Bob Cordell Interview: Power Supplies

Status
Not open for further replies.
Jump to Latest
#521
ingrast said:



Just thinking aloud Bob.

From what you post it seems you are not certain whether the spray is conducted all the way from the rectifier, or whether it is radiated from the transformer and picked up.

Perhaps removing the transformer and connecting with a long thick cable may help, moving it around, to check if something changes.

If it is conducted - which seems more likely - perhaps the culprit is the lower leackage inductance inherent in a better confined magnetic circuit as is the case for a toroid in comparison with a I-E or E-E lamination. If that were the case, then a small input choke will also make for a measurable difference, just to make sure.

Just a cent, and keep us posted on your results.

Rodolfo
Click to expand...


I'm pretty sure the spray is radiated magnetic, as it appears to be getting into the I/O leads running to the amplifier. If I move the leads around, I can make it better or worse. I'm talking about pretty low levels here, on the order of 110-120 dB below 10 watts.

Bob
 
#523
You can move a toroid around, orienting it various ways and
get some improvement, but there's always some noise, and
unfortunately the noise is not necessarily consistent from
transformer to transformer for even the best manufacturers.

The most certain solution is raw distance.

😎
 
#524
Conrad Hoffman said:
A thick aluminum plate between the amps and tranny did the trick, when little else was effective.
Click to expand...

Then it seems that in a the average AB class amplifier, aluminum heatsink can be placed as a divider between a transformer and the PCB, so it should reduce noise to some extent. Unfortunately, this will be hardly applicable in a class A amps due to big heatsinks which won't be placed inside the enclosure 🙁

On the other hand, raw distance works best. Just look at the Gaincard amp. Transformer is in the separate box. Still, for most manufacturers, this is not the most practical solution. Imagine two monoblocks with separate PSU-s. A stereo amp will will have to be sold as a four-piece gadget. Some will either like or hate it.

Regards,

Vix
 
#525
Once I had an amp where the input signal wires (L/R channels) passing near to the toroidal transformer. I got lousy noisy sound. I tried to move the wires somewhat away from the transformer, and it improved. At the same time, I twisted two wires, and it improved even better. Very quiet! Nevertheless, I covered part of the transformer with a piece of thin copper sheet. Finally, I grinned with my full satisfaction.

:darkside:
 
#526
Nelson Pass said:
You can move a toroid around, orienting it various ways and
get some improvement, but there's always some noise, and
unfortunately the noise is not necessarily consistent from
transformer to transformer for even the best manufacturers.

The most certain solution is raw distance.
Click to expand...
Nelson,

I don't mean radiation saying about vertical placement.
 
#527
anli said:
BTW, if we tell about toroid transformers, are there any drawbacks of a toroid transformer vertical placement? I have not tried yet. Anybody?
Click to expand...
No, I do not see any problem to it. I did it in my recently finished amp:

Here is what I did to solve the problem of wiring that needed to go across the whole length. The wires are outside covered with aluminum channel. Dead quiet.
 

Attachments

  • powersupply.jpg
    powersupply.jpg
    80.6 KB · Views: 1,002
#530
Here is the wiring going out of the box right from the AC connector. I was playing with liquid plastic/rubber - yellow portion around the hole and I was very happy. It serves the purpose of protecting the wires from the sharp edge of aluminum base.
 

Attachments

  • acwiring.jpg
    acwiring.jpg
    79 KB · Views: 930
#534
Nelson Pass said:
You can move a toroid around, orienting it various ways and
get some improvement, but there's always some noise, and
unfortunately the noise is not necessarily consistent from
transformer to transformer for even the best manufacturers.

The most certain solution is raw distance.

😎
Click to expand...


Thanks Nelson. What has your experience been with using different types of rectifiers?

Bob
 
#536
Bob Cordell said:
OK, suppose we are stuck with a given transformer, which may often be the case. How can we reduce this tendency to radiate a spray of 120 Hz harmonics? Is there some kind of better filtering we can do at or near the ac side of the rectifiers? Will soft-recovery diodes help this? I've tried snubbing and shunt capacitance with limited improvement. Its just surprizing that the toroid was so much more prone to this than a conventional transformer.

Bob
Click to expand...

In the Gary Galo "Pooge 5.5 DAC960 Modifications" (TAA 1/94) Rick Miller described some work he did on 1n4xxx diodes contrast with BYV28-100 and GI851 -- he didn't do any empirical work on the HexFreds -- the work showed considerably less RFI noise for the high speed/soft recovery diodes.

He used an AD811 cfb opamp at +20dB into 50R coupled with 100nF -- and a Tek 2710 spec analyzer -- although for this application you can probably see the noise over a few MegaHertz.

...and the XYL tells me I have to get rid of these magazines...
 
#538
Bob Cordell said:
What has your experience been with using different types of rectifiers?
Click to expand...

Fast rectifiers and high frequency filtering measurably improve
radiated noise, enough to make the difference between passing
and failing EMI testing in many countries.

Historically I have remained sceptical about the sonic benefits,
but I have a couple of people who can hear a difference in a
single-blind test. I haven't gotten too excited, as the decision to use the parts is moot (as above).

😎
 
#539
Nelson Pass said:


Fast rectifiers and high frequency filtering measurably improve
radiated noise, enough to make the difference between passing
and failing EMI testing in many countries.

Historically I have remained sceptical about the sonic benefits,
but I have a couple of people who can hear a difference in a
single-blind test. I haven't gotten too excited, as the decision to use the parts is moot (as above).

😎
Click to expand...


Thanks, Nelson, good point. EMI requirements can indeed be a tough pain. We recently had a 16-layer daughterboard where we had to plate the board edges to reduce emissions at 6 GHz. That stuff can squeek through a crack like a termite.

Cheers,
Bob
 
#540
Hello all
Please help me with this PSU design. I'm not sure if the common 0v for VAS and LTP is correct connected to the main GND
The SR for first stages will be standard zener + bjt serial regulators.
The last 1000uF low ESR capacitors will be as close as possible to the output BJTs and the amp's PCB will be designed with internal starground
All suggestions are welcome

Mihai
 

Attachments

  • psu.jpg
    psu.jpg
    36.8 KB · Views: 811
Status
Not open for further replies.
Top Bottom