Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Oranged]

There doesn't seem to be any regular pattern either in the amount of SNR drop or the length of time of the drop or the actual time of day of the drop so you may struggle to identify anything with an MW radio.

If it was the PSU I'd expect some regularity associated with length of time the computer had been on or when the cooling fans start up. Have you opened the case and cleaned out the dust etc ?

Is it the same symptom from the test socket ?

Perhaps you could perform a properly structured check system starting at the test socket rather than thinking of something it could be and checking that.

[kapt69]

Hi
Not sure if I'm being unclear - the drop in SNR occurs exactly following switching on the PC and recovers exactly as it is switched off. I had it on cehecking emails between 08.42 & 09.33 - hence the attached graph in previous post. Have tried in test socket - no difference.
So I do not have much doubt it is the PC which is causing the drop. (shown on routerstats run on laptop)
HP : thanks for advice re psu. I'm afraid the PC is not a 'brand' name - it was bought from a firm which specialised in building PCs. I will have to look inside to see if I can identify the PSU.
Thanks to all for suggestions.
Kerris

[Oranged]

the drop in SNR occurs exactly following switching on the PC and recovers exactly as it is switched off.

Apologies, re-reading the whole thread I now realise that I misunderstood.

So you're saying the length of time the computer is on (heat build up) makes absolutely no difference, rather it's the on/off process that triggers the drop.

[HPsauce]

the PC is not a 'brand' name - it was bought from a firm which specialised in building PCs.

That's good. 
(it will have a generic, easily-replaced PSU)

[kapt69]

Hi Oranged
Yep that appears to be so. Length of time on doesn't seem to affect 'drop' - although must admit I have kept it off as much as possible.
It may be that I misunderstand - but - by carefully monitoring routerstats, keeping router on all time, shutting off laptop when/if SNR drops to near 9db I seem to manage to keep sync speed 'up' to (currently) 1408 (probably best so far) and download speed around 940 kbps/IP Profile 1000.
If I mess around with router (on/off etc) it obviously affects profile then speed. That is why I think it is so critical that I don't lose the 1.5 db caused by apparently switching on the PC. At moment (a.m.) SNR is 11.8db and falls in evening - you can see that a 1.5db drop brings it down considerably & close to what I think may be the 'limit' before - is it the dslam??- decides enough is enough & reduces speed to improve SNR.
I stand to be corrected - & no doubt will be!!
Cheers
Kerris

[sevenlayermuddle]

shutting off laptop

... I take it you mean 'shutting PC with suspect PSU ?

- you can see that a 1.5db drop brings it down considerably & close to what I think may be the 'limit' before - is it the dslam??- decides enough is enough & reduces speed to improve SNR.

The SNR should be able to drop to 9, and well below it, before anything gets disconnected.  Some people even report brief spells where the SNR has gone -ve, and survived.

However, if your SNR was (say) 8dB, and you got a short burst of interference that reduced it by 9dB then you'd probably drop the connection.  I'm not sure how long a burst of interference would be required for the connectiuon to drop, maybe some of the Gurus can shed light... a second?  more?  less?

Be aware that such a burst of interference wouldn't necessarily show up on routerstats, which just takes some atomic samples at regular intervals, and joins up the dots.  If a burst of interference occured between samples, routerstats won't show it.

This is making me think, however...  didn't you previously report routerstats sometimes showing sharp dips, which we probably attributed to a glitch in the wireless connection, or something?  Suppose these dips were real, and happened whenever routerstats got lucky, and took its sample just when the intererence burst was happening?

If you still have the energy to investigate, you could try reducing the routerstats sample time, which would increase the chances of a sample being taken during 'bursty' interference?    I've many times confessed to not being a routerstats expert, so somebody who is may correct me.

[roseway]

I think 7LM has got it right - the connection shouldn't normally be dropped unless the noise margin drops close to zero, and a short pulse of interference sounds quite a likely explanation.
 

[Oranged]

although must admit I have kept it off as much as possible.

   Perhaps that's the solution....keep it on as much as possible !

[kapt69]

7LM
Just to clarify - no - actually shutting down laptop. As I suggested, for present I try to leave PC off most of time until I can get PSU sorted.
It may be my wishful thinking, but the little green blinky things on the Netgear seem to be a lot calmer when everything (except router) is shut off - so in order to let it rest when SNR is low I switch off laptop as well and 'get a life'.
I accept that this taken to its logical conclusion would lead to possibly having a faster download speed - but with no actual downloads!!
Hope I haven't upset any sacred cows . . .  will investigate more frequent sampling via routerstats.
Oranged:

Perhaps that's the solution....keep it on as much as possible !

I take it you are joking?? - just making sure.
Thanks to all for advice.
Cheers
Kerris

[kapt69]

Just to say I have found the sample rate setup in config in routerstats and set it to 10s.
OK - here I'm pushing my luck - am I wishing in vain to 'get greedy' & see if I can improve speed by trying to keep SNR up as much as possible by trying to maintain as 'clean' a record as I can? Didn't I read somewhere that you need a sync speed of 1450kbps to get into next category - but then am I thinking of fixed rate lines???(which mine isn't)  I'm sure you get used to all this - but it's very confusing. Apologies if you're all going

[roseway]

There's a table of IP profiles versus connection speeds here.

[sevenlayermuddle]

Just to say I have found the sample rate setup in config in routerstats and set it to 10s.

If we're chasing a noise burst that might last only (say) a quarter of a second, that still leaves odds of 40 to 1 against detecting it.  I'd have been inclined to go right down to 1 second sampling or less, if routerstats is cabable of it? 

OK - here I'm pushing my luck - am I wishing in vain to 'get greedy' & see if I can improve speed by trying to keep SNR up as much as possible by trying to maintain as 'clean' a record as I can?

One problem with conspiring to get the best speed is that the higher your connection speed, the more vulnerable the connection will be when the SNRM drops (or is hit by a noise burst).  Your router will then reconnect, of course, but it will quite likely then be doing so when the background noise levels are already high (that's what caused it to disconnect, just a moment earlier), so when it reconnects it may achieve only a lower-than-ever speed, which affects your IP profile for several days.

To cite my own line as an example... I know that on a good day, with a strong tail wind, and I choose the right moment, it'll connect at 4300 or more, which in theory would give me a profile of 3500.  But if I connect at that speed, the SNRM will drop a long way in the evenings, during which it may drop the connection and then it reconnects at around 3000 or so, lumbering me with a 2500 (or sometimes 2000) profile for a few days.  However, by watching the SNRM variations, and manually connecting when the SNRM is in the middle - rather than at the peak - of it's wanderings, I get a speed of 3500 which can be sustained for ever and a day, with a reliable profile of 3000.

[kapt69]

7LM
Oh Joy!
Just had to reply quickly to say thanks.
Only read it once - but sounds like you're doing what I would like to achieve. Always sensible to aim for the middle ground.
By 'manually connecting', I take it you mean powering off router then back on so it re-syncs?
Also - my idea was that if PC & laptop are off - then although router is still connected (hopefully!) during late evening & night there will be 'zero' traffic for exchange  to 'sense problems' - thereby making use of higher speed during day which may be maintained by trying to avoid 'vulnerable periods'.
Is this completely daft??
Will re-alter sample rate to one sec - but it does warn something about cpu loading or some such.
Thanks
Kerris

[Oranged]

If we're chasing a noise burst that might last only (say) a quarter of a second, that still leaves odds of 40 to 1 against detecting it.  I'd have been inclined to go right down to 1 second sampling or less, if routerstats is cabable of it?

Unfortunately, as far as I know, Routerstats minimum sample rate is 5 seconds

[sevenlayermuddle]

If we're chasing a noise burst that might last only (say) a quarter of a second, that still leaves odds of 40 to 1 against detecting it.  I'd have been inclined to go right down to 1 second sampling or less, if routerstats is cabable of it?

Unfortunately, as far as I know, Routerstats minimum sample rate is 5 seconds

Then another option is to use the Netgear administration GUI, which allows 1 second sampling.  Unfortunately of course there's no way of recording the figures, and I won't be responsible Kapt's wife's reaction when she discovers he's taken to spending hour upon hour staring unblinkingly at something on the computer screen, which turns out to be... the Netgear status page.