I'm having some issues with some really bad upstream noise on a CATV plant. I'm looking for ideas/direction to go to fix this thing. It is not a constant thing, it comes and goes. Watching a graph of the SNR on the CMTS, sometimes there is a sharp drop, and other times it is a gradual decline. SNR is typically around 36, and goes down into the low 20s when this happens, and when it's really bad, drops most all the modems on the system. It seems to be correlated with cold weather, but this may just be a coincidence. On the spectrum analyzer, I believe it shows up as large spikes every 5mhz or so. But I have seen it look differently on the spectrum at different times, so maybe multiple problems?
We have tried pulling the output signal from the headend, while watching inside the headend, and noise drops while the cable is unplugged, but comes right back once it's plugged in.
Same result with turning power to the amps off.
We can "fix" the issue by pulling a tap/splitter that cuts off 75% of the plant, and the noise _does not_ return when you put the tap/splitter back in. Eventually it comes back, but it may be hours or days. We've pulled taps further out in the plant that seem to have either no effect, or sometimes a small effect.
Any thoughts or ideas of how to track this thing down?
Thanks
Have you checked on your reverse amps now that it is cold? Could be running into overdriving with the cold weather and the reverse plant needs to be adjusted down. Do you run any thermal pads in the reverse equipment gain positions?
I haven't checked the reverse amps, I can make sure they're not running close to peak levels. And we don't run any thermal pads. But if it's temperature related, I wouldn't think you would see the noise not there one second, and there the next.
Sounds like you may have already found the problem. Have you tried replacing the face plate? Do you see noise past that point?
Leo
I should add, that I'm not sure that the issue is actually at this point. In the past, we've pulled taps further down the line, and it also went away and didn't return when the tap was put back in. We've seen this at different points in the network. It seems like something is generating some noise, and the noise is bouncing back and forth between different points in the plant, and breaking the line stops it.
We haven't actually watched the spectrum analyzer anywhere but the headend, it's not the most mobile of units. But Maybe we'll try that.
Josh,
I have seen this issue before in our network as well. What I assume happens is when you pull the faceplate you also disconnect power to that area "Either RF power and or AC power. If you do not have a mobile meter with a built in return I would recommend getting one. It makes a huge difference in finding noise. DSAM 3600 is a great one and can be picked up at a decent price even on ebay. One suggestion to try is instead of pulling a faceplate or killing power to an area try going mid way and pull the return pad in a Trunk Amp. This will keep forward RF and AC power passing but everything in the return will be killed. What we have seen in the past is an amp running in X-Mod or a faulty AGC. What you are describing spikes at around every 5Mhz sounds almost like echo's of channels in the return which X-Mod or an Amp running to hot will do. Pulling a faceplate killed the amp and like a computer reset it until it happens again. Pulling the Pad will not kill the amp and in turn you should see the noise go away but come back when you put the pad back in. Let us know how it goes. Feel free to PM me if you need any additional help :)
It came back again this morning, and it's -1 out. I attached an image of it happening. Actually the spikes are less than 1mhz apart. The divisions are 1mhz wide. The spikes are about 10db lower than the modem's signal. This time, we pulled a tap to a different section of the plant. It is close to the other tap we pulled, but a different leg of the network. Pulling it and putting it back in made the spikes go away.
Our amps have no reverse pad to pull. But we can pull the reverse amp out of them. That may invalidate the test though. We can try pulling the input pad to the amp. But the curious thing is that simply cutting power to the plant doesn't make the problem stay away. It comes right back when power is restored. Maybe it needs to be left off a bit longer for capacitors to drain. I would think that pulling a tap to sever a leg, and shutting off power should be essentially the same thing, but evidently not.
I've seen noise like this from broken cables, I assume you've gone for a ride with a leakage meter? Last one I had that looked actually worse than that, we found a feeder chewed wide open by squirrels. I've had several more that were broken at a tap connector, under heatshrink, not obvious at all. I can usually nail those down with leakage meter up in bucket, shake the cables around and leakage level will vary with movement.
Disclaimer: I'm still new. Just bought into my first cable system 2 months ago, sure have learned a lot...like...Trilogy MC2 is evil. yeah it has lower loss, but it breaks all the damn time. We're moving back to P3 just as soon as I get it in.
Hi Josh,
Frequently this behaviour can be seen when you have some amplifiers right on the edge on what they can output on the downstream.
Say that a amplifier is able to do 52dBmV tops and its hovering near that area of 48-52, if your amplifier only can do 52, safe forward might be somewhere around 46dBmV .
Some amplifiers insert broandband junk/noise in the return path when the forward runs out of headroom. Have you done a full plant alignment recently?
Keep an eye on unstable modem values to help narrow down where in your plant this issue occurs.
Is this a buried or hanging plant? a buried plant is quite temperature stable but a hanging plant will have a large change in attenuation over temperature. Also amplifiers will exhibit some change with temperature, but only within 1-3dB's. Make sure that any AGC's are working properly.
Regards,
Philip