@davidair and anyone else with reoccurring service drop outs, my first thought would be to look at the upstream signal data in the modem, specifically looking for the presence of an OFDMA channel running. For the CODA-4582 modems that's fairly new. I don't know if the XB6 modems (Arris TG-3482ER and Technicolor CGM-4140COM) also have OFMDA upstream channels running as well on the same CMTS, but, that would make sense.
I haven't seen any evidence that Rogers has started to employ an upstream mid-split, that is to say, that the upstream frequency range which is historically in the 5 to 42 Mhz range is extended, ranging from 5 to 82 Mhz. That would mean that the upstream frequency range is still using the 5 to 42 Mhz range which has higher, unpredictable noise levels to deal with. OFDM and OFDMA use hundreds or thousands of lower power sub-carriers to transmit data, whereas the traditional QAM channels use higher power 6 Mhz wide channels to transmit data. The QAM structure has been used for several years and will continue in use for years to come, however, OFDM and OFDMA is the future if cable operators are going to increase data rates over hybrid fibre / cable systems, which is what Rogers and numerous ISPs currently use. Getting that technology out of the lab and into the street might be rather painful, and I suspect that's what we're currently seeing.
The following quote comes from: https://broadbandlibrary.com/ofdma/
".......
When cable operators begin utilizing DOCSIS 3.1 OFDMA in the upstream they will require advanced software tools to have clear visibility into the OFDMA channel. OFDMA requires CMTS/CCAP integration in order to have full visibility into the OFDMA upstream. To do so, CableLabs has created a special set of management information base (MIB) called the upstream triggered spectrum capture (UTSC) MIB which allows for precision control and visibility of OFDMA carriers. This provides for unprecedented upstream spectrum analysis and the ability to see ingress, impulse noise and more within the OFDMA channel in the upstream. This cannot be done with a traditional external monitoring system without proactive network maintenance (PNM). An operator can use a stand-alone PNM product or one that has been integrated with PNM. Either way you need PNM to see the OFDMA upstream.
The situation only becomes more aggravated with DOCSIS 4.0 where full duplex DOCSIS and/or extended spectrum will be implemented. As operators turn up DOCSIS 3.1 in the upstream they will quickly realize their hardware-based return path monitoring systems are not what they used to be. Of course, not every operator has turned on DOCSIS 3.1 OFDMA but OFDMA channel usage is becoming more common. Technology evolution is normal and with the improvements of OFDMA it is in with the new and out with the old......."
My current thoughts are that OFMDA is now being enabled on a CMTS by CMTS basis, and, combined with noise problems in that lower 5 to 42 Mhz range, a number of customers are being affected by those noise issues. That isn't to say that there aren't other potential problems such as time and frequency scheduling issues at the CMTS. As noted above, cable operators are going to have to implement newer monitoring systems to determine if there are problems with the upstream path. My concern is that the front line tech, and field techs might not have access to the test equipment and data to determine if there is a problem for OFDMA operation in a given area where problems are seen. End result, tech support and the field techs declare that all is well, when in fact they don't have the tools or training to really determine what the problems are with respect to OFDMA operations. I know that tech support doesn't have access to all of the OFDM signal data on the downstream side, so, I have no illusions that they have access to the necessary data to determine upstream performance and look for problems.
So, fwiw, if your having problems as outlined in this thread, have a look at the signal data in the modem. For the CODA-4582 that is contained in the STATUS .... DOCSIS WAN tab. The very bottom section is the OFDM/OFDMA Overview. That section will show two upstream OFDMA channels, one of which might be active. For the XB6 modems, that same "active OFDMA channel" indication should show up in the upstream signal data, but I'm not sure where that is located in the modems user interface. So, if you have an OFDMA channel active, that potentially translates to a wide band low power channel running across a large portion of that 5 to 42 Mhz range, co-existing with the QAM upstream channels which will also be running. The other question that arises is whether or not the OFDMA channel takes precedence, with the QAM channels being available but not in use. If the modems are set to use the OFDMA channel exclusively, and a noise event occurs, that would be a potential explanation for the dropouts. The QAM channels might not be affected but the OFDMA channel could be. It would take a tech, with the right training, right test equipment, or access to the right CMTS receive data to really understand what the problem is and how its affecting the OFDMA channel performance. If there's a timeslot allocation problem which occurs at the CMTS, I have my doubts that any tech would be able to determine that as its really a network engineering problem. Again, with the right test equipment, a tech might be able to determine if a given modem is receiving enough upstream transmit timeslots to be useful to the modem and its user, but, I have my doubts.
So, fwiw, this is conjecture at this point. For anyone running a CODA-4582 with OFDM downstream running and no OFDMA upstream, on a neighborhood node and CMTS that isn't overloaded, you should see good performance without any dropouts. If you're in that situation, have a look at the OFDM/OFDMA Overview section of the DOCSIS WAN tab and have a look at the channel status. Take note of its current status and keep an eye on it to determine if and when either channel flips to enabled status. If the only thing that changes one day is a change to enabled status, and you start to have service dropouts, that points to a problem in that lower 5 to 42 Mhz range of the cable spectrum, at which point tech support and the field techs declare that all is well.
Its worth noting that the moderators don't appear to be commenting on this situation. I've asked for help from the engineers, pointing directly to this thread, and haven't received any useful feedback or comments which would help troubleshoot the problems, which I believe are beyond the capabilities of tech support and the field techs. That leaves them and the customers to flounder without any useful assistance or direction. A rather unfortunate situation to say the least 🙂
I'd say that they haven't looked closely enough, or, they don't have the necessary equipment to really look at the data that they need to examine. We'll never know the real truth.
Fwiw, yup, you have an OFDMA channel running. Do you have underground or overhead cabling in your neighbourhood?
@88dalejrfan with overhead cabling in place, that would mean that the cabling and its data would be subject to noise injection from other sources. End result, possible service dropouts. I'd be looking at the lower spectrum (5 to 42 Mhz) from the modem all the way to the neighbourhood node. Noise problems are hard to detect and can take a very long time to resolve. That would probably take a specialist of some type to track down the noise source, if in fact that's the issue. The question at hand is whether or not the tech has the necessary equipment to look at OFDMA data and/or has access to same type of data from the CMTS.
Here's an extreme example of noise problems on a cable system:
https://www.bbc.com/news/uk-wales-54239180
@MisterPinst yes, you are running an upstream OFDMA channel. If you look at your upstream signal data, you can see 4 QAM channels plus one OFDMA channel in operation. Its interesting that you're not seeing the dropouts with the Technicolour modem in operation. In theory, the upstream OFDMA transmit levels should be the same for all modems, but, no one would know except for the engineering staff. The other question is whether or not the modems are running a combination of QAM and OFDMA channels simultaneously, or, if the modems are only using the OFDMA channel, with the QAM channels idling, not in use for customer data. That's a network engineering question. The QAM channels show up, but, are they used for anything beyond timekeeping and modem status? Don't know.....
Noise in a cable system is hard to track down, but, newer modems are far more capable of detecting problems like mico-reflections and noise. The only problem is that you have to have a back end system capable of using that data and engineering/technical staff capable of using such a system.
@davidair Rogers won't allow customers to use their own modems on the cable system. If you were a TPIA customer you could use that modem. Its an interesting question of whether or not TPIA customers are seeing the same problem with their TC-4400s? I don't recall seeing multiple, continuous posts about this problem on the DSLReports TPIA pages.
Edit: @davidair yes, you do have an upstream OFDMA channel running as well.
-
227894 views
-
200 Likes
-
-
