Locating diodes close to the M3875

[Greg Erskine]

Hi all,

I have see a few examples of people locating the diodes close to the chip rather than close to the transformers.

I was wondering the pros and cons of such an idea.

[analog_sa]

I don't think diode location will make any difference.

cheers
peter

[Mad_K]

Originally posted by grege
Hi all,

I have see a few examples of people locating the diodes close to the chip rather than close to the transformers.

I was wondering the pros and cons of such an idea.

pros: dunno

cons: higher impedance between the transformer/rectifier. This could make some unwanted effects in your PSU.

[Nuuk]

Probably down to space constrictions but I would prefer to keep these items as far away from the signal as I can!

[Peter Daniel]

I built separate supply for my phono stage and decided to use diodes on main board and only the transformer in a separate enclosure. The wires from transformer were probably 3 feet long (4 per channel) and when I used to run them close together (in a one sleeve) they got really hot. Had to remove sleeve and had the separated.

I would imagine that it's always better to have bridges closer to transformer, with some local caps (I'm using 4.7u BG N), and run DC to the amp. Any possible AC noise is reduced this way.

[analog_sa]

The wires from transformer were probably 3 feet long (4 per channel) and when I used to run them close together (in a one sleeve) they got really hot. Had to remove sleeve and had the separated.

Can't think of any reasonable explanation for that. Poor insulation led to a short?

[Peter Daniel]

I didn't measure any short. But the cable was made out of Audioquest speaker cable (Midnight, IIRC). It was warm to touch. Later I used 8 pcs of teflon insulated, loose run wires and it was fine.

[goudey]

Audioquest speaker cable....

"...and when I used to run them close together (in a one sleeve) they got really hot. ..."

"...the cable was made out of Audioquest speaker cable..."


An interesting and sometimes important note on some of the Audioquest speaker cables: The cable is made up of individually "insulated" solid core conductors. The conductors have different color insulation, but often there is more than one wire with black "insulation". If you use an ohm meter to measure resistance BETWEEN "insulated" conductors, you will discover that the black insulation does not insulate (somewhat conductive). That could explain the mystery of the hot cable when used to carry signal between two black wires.

[Greg Erskine]

Thanks for the responses.

Originally posted by Mad_K
cons: higher impedance between the transformer/rectifier. This could make some unwanted effects in your PSU.

What effects would you imagine besides a little voltage drop due to resistance.

Originally posted by Peter Daniel
I would imagine that it's always better to have bridges closer to transformer, with some local caps (I'm using 4.7u BG N), and run DC to the amp. Any possible AC noise is reduced this way.

The AC to the rectifiers would be say 22VAC @ 50 or 60 Hz, but the AC ripple after the rectifiers would be a few volts? @ 100 or 120 Hz or do those small 4.7uF caps actually remove all the ripple. What is the predominant factor for AC noise, voltage amplitude or frequency?

Peter, so you are using 1000uF BG at the LM3875 and 4.7uF BG at the rectifiers. Have you found this really necessary? I was wondering if the long power cable was acting as a resistive element in an RC filter thus help smoothing the PSU.

Nobody has mentioned the noise from the diodes switching. Is this significant?

[Peter Daniel]

Originally posted by goudey


An interesting and sometimes important note on some of the Audioquest speaker cables: The cable is made up of individually "insulated" solid core conductors. The conductors have different color insulation, but often there is more than one wire with black "insulation". If you use an ohm meter to measure resistance BETWEEN "insulated" conductors, you will discover that the black insulation does not insulate (somewhat conductive). That could explain the mystery of the hot cable when used to carry signal between two black wires.

It is indeed the case. I measured something like 120ohm between some of the black wires, I know there was no short, but I have no clue where this comes from. It may indeed be the reason for cable getting warm.

Originally posted by grege
Peter, so you are using 1000uF BG at the LM3875 and 4.7uF BG at the rectifiers. Have you found this really necessary? I was wondering if the long power cable was acting as a resistive element in an RC filter thus help smoothing the PSU.

My cables are 2' long and this is how I use the caps. I'm using those small 4.7u BGs as they improve sonics. It was also suggested to me by Fred to use some capacitance with the rectifiers. I tried what I had on hand and I liked it. I also tried additonal 1000u at the bridges, but there was no doubt it sounded worse.