@RogersCorey which model of Optical node is installed, the older LBON300AC or newer LBON320AC?
@-G- where are the line stats? I see the signal levels, but don't see the codewords in the above posts from @glanchbery Those stats should show consistent signal levels across the board, which they don't, and there should be no noise or errors as this is a fibre system with a very short cable run to the modem. Any noise or loss of codewords would imply a bad fibre splice somewhere in the system between the neighourhood node and the optical node, or a noise source located very close to the RG-6 run into the basement (?) and then up to the modem. This gets back to my earlier statement where, if there is a splitter in the system to drop the signal levels, then all of the other house RG-6 cables should be disconnected from the splitter. The only cable that remains connected to the splitter will be the cable that runs up to the modem.
@Datalink wrote:
@RogersCorey which model of Optical node is installed, the older LBON300AC or newer LBON320AC?
@-G- where are the line stats? I don't see them in the above posts from @glanchbery Those stats should show consistent signal levels across the board, which they don't, and there should be no noise or errors as this is a fibre system with a very short cable run to the modem. Any noise or loss of codewords would imply a bad fibre splice somewhere in the system between the neighourhood node and the optical node, or a noise source located very close to the RG-6 run into the basement (?) and then up to the modem.
We only have downstream stats. Was saying that we need to also see the error stats at the bottom of that page.
As for RFoG, that is still using analog optical electronics along the fibre path all the way back to the headend, that Maintenance needs to tune/optimize, otherwise you can (and we do) see inconsistent power levels and SNR across channels at the different frequencies.
Looks like we still have a good Downstream D3.1 channel at 350 Mhz, and the other D3.0 channels are below 1GHz.
We could also potentially have a bad tuner on the cable modem. However, if the stats on the modem look good and Rogers sees no problems from their end, then the focus should then switch back to the in-home equipment.
@Datalink I scoured through every data point I have and the node model looks like it may be hidden from me... let me get back to you on this!
Regards,
RogersCorey
Looks like @glanchbery will have to get back to us on the Optical Node model. If it turns out to be the older LBON300AC, then the question is, does it fully support DOCSIS 3.1? There is nothing in the online documentation that indicates this. I for one wouldn't assume that it does. Perhaps, for some reason, it just doesn't provide the performance required for 1.5 Gb/s ops.
What do the Optical Node line stats show? Are they within normal levels, or is there anything that stands out, indicating a fibre cabling problem? You should be able to look at other Optical Modems in that neighbourhood to compare numbers with.
@Datalink wrote:
Looks like @glanchbery will have to get back to us on the Optical Node model. If it turns out to be the older LBON300AC, then the question is, does it fully support DOCSIS 3.1? There is nothing in the online documentation that indicates this. I for one wouldn't assume that it does. Perhaps, for some reason, it just doesn't provide the performance required for 1.5 Gb/s ops.
To be fully compliant with DOCSIS 3.1, you need to be able to support Extended Downstream up to 1.2 GHz, as well as support mid-split and high-split. If the neighbourhood is 5–42 MHz/ low-split and using spectrum that is all below 1 GHz, then I'm pretty sure that either of those Lindsay RFoG ONUs can be used.
Re: bandwidth requirements, I'm also in a similar channel configuration, with a single downstream OFDM channel at 350 MHz and all other QAM channels below 1 GHz.
On the upstream side, my active channels are all below 42 Mhz. The Rogers network team has played with trying to enable mid-split with an OFDMA channel using spectrum above 42 MHz, but it appears that they cannot get that configuration stable, nor can they enable another downstream OFDM channel at the high-frequency end of the spectrum.
@glanchbery Those error stats are fine and your OFDM channel, which should be carrying all your traffic, is totally clean.
I don't know what the power levels were like before but an 8-way splitter would be the equivalent of something like an 11 dB attenuator, which is excessive. The techs these days also carry proper 3 and 6 dB attenuators in their trucks, and they probably should have used a 6 dB to bring the average power levels across all channels closer to 0 dBmV... and I would rather see power levels on the higher side of 0 rather than below.
I believe there is one way to test to see if your devices are indeed capable of speeds in excess of 1Gbps (typically 940 Mbps). That is to run the speed test on several devices at the same time. Then add up the speeds on the various devices. If the result is closer to the 1.5+ Gbps when added up, then the device itself (or software) is usually to blame, as mentioned in my previous link - posted again below.
https://communityforums.rogers.com/t5/Internet/Slow-Speeds/m-p/510588#M75402
P.S. I'm on the 500/30 Mbps plan and regularly get download speeds of 1.2 -1.4 Gbps on my Mac Mini M2 10G Ethernet.
@glanchbery no apologies needed. These things take as long as they need to.
Ok, so the stats are rather interesting. I would expect the signal levels to be consistent across the board. These look like levels one would see in a copper system, not in a fibre system. That makes me wonder what the original signal levels are, at the neighbourhood node. There shouldn't be any reason to ramp the signal levels like a typical copper system, unless of course Fibre has its own signal level peculiarities. But I don't believe so. That's a question for @RogersCorey.
Fwiw, you could stand to gain at least 3 dB or more across the board. Its interesting that you have an 8 port splitter followed by an attenuator. Have a look at the attenuator that is connected to the modem's cable. If its a 3 dB attenuator, I would personally remove it, and just run the cable from the wallplate. That would put you closer to the desired 0 dB average across the board. Signal attenuators typically run in 3 dB steps, 3, 6 9, 12. So, you can get creative by using two or more attenuators to achieve a desired signal loss.
Your OFDM downstream levels will be low. That's due to the low frequency ramp up, which I can't explain. If you look at your signal levels, you can see the OFDM channel running at -5.6 dB. That might not be helping you. Unfortunately, Rogers has never indicated which channels the modem uses, OFDM only, if available, or OFDM and QAM mix. Knowing that little tidbit of info would really help in cases like this, and I have asked the question.
Your upstream levels don't make sense. While they look ok, an 8 port splitter should drive them up to something like 47 to 48 dBmV, assuming that the modem is running in its typical 36 to 40 dBmV output range. An 8 port splitter has an 11 to 12 dB loss thru the ports, depending on the frequency. So, with an 11 to 12 dB loss on the upstream, thru the splitter, the modem should be running around 47 to 48 dBmV or higher on the upstream side to make up for that splitter loss enroute to the CMTS.
The upstream signal levels are all driven by the Cable Modem Termination System (CMTS). The CMTS determines the required output of the modem, so, given your numbers, that would put the output to the CMTS, after the splitter, around 28 to 30 dBmV. That is very low, but, perhaps, with Fibre cabling, the CMTS is satisfied with that. That is something that @RogersCorey would have to check. If the CMTS is happy with that, ok, but, its worth checking.
The codewords are a little interesting as well. I wouldn't expect to see any corrected or lost codewords in a fibre system. So, the question at this point is, what does the data for the Optical Node show? Is the data within spec, and, is there more data that the moderators can't see with their system? Does this require a Level II tech to check the data for the node? Given that you have losses, even though they are low, that raises the question as to where those losses are, in the Fibre system or in the house cable system? I wouldn't call the OFDM channel as running clean. If you look at the numbers. 2/3 are unerrored, 1/3 is corrected. I personally wouldn't call that clean.
Where is the Optical Node located? Is it anywhere near a possible electrical noise source? I'm thinking of the cable connection to the modem and any possible noise impact on the cable.
If you look at the cable jacket, you should see the cable data printed on the jacket. If you look at both cables, connecting the node to the modem, and the modem to the upstairs wallplate, you should see RG-6 printed on both cables. The choice is either RG-59 or RG-6, with RG-6 being the desired cable type. RG-6 has lower signal losses and more shielding to protect it from outside noise sources.
At this point I'm wondering about the upstream signal level arriving at the CMTS, and, whether or not your node can actually support 1.5 Gb/s, given that there's nothing in the documentation which indicates that it supports DOCSIS 3.1.
Ok, can you get back to us with:
1. The signal attenuator number from the side of the attenuator; and
2. The cable type, RG-59 or RG-6
3. Are there any coax terminators on the 8 port splitter? They look like this: https://www.homedepot.ca/product/ideal-cable-tv-coax-terminator/1000449550
@57 does raise an interesting point, and that is to run a simultaneous speed test with two or more devices just to see what the results are.
@57 wrote:
I believe there is one way to test to see if your devices are indeed capable of speeds in excess of 1Gbps (typically 940 Mbps). That is to run the speed test on several devices at the same time.
It's really difficult to accurately measure network capacity this way with multiple computers. -- speed test sites were not designed to be used this way. However, you can get pretty close with multiple concurrent sustained downloads and adding up the bandwidth consumed at a point-in-time on all computers involved in the test.
P.S. I'm on the 500/30 Mbps plan and regularly get download speeds of 1.2 -1.4 Gbps on my Mac Mini M2 10G Ethernet.
Consider yourself extremely lucky. Both my parents and I had 500u and I think the highest speed that I was ever able to attain was 850 Mbps. Still very good.
@-G- wrote: It's really difficult to accurately measure network capacity this way with multiple computers. --
I don't believe it needs to be that accurate, just good enough. For example.
1. Just ran my Mac Mini M2 10E by itself several times and got about 1.3-1.4 Gbps down and 35 Mbps up via Ethernet. This is what I get almost any time, day or night and yes, I do consider myself very lucky since I'm paying for 500/30.
2. Ran my iPad Air 2020 by itself several times on WiFi - got 500-600 Mbps down (in the same room as the Gateway)
3. Ran both together and got 1.0 on the Mac Mini and 0.4 on the iPad download (rough numbers)
4. Got 25 up on the Mac Mini and 10 up on the iPad (together).
So, this indicates that my computer has the capacity to exceed 1 Gbps, that my overall download speed is about 1.3-1.4 Gbps and upload is about 35 Mbps.
If someone runs a similar test with several devices and the total of those devices goes above the 940 Mbps, then this indicates that there is enough speed at the gateway. If the above is true and one device is limited to the 940, then there is an issue (or limit) with that one device. That is why I recommend the test. It doesn't need a lot of accuracy.
BTW, the 1.3-1.4 above are mainly due to Speedboost. If I run a large download, the download usually ramps down to the speed I pay for after some period of time (many seconds). It's difficult for me to test with the new speed, since the download is often over by the time the ramping down occurs when Speedboost cuts off. It was certainly the case when I had 150 down - the download would ramp down to exactly that after a number of seconds (less than a minute)
@glanchbery wrote:
@57 I've covered this with Rogers but I've done the tests a few ways.
1) Directly using a 2.5gb device syncing at 2500/2500 in windows and on the modem page.
2) Directly with 2.5gb and gb devices.
3) Multiple gb devices loading a large file download to help saturate the connection.
In all cases it would not exceed 940ish mb.
Interesting. This would seem to indicate that you may have a Gateway that is improperly provisioned, or that there is something amiss upstream of the gateway - see my post above for "reasoning"
@glanchbery thanks for the stats post. That confirms my thoughts on the low upstream levels, but, if the CMTS is satisfied with the arriving signal levels, ok then.
Is there a reset function on the stats page? I have an XB7 but my screen shots don't show a reset button.
What you need at this point is another 6 dB attenuator. I'd ask @RogersCorey to arrange for a tech to drop by and give you another 6 dB attenuator, or two 3 dB attenuators if that is all that is available. That would drop the downstream signal levels to just above 0 dBmV.
The other question for @RogersCorey is what do the CMTS receive levels look like. At present the inbound CMTS levels should be in the 26 to 28.8 dBmV range, which to me is very low. Is the CMTS ok with those levels?
At this point, my question still stands. Does the LBON300AC Optical Node fully support DOCSIS 3.1 ops as required by Rogers network? Yes or no?
And, why are you seeing any corrected codewords and uncorrected codewords on what is supposed to be an almost complete fibre system? Where are the errors coming from? If the Optical Node doesn't fully support DOCSIS 3.1, in all of its nuances, perhaps that is where the problem is. This would require a DOCSIS 3.1 test set, connected to the optical node output to determine what errors are seen on the output, if any.
On the other hand, perhaps an email to the engineers at Lindsay Broadband might resolve the question.
Did any of the techs run a test RG-6 cable from the optical node to the modem, bypassing the house cable system? That would be a worth doing, just to see if the results are the same. That would eliminate the house cable system as a possible noise source.
Questions, questions ..... There's a logical answer here somewhere.
Edit; ok, question regarding DOCSIS 3.1 ops submitted to Lindsay Broadband.
Other food for thought is grounding the RG-6 cable that runs to the modem. In a typical copper installation, the external demarc holds a cable ground block. That serves to protect the house cabling and connected equipment from lightning strikes and to some extent, from external noise, bleeding away any noise from the braided shield that surrounds the center core and the dielectric insulation layer. That ground block is probably not installed in cases such as yours but, I wonder, in this specific case it would be worth installing one to get rid of any noise that the cable is subjected to. The ground block would have to be connected to a good ground point such as the copper water piping that enters the house. This is where a test with a direct node to modem cable would indicate whether or not it might be worth installing.
@Datalink wrote:
@glanchbery thanks for the stats post. That confirms my thoughts on the low upstream levels, but, if the CMTS is satisfied with the arriving signal levels, ok then.
Is there a reset function on the stats page? I have an XB7 but my screen shots don't show a reset button.
No "reset" that I know of. You need to reboot the gateway to clear the stats.
What you need at this point is another 6 dB attenuator. I'd ask @RogersCorey to arrange for a tech to drop by and give you another 6 dB attenuator, or two 3 dB attenuators if that is all that is available. That would drop the downstream signal levels to just above 0 dBmV.
I dunno. There are times when you need to attenuate both the forward and return path, and splitters do just that across the entire spectrum. In this case, it may be better to use a 2-way splitter (with one of the outputs capped) + a 6 dB FPA.
The other question for @RogersCorey is what do the CMTS receive levels look like. At present the inbound CMTS levels should be in the 26 to 28.8 dBmV range, which to me is very low. Is the CMTS ok with those levels?
The CMTS will see the same power level, regardless of whether the modem's output power is on the low side OR higher but attenuated. The question is whether the modem works better and whether the upstream channels have a lower error rate when the output power is at around 40 dBmV.
I'm okay with those signal levels, downstream around 3 dBmV, upstream around 36 dBmV. It will be interesting to see what happens with the codeword numbers in 24 hours.
Just to point out, there's almost a 3 dB gain from 279 to 350 Mhz, where the OFDM channel is starting and another 2.2 dB gain up to 579 Mhz The OFDM channel runs from 337 Mhz up to ~ 448.95 Mhz, so there's a continuous upslope for the entire frequency range of the OFDM channel. If I had access to the Management Information Base (MIB) data, I could probably tell you if that upslope is ok or not.
Hi @glanchbery. The response from Lindsay is that the Optical Network Terminal should not cause any degradation of the data rate. The terminal simply converts the optical waveform into an electrical waveform, so, for the frequencies that we're talking about here, there shouldn't be any issue.
Having said that, I came across a post in the DSLReports Comcast Xfinity forum that was rather interesting. Here's the post:
https://www.dslreports.com/forum/r33782966-
That post is part of a longer discussion:
https://www.dslreports.com/forum/r33782446-Speed-Our-new-bulk-account-gives-us-FTTH-but-it-is-RFoG
So, I'm going to ping off of @DocDrew to find out what his post is alluding to.
Maybe there's an issue with DOCSIS 3.1 over RFoG that I'm not aware of that limits the data rate thru the optical terminal.
One problem with the discussion that I had is that Lindsay only pushes about 900 Mb/s or slightly higher from their CMTS to the terminal and then onto the modem. That is due to the limitation of their lab CMTS. That's unfortunate as their observed data rate falls in line with what you're seeing, and they have no way of testing higher data rates with their current CMTS. Very unfortunate indeed.
Can you repost your codeword data, just to see how its doing after one week.
Have you tried a simultaneous speedtest with two or more devices just to see what the aggregate result might be?
Have you had any discussions with any of your neighbours to see if they are experiencing the same slow data rates? They should be in the same situation as you, running DOCSIS 3.1 thru the same optical terminal model, and thru a Comcast XB6, XB7, or XB8. If this is a class issue, DOCSIS 3.1 and an RFoG optical terminal, everyone should be experiencing the same result. A quick discussion with your neighbours would determine if its just you, experiencing slow data rate, or, if its all of your neighbours.
-
22941 views
-
36 Likes
-
-
