 EGThe wings of love
Premium Member
join:2006-11-18
Union, NJ |
EG to pflog
Premium Member
2013-Mar-28 6:57 pm
to pflog
Re: [Signals] Balancing my own signalMany times they quote the stuff polled from the modem, not the CMTS. |
|
 pflogBueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA |
pflog
MVM
2013-Mar-28 7:07 pm
said by EG:Many times they quote the stuff polled from the modem, not the CMTS.
Ahh ok. |
|
EGThe wings of love
Premium Member
join:2006-11-18
Union, NJ |
EG
Premium Member
2013-Mar-28 7:12 pm
To me this is not surprising considering the typical training (or the lack thereof) of some of their employees...  |
|
| |
to pflog
I don't know what kind of modem you have, but Motorola modems show counters for errored and unerrored codewords. Those counters would give you a much better idea of whether or not your packet loss is occurring on the coax or not.
My strong feeling is that the issue lies elsewhere. Your SNR is nowhere near bad enough that you should be seeing issues. |
|
pflogBueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA |
pflog
MVM
2013-Mar-28 8:01 pm
said by Jeremy W:I don't know what kind of modem you have, but Motorola modems show counters for errored and unerrored codewords. Those counters would give you a much better idea of whether or not your packet loss is occurring on the coax or not.
My strong feeling is that the issue lies elsewhere. Your SNR is nowhere near bad enough that you should be seeing issues.
Unfortunately, Comcast has dummied down the customer-facing side of the SMC and Netgear (with the same web UI) so this info isn't visible. The telnet interface probably has a way to query this, but customers don't have access to that either. |
|
EGThe wings of love
Premium Member
join:2006-11-18
Union, NJ |
EG to Jeremy W
Premium Member
2013-Mar-28 11:43 pm
to Jeremy W
said by Jeremy W:Your SNR is nowhere near bad enough that you should be seeing issues.
Have you taken in to account the uSNR ??? The Codeword errors as read at the modem are for the downstream only.. |
|
 beachintechThere's sand in my tool bag
Premium Member
join:2008-01-06 |
the Upstream SNR isn't that bad, if the numbers were correct. |
|
| |
flwpwr to pflog
Anon
2013-Mar-29 10:31 am
to pflog
down_snr: 34.0 34.7 34.5 34.3 33.9 33.7 34.2 33.6
35 is the cutoff for MY ffo, not sure about yours, but that would be of concern on my jobs. Now I am not saying it is for you, some plant just runs different but generally you want 35+ on DSNR, and 30+ on USNR. It is extremely uncommon to have a better USNR than DSNR. You are reporting 36 USNR and 34+- DSNR, that's not normal. I have however seen modems that report their SNR just flat out wrong [freaking zoom and cisco], whether that is your case or not, is hard to say it could be, or it may be an actual issue. If you take the forward only attenuator off, what happens to DSNR? The device may just be introducing noise, normally I would suspect a active device like a house amp [hate you so much on-q 8/16 port amps] but bad wiring is bad wiring, inside a passive device or not. Cable simulator I believe is what they call those things, never had to use one, and as to the splitter question yes but you need to TERMINATE the unused ports.
As to testing the packet loss becasue that issue may very well have NOTHING to do with it, ping the gateway address [ipconfig /all] with -t, and leave it for a while, in the back ground, and then see if you are getting packet loss just on your side of the network to the modem or not. If you are, something else is causing the issue other than cable because that pings on the Ethernet/wireless side only up to the modem itself, which could be faulty.
You can also use the line quality test here to test just to the modem on cable side IIRC as it uses the public IP to determine its target.
|
|
pflogBueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA |
pflog
MVM
2013-Mar-29 11:53 am
said by flwpwr :down_snr: 34.0 34.7 34.5 34.3 33.9 33.7 34.2 33.6
35 is the cutoff for MY ffo, not sure about yours, but that would be of concern on my jobs. Now I am not saying it is for you, some plant just runs different but generally you want 35+ on DSNR, and 30+ on USNR. It is extremely uncommon to have a better USNR than DSNR. You are reporting 36 USNR and 34+- DSNR, that's not normal.
I have however seen modems that report their SNR just flat out wrong [freaking zoom and cisco], whether that is your case or not, is hard to say it could be, or it may be an actual issue. If you take the forward only attenuator off, what happens to DSNR? The device may just be introducing noise, normally I would suspect a active device like a house amp [hate you so much on-q 8/16 port amps] but bad wiring is bad wiring, inside a passive device or not. Cable simulator I believe is what they call those things, never had to use one, and as to the splitter question yes but you need to TERMINATE the unused ports. The techs that have come out in the past didn't seem concerned at all, even by 33 dB, even when I flat out say that I think it's low. I'm actually quite happy that I'm at ~35 right now instead of 33. I was at over 37 and a while back they did something at night and I dropped over night from 37 to 34/35 and it's been there since. When I remove that pad, the SNR does go up by 1-1.5 dB. So even though my DS power is > 10 at the moment (due to colder temps over night), the connection seems fine. The CMTS reported DS power was actually 2 dBmV lower than what my modem says, so I'm hoping my "11 dBmV" right now is actually 9 dBmV. As to testing the packet loss becasue that issue may very well have NOTHING to do with it, ping the gateway address [ipconfig /all] with -t, and leave it for a while, in the back ground, and then see if you are getting packet loss just on your side of the network to the modem or not. If you are, something else is causing the issue other than cable because that pings on the Ethernet/wireless side only up to the modem itself, which could be faulty.
You can also use the line quality test here to test just to the modem on cable side IIRC as it uses the public IP to determine its target.
I was using smokeping to gauge but unfortunately realized it had stopped running on the 20th. So I restarted it on my end now that I have the pad removed to see how the smokeping graphs look. I guess I can have someone out, but the installers left my modem running with a SNR of 32.5 and I had to have a trouble call out a week later. Then a year later when the temps started warning up, I started having a low SNR and drops and it took many visits and escalating to business escalations to get them to come out. They rebalanced something and the down SNR was better, but still I had issues with random drops (which turned out to be upstream noise bursts). And during the latest visit, the gut rewired my point of entry connection to give the cable better curves (it was tightly almost kinked before) and added this pad, the sum of which was a worse signal (at least according to the numbers reported on the modem). So you can see why I'm hesitant to involve them, especially if it means I'll have to pay for a truck roll that finds nothing wrong.  |
|
|
| |
news to flwpwr
Anon
2013-Mar-29 1:12 pm
to flwpwr
Personally, I would'nt install any attenuator if your SNR is better without. It doesn't really matter if you have +9 or +10 dB downstream level with a SNR about 35-36 dB. It's not really different from having a -9 or -10 dB downstream level with a 35-36 dB SNR. Docsis specifications are +15 to -15 dBmV. If you have +10 downstream level but 33-32 SNR, it's another story, there's probably an overdrived amplifier somewhere on the cable plant...
I remember the hold Motorola proprietary cable modem (not docsis) with a downstream specifications between 0 and -10 dBmV...
I just cheked my signals right know, my downstream power level is +5 with 36 dBmV SNR. This winter, my signal was +11 dB with also 36 dBmV SNR, there's probably an amplifier that isn't in Automatic Control Mode (AGC), who cares? My signal are within ranges anyway.... |
|
pflogBueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA |
pflog
MVM
2013-Mar-29 1:14 pm
I'm fairly sure I have one or more amps on my leg that don't have AGC enabled, as I see a fairly large swing in power levels from the coldest to the hottest parts of the day. |
|
beachintechThere's sand in my tool bag
Premium Member
join:2008-01-06 |
That doesn't necessarily mean it's not "enabled" (Missing the module). It could mean the unit is bad and causing issues. A slight swing is fairly normal. I guess it would depend on the semantics of fairly large.
That is another one of those things that is impossible to tell without a meter and a little ladder work. |
|
pflogBueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA |
pflog
MVM
2013-Mar-29 7:33 pm
said by beachintech:That doesn't necessarily mean it's not "enabled" (Missing the module). It could mean the unit is bad and causing issues. A slight swing is fairly normal. I guess it would depend on the semantics of fairly large.
I've seen as much as an 8-10 dBmV swing over the course of 24 hours. It's gradual, though. The problem is that in this area, we can see huge temperature swings from the hottest point of the day to the coolest. |
|
beachintechThere's sand in my tool bag
Premium Member
join:2008-01-06
1 edit |
A lot of areas see that, it's not very unusual, especially in the "between" seasons. This is referencing temp swings, not signal swings that drastic. You shouldn't be getting charged for trouble calls if you have a verifiable problem that isn't just getting a truck roll for the sake of better numbers. If you can show them you are having a problem that isn't related to your equipment, then you are good to go. When I was a field tech, very rarely did I charge the customer, mostly just the assholes or the ones who told me how to do my job and were wrong. If they were right and I was going to do it anyway, depended on how much of my time they wasted  |
|
pflogBueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA |
pflog
MVM
2013-Apr-4 11:45 am
Well at least according to my (local) smokeping graphs, I have had zero percent packet loss in the last 10 days. |
|
