Would More Ground Rods Help?

Adding more ground rods would probably not help much to reduce the RFI signal problems. I would suspect that the antenna of the AM transmitter has been damaged so that the antenna's radiation pattern is dumping too much energy down to the local area instead of farther away. If your calibration lab had a spectrum analyzer that covered the AM band you could identify the actual frequency and amplitude that was now bothering your testing. Identifying and isolating an interference signal is almost a black art that requires patience, proper instrumentation and a helpful supply of luck.

Good Luck.

If your calibration lab had a spectrum analyzer that covered the AM band you could identify the actual frequency and amplitude that was now bothering your testing.

We have a spectrum analyzer. It can identify some of the frequencies coming in, but cannot tell the amplitude. That would require an antenna that is flat over all frequencies. Such an antenna does not exist. I used a loop of 4 turns with a radius of 18". This is very inefficient at AM frequencies, but shows the AM station nearby to be the largest signal anyway. This is not proof, but is a good guess. If the station would reduce their power at night, I could determine by nighttime measurements, but they do not.

I do not agree that the signal is coming from the floor since technically it simply cannot be as such, most probably concentration of stray signals at floor level is higher than witnessed else ware, as for the radio stations they keep on upgrading their equipment hence now a higher radiation level can be well understood.

It seems you have overlooked the standard earthing requirements; Metal walls of the room are to be grounded to a dedicated earth pit, Electrical systems are to be grounded to another earth pit and finally both the earth pits are to be interconnected with a jumper wire for bonding.

Interconnection of metal walls and the electrical system earthing and later leading it to the earth pit is problematic on various accounts.

I agree that it is probably not through the floor but possibly a constructive reflection off of the floor that is increasing the signal strength.

Similarly, there maybe a new building in the area that happens to be facing the right direction to make a constructive reflection than what existed before.

I do not agree that the signal is coming from the floor since technically it simply cannot be as such

Then explain why the signal is stronger by the outer wall than by the door with a glass window. Dry earth is not a very good conductor, so it doesn't make a good Faraday cage. We are on a hill above the transmitting antenna. This type antennae have extensive underground wires to make a ground plane antenna.

It seems you have overlooked the standard earthing requirements

I haven't overlooked anything, I didn't design the building. Management wouldn't spend the bucks to do it right when it was built. The inner walls and outer walls are tied together with electrical conduit.

You have avoided answering the question.

As I said, it is likely a ground reflection. Have you ever wondered why your car radio picks up an AM station so well with a metal rod antenna just under a meter in length but if you attach the same length rod to the AM antenna input of a tuner you get a lousy signal? The body of the car is acting as a reflector to increase the signal picked up by the rod antenna. Now this is not a focused reflection like a parabolic dish but still a reinforcing reflection. I expect that you will get a similar boost in this station's signal level just outside your shielded room.

There's also another possible reason that you now cannot do your test. Many improvements in transistor fabrication have happened over the years. Most transistor specifications are actually guaranteed minimums and nominal component values, not the actual values. So the new transistors are permitting signals that were previously attenuated from propagating through the circuit. As PWSlack asked, could you tell us about the test that no longer works?

I am at work where I can't log in. It can't be a reflection because the metal walls and roof are shielding it out. I am 100% sure it's coming from the floor.

The test I referred to is calibrating a resistive divider that has 1V input and 1mV output. The problem is at 20kHz where a high frequency ratio transformer is the standard. A DVM reads the standard and then the divider. The reading from the ratio transformer is too high; the divider reads normally.

I have made the readings significantly better by putting a toroid on the BUSS cable between the computer and the test equipment and using shorter BUSS cables. If I put another toroid between the AC source and the ratio transformer, that also improves the readings.

I have measured the field strength on top of a metal bench and found the AM signal to be significantly lower there (about half the voltage). When I ground the bench to the wall with a heavy braid wire, the field strength drops to half again. I plan to put an aluminum sheet on the bench where the problem measurement is, and ground it to the wall. This looks like a better plan than extra ground rods (I have data to show it helps).

StandardsGuy

If you picked up the interfering signal on your AM receiver, you must know the frequency/wavelength of this station. Please tell us. For a specific frequency, a simple parallel inductor & capacitor will make a much better filter than a toroid through which a susceptible cable is passed. This filter would be in a shielded box, connected by shielded cable into the 20 kHz meter input (or into the 20 kHz source output). Source or meter or both are probably susceptible to the AM station frequency and reducing that weakness is better than shielding a whole room. In effect, you have a metal cube - a "shielded room" of which one side of the 6 is open - of course the AM penetrates the floor!! This supposes you have filtered the AC mains incoming to your room. It may be the mains wiring which brings AM into room (and makes a good induction coil to couple to other wiring)- and mains wiring is often at floor level.

<...A procedure in the unshielded room that has worked for 20 years does no longer work properly...>

Please describe the equipment and the way that it no longer works properly.

Yes it can be 'coming up from the ground', as transmission is a capacitance effect.

It sounds to me as if the shed earth has been reduced, or lost, and it's floating over earth.

You might look at the phase difference between shed and floor, and floor to a temporary remote earth stake, to determine if the floor signal is a 'skin effect' or aligned to the transmitters earth field.

If any 'yes' above; ground rods would help.

But also;

Has it been dry for a while? Try soaking the ground around the shed and any areas where 'earths' are known to be.

You might look at the phase difference between shed and floor, and floor to a temporary remote earth stake, to determine if the floor signal is a 'skin effect' or aligned to the transmitters earth field.

How would you propose I do that?

Has it been dry for a while? Try soaking the ground around the shed and any areas where 'earths' are known to be.

Yes, for a long time. This may be the thing that changed. My last measurement for this teat was early August. It rained 2 days ago, and the signal at the east end has dropped 2-3 dB, but the west end has increased about the same, so now both ends read the same. However, I may not have repeated the tests exactly the same, so it may be measurement uncertainty.

I give you a GA for answering the question, and for suggesting wetting the ground around the building (which I am unable to do). I expect lower levels when the snow falls.

-S

"How would you propose I do that?" - dual trace oscilloscope.

Thanks for the GA 'wish'

dual trace oscilloscope

So do I put a metal plate on the floor and measure to that?

Thanks for the GA 'wish'

Sorry, I actually voted now

Most shielded room are made with metal mesh. The best type use solid copper as the shield, the mesh needs to be soldered/welded at the cross-over points. But mesh is cheaper - but may not be done properly

Low grade copper mesh that has not been soldered/welded can corrode, and leave you will many little dipoles = holes in the shield

The floor also needs to be shielded. How was it shielded? A fast test would be to buy a couple of rolls of aluminium foil and cover the floor--see what happens. Then cover a wall, and then another, and so on. These need not be robust or permanent, just to find the pathway in. The cost of the foil for this test is quite low. Once the problem has been located, a permanent solution can follow

The comments are interesting; however, a few points have been overlooked. The site is on "dry" land which is elevated.

Has the ground water table changed? With the ground rods disconnected, what is the ground to ground potential? An effective ground must be less than 0.060V. What is the earth to electrical grid potential? You may need inductors in line with the conductors, to keep RF out of the equipment.

A metal building is probably gfalvanized. Start looking for corosion that can cause the bimetalic diodic effect.

Good Luck in the contest. Sporadic RFI can be tough to trace

Has the ground water table changed? With the ground rods disconnected, what is the ground to ground potential?

The water table has got to be very deep. There is no chance of hitting it with ground rods. The only ground rod grounds the electrical power. I cannot disconnect that. There is a gas line for the room heat that is grounding the building near one corner. It would be very difficult to measure the voltage between there and the ground rod which is inside the equipment room, but as I said before, there is electrical conduit connecting the rod and the building.

1. At the AM transmitting station, obviously have made changes to their antenna or increased radiated power. 2. The AM signals propagation is mainly through ground waves. 3. Please introduce ferrite beads in all test leads, bypassed by suitable value disc ceramic capacitors with minimum lead length, at the entry point of the screened & earthed metal box test jig in use. 4. Knowing the AM frequency will help in deciding the bypass capacitor value.

Have you spoken with the AM station manager yet? Have you called the FCC yet?

If the station is aware they are producing an interfering signal and they refuse to work with you to correct it, they are in violation of FCC rules.

If the AM station is outputting too much power than it is licensed for, or is generating out of band signals in excess of the permitted levels they are indeed at fault with the FCC regulations. However, just because someones equipment does not have adequate rejection of unwanted signals is not the transmitters responsibility. They may volunteer some help but are not bound to by the rules. I am a ham operator and have saw this happen many times. There was even a city near Oklahoma City that tried to pass a law that radio operation could not interfear with any equipment. If it did, the operator was held guilty. Of course the courts slapped it down almost immediately. Some complain about the courts, but I am happy we have protection from politicians with no engineering or scientfic knowledge.

It is highly unlikly the AM station is at fault. The poster already has said the managment got cheap when putting up the lab; very likely something has changed in the lab grounding system.

I'm a Ham too - KG1O - 73!

Have you spoken with the AM station manager yet? Have you called the FCC yet?

I doubt if there is a station manager. They just re-broadcast another stations programming. I have talked to an FCC agent. They haven't investigated yet. I doubt that the daytime power is too high, but they are supposed to reduce it substantially at night, and they do not. I expect that to change within a week or two.

I've seen this in the past. I used to run an AM / FM station. I would suggest checking all the ground connections in facility to include the central ground. This includes the breaker panel, and telephone / Cable boxes ect... Did this progressively get worse over time or all at once? Do you have the AM Station in your telephone as well? Ask your neighbors if they are experiencing problems. There is a possibility that the AM station is emitting out of band signals that are causing your wires, conduit, pipes or rebar to resonate. If the station has recently made changes or upgrades to digital automation or made changes to the station ground screen via construction or digging a shielding issue or a change in the station radiation pattern may exist. At any rate, I would contact the station after you checked your grounds, they may need calibration services and the trip may be mutually beneficial. Steve Donalson

I asked in post #9 "What is the frequency/wavelength you pick-up on your radio". You did not reply. Anyhow, I now know it is all-AMerican AM and somewhere 600 kHz to 1600 kHz. Up to that, I just had "AM Broadcast" 150 kHz to 25 MHz?? And got off track because AGC volts of 40 means thermionic valve/bottle/tube radio to an old Limey like me and I wondered "where is this guy?". Anyhow, at this frequency, interference by direct conduction is more likely than RF field. I feel the guys suggesting corroded bonding between building and power ground are on the right track. If not bonded, the power ground rods are just an antenna stuck into the ground potential from the transmitter! Try bonding the power ground to the metal wall where it enters the room [but this will not work if you do not have an RFI filter in the room AC power in-feed - the gremlins will be on live and neutral too]. Try bonding the chassis of your test gear/device under test DUT to the metal plate you put on the bench. Turning your room into a "second rate" screened room by plating the floor is a good suggestion from several posts - how do you know you have not already had errors with other kinds of test before??? - but not so obviously - surely, with the business you do, the management care! On that subject, aluminium baking foil (I measured some - 0.01 mm thick) is no good at 600 kHz, the "skin depth" is 0.085 mm for copper [proportional to 1/(root f)] BUT for steel, skin depth is 0.001 mm (even a steel food can I measured was 0.1 mm thick - 100 "skin depths").

The strong AM frequency is 1570 kHz and broadcasts 7kW, with 1060 kHz and 1360 kHz at lower levels from other station. You are right, it's a tube radio, A hallicrafters S-118.

I am concerned about another station. NIST broadcasts several frequencies a few miles north of here on WWV:

2.5 MHz @ 2.5 KW

5 MHz @ 10 KW

10 MHz @ 10 KW

15 MHz @ 10 KW

20 MHz @ 2.5 KW

None of these is measurable in the lab with antennas I have used. I can barely pick up 2.5 MHz at home on the radio with an inside antenna, and none of the others.

NIST also broadcasts 60 kHz on WWVB at the same location. It has 70 KW of radiated power. That frequency goes through ground much better than the AM broadcast band frequencies. Yesterday I built a 3' loop antenna with 12 turns. I added a .016 uF capacitor to tune it to 60 kHz today. I have enough amplitude from that antenna to measure the 60 kHz from WWVB at home in my basement on a counter! I will take it to work Monday and make measurements there. Tell me what metal and thickness to use on the benches for 60 kHz.

The shielded room has line filters filtering Line, Neutral, and Ground. It has an isolation transformer. The unshielded room has none of these. Power is fed in a rectangular 'conduit' about 18" off the floor. What do you think of connecting the ground lead for some of the outlets to the wall directly at the outlets?

We have been considering swapping some benches between these rooms. That is not my first choice. After the inputs on CR4 I will probably ask for line filters.

how do you know you have not already had errors with other kinds of test before???

I have had other 'mysteries' that caused changes in hookups that I could not explain. That is why I am really going after it this time.

GA to you.

The local transmitter is 1/4 mile - others are 3 miles (few). That is 12:1 in distance... Inverse square law applies 144 :1. Hence 70 kW, 3 miles away is equivalent 70/144, say 0.5 kW, 1/4 mile away. The local station is far more likely to be the problem. I suggest keeping the RFI in the trunking. Do not use any power outlets except one. Screw a 1 phase RFI filter on the wall - connect its input with (short) screened cable (screen bonded to wall or metal filter case). Take filter outlet to utility socket - a few amps would be ample to power some test equipment. To make sure trunking is not leaking, drill holes where cover and trunking overlap, join with self-tapping screws every foot(also add screws through into wall from trunking). You will need to choose an RFI filter which works at lower frequencies, some have useful loss at 150 kHz and up, others not much! If still trouble, radiation is leaking in through the floor, but remember leakage through a hole depends on dimension in wavelengths, the attenuation through a hole at 60 kHz will be 20 times that 1200 kHz. This assumes all metal wall panelling is still well bonded at joints, but one problem at a time!!

The local station is far more likely to be the problem.

I understand the square law, but it's a much worse frequency for ground penetration. With the 3' loop antenna in the lab, I get 18mV on a voltmeter from WWVB (60kHz). However, with a 50 ohm spectrum analyzer on it, it's -68dBm, 6 db less than 1570 kHz. It's hard to know what level is a problem. The same antenna shows a 40-42kHz signal that is about the same as 1570 kHz. This may be radiated from the 6 computers in the room. The shielded room has a similar reading.

With the spectrum analyzer looking at the power outlets (ground to hot is highest), I see a 65-70 kHz signal that is 30dB higher than 1570 kHz. Compared to 120V, this is a low level, but is it the problem? I like your idea of individual filters for the trunking. I will see what's available.

Quick post! Filters http://www.schaffner.com/components/en/_pdf/Datasheet%20FN2030%20e%2090.pdf etc I will do some calculations ref your loop and posted levels. One thing to mention is that the loop will pick up by magnetic induction from other cables - you can have an M field, not an E-M field -- the loop responds primarily to M field. The way to eliminate the computers is to turn them off. 65-70 kHz may be switch-mode power supply - directly connected to power cables.

OK, Standards Guy, You gave some numbers to do an estimate on pick-up in loop antenna at 60 kHz! For n =12, loop R = 3'/2 = 0.46 m, wavelength w = 5000m, d = diameter of wire bundle in loop say 0.003 m, A = loop area = 0.66 sq m. : induced voltage Vr = 7.12E(nA/w) where E = volts/metre field strength. For given numbers Vr = E x 7.12(12x0.66/5000) = 0.01 E. If emf measured is 18 mV on resonant loop, with magnification Q = 100, then induced emf Vr is 0.18 mV and E = 0.18/0.01 = 18 mV/m. For voltage in 50 ohm resistance (input of Spec-An), inductance of loop must be considered. 0.016 mmF resonates with 440 microhenries at 60 kHz, so reactances for both are 165 ohm. 50//165 has impedance almost 50 ohms, 165 L series 50 ohms R has impedance 172 ohm -- Hence Spec An is seeing 50/172 = 0.3 times induced voltage. Spec An sees -68 dBm = 90 microvolts – induced voltage = 90/0.3 = 300 microvolt (0.3 mV), much the same as assuming Q = 100 with loop resonant. Conclusion is E-M field strength is only 30 mV/m @ 60 kHz - this compares with (E) field strengths of about 50 mV/m for 1 kW at 3 miles for low frequency transmitters with good antennas. I believe about 3 mV/m is a good strength for AM broadcast coverage to meet FCC requirements. At 1570 kHz, owing to inductive reactance increase, say, 165 x 26 = 4300 ohms and parallel 0.016 mmF 165/26 = 6.3 ohms, it is about 90 microvolts + 6 dB = 180 microvolts x 4300/6.3 = 122000. At 1570 kHz, the Vr formula becomes Vr = 0.27E, so E = 122000/0.27 microvolt = 450 mV/m - a possible "open field" value considering the transmitter power and proximity - BUT as mentioned, the emf in the loop could be induction from wiring. YOU WRITE, for power outlets, "65-70 kHz signal that is 30 dB higher than 1570 kHz" BUT YOU DO NOT GIVE 1570 kHz dBm SO posts do not know strength!! DID you try power ground to WALL?? A dBm of -68, plus 30 dB is -38 dBm or about 5 mV. Suggestion from post #32 to hump in a UPS and run it off batteries (unconnected to building power) is a good idea, if you can repeat the test setup of your original problem. For your information, European EMC standards (susceptibility) blast "light industrial" equipment with 3V rms behind 150 ohms on the wires (0.15 to 80 MHz) and 3V/m E-M field ( 80 to 1000 MHz). Regards 67model

Ref. metal and thickness on benches. Skin depth d for copper is 0.27 mm at 60 kHz. Skin depth is the thickness of sheet which has a DC resistance of the same value as the RF resistance at the given frequency of a thick sheet. Because skin depth is also the depth at which current density is [e = 2.718] 1/e times that at the RF exposed surface, having a thickness many times d is unnecessary for conductivity. Not much point in going over 1 mm copper. Skin depth is 1.28 times copper for aluminium, but higher resistivity of Aluminium makes RF resistance is 28% higher. Its a matter of cost, 2 or 3mm aluminium vs 1 mm copper. Copper is much easier to bond to, studs or nuts can be soldered on, thicker aluminium can be tapped for screws.

I likewise with some other respondents would suggest that cabling coming into the unshielded room is the culprit.

Have there been any changes in the vicinity recently to power or phone lines, or even plumbing? Any changes to the building?

While your loop antenna on the Spec-An may not give you a measurable result, it is certainly a useful tool to 'sniff' around the room to find the area of highest signal strength, and to potentially identify the source, be it a slot between roof and walls, a cable penetration, a plumbing penetration or something else - reo in the slab extending beyond the walls .....

Your two biggest clues are:

1 where is the signal strongest?

2 what changes have happened to the building or in the vicinity - anything

Has a repeater been installed in the close by?

Think outside the square as well as inside.

Bill

Have there been any changes in the vicinity recently to power or phone lines, or even plumbing? Any changes to the building?

The only change I know of is the equipment on the other end of the room was rearranged (one rack is now plugged into a different circuit). Keep in mind, many rearrangements have been made in 20+ years, 2 years in this building.

1 where is the signal strongest?

Near the floor, any part of the room.

Sniff around with your antenna on a Spec-An it will be higher in one spot somewhere in the room.

If it appears to be consistently high everywhere near the floor:

Is there under/infloor heating?

also check close to the ceiling.

The RF is entering somewhere, there will be a peak close to the entry point. Do not overlook anything, do not assume anything, check everything, check everywhere.

Perhaps a nearby power line has been moved, perhaps erosion has exposed a water pipe, creating an effective antenna ..... thoughts....

Bill

Have you tried operating your instruments off of a UPS that is not connected to the mains?

If the RFI is being conducted through the power lines this might help.

The FCC has closed the case and will not force the AM station to reduce power at night. The 70kHz signal was coming from the computers. I have put two line filters on each AC bench, one for the computer equipment, and one for the measuring instruments. A signal from the computer would have to go through two line filters to get to the instruments. This didn't solve the mystery. Neither did using a metal bench with it grounded to the wall (which attenuates the AM radio signal). I have to conclude that neither of these is the problem. Meanwhile, many changes are being made in the main building for new products. Thankfully, I have a solution that still works, but the mystery probably will never be solved.