I know this has been discussed some, but I was thinking starting a thread with the desire to have a comprehensive discussion in one place could be a good idea.
I am about to start a new BIIIse/Legato 3 build. It appears there are a few possible ways to deal with the offset from +/- to ground on the Legato3
1. Do not worry about it and hope the amplifier is compatible.
2. Add caps and eliminate the problem entirely.
3. Do not connect the signal ground to anything, and hope the amplifier is compatible, floating.
4. Do not connect the signal ground to anything, and transformer couple the output.
I have heard some suggest that #1 will only work in some cases, and that it may not provide the best sonics (why?).
We know why we would like to avoid #2 if possible.
I have read some suggestions that #3 can work in some cases, but may be an undesirable solution as there is a sonic advantage to a signal ground connection between components even in the case of true differential circuits.
Solution 4 is relatively expensive, the Legato's medium output impedance may not drive a transformer ideally, and some feel that the transformer itself will color the signal (although I suspect this would not be a problem with a really good Jensen or Lundahl properly implemented).
I am interested in hearing of others' experiences and opinions on this topic. Currently, I am using a cap (Mundorf Supreme) coupled Legato2, but I would like to eliminate caps in my new Legato3 build, without other compromises if possible. I have a few different amplifiers: Pass Labs X150.5, PS Audio Classic 250, and nCore 400 based, and I only run balanced.
I am about to start a new BIIIse/Legato 3 build. It appears there are a few possible ways to deal with the offset from +/- to ground on the Legato3
1. Do not worry about it and hope the amplifier is compatible.
2. Add caps and eliminate the problem entirely.
3. Do not connect the signal ground to anything, and hope the amplifier is compatible, floating.
4. Do not connect the signal ground to anything, and transformer couple the output.
I have heard some suggest that #1 will only work in some cases, and that it may not provide the best sonics (why?).
We know why we would like to avoid #2 if possible.
I have read some suggestions that #3 can work in some cases, but may be an undesirable solution as there is a sonic advantage to a signal ground connection between components even in the case of true differential circuits.
Solution 4 is relatively expensive, the Legato's medium output impedance may not drive a transformer ideally, and some feel that the transformer itself will color the signal (although I suspect this would not be a problem with a really good Jensen or Lundahl properly implemented).
I am interested in hearing of others' experiences and opinions on this topic. Currently, I am using a cap (Mundorf Supreme) coupled Legato2, but I would like to eliminate caps in my new Legato3 build, without other compromises if possible. I have a few different amplifiers: Pass Labs X150.5, PS Audio Classic 250, and nCore 400 based, and I only run balanced.
Actually #1 and #3 are the best possible solutions if your amplifier can handle it.
#2 Is as good as the caps you use.
#4 is only as good as the trafo/scheme you use. 🙂
#2 Is as good as the caps you use.
#4 is only as good as the trafo/scheme you use. 🙂
am i missing something?
I have had a "no caps" lagato for quite awhile now and once I set the offset ot zero its been there ever since. I check it about once a month and it has not moved. I use TP power supplies that are stock except for a few bypass caps on the lytcs.
I use a calibrated Fluke 77 to measure.
I have had a "no caps" lagato for quite awhile now and once I set the offset ot zero its been there ever since. I check it about once a month and it has not moved. I use TP power supplies that are stock except for a few bypass caps on the lytcs.
I use a calibrated Fluke 77 to measure.
Good point of clarification. It is very easy to keep the actual offset nulled.
I think Barrows really meant the common mode voltage - which you are right is not really offset.
I think Barrows really meant the common mode voltage - which you are right is not really offset.
Yes,
Good point Russ. I am definitely referring to the DC (generally 7-8 volts) from +/- to ground.
I am really considering the transformer option, I know some will cringe, but good transformers from Lundahl and Jensen (if one can handle the price, but hey it is cheaper than going dual mono, which is of questionable benefit) add very little distortion, and may actually improve performance of the connected amplifier by blocking RF.
One thing I would like to hear from experts on is the option of not connecting the signal ground at all. I have read some reports where some people claim this approach could somehow reduce the performance of the connected amplifier. I do not really understand this notion in the case of an amp with a truly differential input stage?
Good point Russ. I am definitely referring to the DC (generally 7-8 volts) from +/- to ground.
I am really considering the transformer option, I know some will cringe, but good transformers from Lundahl and Jensen (if one can handle the price, but hey it is cheaper than going dual mono, which is of questionable benefit) add very little distortion, and may actually improve performance of the connected amplifier by blocking RF.
One thing I would like to hear from experts on is the option of not connecting the signal ground at all. I have read some reports where some people claim this approach could somehow reduce the performance of the connected amplifier. I do not really understand this notion in the case of an amp with a truly differential input stage?
There should be ~8V DC between Ground and +, and between Ground and - outputs from the Legato. You can only zero out the DC between the + and - outputs.
If outputting to a bridged differential amp, it is a big problem. Unfortunate, because this is an easy way to get 2x to 4x power from an amp while decreasing noise at the same time.
and...
For the purpose of discussion, why would you think that? One of the best digital sources I have ever heard (the Linn Klimax DS, current version) uses Lundahl transformers to couple its output. This unit has been third party tested by Stereophile, and the measurements are quite good, easy to look up at stereophile.com as well.
When you say "good" coupling capacitors, what would you consider "good": one could pretty easily spend >$1000 on four coupling caps…
Here is a link to the Klimax DS measurements:
http://www.stereophile.com/content/linn-klimax-ds-network-da-processor-measurements
note that these are for the earlier version, since then Linn has upgraded to Lundahl's amorphous core transformers, which should have slightly less added distortion.
For the purpose of discussion, why would you think that? One of the best digital sources I have ever heard (the Linn Klimax DS, current version) uses Lundahl transformers to couple its output. This unit has been third party tested by Stereophile, and the measurements are quite good, easy to look up at stereophile.com as well.
When you say "good" coupling capacitors, what would you consider "good": one could pretty easily spend >$1000 on four coupling caps…
Here is a link to the Klimax DS measurements:
http://www.stereophile.com/content/linn-klimax-ds-network-da-processor-measurements
note that these are for the earlier version, since then Linn has upgraded to Lundahl's amorphous core transformers, which should have slightly less added distortion.
Last edited:
I'm basing my conclusion (re: caps vs transformers) based on my experiments comparing interstage transformers versus capacitor coupling in tube amplifiers. The capacitors were cheaper and sounded better.
I've also used capacitors at the DAC's output, and it sounded very transparent. One thing I've noticed is that the smaller the value, the more transparent the capacitor. Now I use as small a value as my calculations allow for, with -3dB at 10Hz.
One place to start would be to try Mundorf Silver/Oil and ClarityCap MR. I think they will beat the transformer solution.
Bridged amplifiers are different from a truly balanced (fully differential) input. In a case like that you would be better off using coupling caps or the IVY-III 🙂
A fully differential input stage does not need a ground reference (at least the signal need not be centered there - DC offset will be ignored as long as it is within limits). For example my Sympatico design. will work fine with common mode DC voltage up to about 10V - maybe even a bit more.
A fully differential input stage does not need a ground reference (at least the signal need not be centered there - DC offset will be ignored as long as it is within limits). For example my Sympatico design. will work fine with common mode DC voltage up to about 10V - maybe even a bit more.
Last edited:
These are very impressive measurements. But amorphous core transformers will not sound neutral. They will sound different, some will like and others will not.
Can you redesign the Legato to have no common mode DC offset? 😀 I prefer the Legato to the Ivy and want to stick to Legato.
I personally have another I/V stage that does just that. It is called the Counterpoint III. It is not available to anybody but me at the moment, and it may not ever be. It is a DIY project I have taken on for myself.
It uses a folded cascode output with a common mode servo.
It uses a folded cascode output with a common mode servo.
Last edited:
I would...
really want my -3dB point to be closer to 2Hz, otherwise there is going to be phase shift within the audible bandwidth. Hence, good caps start getting pretty big and expensive. I have Mundorf Supremes (3.9uF) in my BII/Legato2 build right now.
Also, I wonder of some of the good listening experiences I have had with the Linn are due to the isolation factor of the transformers. I guess without trying them it is hard to say. But I am getting suspicious about RF on the output of DACs, and how this RF behaves in the input stages of some amplifiers.
What line level transformers have you tried? From the published specs, some appear to be way better than others.
really want my -3dB point to be closer to 2Hz, otherwise there is going to be phase shift within the audible bandwidth. Hence, good caps start getting pretty big and expensive. I have Mundorf Supremes (3.9uF) in my BII/Legato2 build right now.
Also, I wonder of some of the good listening experiences I have had with the Linn are due to the isolation factor of the transformers. I guess without trying them it is hard to say. But I am getting suspicious about RF on the output of DACs, and how this RF behaves in the input stages of some amplifiers.
What line level transformers have you tried? From the published specs, some appear to be way better than others.
The Legato design itself with its LPF is already a good barrier to RF as long as it is wired sensibly . 🙂
Yeah...
Russ, I hear ya on that. My build will have considerable output wire length (although balanced), as the only way to properly mount a USB receiver is between the back panel and the Legato/BIII, to keep I2S lines short. I wonder if their is any RF pickup after the filter from airborne RF… This is one reason why I am thinking trannies right at the XLRs might be a good thing. Admittedly, I am picking nits here!
Input impedance depends on the amp used, the lowest would be my Pass at 11K per phase. The 3.9uF caps I have now work pretty well with the Pass, but they are really big physically.
Russ, I hear ya on that. My build will have considerable output wire length (although balanced), as the only way to properly mount a USB receiver is between the back panel and the Legato/BIII, to keep I2S lines short. I wonder if their is any RF pickup after the filter from airborne RF… This is one reason why I am thinking trannies right at the XLRs might be a good thing. Admittedly, I am picking nits here!
Input impedance depends on the amp used, the lowest would be my Pass at 11K per phase. The 3.9uF caps I have now work pretty well with the Pass, but they are really big physically.
Honestly - IMO if you were going to use a transformer in that way - the best possible place to put it would be at the power amp.
I personally have never heard a transformer that did not add some color. Not saying it's good or bad. And I have heard quite a few. 🙂
I personally have never heard a transformer that did not add some color. Not saying it's good or bad. And I have heard quite a few. 🙂
yup...
Everything matters: I agree that a transformer will add its own signature, as will differing output wiring, or caps. From a technical perspective though, the measured distortion added is so minimal as to be inconsequential, and I still think there may be benefits: whether these are to taste/preference will be the question. So far, the two DACs which I have listened to quite a bit which used very nice Lundahl output transformers have both been among the best digital sources which I have ever heard, so, for me, I am not too worried about a good Lundahl (1684 amorphous or 1588 Mu metal) degrading the sound from what I have already heard.
Everything matters: I agree that a transformer will add its own signature, as will differing output wiring, or caps. From a technical perspective though, the measured distortion added is so minimal as to be inconsequential, and I still think there may be benefits: whether these are to taste/preference will be the question. So far, the two DACs which I have listened to quite a bit which used very nice Lundahl output transformers have both been among the best digital sources which I have ever heard, so, for me, I am not too worried about a good Lundahl (1684 amorphous or 1588 Mu metal) degrading the sound from what I have already heard.
- Status
- Not open for further replies.