The major differences are these:
1) Legato 1 has a buffer (usually not necessary) Legato 2 does not.
2) Legato 2 has three FETs per half in parallel, Legato 1 has a single FET per half.
3) Legato 2 uses two resistors in parallel per half for the CCS. Legato 1 only a single resistor per half.
Otherwise the circuits are the same.
1) Legato 1 has a buffer (usually not necessary) Legato 2 does not.
2) Legato 2 has three FETs per half in parallel, Legato 1 has a single FET per half.
3) Legato 2 uses two resistors in parallel per half for the CCS. Legato 1 only a single resistor per half.
Otherwise the circuits are the same.
These are small signal FETs. Small signal FETs are great for linearity but have rather high Rds on values (2-5 ohms depending on the exact make). The linearity we want, and the CFP basically eliminates the RDs on, but the lower we can make it, the less we have to modulate the voltage at the gates. 
Using three of them gets their Rds on down quite a bit without adding too much capacitance, and so we get the transconductance up into the sweet spot. This make for a bit lower distortion. It was already very low. This is more of a *tweak* than a necessary change.
One note, there is a limit on how many FETs to add because of the added capacitance of each FET. That is how I settled on three. Lots of listening as well.
Cheers!
Russ
Using three of them gets their Rds on down quite a bit without adding too much capacitance, and so we get the transconductance up into the sweet spot. This make for a bit lower distortion. It was already very low. This is more of a *tweak* than a necessary change.
One note, there is a limit on how many FETs to add because of the added capacitance of each FET. That is how I settled on three. Lots of listening as well.
Cheers!
Russ
Last edited:
Sure it's a fairly simple mod for anyone handy with a soldering iron. But I would be hesitant to call it an "upgrade". Really its just a different way to do the same thing.
All you would need to do is solder three FETs in parallel were you would normally use one. You don't need the two Rs in parallel at the CCS, but you could also do the same thing there. Just use two 680R where you would normally use one 320 or 340R.
It should not be a problem. You may need to adjust things a little, but it should be just fine.
Can I get a little clarification on the buffer? The output buffer is primarily for the balanced output of the Legato I/V stages and is not needed for loads greater than 4K or so per the manual. But it also does feed the Bal/SE stage, which is buffered on its own.
So it seems that the Bal/SE can be fed from the I/V stage with or without the buffer, but is it better to not use the buffer? The only comment I have seen is that the only benefit to removing the buffer is that there are less parts in the signal path, not that it make a difference in how it sounds or any other important (to me) parameter.
Since the Bal/SE section has a high input impedance, the buffer does not matter and is even left out of the Legato 2 that is coming. But if I am planning to use SE out mostly and my only (current) balanced option, should I choose to use it, has an input impedance of only 2K, I guess keeping the buffer may be better?
So it seems that the Bal/SE can be fed from the I/V stage with or without the buffer, but is it better to not use the buffer? The only comment I have seen is that the only benefit to removing the buffer is that there are less parts in the signal path, not that it make a difference in how it sounds or any other important (to me) parameter.
Since the Bal/SE section has a high input impedance, the buffer does not matter and is even left out of the Legato 2 that is coming. But if I am planning to use SE out mostly and my only (current) balanced option, should I choose to use it, has an input impedance of only 2K, I guess keeping the buffer may be better?
if you are using the BAL/SE stage you definitely do not need the buffer for Legato 1.
Honestly the only reason I added the buffer was to potentially drive highish impedance balanced wired headphones. But really, that is not the best use for this I/V stage.
If you have a balanced load of 2K you still don't need the buffer, but there would be a slight signal attenuation when omitting the buffer. If that attenuation is undesirable then go ahead and use the buffer or simply increase the output swing a little. The choice is yours.
You can calculate the attenuated output by the following formula:
Zin/Zin+Zout. In your case the result is 0.930233. Nothing to really worry about.
The active components of the balanced I/V stage see a constant current no matter the impedance of the target load. This is not true for the SE output.
I hope that helps.
Cheers!
Russ
Honestly the only reason I added the buffer was to potentially drive highish impedance balanced wired headphones. But really, that is not the best use for this I/V stage.
If you have a balanced load of 2K you still don't need the buffer, but there would be a slight signal attenuation when omitting the buffer. If that attenuation is undesirable then go ahead and use the buffer or simply increase the output swing a little. The choice is yours.
You can calculate the attenuated output by the following formula:
Zin/Zin+Zout. In your case the result is 0.930233. Nothing to really worry about.
The active components of the balanced I/V stage see a constant current no matter the impedance of the target load. This is not true for the SE output.
I hope that helps.
Cheers!
Russ
Last edited:
Honestly I would leave it in if it is already there for a little while. See how you like it. Its easy enough to reverse later.
There is not really any negative effect to having it other than drawing a bit more current because the buffer is class A.
Last edited:
I plan on this also. I haven't yet determined which transform to use yet, but am open to suggestions...
btw Russ,
Which "little things" will need adjusting?
Jeff Davison
btw Russ,
Which "little things" will need adjusting?
Jeff Davison