About Ground Loops & Oscillations - Stability

[Logistics]

I'm defenitely no pro, but I've done alot of research when it comes to opamps, whether it be audio, video or just some other funky IC of special usage. One thing which has always been important in many application briefs and datasheets is the need to eliminate ground loops. There are tons of articles on grounding techniques. Ground loops are known to cause oscillations in IC's and noise problems, usually from one subsection of a PCB being forced to find another path to ground that it wasn't meant to use. This is known to cause audible distortion in audio IC's and visible distortion in video IC's. It seems reasonable to conclude that this can also cause instability issues.

I'd like to see someone try to improve their overclocking results by eliminating ground loops in a PC. PC's are a gigantic mess of ground loops. You have earth ground coming into the PSU, being connected to the PSU's PCB and also usually tied to it's chassis. Naturally, you pass ground from the PSU's PCB to the motherboards PCB, but then you also have the chassis of the PSU being tied to the case. The case is tied to the motherboard and all it's add-on cards because they attach to the case. This can only be potential for something bad. This doesn't even take into account ground trying to pass out of or into all attaching cables, sound, video, etc., which throws even more fat on the fire; plus your add-on cards having potential for grounding through eachother. And since you obviously aren't going to lift ground from the spaghetti coming out of your PSU, it should be removed from the case.

Any thoughts?

[Per Hansson]

Re: About Ground Loops & Oscillations - Stability

If you remove grounding from the case and the hot wire comes loose and touches it you will be dead when you touch your computer, maybe not all that fun?

[Logistics]

Re: About Ground Loops & Oscillations - Stability

Since the mains run inside the PSU chassis, as long as the PSU chassis is grounded internally, that eliminates that problem.

[mutant]

Re: About Ground Loops & Oscillations - Stability

I'm no pro either but I think you can get the results your looking for if you remove the case entirely. Like when you benchtest a new system to make sure everything works.
Maybe get one of those clear acrylic cases.
But would it really make any difference? What about EMI?
Cooling might be better.

[Logistics]

Re: About Ground Loops & Oscillations - Stability

What about EMI?

That's why I wanted to leave it in a normal case. You could breadboard everything, but it wouldn't have the case acting as a sort of RF trap. This is why it would actually be ideal to still have the case grounded, but not to anything except for the PSU chassis, so that as much of the mess in the air as possible hits the case and is passed to earth rather than ever trying to find it's way through the cards, drives or motherboard. I need to build a faraday cage for my PC.

Cooling might be better.

Mmm... how does cooling solve oscillations caused by ground loops?

[Per Hansson]

Re: About Ground Loops & Oscillations - Stability

Isn't the negative DC plane in a PSU grounded though the PCB screws?

[mutant]

Re: About Ground Loops & Oscillations - Stability

Mmm... how does cooling solve oscillations caused by ground loops?

I should have specified more clearly. What I meant was for the overclocking. Usually better cooling for most components.

I like the faraday cage idea.

I havent read anything about anyone trying what you suggest but I bet the military has hardware thats designed like that.

[999999999]

Re: About Ground Loops & Oscillations - Stability

The answer is no. Do not think of a computer like you'd think of a power amp with vast differences in low level analog signal input versus high current ground returns from the load.

There are many many different hypothetical ground loops in a PC and it is not feasible to try to eliminate them, and no scientific backing in it being useful to do so for overclocking purposes.

On the other hand, if you wanted to apply a PC scenario to analog audio, it suggests the cleaner analog audio output is obtained by having a slight impedance from the ground plane to the audio card (instead of an integrated audio solution) and some LC filtration on that audio card.

[Super Nade]

Re: About Ground Loops & Oscillations - Stability

A motherboard should not have those problems or it should be minimal at best if signals traces are routed well and if there is optimal use of ground planes. These are 6-8 layer boards, so ...