Diode Isolated Ground

[OperaHouse]

I've been refurbishing some older equipment lately and have really noticed how safety practices have improved. I've seen switches and fuses placed in the neutral line. Many grounds were placed to points that wouldn't always make a good connection. Even HP made a line ground connection to a long circuit board trace.

That got me thinking about an old HP application note I saw in the 70's. To reduce ground loop currents they placed a 35A bridge rectifier in series with the AC ground lead. + & - were shorted together. One AC went to ground and the other to the line ground. This would give over a volt of isolation from the line ground. While I basically like the concept, I wouldn't like to be in a witness stand and being asked questions about it! Anyone else remember this?

I hope we all think more about safety now. Ground connections are the first things I look at when I get a piece of equipment in. Often I add a redundant connection or move the ground. Every piece I connect to my HYPATIA 309 High Current Source for a timed stress test. This has become one of my favorite pieces of equipment with lots of uses besides stress testing.

 

[VEBill]

The obvious problem with a diode in series with safety ground is that it becomes an issue as to what blows first in the event of a short from line hot to chassis: the overcurrent protection device or the 35A diode bridge?

Magnetically-assisted breakers (eg SquareD QO series) and normal fuses would probably pop first (but maybe not).

Slow blow fuses and thermal breakers would probably be sufficiently slow to react that the diode would pop open leaving the chassis hot and ungrounded. Ouch.

 

[Skogsgurra]

I had this discussion once (but I never saw that HP appnote) and my "pure signal" collegues told me that 'diodes almost always fail shorted' - which they thought was enough safety. So, instead of discussing safety, we discussed statistics. How many diodes fail open? How many fail shorted? We solved the problem in a much better way. Heavy copper bar along the rack.

Gunnar Englund

http://www.gke.org

 

[stevenal]

Sounds like some of the galvanic isolators used on boats. Keeps stray DC currents from circulating between boats and causing corrosion when hooked up to marina shore power. I don't know for sure if the commercial UL/ABYC units use diodes for isolation, but the home brew units do. I have plans in one of my marine books.

 

[VEBill]

'diodes almost always fail shorted'

My understanding is that overvoltage (more common) would typically puncture the junction and short the diode, but that overcurrent (less common) would melt something (perhaps a bond or an internal wire) and would open the diode.

In this case we ARE talking about overcurrent as the diode tries desperately to blow the fuse or breaker.

So I wouldn't rely on overall statistics of diode failures.

Perhaps the diode bridge selected was specifically chosen for its robust overcurrent characteristics and its ability to reliably kill off even the slowest fuses and breakers.