Topband: Desktop Power Supply Brand/Noise Question

[ZR]

>> On 7/21/2013 8:49 AM, Tom W8JI wrote:
>>>
>>> I can't understand why throwing some ferrite beads at a problem, or 
>>> changing the supply, are the only two solutions.
>>>
>>> Many times, if not most times, a few .01 uF line voltage rated bypass 
>>> capacitors are significantly better than a sting of cores, or a winding 
>>> through cores.
>>
>> Yes, IF the problem is differential-mode coupling to the power line. But 
>> often it is NOT --
>
> Bypass capacitors will cure common mode also, because common mode cannot 
> be generated without differential mode someplace in the system between two 
> points.
>
> In virtually every situation, the power supply is in a metal box. The 
> trash is between the power line leads, and also between those leads and 
> the case. This is the nature of switching supplies, because they tie a 
> "chopper" with squarewaves across the power mains, and the system is not 
> well balanced.
>
> As a general rule the stuff exiting the dc side is much less problematic. 
> If it is problematic, it needs cleaned up. Bypassing to the cabinet, 
> groundplane, or case fixes or greatly assists in fixing it, too.
>
> If a line is properly bypassed to the case, the system can't produce much 
> common mode or differential mode on that particular line.
> In the rarer case, where two or more lines are involved, they all must be 
> suitably bypassed.
>
> The same things that work for lightning protection work for RFI issues, 
> and vice versa. Payback for doing things right, rather than just tossing 
> beads at wires, is much wider than the initial target.
>
> 73 Tom

I certainly agree about not throwing ferrite at  problem without 
understanding why.
My introduction to serious line and computer noise was after moving here in 
89 and I remembred a QST article by DeMaw a few years earlier where he 
introduced a differential mode noise filter using ferrite rods and caps in a 
pi filter. The line was noisy and I had just put my Commodore 64 out to 
pasture to be replaced on the new fangled DX Packet Cluster by a 386-33 PC 
with 12" monochrome monitor. It was soon also used for various DOS RF and 
other programs, Windoze 3.1 and a long learning curve.
 I built DeMaws circuit in a couple of outlet strips and used them for the 
PC and the TS-940's which were all on seperate circuits.

Most of the noise was gone and using toroids on the keyboard and monitor 
cables took care of the rest. This was before ferrites were included from 
the factory.

About 2-3 years ago when I took an interest in the LF frequencies I realized 
that DeMaws circuit wasnt effective down there and rebuilt the filters into 
2 stage affairs with a 100KHz cutoff.

There are now 6 PC's in the house and each got its own filter as well as the 
modern and boatanchor operating desks. A new level of quiet appeared as if 
by magic by attacking the differential noise first.

Any remaining annoyances were locally radiated coming in via various 
antennas (including portable radio loopsticks and whips) which was 
eliminated by a lot of large 43 and then 31 toroids on power cords, router 
cables and some feedlines.

With a couple of Beverages moved another 600-700' away back in the woods and 
feeding a new switch box and its own 1/2" CATV hardline run back to the 
house, even neighbors noise is mostly tolerable with none....so far....on 
any good DX frequencies.

Carl
KM1H