DSL SNR inconsistencies

[Polemus]

I know ... TLDR ... but this is worth a read.

For the last year and a half I have been struggling with SNR inconsistencies on my DSL line.

Countless calls to Telkom has yielded no results. and ultimately I had to downgrade my line to get a more stable connection.

I made a point of monitoring my SNR and found that my SNR sits stable at around 17db (downstream) upstream is always above 20db. The SNR sits stable at 17db and at random intervals it drops to about 11db, it then slowly climbs back up to 17db. Every now and then, usually during the day, my SNR suddenly drops to -6db, and my DSL line loses sync, and it drops completely. Once it loses sync the first time the line turns into a ball of poo. It re-syncs and then drops to -6db about 4 seconds after it syncs. This goes on for hours, and I eventually switch everything off and leave it like that until around 9PM, when I switch on my router, everything is stable again.

A year and a half is a long time for this to be happening, so many things was tried and tested.

3rd PARTY NOISE

I started off by listening to my Telkom line. Immediately I noticed a tick every second or so, and it was constant. It could only mean one thing, someone's Electric fence was touching copper somewhere. I had Telkom out numerous times, and they confirmed it, however, could not find the culprit. This could most definitely have an effect on my DSL SNR, however, the tick is constant, and my SNR was not, so event though it could cause noise, it would make more sense if the SNR dropped at the same interval as the tick, so something else must be causing my SNR drops. Never the less, I installed a DSL filter to assit with noise coming from the tick. This made my SNR's small fluctuations of .5db go away, but the main issue of intermittent -6db still continued.

Home Noise

I started noticing every so often, that when my borehole switches on, my SNR drops (not always, but sometimes), also when microwave or dishwasher switched on my SNR drops, and WTF.... when my daughter starts watching TV my internet drops. This led me to believe that there might be power issues, so I started with a process of illumination, Borehole tests... my SNR drops by 1.5 (on avg), Microwave, SNR drops by 1 (on avg) , TV had no effect. Dishwasher my SNR dropped by about .5 on avg. All these combined does not have a great effect, and should in theory only drop my SNR by about four from 17db to 13db. Something else had to be the problem.

The Server
This weekend my internet was at bet **** to put it mildly. I was never connected for more than 3 minutes at a time, and my line synced @ less than a meg due to SNR ratios. I have a home server, it hosts all my "bought" movies for my daughter, and does various other things. This weekend I decided to redo it. At first I did not notice it, but as soon as I unplugged it to move it (wanted to do a fresh install) the line synced at 4mb. I took it to my desk and plugged it in, when I started it up, my line was dead, WTF!!!!. So I unplugged it, and my line was fine. BINGO, my server is causing the issue.... but not quite, I plugged another machine into the plug, and the line dropped again... OK, so pc's are the problem. NOPE.. I plugged a water pump into the plug and the line dropped.

I think I found it, it is the plug, so I plugged the server into another socket. Same thing, the line dropped.

SOLUTION
So I started doing some research, and noticed that all the devices I plugged in had power factor issues. When I tried my oil heater, it made no difference on my SNR, this was cause an oil heater is merely a resistor, and sent the same waves back as what it received.

ALL HAIL 2 DEVICES
I took the plunge and got me three new devices.

1. 1x Power Factor corrector.
2. 2x Ferite Cores

The PFC plugs into any socket, and cleans your waves. These are merely examples, my Electrician buddy had a spare one I managed to get installed free, but they are not all that expensive.

The two ferite cores was installed on my router, one was clipped over the power power supply, 5 cm from where it was plugged into my router (two loops), the other was clipped over my Telkom Line about 5 cm from where it plugs into my router (three loops).

RESULT
My SNR went from avg 17db on a 4mb line to 29db, and it was constant, no matter what was plugged in anywhere. I was able to bump my line up to 10mb, and SNR dropped to about 12db constant. The line has now been on for more than 4 days without losing sync once.

CONCLUSION
Who would have ever thought that this could be an issue. Yes, you read a lot about people saying look for stuff that can cause interference, but that is a very broad statement. They also do not give real world examples of how to illuminate interference, other than unplug or move. These two devices I bought made a world of difference.

I hope my little write up helps at least one other person, cause the frustration is real, and one always assumes the problem sits with Telkom, or someone else.

On a side note, according to my line attenuation I am supposed to have a stable 8mb line, I now have a stable 10mb line, and I get max speeds.

... lastly, i figured out why my TV caused the line to drop. My daughter used to watch movies from the server, which would then go out of hibernation, and caused more Power Factor issues...

 

[Mr.Jax]

Very nice write-up and some great advice here Polymus!

EDIT:
Which one made the biggest difference ? PFC, ferrite core on the modem power, or the ferrite core on the phone line ? Do you perhaps know how much of an improvement each made separately ?

 

[AfricanTech]

Thanks for a great write-up.

Hmm, my line used to connect at 10 and has dropped to 7-8 in the last couple of months.

This might be worth checking out - even if only the Ferrite Cores.

Have you checked what has brought about the most improvement? The Mr Power or the Ferrite Cores?

 

[Geoff.D]

Yes a very interesting post and some excellent fault diagnosis! Congratulations.

Of course, something that has crept up on all of us the increased use of CFL lamps!. This may be the factor that broke the camel's back so to speak, making the accumulated PF just out of limits overall. CFLs do contribute a lot to PF issues.

I have noticed a very small change in SNR on my line when I use a cfl lamp in my reading lamp on my desk, but not enough to affect performance.

But I guess with most of using more and more reactive devices in our homes, it was inevitable ..

 

[Geoff.D]

Thanks for a great write-up.

Hmm, my line used to connect at 10 and has dropped to 7-8 in the last couple of months.

This might be worth checking out - even if only the Ferrite Cores.

Have you checked what has brought about the most improvement? The Mr Power or the Ferrite Cores?

The ferrite cores helps to reduce incoming EMI, so they are a good idea anyway .. I have just checked and I notice my power distribution and surge protection unit from Clearline is equipped with ferrite cores on the incoming line. The power cord to my laptop has ferrite cores and so does the power feed to the TV etc in the lounge, where there is another Clearline power feeding unit.

The washing machines etc are on a different CB to the feed in the lounge and study, so presumably that helps a bit.

Would be very interested in any report back of what happens of the PF corrector is removed for a period, with only the ferrite cores still in place.

 

[Polemus]

Thanks for a great write-up.

Hmm, my line used to connect at 10 and has dropped to 7-8 in the last couple of months.

This might be worth checking out - even if only the Ferrite Cores.

Have you checked what has brought about the most improvement? The Mr Power or the Ferrite Cores?

Mr Power is actually just a link to a PFC, you can find various PFC's at better prices if you search online.

Regarding your question on which made the biggest impact.

The PFC gave me a snr boost of about 6db, while the ferrite cores gave me a boost of about 3db. Strangely this amounts to about 9db, but when I used both at the same time I managed to get a 13db improvement. I am by no stretch an electrician, so I wont be able to say why

The two ferrite cores on the router yielded different results. The one on the Telkom line gave me about 1db, and the one on the power cord gave me another 2db. But the ferrite cores are complicated with various results, it all depends on how many loops you choke, as well as how far from the actual connection into the router you put them. The distance to the plug on the router mainly depends on the noise created by surrounding objects, and where it affects your line the most, so one can play around a lot with that to get optimal performance.

The reason why the PFC gave me such a big performance increase, is cause I have a lot of motors in my house. Any device that has a motor causes PF issues, things like borehole, pool pumps, fish pond pumps, gate motors, garage motors, and strangely enough pc's also deals a great bit of wave inconsistencies.

You can look at alternate PSUs to alleviate the problems, GOLD Rated PSUs normally have a PF of 98% which makes a difference.

 

[AfricanTech]

..

Thanks

 

[Privateuser456]

Bookmark this now

 

[Kwerty]

Very interesting. Thanks for the post.