• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Legato 3, Common Mode DC and You. How do you choose to deal with it?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
If you are going to use a transformer, why use a Legato at all?

One reason would be to present the DAC with a far lower input impedance than you could with just a transformer. This makes the DAC behave more like a true current source.

The more the voltage modulates at the DAC outputs (the more it acts like a voltage source) the worse it performs in terms of both IMD and THD. So keeping that impedance as close to zero as possible is very important.
 
OK, I buy that to a point, but wonder if the overall complexities would negate the improvements, and worth the risk of floating the ground or DC imbalances at the balanced outputs.

As a TVC user, I keep worrying about the DC balance drift of the Legato's output saturating the transformer on my volume control. I prefer the body/meat (distortion?) the TVC gives me over amp direct, and have considered trying a transformer I-V in lieu or in conjunction with my present TVC (Bent Audio TAP).
 
Yeah...

The transformer will just reflect the input impedanc of the following stage (amp) so the output of the ESS would be loaded with 10K-200K, or whatever your amp is.
Plus, I have not looked, but is the voltage and current output enough to drive the input of a transformer with low distortion? In other words, what is the output of the ESS in V when loaded with, say, 100K, and how much current is available vs the Legato's output set for 2V?
 
Hi Russ,

I have a question about the loading of the ESS9018 in your Legatos. The two inputs "sit" at half of VACC (around 1,65VDC), and this means, I think, that the ESS9018 does not give out any DC current Only AC current (the music). Using a resistive load on the other hand would draw DC current out from the ESS9018 chip. Using a transformer will not draw any DC current either.

I wonder, what does the ESS9018 prefer - that it outputs both DC and AC current or only AC current?



/S

One reason would be to present the DAC with a far lower input impedance than you could with just a transformer. This makes the DAC behave more like a true current source.

The more the voltage modulates at the DAC outputs (the more it acts like a voltage source) the worse it performs in terms of both IMD and THD. So keeping that impedance as close to zero as possible is very important.
 
Current will flow from it always - it just depends on where the load is sitting.

You can calculate the current using Ohms law.

For example with a AVCC at 3.3V and a virtual ground (0V) at outputs you will get ~ 8ma static current.

My experience about the static/bias current is that the DAC is not particularly picky - it is just that the voltage at the outputs should remain as constant as possible.

I have also found it does not really matter much if the static current is sunk or sourced.
 
Last edited:
The transformer will just reflect the input impedanc of the following stage (amp) so the output of the ESS would be loaded with 10K-200K, or whatever your amp is.
Plus, I have not looked, but is the voltage and current output enough to drive the input of a transformer with low distortion? In other words, what is the output of the ESS in V when loaded with, say, 100K, and how much current is available vs the Legato's output set for 2V?

Well, that depends upon the turns ratio. The reflected load is the secondary load times the squared primary/secondary ratio. Furthermore, if there is a resistor in parallel to the primary, the reflected load will be a combination of the impedances in parallel. In either event, not the 100k described.
The 9018 has plenty of drive to get 2v out via transformer btw.
 
Yes...

Well, that depends upon the turns ratio. The reflected load is the secondary load times the squared primary/secondary ratio. Furthermore, if there is a resistor in parallel to the primary, the reflected load will be a combination of the impedances in parallel. In either event, not the 100k described.
The 9018 has plenty of drive to get 2v out via transformer btw.

Those are the details. My point was, that in any case, the load is going to be much closer to that of the amplifier than the <1 ohm the ESS 9018 wants to see for best performance. And, thanks to Russ' answer which does confirm that the direct output of the DAC chip will produce enough voltage and current, barely (just as the Legato does, barely) to drive the transformer well. But it does appear that if one wants 2V output, and adequate current to drive an amplifier directly, a 1:1 transformer would be the only suitable choice for use with most amps. Of course, we digress! There is already a huge thread on direct coupling the Buffalo's output to transformers:
4est, are you using caps with your L3? If so what has your experience been?
I am leaning more an more to just running with the G unconnected at the output of the L3 into the nCore amp, hoping that the nCore is going to beat out my Pass this time... But I do not think this appraoch will work with my Classic 250, as it has TVS diodes at the inputs...
 
Those are the details. My point was, that in any case, the load is going to be much closer to that of the amplifier than the <1 ohm the ESS 9018 wants to see for best performance. And, thanks to Russ' answer which does confirm that the direct output of the DAC chip will produce enough voltage and current, barely (just as the Legato does, barely) to drive the transformer well. But it does appear that if one wants 2V output, and adequate current to drive an amplifier directly, a 1:1 transformer would be the only suitable choice for use with most amps. Of course, we digress! There is already a huge thread on direct coupling the Buffalo's output to transformers:
4est, are you using caps with your L3? If so what has your experience been?
I am leaning more an more to just running with the G unconnected at the output of the L3 into the nCore amp, hoping that the nCore is going to beat out my Pass this time... But I do not think this appraoch will work with my Classic 250, as it has TVS diodes at the inputs...

I do not think that it is so cut and dried, otherwise there would a clear consensus on what transformer to use, and 10 ohm is closer to 1 than it is to 100k while eliminating a bunch of other parts and supplies, but yes enough of that!

I briefly ran mine with 10uf polyprope/tin before I dared hook it up direct. I needed to know that it wouldn't drift during break in. It still sounded good, but a cap is a cap and I took it out first chance I got, and then in again to briefly test it amp direct with the volumite. I struggle with the idea of lifting the ground and floating 8-9v on the input pins, seems as if begging for trouble.
 
Last edited:
4est: I am recalling correctly that you run JC-1s? If so, can you confirm no issues running the L3 without caps into them?
The JC-1, I am pretty sure, has a servo in the input stage, so it might be OK with a little bit of offset anyway, but I am going by memory here and am not certain.
 
The beauty...

Yes, but ground is everywhere, well er umm, should be! ;)

of floating grounds. I prefer floating grounds, no AC ground reference, no connection of signal ground to AC ground. This way, with balanced connections, all grounds are "virtual". This also eliminates circulating ground currents.
Warning: this is not recommended for people building their own stuff! Not connecting the AC ground might kill you! But I do it anyway, and so do some very well established major audiophile component makers of which I am aware.
With the nCores I am pretty sure this method will work well: BIIIse/Legato3 with no AC ground connection, no signal ground connection to XLR-cable-nCore amp, with pin 1 connected to chassis, no AC ground connection. No loops, no signal ground connection so no offset.
I know floating grounds seems to freak some people out, but it has worked quite well for me for awhile now, silent...
 
DO NOT DO THIS, EVER

of floating grounds. I prefer floating grounds, no AC ground reference, no connection of signal ground to AC ground. This way, with balanced connections, all grounds are "virtual". This also eliminates circulating ground currents.
Warning: this is not recommended for people building their own stuff! Not connecting the AC ground might kill you! But I do it anyway, and so do some very well established major audiophile component makers of which I am aware.
With the nCores I am pretty sure this method will work well: BIIIse/Legato3 with no AC ground connection, no signal ground connection to XLR-cable-nCore amp, with pin 1 connected to chassis, no AC ground connection. No loops, no signal ground connection so no offset.
I know floating grounds seems to freak some people out, but it has worked quite well for me for awhile now, silent...

AC Ground (Earth) SHOULD ALWAYS BE CONNECTED TO YOUR EQUIPMENT CHASSIS. ALWAYS. NO EXCEPTIONS.
 
Yes, but ground is everywhere, well er umm, should be! ;)

That is true to a point, but 0V does not *have* to be DC coupled to say the mains earth. For certain the chassis and Mains GND should always be joined for safety.

From a signal point of view 0V is relative, and when talking about signal GND(0V) that does not necessarily need to be connected to mains GND at all. Though many times it is. :)
 
That is true to a point, but 0V does not *have* to be DC coupled to say the mains earth. For certain the chassis and Mains GND should always be joined for safety.

From a signal point of view 0V is relative, and when talking about signal GND(0V) that does not necessarily need to be connected to mains GND at all. Though many times it is. :)

Yes sir! That is what I meant, but still, the digital grounding gave me cause to wonder. I am running i2s, and unlike SPDIF, there is no transformer to assist in isolation. To be honest, I didn't think it all through, but I didn't want to force my DAC to signal ground >i2s>usb and>cpu. I guess I just assumed that the digital ground was tied to the Legato ground.
 
That's the nice thing about balanced signals. GND is common mode. So there is little to fear about joining a digital ground.

Also, a great option if you want to not join digital and analog GNDs is to use a pair of teleporters. I have see people do this even between components in the same chassis, seems a bit overkill to me, but hey, you can do it. :)
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.