RF Noise

[ccdubs]

Hello,

We have a number of T100MD888+ installed in remote sites that communicate over a VHF radio network. We have always noticed at some sites that there has been some interference to the radio that we have always considered as being from an external source. Recently this level of interference has become problematic and we have employed RF consultants to investigate the issue. The result is that the interference is coming from the T100MD888+ PLC and more accurately their "RF sniffer" identified the noise emanating from the CPU.

Our radios work in the VHF spectrum at around 150MHz. I was originally skeptical to the PLC being the issue but based on their analysis, I tend to agree (i.e., removing power from PLC stops the source of interference).

Are you aware of this product generating RF interference? Do you have any recommendations as to how we could reduce this interference? It is unusual that it would seem that things have degraded over time (although that could be explained by other influences) but is there anything within the PLC that may degrade and cause RF noise?

Any help, as always, is greatly appreciated.

[support]

The T100MD888+ CPU runs at 16MHz. It is not a direct radio emitter but as with all digital electronics the switching signal in the circuit can emit radio frequency and the higher harmonic could coincide with your VHF radio band.  

Is your radio network housed in the same control panel as the PLC? If your control panel is metallic the simplest solution is to move the radio equipment to a different box from the PLC and ground the metalic control panel that contains the PLC. That forms a Faraday cage to block most RFI signal.

If that is not feasible, would it be possible to place some grounded metallic shield over the PLC which can also block the RFI.

[ccdubs]

Thanks for the answer. It isn't that easy to screen the PLC in the way you mentioned (especially considering there are 20 units that need this modification). Our tests show it is just as effective to add an additional sleeved screen over the radio aerial cable within the cabinet and we will be doing this instead.

FYI, we believe that the reason this has only shown up now after four years of acceptable operation is that some other external interference has raised the noise floor which means we have lost some of our signal strength margin and that has pushed the reception tolerance to the limit.

[ccdubs]

An update on this issue....the aerial cable screening wasn't as effective as thought. Further testing provided interesting information. When the PLC is powered from a separate source, like a battery, the interference goes away. This would say to me that the noise is not airborne from the plc directly and is rather being radiated down the power supply cables.

Does this info generate any other solutions? Using a separate supply isn't really an option due to space constraints and backup requirements.

[garysdickinson]

It is not unusual for RF noise to conducted and radiated via power lines. The usual remedies come in the form of filters to block the transmission of the RF noise.

Ferrite RF suppressors that clamp onto the outside of AC power cords can help.  Companies such as Corcom have made a living selling filters to block RF nose from entering/leaving enclosures via AC wiring.

If the problem is on the DC side of the power supply, I would add RF Filters both at the power supply at where the power connects to the PLC.  These are usually in the form of ferrite rings that BOTH power supply leads pass through.

It is possible for the coax that feeds your antenna to be both radiating RF as well as picking it up.  In a properly working system no RF travels on the outside of the coax. It is common practice to slip ferrite beads over the antenna coax to prevent RF currents from traveling on the outside of the coax.

As you've figured out there is a bit of black magic in tracking RF noise issues.  You might want to track down a local seasoned (old) ham radio guy to help put this to bed.

Best of luck,

Gary d

[ccdubs]

Thanks for the response Gary.

This may be worth a go and I'll be sure to post the results.

[support]

Are you using a switching power supply? A switching regulator uses high frequency (100s of KHz) PWM in its operation and if the output filtering circuit malfunction you are going to see a lot of high energy noise that could affect your radio. If you replace the power supply (especially with a linear power supply) and the noise goes away then you can confirm that it is the power supply noise.

[ccdubs]

Well it has been a while but I can now provide an update.

Gary: if you are still there, the RF chokes helped but didn't solve the problem.

Support: you were right. After a lot of trial and error testing we found the noise to be coming from the switch mode DC-DC converter. It wasn't that easy to find and we were mislead by the fact that unplugging the PLC made the noise go away...this still isn't really explained. Perhaps the noise from the DC-DC converter is load dependent (i.e., PLC small load is enough) or perhaps there is an interaction between the two that amplifies the noise to the radio.

[garysdickinson]

Switching supplies are very sensitive to output loading. If you remove the load from a switching supply, it doesn't need to do any "switching" to keep the output at 24v.  A switcher that is not supplying current should be electrically quiet.

In the "old days" switching power supplies were always spec'd with a minimum load.  If you did not draw this minimum load, then the supply could not maintain it's output voltage without big nasty voltage spikes.  

Another issue with switchers is that they can generate electrical noise on the output DC side as well as noise on the input AC side.  Since you are working with sensitive RF receivers, noise on either side of the power supply can be an issue.

If it were me, I'd replace the switching power supplies with linear power supplies.  I'd actually prefer running stuff off of car batteries.  No electrical noise with lead acid batteries.  I do realize that this might be a little inconvenient, but I have done once or twice.

If you stick with the switching supplies, then you will need to fix the RF noise issues.

As for the observations that RF chokes helped but didn't 100% solve the problem, this is actually good news:

1.  It's the 2nd thing that you've done that made a difference.
2.  It's the second thing that clearly points at the power supplies as acting as RF noise generators.
3.  It points up that RF chokes are not an exact science.  There are literally hundreds of different formulations of the ferrite material used for there devices. Each formula is effective at different frequencies.
4. Even if you had the best material, you may need to use more chokes (more inductance)  You may need to pass the wires through the choke more that one time.  The inductance of the choke goes up by the square of the number of turns.

I've got lots of questions/suggestions:
1. Did you put RF filters on the AC Side of the power supply?
2. Did you run both legs (GND and +24v) through the RF choke?  This is an attempt to deal with common mode radiation.
3. Did you run only the +24v leg through the RF choke?  I know this is sort of the same question, but there is another issue.  It is possible to saturate the RF choke with too much DC current.
4. How long is the wires between the power supply and the load?
5. Are the power supply output wires twisted together?  Are the power supply wires running inside shielded cable or metal electrical conduit?  All of these questions are about minimizing common mode radiation from the supply wiring.

If you haven't found an old ham radio guy in your area, it's never too late.

Best regards,

Gary d