Should a computer case have electrical continuity to ground?

themaidenmaniac

Experienced
Hey guys, troubleshooting a lot of EMI from my pc and how to resolve it.

I was wondering if a computer case should have continuity through its panels and through to the ground pin on the PSU ?

(Understand a case should act as a Faraday cage to some extent as well as grounding the unit?)

Thanks :)
 
This is one of those “the thinking has already been done by people way smarter than us” type of deals.

I have yet to run into an EM issue thst a decent UPS cannot solve.
 
How does a UPS solve a PC's EM issue?
It can isolate the system from mains power, which is typically the primary source of EMI.

The motherboard, PCI and PCIe cards also typically ground to the case via their back plate brackets.

The back plates and motherboard standoffs can create multiple paths to ground which can create ground loops in the PC. PC designers often don't care about that since they are more concerned with digital logic signals, not analog audio signals. In theory, electrically isolating the mobo and PCI connections to the case might help with audio noise issues. That would make the PSU connector's ground wire the single path to ground. Nylon standoffs and strips of tape on the backplates would be one way to do it.

It also depends if OP's issue is an inny or an outy problem. Is the PC creating EMI that other devices are picking up, or is it picking up EMI from other external sources? Your approach will likely differ depending on that answer. Noise issues can sometimes be bitch to figure out and solve.
 
Yeah for audio, mains 60 cycle hum is what people tend to fight the most. There's lots of other potential culprits though. Bluetooth, WiFi, cell phone signals, TV and radio stations, microwave ovens, TVs and monitors, etc. Basically anything that produces electromagnetic radiation (light, radio waves, micro waves, etc.) or electromagnetic fields (magnets, current through wires, coils, transformers, etc.) has the potential of causing some kind of interference.
 
In my experience, electromagnetic interference (EMI) is caused by electromagnetic radiation, and it’s picked up by the most EMI-sensitive part of the system (your guitar) that’s hooked up to the most downstream gain (the rest of your gear).

It’s true that there are unwanted signals that can get in through the power line, but most EMI can be stopped in its tracks by unplugging your guitar (not a very practical solution, but it does show where the EMI is getting in).
 
Thanks for your ideas guys, interesting idea from you as well Mr Fender!

I am going to isolate the PC on a different power ring tonight, but I strongly suspect the EMI is being generated internally on this ocassion.

It's a bit annoying as it is a new build PC, with a sturdy fractal design (unrelated company!) case. I wonder how well the fractals actually shield EMI. I did a bit of reading online and apparently thermaltake specifically link all the case panels on their mid to higher end models (not sure if it helps matters though!).

It can isolate the system from mains power, which is typically the primary source of EMI.

The motherboard, PCI and PCIe cards also typically ground to the case via their back plate brackets.

The back plates and motherboard standoffs can create multiple paths to ground which can create ground loops in the PC. PC designers often don't care about that since they are more concerned with digital logic signals, not analog audio signals. In theory, electrically isolating the mobo and PCI connections to the case might help with audio noise issues. That would make the PSU connector's ground wire the single path to ground. Nylon standoffs and strips of tape on the backplates would be one way to do it.

It also depends if OP's issue is an inny or an outy problem. Is the PC creating EMI that other devices are picking up, or is it picking up EMI from other external sources? Your approach will likely differ depending on that answer. Noise issues can sometimes be bitch to figure out and solve.
 
I wonder how well the fractals actually shield EMI
Can you describe the symptoms of the issue? Unless something is broken/unplugged/cold soldered in the inside of your axe-fx, your guitar is the thing sucking in EMI.

By the way a UPS can/could fix ground loop issues for sure. I really doubt it helps against EMI, not on a "I can hear those with a guitar" order of magnitude.
 
Hi Gira- not a Fractal audio issue there- i was referring to my case which is a "Fractal Design" R6 :)

Also- just in case anyone googles and finds this thread in future- this is a useful PC EMI reduction guide for even the most savvy builder:

1. Spread Spectrum Clocking
AMD processors are designed to run with spread spectrum clocking enabled. Ensure that the
motherboard BIOS has enabled the spread spectrum feature of the system clock generator.
Enabling the spread spectrum setting often lowers frequency amplitudes by more that 5 dB.

2. Disable Unused Clocks
Clock signals that have no load can have high levels of ringing that can lead to EMI problems. The motherboard BIOS firmware should be programmed to detect and disable unused memory DIMM and PCI clocks.

3. Processor Heatsink Fan Cable Routing
A problem sometimes encountered with the processor heatsink DC fan cable is the large loop formed in its routing to the motherboard connector. Shorten this cable length to reduce the loop area as much as possible.

4. Power Supply Cable Routing
Historically, the system power supply cable has been found to be very susceptible to picking up EMI energy from within the system and coupling into the power supply and then onto the AC power cord. Keep the power supply cable against the metal chassis and as far away from the processor, memory DIMMs, and VRM components as possible. Fix this cable routing in place with plastic cable ties.

5. Other Internal Cable Routing
Cable routing inside the system should generally be routed along the metal chassis and away from EMI sources such as the processor heatsink, clock modules, memory DIMMS, VRM components, and high speed VLSI modules. Internal cables that connect to front I/O ports such as USB and Audio are particularly sensitive. The use of a shielded cable or a ferrite core or both over these internal cables can be effective at reducing EMI.

6. Rear I/O Connector Shield
One common problem in many computer systems is poor electrical contact between the I/O connector metal housings, the metal I/O connector plate, and the cut out in the system chassis wall. This problem can be due to soft metal being used in the I/O connector plate or to an insufficient number of spring-finger contacts. A solution is to use a hardened stainless spring steel with a sufficient number of contact points to the I/O connectors and the wall of the system chassis. Each I/O connector housing should have at least two contacts and as a general rule, there should be a contact point at least every 1 cm between the I/O connector plate and the chassis. As a quick remedy if this condition exists, a die-cut, conductive, foam gasket matching the I/O connector pattern can be added to improve connector grounding to the chassis.

7. Chassis Shielding
All chassis designs have gaps and seams to enable assembly and option installation. From an EMI standpoint, however, some gaps are worse than others. The important dimension of a gap or seam is the longest dimension. If you can slide a piece of paper for several inches along a seam, that seam could cause an EMI problem. Spring fingers or foam EMI gasket can be used to seal these gaps or seams.

8. Processor Heatsink Grounding
Although grounding of the processor heatsink has not yet been required on any AMD Athlon™ 64 processor-based systems, grounding of the processor heatsink can further lower the harmonic EMI levels of the processor. Many AMD Athlon 64 processor-based motherboards contain grounding pads around the footprint area of the processor. These grounding pads can be utilized to ground the heatsink to the motherboard ground.

If excessive system level EMI radiated emissions exist after attempting all the listed EMI reduction techniques, then more extensive remedies may be necessary. First, determine if the emissions emanate from the system I/O cables (including the AC power cord) or from aperture leaks in the system chassis. If EMI emissions emanate from a particular I/O cable, then improved filtering or cable shielding may be required on that cable. If EMI emissions emanate from slots or seams in the chassis enclosure, use copper tape across the apertures to improve shielding effectiveness. If copper tape reduces emission levels to a satisfactory level, then chassis sheet metal changes or conductive EMI gasketing can be added at that location
 
Ripped the machine apart tonight and tested a few things.

1. Unscrewed the plug and removed ground. Issue unchanged

2. Removed all cards from machine. Issue unchanged.

3. Tested various points and components around the case for continuity. Largely pretty good.

I guess the board or the PSU are causing the chaos. It's a seasonic psu which as far as I knew have great parts and the mobo is an Asus. However, fairly cheap b350 board.

2 options I see are flipping the board, or working out a way to copper line or EMI foam the case with connectivity to the cases ground. Easier said than done as the sides of the case are coated.
 
Back
Top Bottom