the same cable as the RCA's(Cardas Cross). My intention was to get a pair of XLR's from Blue Jeans Cable, and see how the balanced connection sounds. Has anyone heard the BJC XLR's? For about $65/pr., they're about the same price as 2 pairs of XLR caps which I like to use for aesthetics and possibly some noise blocking. I realize that the comaprison will also take into account differences in the cables, but it should uncover any startling disparity. I'll report back in a few weeks.

 

[Panelhead]

But Lavacable will make whatever you want. The Sommer Carbokab are the best balanced cable I have found. They will terminate with any connections you require.

 

George

 

[DiminishingReturnsOfficer]

I decided to take Panelhead's advise and get a pair of the Sommer Carbokab XLR cables. They seem to be a good pro-audio / industrial microphone cable with low capacitance, excellent shielding, and fairly low resistance. I contacted Mark at Lava Cable and he's working on a 4' pair for me. The single ended cables are Cardas Cross; a good audiophile cable, but nothing super exotic. I'm hoping that the Carbokabs will sound similar enough in comparison that I'm not simply doing a A/B comparison of cables. My intention is to compare the balanced outputs of a true balanced DAC (Bel Canto DAC3) to the single ended outputs of the same DAC. As mentioned earlier, the preamp (Bel Canto Pre1) accepts balanced inputs, but converts them to single ended. This is in no way a scientific experiment because the results will be subjective (to my ears), the cables are not from the same manufacturer (or even the same intended use), and the preamp isn't internally balanced. I just want to see if I can hear a difference. I would end up spending $30 ea. for a pair of Cardas XLR caps to block off the jacks anyway, so for $40 more I've got something to compare. I am excited about the Carbokabs. They look like a well designed cable that makes use of good quality materials. This cable is more than twice as expensive as the Belden cable used in the Blue Jeans Cable balanced cables. Its been my experience as an electrician that; when you're talking about a 2x price difference in an industrial product, this usually translates to a significant quality difference. I'll make one final entry when I get the cables, run them in a bit, and have some time to compare.

 

[DiminishingReturnsOfficer]

The balanced cables arrived today. First let me say that the Carbokabs are truly a first rate interconnect. They don't seem to require much break in, or if they do and it gets much better than this I'm going to have some interconnects for sale on Audiogon real soon! I don't know if Lava makes single ended IC's or not, but I can strongly recommend the Carbokab to anyone looking for an inexpensive interconnect that doesn't sound inexpensive.

In comparing the balanced output to the single ended on my DAC3, I feel it necessary to mention that I'm using it with a Bel Canto Pre1 and the output to my amplifier (CJ MF2250) is single ended. From what I found out through Matthew at Bel Canto, the Pre1 converts all input signals to single ended even if you're using the balanced outputs. There was some question raised as to how this conversion was done within the Pre1. One possibility was that it just dropped one leg of the balanced signal. Another is that there would be either a transformer or some sort of balanced circuitry to accomplish this. Matthew indicated that this was accomplished with additional circuitry. The volume on the balanced input is about 5dB higher than the single ended which would indicate to me that its a stronger signal as would be typical of a balanced circuit. Due to the difference in volume, the first thing I had to do is balance the levels. I have no equipment to work with other than my ears which are fairly sensitive. I determined that nearly equal levels can be obtained by setting the volume 5dB higher on the single ended input. Each half step on the Pre1 is supposed to equal 0.5dB, so I'm saying 5dB.

I'm notcing a definite preference for the balanced input: even with brand new cables it sounds fuller with more depth and more "air" around individual instruments. Vocals are rendered very effectively and convincingly. The single ended input by way of Cardas Cross RCA interconnects sounds slightly more forward and a bit colder.

Matthew of Bel Canto warned that while using the XLR input on the Pre1 might sound better due to the fact that the DAC3 is internally balanced, the extra circuitry in the Pre1 might negate it's effect. There was also the issue of DC offset on the single ended output of the DAC. My original thinking was that I could sidestep that by using the balanced outs. I've had several responses challenging this notion asking why the DC offset would be present on the RCA's and not the XLR's. My only response to this the quote from the owner's manual that says it to be so. It says "due to the DC coupled nature of the DAC3", so I'm saying the reason for it is that the DAC is DC coupled. Originally, I was under the impression that DC offset could be cancelled by common mode rejection inherent in a balanced circuit. I learned that this is not the case, and I don't really know what role the DC offset plays in the equation. In fact, all I know for sure is that #1) I prefer the balanced signal, and #2) The Sommer Carbokab as made by Lava Cable is one fine cable; worth every penny of its asking price. Thanks to all who have participated.