Westinghouse Radio Frequency Monitor

[ScottyUK]

Do any of you have experience of a device called a Radio Frequency Monitor, specifically the RFM Series II, manufactured by Westinghouse? Ours dates from the early 90's. The purpose of the monitor is to detect a change in the background RF level from the generator as a warning of arcing activity somewhere on the generator.

The unit has only been recently commissioned after solving a design problem with the power supply which affected the self-calibration feature and made the unit unusable. It's one of those jobs which a lot of people have fiddled with over the years but now it is working after a concerted effort to fix it properly. Unfortunately there are some strange things happening.

For the benefit of those who haven't seen one of these units, it consists of the following main parts: a heavily insulated RF CT on the generator earthing conductor at the neutral point; a head end pre-amplifier local to the RF CT; about 100m of co-ax in a grounded steel conduit; a narrow band filter at either 3MHz or 4MHz; an averaging circuit and logarithmic amplifier driving a panel meter.

Every day we see an increase in RF activity of about 40dB at roughly the same time and which lasts for a period of about eight hours, then a lull for about four hours and another noisy period of about an hour. This sequence starts at about 1600hrs. Through the day there is nothing but the RF background level detected from 0500hrs through to 1600hrs. This pattern is repeatable and has been trended.

I've pretty much eliminated the mains supply as a source of the noise with an input filter giving 70dB attenuation in the frequency range of interest. The accessible equipment appears to be behaving correctly and has been calibrated, and when a signal generator is used to inject a signal the unit behaves as would be expected. The RF CT and head amplifier are within the neutral cubicle of a running generator and are thus inaccessible. No machine parameters correlate to the RF noise. We can not detect any significant discharge activity using TEV techniques, and the predictable nature of the RF signal does not really support discharge as a likely source. Changing detection band from 4MHz to 3Mhz has no affect on the signal; it is present at both frequencies.

I'm looking for ideas about how to track down the source of this RF signal. EMC type work isn't really my field of expertise, and neither is radio transmission. It has been a long time since I studied antenna theory!


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

Hey Scotty, Hi!

Here are just some wild speculations that might be useful.(or not)

1) Could it be the fact that during the day millions of people are flipping switches to the extent that it is continuous. During the wee hours customer switching drops substantially to the point that they become nearly single detectable events.

2) Voltage rises slightly allowing corona that doesn't normally occur to happen.

3) That is very close to the period that the ionosphere changes radically getting lower and greatly changing the propagation of radio waves. Perhaps this allows your generator to start receiving something.

4) There is a major shift in the earths magnetic field every time the Bay Area Rapid Transit (BART) changes its load dramatically like during the weekend. Could you have something like that nearby? Electric public transport? Or a radar station changes modes?

It might be interesting to have a radio person bring in a receiver and see if it's possible to detect if 'something' can be received.

Any way you cut it.. It's a cool mystery!

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[ScottyUK]

Hi Keith,

Terminal voltage is pretty much rock steady at 16kV. We're base loaded at the moment and the grid voltage is solid at our location so we don't see anything change much. Terminal voltage movements would show up either directly or in the form of a change in excitation or change in reactive load. Three instruments say it isn't happening.

We have some horrible neighbours including a massive chemical plant and a steelworks. We think can see the disturbance caused by the massive rolling mill drives, albeit at a very low level, by the response of our AVRs. This doesn't correlate to the RF disturbance. There is an arc furnace across there but we don't think it is in cyclic operation on a daily basis. It would be a damned good candidate otherwise. God knows what goes on in the chemical plant, but although we're aware of a number of very large drives we don't know of anything likely to be pushing out RF.

I am inclined to think this might be a radiated problem rather than conducted. I am struggling to believe that RF would propagate through at least one distribution substation and two transmission substations between us and whatever is causing this. I could however be totally wrong about this. I wonder if my old longwave radio receiver will pick up anything?

The changes in the atmospheric conditions sounds like an interesting theory. My only efforts relating to atmospheric conditions was a brief idea that cooler evenings and high rel humidity might be provoking partial discharge activity, but that appears to be a blind alley. Any idea where I can read up on how the atmosphere affects radio transmission or what to Google for?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[lz5pl]

Looking on the time period the reason may be receiving of some strong radio station. Nature of propagation of radio waves is that during the night exactly this band - up to 4-5 MHz - opens, but in daytime it closes. Please check whether you have some strong HF radio station nearby - up to 50 km from your site.
One of the most popular radio amateur's band is 3.5-3.8 (in USA 3.5-4.0) MHz and best contacts are possible during the winter and between 20.00 and 07.00 o'clock.
Anyway for me is strange how such signals enter the input of RF amplifier. One possible reason may be bad grounding of the shield of coaxial cable.
I am just radio amateur, not RF engineer. I would suggest to put this question also on Communication & Signal Processing engineering Forum.

------------------------
It may be like this in theory and practice, but in real life it is completely different.
The favourite sentence of my army sergeant

 

[itsmoked]

"radio propagation times night"

I would be tempted to contact a local HAM group. A lot of these clubs have some exceptional talent that would jump at the chance of helping "a powerplant". Fodder for club talking.

I'd explain this to them and see what happens. Someone could have a clear idea of what causes it because it's a spectrum they frequent and understand. They could even have a nice fix for you. Donate something to their club if they can help you.

It could be your 16kV cable is a tuned length or something.

Could be none of this at all... But many times using high gain amps hooked to obnoxious transducers, I've had radio reception that caused perplexing errors until understood.

When our area had an early warning radar I also caught it many times, finally recognizing the every 5-6 second sweeping blast in my circuits.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[ScottyUK]

From my fumbling through old university notes it seems that if my co-ax is acting as an antenna then I have, at 3MHz, a cable run of roughly one wavelength. Presumably any interfering signal on the co-ax is common mode where the signal I'm interested in is differential. I wonder if I can raise the CM impedance or detune the 'antenna' to weaken the signal? Ferrite ring on the cable?

I guess the best solution could be to mount the equipment up near the head end amplifier where I can probably reduce the coax length down to 5m or so. I may try that in the New Year. I am open to any ideas at all so please share any thoughts you might have.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[slavag]

Hi Scotty.
May be possible use some coax cable with double shield
as used for the Rogovski coil sensors.
I try found some information.
Regards.
Slava

 

[slavag]

Scotty next simple idea.
It's remaind REG216 earthing, that used against HF, it's help me several time.Try for used eq and cubicles, braided copper strip 20mm. according to REG216 cubicles.
Please pay attention on connection of special cables to REG216 AD card or provide something like to this.
Please see attached.
Regards.
Slava

 

[ScottyUK]

Hi Slava,

Thanks for the comments.

The coax screen at the generator end is connected to a dual 240mm² main frame earth direct into the station earth grid. At the other end it is tied to one of the building earth bars which is only a 95mm², again direct to the earthing grid. The conduit is earthed along its length by virtue of being fixed to the structural steel work. I can't get at the screen of the coax at the generator end right now because of the adjacent live parts - I will verify the condition of the connection at the next outage. The other end looks in good condition.

Regarding double screened cable, I think the interference might well be a common mode signal which is being coupled into the screen and core. Replacing the cable is an awkward job because of the high level route - I want to avoid it if at all possible. Maybe I can run an alternate cable along a more accessible route and try that, but I will lose any benefit the steel conduit is giving me.

I like tha copper earthing clamp - must be a later design than our elderly (early 90's) REG216 system.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[dpc]

I guess I'd be more suspicious of any other wiring entering the device and less so the coaxial cable. Any wiring can act as an antenna, including power wiring, output signal wiring, etc. Ferrites on any other cables could help with common-mode noise. The filters on the mains circuit will take care of a noisy sine wave, but may not help with common mode noise rolling down both conductors.

Ferrite beads can be also be placed on the outside of the coax to help with any common mode noise. Google for "coax choke balun". (Baluns are normally used to transition between balanced feedline and unbalanced coaxial cable, but the choke balun will help to eliminate any common mode signals.)

http://www.palomar-engineers.com/1_1_Balun_Kits/1_1_balun_kits.html

Any military radar installations nearby? Airports?

I'd suspect a fairly nearby source.

This may be too simplistic, but if you disconnect the coax from the amplifier, does the noise go away?

 

[ScottyUK]

if you disconnect the coax from the amplifier, does the noise go away?

Yes. I've had it running quite happily in-situ with a signal generator connected in lieu of the coax and all other wiring undisturbed. No problems. I've also tried connecting a few yards of un-terminated coax to the unit and holding it in various orientations. This also didn't pick anything up. From this I conclude that whatever the signal is, it is coming down the coax, either picked up by the cable or the signal is somehow being coupled into the pre-amplifier.

There is a tower on top of a nearby hill which has a number of what I think are microwave dishes, but I didn't really think microwave stuff would cause a problem: I'd expected the beam to be too narrow to affect us. Could the modulation on the microwave signal be in the frequency range I'm having problems with?

The nearest airport is civilian and is a fair distance away - 20 miles or so. Is radar affected by the atmospheric phenomenon Keith mentioned earlier?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[lz5pl]

Microwave dishes and radars should not disturb your equipment - their frequency range is much higher than mentioned 3-4 MHz.

Source should not be very close to the site, as problems start during the night - when propagation of radio waves allows receiving of far signals.

By the way - if you can change tuning of narrow band filter to let say 14MHz (I doubth it is possible} -
I am almost sure you will face the opposite case: RF noise in daytime and no problems in the night.

After re-reading of your first post I think that we could reject coaxial line as a source, as it is in grounded metal conduit. For me the most probable point is the sensor. I never have seen such system, so I have no idea how such RF CT is looking like. But I suppose it picks-up RF noise not only from the generator, but also from outside - so the CT works as an antenna for your system. Would you clarify whether this RF CT is located inside of screened structure together with the generator, or radio waves outside of the building can reach it too. Also - when this problem became - from the beginning or recently (possibly after some works around the sensor)?

Can you post some oscillogramms of the RF signal? There are guys who could understand it's nature - pure "white" RF noise or generated by some communication system.

------------------------
It may be like this in theory and practice, but in real life it is completely different.
The favourite sentence of my army sergeant

 

[ScottyUK]

Hi lz5pl,

I off work until January now using up vacation time so I can't get waveforms at present. I'll put some up if I'm passing work. I will scan some of the diagrams too. The unit has only just been commissioned so I don't know if this is a long term problem.

The RF CT is within the earthing resistor housing so it is screened to some extent by the outer casing of the NER. The sensor has not been inspected for a long time - I need to do this.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[dpc]

If you can access the coax terminations, the ferrite bead approach wouldn't cost too much to try and I can't see it causing any major side effects. Might try it on both ends.

 

[Hoxton]

Scotty.

If you are where I think you are, you are just north of a very large radar station......

I'm not surprised at what seems to be pick up, but the time factor seems curious. Is it related to any shift patterns at the steeel works or chemical works? I see I a possible link here.

My view is that its a resonance. Somehow you have a "ring" and the time thing is exiteing it.

 

[ScottyUK]

Hi Hoxton,

We haven't been able to pin this down to any shift pattern, although the 1600hrs - 0000hrs does look suspiciously like an 8 hour shift. The one hour block from 0400hrs to 0500hrs doesn't fit though. The times don't fit well with the shifts we're aware of.

We're located here:

http://www.flashearth.com/?lat=54.556488&lon=-1.120263&z=15.2&r=0&src=ggl.

Where's the big radar installation?
I know about the (supposed) microwave relay tower about a mile south of us here on Eston Nab:

http://www.flashearth.com/?lat=54.556488&lon=-1.120263&z=15.2&r=0&src=ggl

Any idea how easy it is to reduce the effect of a radar signal if that is the cause of our problems?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[Hoxton]

Scotty

How far is it from Eston Nab to Snod Hill?

Actually, I do not think they are bothering you, since they seem to have gone passive.

Looking at your timing, and the time of year, Heaviside layer? Are we looking at atmospherics?

 

[ScottyUK]

Oh, RAF Fylingdales! I'd forgotten about that place completely. 'Snod Hill' had me wondering for a moment or two: seems in bygone days the RAF couldn't bring themselves to name a base "RAF Snod" so they chose the name of a neighbouring moor instead. It isn't that far away in a straight line, only about 24 miles. I have no idea if it is active or not. It's not the kind of place I would ring up to ask either: who knows what Government lists of subversives and threats to national security I might end up joining!

I will have to go to Google to investigate the Heaviside Layer and its ilk. Totally out of my field of expertise. I will be back with more questions I suspect!


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[electricpete]

On the remote possibility that it might be a real indication (vs noise), do you have daily temperature excurions in a similar same pattern (I realize it's a pretty bizarre pattern).

I see fairly dramatic change in pd with temperature on our air-cooled motors. As temperature goes down, the pd generally goes up.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[ScottyUK]

Hi ePete,

Thought you were quiet!

We do expect to see a PD increase as night falls and the temperature drops. The RH and dewpoint both change at this time. At first I thought rain might be the cause, finding its way in past a damaged seal on the isophase bar and causing PD on one of the resin insulators. This has happened once before. When the rain theory was disproved I started wondering about the effects of the evening drawing in. In this instance I think we can discount PD activity because it is so predictable. The problem occurs at a couple of minutes before 1600 hours, regardless of the ambient weather conditions which have included mild and dry, bitterly cold and dry, cold and wet, mild and wet. I had even considered things like the backup processes which run on the IT equipment but none fit the timeline.

It is puzzling, but probably also has a very simple explanation. Unfortunately I'm just perplexed at the moment!


----------------------------------

If we learn from our mistakes I'm getting a great education!