NCD questions

[ackcheng]

Gertjan,

Are you still using SMPS for the UCD? Leaving the PSU ground floated would be dangerous??

[soongsc]

Originally posted by ackcheng
Gertjan,

Are you still using SMPS for the UCD? Leaving the PSU ground floated would be dangerous??

How did UcD sneak into this thread?

[classd4sure]

Originally posted by Lars Clausen
The best GND'ing setup for balanced input depends largely on the way the balanced output was implemented in the preamplifier.

Some use two active output with a phase reversal, and GND of both connected to pin 1. Others have a transformer with a center pin connected to pin 1. Others again have only connected pin 1 to the chassis, and have the signal pins more or less floating in relation to pin 1. The professional standard is a further different variant with several twisted pairs in one shielded cable, where the pin 1 carries only shield drain. In this case transformers in the input is more or less a 'must have'.

I can't see there is one GND setup in the power amplifier that can take all cases into account optimally.

So to make the best solution, you must first look at the way your preamplifier has implemented balanced output. But i guess this discussion is not realted to NewClassD, but balanced signalling in general.

Hi Lars,

You claim it optimal for DC to be taken into account elsewhere, at the source, but the source can't account for the DC offset of your module. In actual fact it probably only saves you a few pennies.. but you should be really make it very clear that it's up to the user to ensure the input is DC free.

As per the grounding:

I seriously can't recommend only having a ground lift on just one channel. To preserve all the good spatial info contained in the microdynamics they should be as error free as possible. Using the ground lifts has some smoothing effect because it creates an error voltage across them, so what the input stage see's as a signal reference ends up differing from the signal reference at the source.

Any level of hum induced by groundloops should certainly be one area, if you're at all interested in preserving signal integrity, that should most certainly be taken care of elsewhere.

In my case removing them had no difference on the level of HF noise, low frequency hum (I took care of that elsewhere, by removing your other ground lift resistor in the supply), it only helped give it a more accurate sound by ensuring the amp's signal ground shares the same reference as that of the source.

As you've said there's no possibility of a one size fits all cure, so then why would you attempt to do so with ground lift resistors when it is only going to ruin signal integrity? I think it definatly one area where you should have kept the same philisophy as you had used for the DC on the inputs problem.

By the way, I said this to Sander yesterday:

With the FC input stage decoupling caps used that you sent me, I'd really never before heard such good seperation of instruments, as if each one was carved in the air. It's too bad they have such a distasteful coloration to them, which in my view makes an otherwise excellent cap unusable.

[Lars Clausen]

ClassD4sure: Thank You for your comments.

[Bgt]

Originally posted by ghemink
Thanks for the info, however, this does not look like the Hypex recommendation, they recommended to keep the power supply floating. I`m also using XLR balanced inputs while the GNDing scheme in your link is for RCA input. It may work pretty good though as those UcD differential inputs maybe quite forgiving.

Gertjan, I've noticed the Hypex way serves well in a dual mono config(have an amp. like this also). With 1 transformer it is a totally different story. For me there was no way of getting the crosstalk and capacitive sensitivity of the input cables down. Remember I use 35db gain which makes the input more sensitive of EMI and bad grounding. And whatever Chris suggests maybe OK for him but for me it was a dead end.
I've spent hours and hours to get it right and this for me is the best setup. Scope and ears will tell you.

[classd4sure]

Originally posted by Bgt


Gertjan, I've noticed the Hypex way serves well in a dual mono config(have an amp. like this also). With 1 transformer it is a totally different story. For me there was no way of getting the crosstalk and capacitive sensitivity of the input cables down. Remember I use 35db gain which makes the input more sensitive of EMI and bad grounding. And whatever Chris suggests maybe OK for him but for me it was a dead end.
I've spent hours and hours to get it right and this for me is the best setup. Scope and ears will tell you.

and yet, I've done the exact same thing with the NCD, far less forgiving inputs, and a gain of 45.

You can't compensate for EMI induced crosstalk with groundlift resistors. To me that's like putting the bandaid on the floor to cover up the spilled blood, while allowing the wound to continue bleeding.

[Bgt]

Originally posted by Lars Clausen
The best GND'ing setup for balanced input depends largely on the way the balanced output was implemented in the preamplifier.

Some use two active output with a phase reversal, and GND of both connected to pin 1. Others have a transformer with a center pin connected to pin 1. Others again have only connected pin 1 to the chassis, and have the signal pins more or less floating in relation to pin 1. The professional standard is a further different variant with several twisted pairs in one shielded cable, where the pin 1 carries only shield drain. In this case transformers in the input is more or less a 'must have'.

I can't see there is one GND setup in the power amplifier that can take all cases into account optimally.

So to make the best solution, you must first look at the way your preamplifier has implemented balanced output. But i guess this discussion is not realted to NewClassD, but balanced signalling in general.

Absolutely agree, 1 should look at his setup.

[Bgt]

Originally posted by classd4sure
:and yet, I've done the exact same thing with the NCD, far less forgiving inputs, and a gain of 45.

You can't compensate for EMI induced crosstalk with groundlift resistors. To me that's like putting the bandaid on the floor to cover up the spilled blood, while allowing the wound to continue bleeding.

Utter bs, if you look at the wiring diagramm you'll see the output from the preamp, input at poweramp, going directly to the + and - input of the opamp, so where is the difference in signal????? The groundlift resistor has nothing to do with purityy of sound

[Lars Clausen]

.....cover up the spilled blood, while allowing the wound to continue bleeding.

Is it halloween already ... ??? [joke]

[ghemink]

Originally posted by ackcheng
Gertjan,

Are you still using SMPS for the UCD? Leaving the PSU ground floated would be dangerous??

Yes, but only for the woofer amps at the moment, I keep them floating, I know that may not be 100% safe so I plan in the future to use a balanced transformer in which the midtap is used as a kind of floating GND.�@This should be safe and give me balanced power at the same time.

Best regards

Gertjan