Topband: Chassis Bonding

[Jim Brown]

On 1/29/2014 5:15 AM, Tom W8JI wrote:
>>> 2.) #10, ignoring connection imperfections, has about .01 ohms per 
>>> 10 ft. That's up to 0.2volts with 20 amps of power supply loop current.
>>
>> Not a problem unless you have bonded DC- to the power supply chassis.
>>
> Bingo.
>
> Virtually EVERY piece of gear has a bonded negative to chassis. The 
> whole concept is screwed up, especially negative lead fuses, but it is 
> what it is. No one can force the industry to float negative grounds 
> from chassis. Even RF transistors have the metal bases tied to 
> emitters, which forces a negative ground to heatsinks unless the 
> designer can seriously compromise cooling.

There is no problem with V- bonded at load equipment (the rig). The 
problem is bonding at the power supply, and that is EASY to fix -- 
indeed, many (most?) power supplies are built  either with V- NOT 
bonded, or with nothing chassis-referenced and a removable jumper at the 
output. All Astron supplied I've looked at are built this way -- indeed, 
all I've looked at had the bond to the mounting stud of terminal strip 
that was insulated from the chassis by paint. So they were NOT bonded, 
and neither was the green wire, which went to the same lug.
>
>>> Also, almost every Ham radio manufacturer in the world isolates the 
>>> low level audio shields inside radios except at one point because 
>>> they are well aware of issues with the large currents in 12V solid 
>>> state high power systems.
>>
>> Yep. It's called "The Pin One Problem," and is a well documented 
>> cause of hum, buzz, and RFI.
>
> ...and no matter how much we idealize the situation in our minds, we 
> cannot change the population and things we cannot change just because 
> we dislike a $2 transformer for some reason.

Indeed we can -- the pro audio world changed over a period of ten years 
or so. Sadly, the consumer world never got the word, and hams, with 
their superior "I grew up in broadcast and know the way its done" 
attitudes haven't bothered to learn anything new in 40 years.
>
> We cannot force the world to do things that are difficult, expensive, 
> and/or inconvenient.

Huh? Ground bonding is required by LAW in most of the civilized world, 
and part of the National Electric Code that has been adopted as an 
electrical Building Code almost everywhere in the US. And in those 
places where NEC has not been adopted directly it is the model for local 
codes (LA, Chicago, for example).  Beyond that, it is good engineering 
practice.

>
> Because of that, we have to move away from the theoretically ideal 
> world of zero ground potential difference and face the way the world 
> really works.

Who's talking about theoretical?  I'm certainly not.  The bonding 
methods for a ham station that I have outlined work for ham stations 
because they are  "good enough" -- they provide a transmitted signal to 
noise ratio of 50 dB, not good enough for broadcast, but plenty good for 
us. Indeed, I view the knee-jerk use of broadcast techniques, including 
transformers where they are not needed, as "purist."

>
> Very few people are going to bond grounds properly, 


> and fewer still will drill into a computer or other equipment to add a 
> ground lugs 

It's trivially easy for most equipment -- DB connectors are bonded to 
the chassis of laptops, the chassis of most computers have exposed 
screws, all ham gear has a chassis lug.

> or modify equipment to float negative returns from chassis.

No need to do that except at the power supply,
>
> The only thing eliminating the isolation transformer does, other than 
> saving $2, is ruin systems in the real world. It is a terrible 
> suggestion. No accessory manufacturer will ever follow your directive.

Bonding cable shields to the chassis has been an AES Standard for nearly 
ten years, and it's been widely implemented in pro audio since the late 
'90s.

73, Jim K9YC