UCD input configuration

[Bgt]

Measuring an UCD 180 I was just testing if my crosstalk was still ok and I was bit suprised to see that touching the wire tree made the crosstalk differ from 90db to 60db(@15khz). So I made a different input cable.
I think this only applies to people using 1 transformer with 2 power supplies and using single pole inputs.

[wytco0]

BGT, I am sure its obvious to others what you have done and why but just in case there are any other ignorant DIYer like me could you explain your thinking and the solution?

Thanks.

[Bgt]

trying to:


explain that in such a setup, 1 transformer 2 bridge rectifiers, 1 for each channel, 4 caps, 2 for each channel the original Hypex input cable says you have to connect the - input wire to shielding of the input cable at the input device which is on the back of the housing. This not a sym. input device but a standard cinch device. In that case with the Hypex recommendation you can expect microphonic cable sensitivity. It means it is very susceptable for EMI(electr.magn. interf.) hence your crosstalk will only depend how you position your cabling. What I did was to put the - input on the ground pin with a 10 ohm resistor in series. Also the input cable screening is connected to the 10 ohm resistor.
This makes the input cable almost insensitive for EMI hence your crosstalk goes from -60 to -90Db@15khz.
BTW case earthpoint is just at the 0V position in the middle of the Caps.
Bert

[Bgt]

sorry, mod. one



Twisted the output and +and- psu cables also. Quite difficult, they are thick ones. 4mm
It does sound different now but I don't know if its the twisting or the cable mod? It sounds more friendly now..........softer...ah well, different.

[Bgt]

here is a scope picture, noise level is about 10mv, than crosstalk is about 10mv while the active channels pumps out 100V at 15khz on 8 ohms. LF filter is 50khz. Input is loaded with 2k5(my passive pre).
When I did a long listening test the soundstage is much better, even the radio sounds better.

20mv/div.
15khz crosstalk



noise level

[Bruno Putzeys]

Could you explain just how you originally connected the input cable to the RCA chassis parts? (i.e. before you made this mod). Please do not reply "As recommended by Hypex" but draw your own diagram and explain in your own words how the original configuration was.
It seems to me there must have been a problem there to begin with. Opening the shield connection increases RFI sensitivity because whatever happens to the shield efficiently couples into the signal conductors. Likewise, preventing the flow of HF current through the shield effectively allows external magnetic fields to create a voltage differential between the shield and the signal wire.
The mention of microphony in the discussion is indicative of a certain degree of confusion, one which I'd like to clear up.

[Bgt]

Originally posted by Bruno Putzeys
Could you explain just how you originally connected the input cable to the RCA chassis parts? (i.e. before you made this mod). Please do not reply "As recommended by Hypex" but draw your own diagram and explain in your own words how the original configuration was.
It seems to me there must have been a problem there to begin with. Opening the shield connection increases RFI sensitivity because whatever happens to the shield efficiently couples into the signal conductors. Likewise, preventing the flow of HF current through the shield effectively allows external magnetic fields to create a voltage differential between the shield and the signal wire.
The mention of microphony in the discussion is indicative of a certain degree of confusion, one which I'd like to clear up.

Original was the standard Hypex cable on the UCD end and on the cinch part I've put the red cable end in the middle of the cinch bus, black one to the the cinch part housing(ground side) and the cable shielding went here also in series with a 10 ohms resistor.
If I would connect the shield to the cinch parts housing directly I had a humm.

[Bgt]

Bruno, its a pity you did not use a small resistor between the analog opamp ground and power amp. ground on the UCD PCB's like they do in Class AB amps. This will reduce the humm part when using 1 transformer for a stereo amp. having 1 ground plane in the PSU.

[Bgt]

this is the old situation

[Jocko Homo]

You are certain about this ground connection SNAFU, then. Eh???

Not sure that I agree with you.

Jocko