dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
9166
share rss forum feed


TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26

1 edit

[rCable] RF problems lately

Hey all,

I'd like to get some extra feedback from you guys to see where all of these RF issues are coming from. It appears that what has been done is to go from 4 x 3.2 MHZ wide upstream channels to 2 x 6.4 MHZ wide channels plus 1 x 3.2 MHZ channel.

So at a glance it appears that there is in effect more upstream bandwidth now than before. With the 4 upstream channel before there was about 12.8 MHZ of bandwidth, now with 3 channels there's 16 MHZ of bandwidth.

This is a pretty big change though and we have been seeing many RF problems. Can anybody comment/confirm/deny that this happened?
--
Marc - CEO/TekSavvy



TypeS

join:2012-12-17
London, ON
kudos:1

I was 4x4 before and now I'm 4x3 on channel bonding.

Here's a shot from the signal stats page on my SB6121:



TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26

Right so on the upstream you have 1x2.3 plus 2x3.1 for a total of 9.4MHZ...
--
Marc - CEO/TekSavvy



TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26
reply to TypeS

What speed are you on?
--
Marc - CEO/TekSavvy



TypeS

join:2012-12-17
London, ON
kudos:1

I'm on 28/1 right now, hopefully 45/4 sometime next week.

Though I know on my node specifically my neighbor has had 150/10 for a good number of months. Seems like the upstream bandwidth tinkering may be for speed increases later in the year.



TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26

Yeah, major areas appear to me Bloor, 475 Richmond, Newmarket and Markham.. Probably others but those are areas where we have master tickets open.
--
Marc - CEO/TekSavvy



zed173

join:2010-07-17
Mississauga, ON

FYI - in general, speed tiers are not tied to what the modem actually "syncs" at with the CMTS.

ie. if the CMTS is 4 downstreams and 3 upstreams all bonded...then all modems on that physical node are the same, regardless of speed tiers.


fuzzby4

join:2008-08-31
Toronto, ON
reply to TSI Marc

Click for full size
Here's the screencap for mine. I am in the Bloor area. I've called support, opened a ticket and went through troubleshooting.


Teddy Boom
k kudos Received
Premium
join:2007-01-29
Toronto, ON
kudos:20
reply to TSI Marc

Very interesting! I'm still at 4x 2.56Msym/sec here at Bathurst and College (formerly DuPont POI).

Is there a way to tell the symbol rate on a DCM475/6? I'll poke around, but I don't see anything...
--
electronicsguru.ca


tross

join:2010-06-11
kudos:1
Reviews:
·TekSavvy Cable

1 edit
reply to TSI Marc

Marc, you seem to be looking at the frequencies of his three upstream channels. That won't give you any indication of how much upstream bandwidth he has.

For cable, in North America, DOCSIS standard says 25% roll-off (on the upstream side), so the bandwidth (in MHz) will be the symbol rate (in Msym/s) x 1.25. So for TypeS, he has one channel at 5.12 Msym/s ( = 5.12x1.25 = 6.4MHz bandwidth), and two channels at 2.56 Msym/s ( = 2.56 x 1.25 = 3.2MHz). So, 1x6.4MHz and 2x3.2MHz channels.

At least one channel has to be 3.2MHz wide for compatibility with previous versions of the DOCSIS standard.

Hope that helps.


nbinont

join:2011-03-13
kudos:2
Reviews:
·Start Communicat..
·TekSavvy Cable
·Bruce Telecom
reply to TSI Marc

said by TSI Marc:

Hey all,
It appears that what has been done is to go from 4 x 3.2 MHZ wide upstream channels to 2 x 6.4 MHZ wide channels plus 1 x 3.2 MHZ channel.

Marc: you seem to be confusing data rate with channel bandwidth.

Channel bandwidth is measured in MHz and represents the physical frequency range of the analogue signal. Channel bandwidth is controlled by filters in analogue side of the modem, and is currently always 6Mhz/channel (or 8Mhz/channel in europe), so that the channel sizes and spacing matches the TV channel specs. As others have mentioned, the filters specify a specific roll off, but this only represents the "sharpness" of the channel, not the channel width.

The data rate is determined from = symbol rate (symbols per second) X symbol size (QPSK(4)/QAM(16/64/256...),etc) X coding rate

What you're seeing is Rogers increasing the symbol rate from 2.56 to 5.12 Million samples per second). This has the immediate effect of pushing more symbols (and bits / Mbps) through per second.

Increasing the symbol rate doesn't increase the channel bandwidth assigned - it just tries to push more samples though the same channel. What this DOES do is make each sample more likely to be affected by noise, and reduce the upstream SNR as observed by the CMTS side.

As usual the signal strength issues at a higher symbol rate can be overcome in a variety of ways: higher transmit power (are customers near max tx power?), lower symbol size, or a different coding rate.

HTH.


Teddy Boom
k kudos Received
Premium
join:2007-01-29
Toronto, ON
kudos:20

1 edit
reply to Teddy Boom

said by Teddy Boom:

Is there a way to tell the symbol rate on a DCM475/6? I'll poke around, but I don't see anything..

In factory mode I'm able to see details on one upstream channel through the serial console. Unfortunately, Rogers appears to block the modem from getting an HFC IP in factory mode, so that's as far as I can get.

(that single channel is most likely the backwards compatible channel fuzzby4 is getting a single wide channel..)
--
electronicsguru.ca

Bootly

join:2005-08-22
Toronto, ON

Factory mode? Please tell me more...



Teddy Boom
k kudos Received
Premium
join:2007-01-29
Toronto, ON
kudos:20

said by Bootly:

Factory mode? Please tell me more...

It's all in the DCM475 firmware upgrading thread.. Here's the details:
»Re: Upgrading Firmware on Thompson DCM475

It might be possible to load it up with a noisy bootloader and get the details that way... I'll look into it eventually
--
electronicsguru.ca


TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26
reply to nbinont

Thanks tross and nbinont..

Ok. So more RF problems... We're seeing a ton of them so this is why I'm asking...

So. Basically if, as a result of this change, there are RF issues... All we can do is wait until they clean up the noise and go from there.
--
Marc - CEO/TekSavvy



Teddy Boom
k kudos Received
Premium
join:2007-01-29
Toronto, ON
kudos:20

said by TSI Marc:

Ok. So more RF problems... We're seeing a ton of them so this is why I'm asking...

When people come for firmware upgrades, I often get them to send me a signal stats screenshot. I have definitely noticed Toronto area customers with only three upstream channels lately. Thing is, they are mostly DCM475/6, so I can't tell for certain if this is what's going on. I'll keep watching for it though.
--
electronicsguru.ca


TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26

Ok. Yeah. Looks like we'll be into RF issues for months and months unfortunately.
--
Marc - CEO/TekSavvy


tross

join:2010-06-11
kudos:1
Reviews:
·TekSavvy Cable

1 edit
reply to TSI Marc

I never paid much attention to how things were before on Rogers. DOCSIS1/2 allowed upstream channel widths up to 3.2MHz. DOCSIS3 adds 6.4MHz. So it really depends what's changed. If we are now pumping more symbols through, but increased channel bandwidth accordingly, then it shouldn't be a big problem. If they're using the same channel widths, but have decided to pump more data/symbols through but with the same bandwidth, then a the link will need a better SNR. For people who were already borderline, then they'll have trouble. I suspect the latter (though it could be both), since personally, I used to have 4x16-QAM channels on the upstream, now I seem to have 3x64-QAM channels. 64-QAM requires higher SNR to maintain the same error rate, all other things being equal.

So long story short, you're right: unfortunately, those borderline people will have RF problems until the quality of the link can be improved and they can get a better SNR.


nbinont

join:2011-03-13
kudos:2
Reviews:
·Start Communicat..
·TekSavvy Cable
·Bruce Telecom

1 edit
reply to TSI Marc

said by TSI Marc:

So. Basically if, as a result of this change, there are RF issues... All we can do is wait until they clean up the noise and go from there.

Well, that's one approach.

Another idea: Taking a look at fuzzby4's stats we observe that the upstream power level is at max, meaning that the CMTS is having trouble hearing the modem. Yet it is still assigning QPSK and 64QAM rates. The QPSK coding is much more robust and the CMTS can probably hear this just fine. But it's likely that the CMTS can't hear the 64QAM rates. With this configuration the connection will probably be spotty.

Now the 64QAM probably worked fine at the lower symbol rate, but we need approximately twice the SNR (3db more) to operate the 64QAM at twice the symbol rate. At the higher symbol rate, it may not work.

Now the CMTS should be monitoring and assigning appropriate rates to the modem. I'm concerned that this is not working correctly. Probably fuzzby4 should only be assigned QPSK upstream when operating at 5.12Msps. Rogers would need to verify their CMTS is working right.

nbinont

join:2011-03-13
kudos:2
Reviews:
·Start Communicat..
·TekSavvy Cable
·Bruce Telecom
reply to tross

Ok, so it looks like tross may be right about the physical channel bandwidth: »volpefirm.com/docsis101_upstream-rf/ It appears that, for DOCSIS at least, the upstream (but not downstream) physical channel bandwidth scales with symbol rate.

If this is the case, then situations like fuzzby4's really should not happen unless:

1. There's a bug in the modem tx filter that transmits with the wrong filter bandwidth. Marc Question: are all your problems occurring on a specific brand/firmware modem?

2. There's a bug in the CMTS rx filter (this should impact all modem brands equally)

3. Some other RF issue (overlapping channels, uneven interference, etc).

But at the end of the day, the CMTS should be assigning modulations that work for the modem...


fuzzby4

join:2008-08-31
Toronto, ON
reply to TSI Marc

My cable modem:
Model - Motorola SB6120
Firmware Name: SB_KOMODO-1.0.6.1-SCM00-NOSH
Boot Version: PSPU-Boot 1.0.0.4m1
Hardware Version 3.0
Firmware Build Time: May 12 2011 15:15:14



TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26
reply to TSI Marc

Looks like Kitchener/Waterloo was added now too...

»[Outages] Outage Report July 20th 2013 @ 5:20PM
--
Marc - CEO/TekSavvy



TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26
reply to nbinont

I don't know yet if its the same brand or firmware. Doesn't seem to be yet anyway.
--
Marc - CEO/TekSavvy



TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26
reply to tross

So, what signal to noise ratio should we expect to see if its healthy?

I'm sure my team knows all of this.. But I'm getting rusty since I haven't been doing it every day for a few years now.
--
Marc - CEO/TekSavvy


resa1983
Premium
join:2008-03-10
North York, ON
kudos:10
reply to TSI Marc

Click for full size
My local node was split around the beginning of June. I'm on the new node, as my IP changed and I was stuck with no upstream channel bonding for 2 weeks. No DHCP issue either, shockingly enough.

Since they re-enabled upstream channel bonding, this is what I've been seeing.

I've had no issues on York Mills POI since then. Bathurst/Steeles, 35/3 package.
--
Battle.net Tech Support MVP

nbinont

join:2011-03-13
kudos:2
Reviews:
·Start Communicat..
·TekSavvy Cable
·Bruce Telecom
reply to TSI Marc

said by TSI Marc:

So, what signal to noise ratio should we expect to see if its healthy?

Well, you won't see it from the modem side, as only the CMTS can measure the upstream SNR. However, the CMTS will ask for different upstream powers from the modem to maintain it's desired SNR. So you can get an idea. You can tell the CMTS has/or is close to having a bad SNR when the upstream power level is at (or almost at) max.


TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
kudos:26

right. of course, that makes perfect sense.

and max power is what?
--
Marc - CEO/TekSavvy


fuzzby4

join:2008-08-31
Toronto, ON
reply to TSI Marc

I have access to a Rogers cable modem that is currently active in service. The owners are gone for the week; is it possible to take that Rogers modem and have it installed on my TSI account? I've also ordered a TSI DOCSIS3 modem but Customer Service said that's 3-5 days out.


tross

join:2010-06-11
kudos:1
Reviews:
·TekSavvy Cable

2 edits
reply to TSI Marc

On the upstream, I think the spec calls for a CNR (SNR is a similar concept but will be slightly worse than this number) of at least 25dB... but we are talking upstream, so I don't think any cable modems are going to be able to report it (the noise all gets added after it leaves the customer's modem, after all). I think you'd only be able to measure it on Rogers's end, as nbinont said.

As for not dropping down to a lower-order modulation, it could be Rogers has disabled that.. QPSK would lead to someone using the wire a bit longer, in order to transmit the same amount of information, so they might not allow dropping down there, to prevent congestion... I don't know.

You have to be careful about just looking at whether the modem is at maximum power or not, as well. If a modem is seeing a lot of reflected power, it should drop its output power down to a safe level, so it doesn't kill itself... even if the other end wants more signal...


nbinont

join:2011-03-13
kudos:2
Reviews:
·Start Communicat..
·TekSavvy Cable
·Bruce Telecom
reply to TSI Marc

said by TSI Marc:

and max power is what?

53 or 54dBmV This is a bit dated, but mostly still accurate: »Adelphia High Speed Internet »FAQ: Acceptable cable modem signal levels.