4G TD-LTE strong self interference - grounding

Hi,

I operate a small TD-LTE network on 3.5GHz, and on one of the sites I can detect strong self interference from the same site (no nearby sites, no external interference). And I have a few grounding related questions as I am fairly sure the reason for this is a ground fault (likely due to the incorrect installation):

1. If I have ~12 meters of feeder between the RRU and the antenna, how many ground clamps do I need to install on the RF feeders?

2. If I only have ~2 meters of feeder between the RRU and antenna, how many ground clamps do I need to install on the RF feeders?

As the antennas does not have a ground point, I suspect that in both the above cases, at least 1 ground clamp is necessary on the feeders close to the antenna ports. I also suspect that for the ~12 meter cable I need at least 2 ground clamps close to each end of the feeders.

Problem is: this is an 8T8R site, so the 3 sectors have a total of 27 feeders, so I just want to make sure we have the minimum amount of ground clamps which is enough.

I did some lab tests with these RRUs, when the antenna ports were terminated to 50Ohm loads, but the RRU itself was not grounded, there was massive RF leak from the RRU itself. I have a suspicion that neither the RRUs, nor the feeder lines are grounded, and as the RRUs are relatively close together (60-70cm), the RF signal cross couples between the RRUs, and thats the source of the self-interference.

My above grounding questions are more related to RF interference, than lightning protection.

Thanks for any and all help.

HI!
The question of grounding (more precisely, lightning protection) is connected with the need to remove the voltage caused by atmospheric electricity.
1. Antenna feeder (jumpers) between the antenna system and the radio unit, up to 3 meters in length, is not grounded.
2. An antenna feeder between the antenna system and the radio unit, up to 12 meters long, is grounded. Install a thunderstorm protection device in the amount of two pieces, on the antenna feeder. At the same time, it is important that the thunderstorm protection devices are grounded to a ground bus with a resistance of at least 4 Ohms.
It is also necessary to ground the RRU units. Use shielded power cable, which must also be grounded.
These activities provide protection against atmospheric electrical quantities.


Causes of interference is a research question.
1. It is necessary to investigate a part of the frequencies with a spectrum analyzer, with the site turned off for local interference.
2. It is necessary to check the SWR of the system (antenna feeder + antenna system). In your case, this is 27 feeders.
3. It is necessary to examine the antenna system for mechanical damage.
The result should be

The question arose: why do you have 27 antenna feeders?
Site configuration 3 sector, 8T8R each?
There should be 24 antenna feeders.

Quote:

---

1. Antenna feeder (jumpers) between the antenna system and the radio unit, up to 3 meters in length, is not grounded.

---

Thanks! I kinda suspected that for short cables (RRU is on the back of the antenna type scenarios), it is enough if the RRU itself is grounded.

Quote:

---

2. An antenna feeder between the antenna system and the radio unit, up to 12 meters long, is grounded. Install a thunderstorm protection device in the amount of two pieces, on the antenna feeder. At the same time, it is important that the thunderstorm protection devices are grounded to a ground bus with a resistance of at least 4 Ohms.

---

We need to get into a bit more details here. When you say "thunderstorm protection device" you mean ground clamps on the feeder, or a lightning arrestor between the feeder line and the RRU? In our case, eve though the RRU is lower, it is still outside, mounted on the tower, not in a cabinet/CO. In this case, I think there is no need for lightning arrestor, but ground clamps on the feeder might be needed.

Quote:

---

t is also necessary to ground the RRU units.

---

Yep, that was quite clear :)

The rest of your questions:

1. It is definitely us who generates the self interference (intra-site), not an external source.
2. The VSWR is constantly monitored by the enodeb, and on all chains we have a VSWR of 1.1 - 1.2, so this is not a VSWR issue.
3. It is unlikely that the antenna or the cabling is mechanically damaged.

Quote:

---

The question arose: why do you have 27 antenna feeders?

---

Each sector is 8T8R plus a calibration port per sector. Thats 3 x 9 = 27 feeders. No passives, couplers, diplexer nor duplexers of any sort between the RRU and antenna. Just direct feederlines. By the way antenna port calibration (for phase and amplitude) are also succeeds on all sectors.

@We need to get into a bit more details here. When you say "thunderstorm protection device" you mean ground clamps on the feeder, or a lightning arrestor between the feeder line and the RRU? In our case, eve though the RRU is lower, it is still outside, mounted on the tower, not in a cabinet/CO. In this case, I think there is no need for lightning arrestor, but ground clamps on the feeder might be needed.@

The device is a thunderstorm protection, it is a clamp on the antenna feeder (sealed from moisture), which is connected to a common ground.
Question: Do you understand that there is interference?

@The device is a thunderstorm protection, it is a clamp on the antenna feeder (sealed from moisture), which is connected to a common ground.

OK. From impedance and ground loop point of view, I think on the 12 meter feeders I will need ground clamps on the antenna end, on the RRU end the RRU itself will have decent grounding, so I think I can avoid the second set of clamps there (the surface area of the 9 feeders will form a very good connection to the RRU's chassis, which will have direct ground connection with high cross section conductors).

@Question: Do you understand that there is interference?

If I set up a test location, line of sight exactly at the middle of the primary beam of one sector (outdoor CPE), I can lock it to another sector which faces exactly the other way around, and still receive -95dBm RSRP. Thats a lot of RF energy on the wrong direction --> there is clear coupling somewhere.
Furthermore: on this test location I can get 42-43dB SINR, but there is large fluctuation (between 15 and 42dB SINR), especially in the busy hours, that is also a clear mark of self interference.
Last but not least: when I set all 3 sectors to different frequencies, all these issues are gone.
FYI, PCI and root sequence planning is also correct, sector separation is also.

Just received a few photos, and it is clear there is zero grounding on the RF parts (antenna, feeders, RRU), the tower is also painted, so there is literally zero ground... Bunch of idiots...

Will get back with the results after they fixed the complete site.

Thanks for the help!

You need to consider that part of the antenna energy will always be present at the back of the antenna, due to the back lobe of the antenna pattern. Here the antenna characteristics are important. possible damage, visual inspection sometimes does not give results. It may be necessary to replace the antenna.

Hi Slava,

Yes, I know that, but not this much. The antenna has a 25dB front to back ratio with 90 degree 3dB, and 120 degree 10dB spots. In beamforming, it is configured to 65 degrees 3db, 100 degree 10db. The total output power of a sector is 128 watts RF @3500MHz and 8 chains combined. A 25dB front to back ration is about 1:300 (a bit more), in this case from the total RF energy of 128 watts, 0.4 watts is (should be) radiated backwards. You are not going to measure this with -95dBm from 1.5km-s on 3.5GHz exactly the opposite direction :)