Can signal be too good?

Recently I’ve run into a few customers with really good signal, SNR and modulation but constant loss of registration and drops in the SNR graph. I move them to an adjacent sector and their signal drops, but the problem goes away. The most recent case was a customer on a 450b high gain at 2 miles with a -54 signal and really bad dropouts. Between the dropouts though they were at 8X and really good SNR on both HPOL and VPOL.

I didn’t think the signal could be too strong until it was in the 30’s or 40’s. Could -54 be too hot?

What about the SM receive target, can there be a problem is the signal at the AP is above or right at the target? Usually I leave the default of -52, sometimes if I remember I change it to -62.

1 Like


Its called receiver over feed (some use the term saturation). It happens when the received signal is higher than what the rx amplifier was designed to handle and can not drop its output below the rx input gain threshold. This means that higher coding schemes get distorted as it is akin to having someone screaming in your ear while giving detailed complex instructions to do a dificult task.

To mitigate this, Cambium (and others) use an AP controlled SM-tx power control to reduce SMs that are received too strongly in comparison to other SMs. On the SM side, if you are receiving the AP too loudly, you can detune to SM a bit, usually downward is good, and this will still keep a reasonable SNR.

This is the part that is not taught enough to new techs and mostly forgotten by us older guys.

But could -54 be too hot? I assumed if the default SM receive target was -52, the overload point was quite a bit higher.

Also I thought the point of ATPC was to avoid unnecessarily interfering with another AP pointed in the opposite direction, which will only have so much antenna F/B ratio.

IMHO, -54dBm isn’t too hot a signal. I usually set my SM receive target to -60. Do you have this SM onboarded into cnMaestro so you can look at the SnR, modulation, etc. over time?

-54 is hot but not maxed, I had a link at -45 but it was not a good performing link until it was toned down a bit. -60 is usually the hottest you want with -65 being a great strength.

The ATPC is only partly involved for frequency reuse. Its main function is to equalize the AP received signal strength of all the SMs on a sector. It has a side effect of ensuring the SM of the back sector has a hard time connecting to the front sector, but that is not the main intent. Think of a sector like an auditorium, if you have someone upfront chatting away but an attendant in the back asks a question, the presenter will have a hard time hearing the question. Now the chatter up front keeps it quiet and to about the same level as the querant then there is no problem.
The way the manual teaches howbthis all works isnt wrong, its just not fully correct as it does not explain in any detail how these systems work together to mitigate interference and self-interference.
The fact is an SM will connect to an AP which has the better SNR if the signals strengths are reasonably close to each other. Thats why we have the preferred ap list.

Frequency reuse and self-interference are both a function of setting up the GPS timing system and making sure that all APs have the same setting other than what has to be kept different on each ap. This and the fact that your antennas should have a 30dB f/b ratio means that the two APs have 60dB of isolation but since they are transmitting at the same time this is moot and only the 30dB loss is considered for SM connections. If you had untimed APs then the 60dB isolation would limit your tx power so you wouldnt exceed the isolation threshold.

See if you can get a copy of the old Motorola MWFE and MWSE books, this has a very good explanation of RF mechanics short of an EE degree.

1 Like

Eric, this particular SM is not in cnMaestro, at this time we have only onboarded our CBRS radios. But I do have graphs via SNMP from the SM. It is harder to get AP side data.

The signal level is absolutely steady, so we’re not seeing multipath or trees or anything like that. I do see variations in the SNR, which led me to believe it was an interference problem, but I’m doubting that. And of course these radios don’t really measure SNR but rather BER from which they extrapolate SNR. I’m reluctant to say the SNR suddenly nosedived and caused the SM to deregister, and it’s difficult to tell if it’s an uplink or downlink problem causing loss of registration.

450b complicates things because of its higher xmt power, so the rcv signal at the AP is usually several dB higher than at the SM. But we have ATPC do deal with that. If the signal at the SM is too hot, there isn’t much we can do except use a smaller antenna, tilt the antenna up, or connect to a different AP on the tower.

Try lowering the target receive down to -60 or lower. Read on a forum post or the docs somewhere that greater than -60 can self interfere. Generally we set ours to around -63 with exceptions for APs with very long shots.

What version of software are you running on these? I’m seeing the same thing on 16.2.3 where didn’t have a problem.

The problem I’m working on now, both AP and SM are still on 15.2.

I’ve also seen it on a CBRS sector and all our CBRS stuff is on 16.2.2. Actually we’re trying to take everything to 16.2.2 but haven’t gotten to all of them. With the situation during the pandemic, people get mad if we do firmware updates even at 3am.

But what I’m seeing is definitely not a 16.2.3 issue.

1 Like

I can report that changing the rcv target from -52 to -62 has resolved this problem.

1 Like

Thanks for reporting back Ken. This information might help another community member with similar issues. I wouldn’t have thought that -52 is “too hot” either… and will ask engineering about this just as a follow up.

Yep, now a retired wireless tech, this is not uncommon with high signal levels. HF or Microwave, receivers can be overloaded and give odd effects, including making the SNR ratio worse, not better. A -54 is very high. In old schools terms, it’s almost 500 microvolts (into 50ohms). Taking it down 10-20db would be perfect. Note the high traffic point to point microwave systems are designed for these levels and higher, but seldom the cheaper receivers used at the subscriber end.

I’m going to somewhat disagree with you, at least with respect to Cambium PMP products. If you want 6x or 8x modulation (64 or 256QAM), -74 will be too low, especially in 5 GHz where there’s always some interference. -62 should be good, and I think even Cambium expects that -52 shouldn’t be overloading the receiver front end.

The problem was fixed for me by changing the rcv target at the AP. The SM was OK with its rcv signal. But also the 450b SMs are prone to this because they are capable of a higher xmt power than the AP.

Another thing, the symptom was re-regs. 99% of the time, the SNR and modulation readings were excellent, but then the SM would lose registration.

Lots of equipment based on cheap WiFi chips, including routers and laptops, are fine with signals as hot as -30. If someone has their laptop or tablet on the desk next to the router, that’s going to happen. I would criticize them for using WiFi rather than a 3 ft Ethernet cable, but most devices these days don’t have Ethernet ports.


Wifi devices will turn their power down to not overload the AP rx gain amplifier.

The Cambium in the user manual suggests -65 for a very good reason. The system uses a heartbeat system and if the SM does not decode this properly due to too high of power then the SM will deregister simply because they think there is a bad connection or is out of sync which re-registration will correct for a little bit. The SM receivers are tuned for about 125uV which is fairly hot as it is considering your average FM station is between 3 and 5uV within the coverage zone. The AP uses ATPC to ensure that no radio comes in too hot and thus the AP will be able to decode the SM heartbeat response.

If everyone turned the power out levels down 3dB then 90% of interference would be gone but every dual band wifi router comes set to the highest level they are allowed thus more interference.

There used to also be a factor that best practices were to make sure the receive power levels of all the SM units that are trying to connect are within a reasonable range. Vaguely I remember 10dBm of variance being a recommended number but that was back in the 100 series days. I think I was able to keep most my APs within 13dBm form the normal hottest signal to the normal weakest signal. The receive power level if you are not used to looking at that can be viewed on the AP if not through other management methods on the web GUI via the Home->Session Status->Power tab. You can sort by that also changing ascending and descending if you want. Getting feedback from Cambium or others with best practices would be good. I still try to target 10dBm when I can. Balancing modulation rate practical things like mentioned here and trying to keep the range as minimal as possible depending on how well different radio chipsets handle hot and “cold?” signals in practice is often necessary. Say the highest is -50 dBm at the AP and the “coldest?” is -70dBm. Tweaking the AP under the Configuration->Radio->SM Receive Target Level is where in this example if you put in -60 in that field the radios that are at -50 could drop down their power output levels until the AP is receiving at -60 and then the range would be -60 to -70 down from a difference of 20 dBm to 10dBm. I am not stating this is a number you should put in. Real world practice even considering things like how this affects each SM and even local interference they could be having, etc. It used to be a very important thing to get best overall AP throughput to minimize within reason trying to keep up modulation rates the range of worst receiving signal power levels at the AP with best. I’m not sure if I could find that document anymore or if Cambium has an updated on for the newer radios but the ideas I bring up are important to consider from my experience.

A example as I am around deafness a lot would be trying to listen to someone scream and then very quickly trying to listen to someone whisper and then back and forth without even knowing if the next speaker will be screaming or whispering or in between. Even if they are not doing that at the same time that would be very very hard. Radios, especially access points, deal with this issue too. I expect the newer ones are better but doubt the problem has not gone away as it is likely just a reality of physics. I’m not sure that is what you are experiencing or not but it could be. Another factor is, again, within reason, factoring in real world realities and trying to come up with reasonable policies since telling a customer they cannot get service at all or at a particular speed is often in play, trying to keep access points at a particular site as close to the same range as possible but I would expect practice might show this is not as important as for each access point.