Community Forum

Expecting better speed

Contributor

Expecting better speed

I live in SF bay area, and am paying for 400 as part of triple play.

There was a time when I was getting 450+, maybe 480, but it was close enough not to bother me.

I was renting TG1682G, but I decided ot buy after my bill went up in Jan.  I work from home, often pushing large zip files to AWS, and I have been considering going to gig.

So I bought a CM1150v, Docsis 3.1 with EMTA for phone.  No wifi, but that's OK since I have a dedicated router and a dedicated AP.

 

Setup was simple,  both Internet and phone worked immediately.

But speed tests were all well below 400, so I called comcast support, and after playing games with old modem (slightly worse speed) and swicthing back to new, they suggested a tech appointment.

 

The tech showed up, reversed a "coupler" back near the demarc, and then we did some more speed tests.

There might have been a slight improvement, but still only mid 300's.  Then we tried another laptop and got high-300s/low 400s.  The laptop switch left me puzzled because I am certain that I have measured north of 400 with the "slow" laptop it in the past.  Very odd...

 

Then the tech suggested that I consider downgrading to 250 using the argument that I would be better off paying for 250 and getting 280.  He had been insisting that the speed plan is "up to 400", so I asked him what the odds were that I would end up with high 100s.  He had no answer.

While I'm no docsis guy, the signals and SNR look OK to me.

 

The tech did indicate that he was going to schedule someone to review the situation on the pole which is immediately across the street.

My questions:

  1. Given clean signals, does it make sense that there might be a problem on the pole or somewhere else in the Comcast network?
  2. What is the deal with "up to 400"?  Should I be expecting 480, or maybe demanding 480?
  3. Is there a reasonable technical reason that would prevent Comcast from being able to provistion to 480?
  4. Does it make sense for a technical challenge somewhere in Comcast network to present more on one laptop than another (my "slow" laptop), or can I rest assured that the slow laptop/dongle is truly slower.  Again, I know I measured speeds significantly higher in the past.
  5. Additional thoughts/advice would be hugely appreciated.

Thanks!

 

Version:1.0 StartHTML:000000191 EndHTML:000040173 StartFragment:000016538 EndFragment:000040141 StartSelection:000016538 EndSelection:000040141 SourceURL:http://emta/DocsisStatus.htmNETGEAR Gateway CM1150V

Cable Connection
 
 
Frequency start Value
This field below allows you to modify the frequency the cable modem start with its scan during initialization and registration. Enter the new start frequency and restart the cable modem for it to take effect.
Starting Frequency
 
Startup Procedure
Procedure Status Comment
Acquire Downstream Channel 663000000 Hz Locked
Connectivity State OK Operational
Boot State OK Operational
Configuration File    
Security Disabled BPI+
IP Provisioning Mode Honor MDD honorMdd(4)
 
Provisioned Downstream and Upstream Throughput
Type Throughput
Downstream Provisioned Rate  
Upstream Provisioned Rate  
 
Downstream Bonded Channels
Channel Lock Status Modulation Channel ID Frequency Power SNR Correctables Uncorrectables
1 Locked QAM256 31 663000000 Hz 10.4 dBmV 41.3 dB 0 0
2 Locked QAM256 1 471000000 Hz 8.9 dBmV 41.3 dB 0 0
3 Locked QAM256 2 477000000 Hz 9.1 dBmV 41.3 dB 0 0
4 Locked QAM256 3 483000000 Hz 9.3 dBmV 41.5 dB 0 0
5 Locked QAM256 4 489000000 Hz 8.2 dBmV 41.2 dB 0 0
6 Locked QAM256 5 495000000 Hz 8.2 dBmV 41.2 dB 0 0
7 Locked QAM256 6 507000000 Hz 8.4 dBmV 41.4 dB 0 0
8 Locked QAM256 7 513000000 Hz 8.5 dBmV 41.4 dB 0 0
9 Locked QAM256 8 519000000 Hz 8.4 dBmV 41.4 dB 0 0
10 Locked QAM256 9 525000000 Hz 8.1 dBmV 41.2 dB 0 0
11 Locked QAM256 10 531000000 Hz 8.2 dBmV 41.5 dB 0 0
12 Locked QAM256 11 543000000 Hz 7.9 dBmV 41.3 dB 0 0
13 Locked QAM256 12 549000000 Hz 7.8 dBmV 41.1 dB 0 0
14 Locked QAM256 13 555000000 Hz 7.9 dBmV 41.1 dB 0 0
15 Locked QAM256 14 561000000 Hz 7.8 dBmV 41.2 dB 0 0
16 Locked QAM256 15 567000000 Hz 7.9 dBmV 41 dB 0 0
17 Locked QAM256 16 573000000 Hz 7.7 dBmV 41 dB 0 0
18 Locked QAM256 17 579000000 Hz 7.8 dBmV 40.8 dB 0 0
19 Locked QAM256 18 585000000 Hz 7.9 dBmV 41 dB 0 0
20 Locked QAM256 19 591000000 Hz 8.1 dBmV 41 dB 0 0
21 Locked QAM256 20 597000000 Hz 8.1 dBmV 41.1 dB 0 0
22 Locked QAM256 21 603000000 Hz 8.2 dBmV 41.2 dB 0 0
23 Locked QAM256 22 609000000 Hz 8.5 dBmV 41.1 dB 0 0
24 Locked QAM256 23 615000000 Hz 8.4 dBmV 41.1 dB 0 0
25 Locked QAM256 24 621000000 Hz 8.7 dBmV 41 dB 0 0
26 Locked QAM256 25 627000000 Hz 9.1 dBmV 41.3 dB 0 0
27 Locked QAM256 26 633000000 Hz 9 dBmV 41.1 dB 0 0
28 Locked QAM256 27 639000000 Hz 9.2 dBmV 41 dB 0 0
29 Locked QAM256 28 645000000 Hz 9.7 dBmV 41.3 dB 0 0
30 Locked QAM256 29 651000000 Hz 10 dBmV 41.2 dB 0 0
31 Locked QAM256 30 657000000 Hz 10.3 dBmV 41.5 dB 0 0
32 Not Locked Unknown 0 0 Hz 0.0 dBmV 0.0 dB 0 0
 
Upstream Bonded Channels
Channel Lock Status US Channel Type Channel ID Symbol Rate Frequency Power
1 Locked ATDMA 1 5120 Ksym/sec 17300000 Hz 43.3 dBmV
2 Locked ATDMA 2 5120 Ksym/sec 23700000 Hz 42.3 dBmV
3 Locked ATDMA 3 5120 Ksym/sec 30300000 Hz 41 dBmV
4 Locked ATDMA 4 5120 Ksym/sec 36700000 Hz 40.3 dBmV
5 Not Locked Unknown 0 0 Ksym/sec 0 Hz 0.0 dBmV
6 Not Locked Unknown 0 0 Ksym/sec 0 Hz 0.0 dBmV
7 Not Locked Unknown 0 0 Ksym/sec 0 Hz 0.0 dBmV
8 Not Locked Unknown 0 0 Ksym/sec 0 Hz 0.0 dBmV
 
Downstream OFDM Channels
Channel Channel ID Indicator Subcarrier Zero Frequency Active Subcarrier Number Range Total Codewords Uncorrectable Codewords
1 48 Primary 611600000 Hz 1348 ~ 2747 268619728 0
2 0 other 0 Hz 0 ~ 0 0 0
 
Upstream OFDMA Channels
Channel Upstream Is Muted Ranging Status Subcarrier Zero Frequency Active Subcarrier Number Range Symbols Per Frame Tx Power
1   Other 0 Hz 0 ~ 0 0 0 dBmV
2   Other 0 Hz 0 ~ 0 0 0 dBmV
 
Current System Time:Wed Feb 06 10:13:19 2019
System Up Time:1 day 01:20:27
 
Regular Contributor

Re: Expecting better speed

I just browsed Comcast's residential plans and they all had "up to" in them.  Maybe business plans are different.  It does seem that they provision somewhat conservatively so on a good day you can get ~20% over your "up to" speed, but I guess they fall back to the "up to" wording when it's not such a good day.

 

Silver Problem Solver

Re: Expecting better speed


@strega7 wrote:

I just browsed Comcast's residential plans and they all had "up to" in them.  Maybe business plans are different.  It does seem that they provision somewhat conservatively so on a good day you can get ~20% over your "up to" speed, but I guess they fall back to the "up to" wording when it's not such a good day.

 


FYI:

https://www.dslreports.com/faq/15643

I am not a Comcast Employee.
I am just a customer, volunteering my time to help other customers here in the Forums.
Was your question answered? Mark the post as best answer!
Contributor

Re: Expecting better speed

I upgraded to gigabit, and I'm really not seeing a dramatic improvement in speed.  After doing some reasearch, I've opted to focus on speedtest.net and speedof.me, both of which are html5 based.

I've run tests throughout the day with widely varying results.  I see nothing odd going on in my modem, so while I'm not knowledgeable on docsis, I have to assume that the downstream OFDM channel is OK.  I'm uncertain on what is going on with the 32 downstream bonded channels that continue to show up post upgrade to gig.  I assume that they aren't being used now that I have subscribed to gig? 

My conclusion (maybe hypothesis) is that all of the various speedtest sites can only be as accurate as the Internert path between Comcast and the test site.  I now think that the week speed/bandwidth link is not the connection between my modem and the  headend; rather it is inconsistencies associated with the wild west of the Internet itself.  Reasonable or silly?

 

Unfortunately (or maybe telling?), the Xfinity speed test site has always registered the worst scores.

I've been focussing on upload speeds for the time being, as I should be able to test them reasonably accurately without connecting directly to the modem.  My wifi AP (R7000) and router (erlite-3) should not be taxed by the relatively low upload speeds offered even with gig.

Assuming I am correct that the connection to the headend is healty, then it stands to reason that the speed between headend and my modem is at or near 1000 gpbs.  Right?  Or wrong, or it depends?

 

From what I can tell, it's going to be difficult to acheive say a consitent 800 Mbps down and 30 up.  Given that my motivation right now is improving upload beyond an extremely consistent 12mbps with Extreme Pro 400, I'm thinking I might be better off with Extreme 300 at this time, which offers speeds "up to" 25 Mbps.  My understanding is that, as a triple play customer,  I would continue to pay for Extreme Pro 400, but this is still $16 cheaper than gig.

 

Does my reasoning make sense?

Silver Problem Solver

Re: Expecting better speed


@JJNorCal wrote:

I upgraded to gigabit, and I'm really not seeing a dramatic improvement in speed.  After doing some reasearch, I've opted to focus on speedtest.net and speedof.me, both of which are html5 based.

I've run tests throughout the day with widely varying results.  I see nothing odd going on in my modem, so while I'm not knowledgeable on docsis, I have to assume that the downstream OFDM channel is OK.  I'm uncertain on what is going on with the 32 downstream bonded channels that continue to show up post upgrade to gig.  I assume that they aren't being used now that I have subscribed to gig? 

My conclusion (maybe hypothesis) is that all of the various speedtest sites can only be as accurate as the Internert path between Comcast and the test site.  I now think that the week speed/bandwidth link is not the connection between my modem and the  headend; rather it is inconsistencies associated with the wild west of the Internet itself.  Reasonable or silly?

 

Unfortunately (or maybe telling?), the Xfinity speed test site has always registered the worst scores.

I've been focussing on upload speeds for the time being, as I should be able to test them reasonably accurately without connecting directly to the modem.  My wifi AP (R7000) and router (erlite-3) should not be taxed by the relatively low upload speeds offered even with gig.

Assuming I am correct that the connection to the headend is healty, then it stands to reason that the speed between headend and my modem is at or near 1000 gpbs.  Right?  Or wrong, or it depends?

 

From what I can tell, it's going to be difficult to acheive say a consitent 800 Mbps down and 30 up.  Given that my motivation right now is improving upload beyond an extremely consistent 12mbps with Extreme Pro 400, I'm thinking I might be better off with Extreme 300 at this time, which offers speeds "up to" 25 Mbps.  My understanding is that, as a triple play customer,  I would continue to pay for Extreme Pro 400, but this is still $16 cheaper than gig.

 

Does my reasoning make sense?


I have Gigabit and both Xfinity and DSLReports show 950/35 so it is possible to get what is advertised (even though my Comcast Plan is higher)

Have you seen this?

https://forums.xfinity.com/t5/Your-Home-Network/Troubleshooting-Gigabit-internet-speeds/m-p/3145413#...

I am not a Comcast Employee.
I am just a customer, volunteering my time to help other customers here in the Forums.
Was your question answered? Mark the post as best answer!
Contributor

Re: Expecting better speed


@RobertWy wrote:

I have Gigabit and both Xfinity and DSLReports show 950/35 so it is possible to get what is advertised (even though my Comcast Plan is higher)

Have you seen this?

https://forums.xfinity.com/t5/Your-Home-Network/Troubleshooting-Gigabit-internet-speeds/m-p/3145413#...


I had not run into that link.  Thanks for pointing me to it.  Much appreciated.

 

My upload speeds are solid today.

 

I ran a few new tests.  I'm using a 1 year old high end HP Spectre.  I am connecting using a USB-C 3.0 adapter which is rated at 5 Gbps.

First, I attempted to validate the Ookla based speed test built into my r7000 wifi router which consistently reports 380/42.  It is connected as: r7000 > erlite-3 > cm1150v using cat6.  I unplugged it's uplink cable and connected it to the laptop: xfinity= 213/39, dslreports= 613/42, ookla= 501/42.  These results were consistent over multiple tests and are very similar to what I measure over wifi; xfinity always significantly slower, with the other two in closer agreement.

 

Then I reconnected directly to the modem: laptop > cm1150v using short short cable with unkown rating: xfinity= 665/41, dslreports= 680/43, ookla=317/42.  I ran multiple tests and ookla was consistently odd man out.

 

Lastly, I replaced the ethernet cable with the one that shipped with the cm1150v and it did not make a difference.

 

I really don't know what to make of the odd man out swap between xfinity and ookla, but my takeaway is that I'm probably getting mid to high 600 speed today.

 

I guess I can't prove it, but I have every reason to think that the USB-C adapter can acheive gigabit speed.

Contributor

Re: Expecting better speed

Netgear level 2 support set up a remote session and reviewed the signals and event logs.

 

They indicated that the downstream power levels should be between -7dB and +7dB, saw no indication of any problem with the modem, and expressed confidence that there is an issue somewhere between headend and modem.

 

Netgear also pointed to numerous T3 timeouts in the event log, indicating that I should never see these, even on startup.  There were also large numbers of uncorrectables showing up today on every channel.

 

The signals are higher now than in my original post with all channels greater than +10dB and some north of +11.  

 

So now back to Comcast.  A tech is coming out again, and my hope is that I can get Netgear back online with me so that they can work together.

Contributor

Re: Expecting better speed

I received a call today from "advanced troubleshooting".  A bit difficult to have a clear discussion, but something was done somewhere in the comcast network.  First try resulted in same mid 500 speeds, and second try actually slowed things down, most clearly manifesting in slower mid 30 upload speeds.

 

Then a call from somebody else at Comcast.  I mentioned that Netgear indicated that the power level should be between -0.7 and +0.7 dB.  He then asked me if I "wanted" to have the signal attenuated.  I told him that , given that Comcast is an expert delivering 1Gbps over copper, and that my CM1150v is on their official retail list, my expectation is that Comcast should be capable of addressing the problem.  I don't "want" the signal attenuated; I "expect" Comcast to leverage their expertise to resolve my problem.

 

I was then told that they would bring equipmment which would be capable of measuring the delivered speed.  That would be welcome.  I have reviewed all specs on my high end laptop, and given the USB-C thunderbolt connection, there should not be an issue.  Both Netgear and Comcast asked what Windows is reporting.  The answer is 1Gbps.

 

I have seen other posts where folks seemed to have similar problems, many of which have no clear resolution.  If Comcast can't deliver gigabit to my home, for whatever reason, would they tell me this in a straightforward fashion?  At this point I'm having my doubts.

 

Any and all thoughts would be greatly appreciated!