02-04-2013 02:26:03 PM
We upgraded from 768kbps DSL to 12Mbps U-verse 2 months ago (just phone and Internet, no TV). AT&T installed the i3812V system, which as I understand it, combines 2 phone lines since our distance is relatively far (pair-bonding).
This has been intermittent, but we had a total outage (no dial tone) about 2 weeks ago. AT&T themselves fixed this. Apparently, 1 of the 2 outside pairs was not properly connected by the original installer at our iNID.
Then we were solid for about 1.5 weeks, and now the Internet connection is super intermittent again for about 4 days. Countless power-cycles are not helping. Great for 1 hour, terrible for the next.
My question is this. When my internet connectivity gets flaky like this, so does my connection to the router/gateway. Pings to 192.168.1.254 start getting dropped, loading the 2-Wire (same IP address) webpage takes forever and sometimes fails, and UV Realtime becomes unusable since it can't connect, either. Why is this? I would not expect this. Have others had a similar experience?
I've reproduced this from 3 different computers/laptops to make sure it's not the port on the i38HG or my CAT5 cables or PCs. Could it be the connection from the internal i38HG access point to the external i3812V iNID box on the outside? Is the router actually out there, or in the house in the i38HG? Is it getting to "busy" doing something else? Makes no sense.
Attached are my typical screenshots (no real changes between good times and bad). Please help, or help me to guide the tech support folks in the right direction. Since this problem comes and goes, I'd hate to have tech support come out while all was well...
(I believe the ping "General failure" lines above were due to closing the lid on the laptop and it going to sleep. Just look at totals.)
Solved! Go to Solution.
02-05-2013 07:05:44 AM
It was rock solid last night (0% ping loss), but unusable this morning (50% ping loss!). Go figure. Nothing changed, not even the weather by much.
I tried various permutations of connecting the i38HG into different phone jacks, using different phone cords. The installer had used 2 of the 6 pairs in the ~20-foot cable from the iNID to my in-house patch panel for bringing in the primary phone line. One pair went to 5 phones, the other pair went straight into the i38HG (which was a pretty clean way to do it, I thought). I tried different permutations of these connections as well, but nothing eliminated the ping loss.
Then I enabled the external Ethernet port on the iNID (after cursing for ~15 minutes while trying to get the i38HG's web page to load under these intermittent conditions). I then disconnected the i38HG altogether. I ran a long CAT5 from the iNID Enet jack, through a window, and into my 10/100 switch - and crossed my fingers.
Fixed! 0% ping loss. Router web pages load instantly. Internet connection is fantastic! But I'll really have to wait a few months before declaring this as fixed for good. I love that the iNID has this Fast Ethernet jack!!
So why is the HPNA data connection so crappy for me? Is the iNID or i38HG to blame? Has anyone fought this before??
I guess I don't care. Just wanted to share in case others are in a similar world of pain. I'll try establishing Wi-Fi tonight with my old Cisco router left over from my DSL days (after disabling its DHCP, etc.). I'll also re-purpose 2 of the unused pairs from the iNID to my closet for the 10/100 Ethernet connection so I can get rid of the long CAT5 cable and close my window tight.
02-05-2013 07:33:13 AM
If switching from HPNA over the existing wiring to Cat5e Ethernet fixed your problems, then there must have been some issues with the connection, i.e. that should be a fine solution. However, I will say that gateways, as a rule, can go unresponsive when the upstream connection isn't there. This was true of my Motorola Surfboard from days of yore.
Your Bitloading graph indicates some issues with your connection between the VRAD and the iNID. The big null at 1.6 MHz is probably a local AM 1600 KHz station transmitter near your home. Quite a few other nulls are showing up, so I'm wondering if you have a bridge tap, or just that many discrete sources. Since both pair seem to have about the same pattern, that would mean both were affected by a similar bridge tap, which seems kind of odd.
It's worth calling AT&T Technical Support at 800-288-2020 and having them take another look at your connection. At 2600' not being able to sustain a bonded-pair for 12 Mbps HSI is disappointing.