dslreports logo
 
    All Forums Hot Topics Gallery
spc
uniqs
11

Sunny
Runs from Clowns

join:2001-08-19

Sunny to koitsu

to koitsu

Re: Firmware update.

I actually posted that mostly to poke at my "almost" neighbor, tshirt.

Thanks for the reply, though. Hindsight told me I probably should be quicker to powercycle when I run into issues like that instead of futzing with browser configuration which had always been fine before and your post confirms that.

Edit: I really didn't think my signal levels had anything to do with slow page loads.

koitsu
MVM
join:2002-07-16
Mountain View, CA
Humax BGW320-500

koitsu

MVM

said by Sunny:

Edit: I really didn't think my signal levels had anything to do with slow page loads.

In my experience power level doesn't, but SNR can. If SNR gets too bad, the modem has to retransmit frames back to the CMTS (or vice-versa) which causes slowdowns. There's actual proof of this happening, because it's something I dealt with for many weeks (the official Comcast Forum thread is more indicative of the behaviour -- look closely at the mtr/traceroutes).

Sunny
Runs from Clowns

join:2001-08-19

Sunny

Good to know, SNR on all four downstream channels was 36-37 and still is. It was downstream power that improved after the upgrade/powercycle.

tshirt
Premium Member
join:2004-07-11
Snohomish, WA

1 recommendation

tshirt to Sunny

Premium Member

to Sunny
Actually, I don't think it mattered, I got a forced powercycle last night, with no apparent problems, so I think they are working through area by area until everyone has it

The effect I'm hoping it solves is at the DHCP renew switching to only the first channel(only that channel acrews codewords) for a while (2 hours to a couple days) before all channels begin working again.
I believe this is part of the cause of slowedpageloads/"page not found" (which loads quickly if refreshed) but now I need it to stay active(no reboots or powercycles) for at least 2 DHCP renews to prove it.

I did find under the preivous firmware a powercycle temporarily improved it, and the problem did not appear to be signals or SNR related (always very good here)
Anyway CC must have found enough advantage to having everyone on 1.0.6.8 to push it, in spite of a few remaining issues
maybe in 2 more years...
tshirt

tshirt to Sunny

Premium Member

to Sunny
said by Sunny:

I actually posted that mostly to poke at my "almost" neighbor, tshirt.

Thanks for the reply, though. Hindsight told me I probably should be quicker to powercycle when I run into issues like that instead of futzing with browser configuration which had always been fine before and your post confirms that.

Edit: I really didn't think my signal levels had anything to do with slow page loads.

Really just meant power and SNR readings are centered in the good range and consistantly stable.
Someone with unstable or borderline power might need to consider those as a possible causes of dropouts, reboots or packet loss.