dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
2237
share rss forum feed


JeepMatt
C'mon the U
Premium
join:2001-12-28
Wilmington, DE
kudos:2
Reviews:
·Comcast

4 Upstream Channels

So since D3 modems can support at the minimum 4x4 with Downstream / Upstream - is there any reason why Comcast decided to allocate 8 QAM's to downstream (in the most updated areas) but only 3 upstream?

First thought would be to at least do 4 - but I wondered if there are technical reasons (noise) for this that prohibited them from adding that 4th channel??



gar187er
I do this for a living

join:2006-06-24
Dover, DE
kudos:4

5-42mhz is the sub band......take into account diplex filter rolloff in amps, and your at 5-37ish.

Then subtract the 5-18 garbage you cannot avoid, so your left with 18-37mhz. 19 usable mhz worth of space.

6.4mhz per 64qam, 3.2 for a 16qam, and dont forget the box driver!!!

so approximately 16mhz is used for 3 channels, 2 64s and a 16....and a couple db for the qpsk box driver and youre at 18mhz.

can you throw another 16qam down in the garbage of 5-18? sure, but no modem will use it...
--
I'm better than you!



telcodad

join:2011-09-16
Lincroft, NJ
kudos:3
reply to JeepMatt

As gar187er said, fitting 4 upstream HSI QAM carriers in that 18 to 37 MHz usable spectrum area is a challenge.

However, in some field trials back in 2010, Comcast had managed to obtain an upstream data rate of 75 Mbps by bonding two 6.4MHz-wide 64QAM channels with two 3.2MHz-wide 16QAM ones in that space (»www.lightreading.com/document.as···lr_cable). I suppose this is how they figure they will be able offer the 65 Mbps upload speed for the new 305/65 "Platinum" HSI tier.

Looking through all the modem signal screen captures posted in Johkal's »[Signals] Evidence of Upstream Channel Bonding - Post Here stickied thread, the majority of the systems do just show having the usual two 64QAM channels along with one (legacy DOCSIS 1.X) 16QAM one, at the time they were captured.

However, there are actually a few systems that already show 4 upstream channels, similar to Comcast's field trials. In those few cases, a second 16QAM channel had been placed just inside in the sub-20MHz area, where the greater noise level tolerance of a 16QAM carrier was utilized.

Two notable exceptions to that 4-channel scheme can be seen, however, where in the first (Johkal's) post, all 4 upstream carriers were placed between 22 and 37 MHz, and in another case from Hillsborough/San Mateo, CA (»Hillsborough/San Mateo, CA), where they managed to fit three 64QAM channels (at 23.7, 30.6 and 37 MHz) and one 16QAM one (at 18.9 MHz). That arrangement should definitely support that new 65 Mbps upload speed.



EG
The wings of love
Premium
join:2006-11-18
Union, NJ
kudos:9

FWIW, might it also be that the local engineers decide how many return QAM's to implement to satisfy the current upload capacity needs of any given local system, so it may be deliberate decision ?

Expand your moderator at work


JeepMatt
C'mon the U
Premium
join:2001-12-28
Wilmington, DE
kudos:2
reply to telcodad

Re: 4 Upstream Channels

Telco-
Thanks for posting those links. I wasn't aware that some areas had previously dabbled in it. Very interesting to see the "combinations".
--
"ONE team - ONE city - ONE dream!!"


noisefloor

join:2010-05-09
reply to EG

said by EG:

FWIW, might it also be that the local engineers decide how many return QAM's to implement to satisfy the current upload capacity needs of any given local system, so it may be deliberate decision ?

In your area they probably still have local engineers (East div). In mine (West div) everyone is now centralized in CO. There are no more local engineers :-(
Just some hub techs that can run in and replace a downed SPA card if needed.


EG
The wings of love
Premium
join:2006-11-18
Union, NJ
kudos:9

That's sad. Hope that the headend techs are well trained and qualified.



gar187er
I do this for a living

join:2006-06-24
Dover, DE
kudos:4

HE Techs cant make calls like adding channels, and what not....all of those decisions are done at region/division....
--
I'm better than you!



EG
The wings of love
Premium
join:2006-11-18
Union, NJ
kudos:9

Figured that, thanks gar187er See Profile !



cypherstream
Premium,MVM
join:2004-12-02
Reading, PA
kudos:3
reply to gar187er

In Motorola systems, box return is usually around 8 - 16 Mhz. Though in SA/Cisco plants, I've never seen a DHCT return lower than 18 MHz. Weird. Wonder if the Moto aloha protocol is more robust with noise than SA's DAVIC? Eh, hopefully one day they will all just run on DOCSIS and that space can be free'd up.



telcodad

join:2011-09-16
Lincroft, NJ
kudos:3
reply to JeepMatt

An article on the CED Magazine site today about some of the ways cable operators can increase upstream data speeds:

Options abound for boosting cable upstream speeds
The return path lags behind downstream speed, but cable is working on its game plan.

By Mike Robuck, CED Magazine - August 20, 2012
»www.cedmagazine.com/articles/201···m-speeds



RR Conductor
NWP RR Inc.,serving NW CA
Premium
join:2002-04-02
Redwood Valley, CA
kudos:1
reply to JeepMatt

Click for full size
We just have two upstream channels here, though 8 downstream channels.


Ripper
Premium
join:2002-02-07
West Palm Beach, FL
Reviews:
·Comcast

1 edit
reply to JeepMatt

Click for full size
All areas are different,I now have 8 down, 2 up.


cypherstream
Premium,MVM
join:2004-12-02
Reading, PA
kudos:3
reply to JeepMatt

But can you get more than 27.5 dB upstream SNR on Older 750 MHz Antec nodes?



telcodad

join:2011-09-16
Lincroft, NJ
kudos:3

1 edit

said by cypherstream:

But can you get more than 27.5 dB upstream SNR on Older 750 MHz Antec nodes?

I'm assuming you're talking about, in reference to that article, the greater uSNRs that are needed to support higher-order QAM modulations (than the current 16- and 64-QAM) on the upstream, such as 256-QAM.

From this previous thread: »QAM Questions

said by DrDrew:

said by pclover:

What is the SNR required for 128 QAM?

It's a 3dB increase for every doubling of the QAM number.

Min. Good Modulation
24dB 27dB 64QAM
27dB 30dB 128QAM
30dB 33dB 256QAM
33dB 36dB 512QAM
36dB 39dB 1024QAM


nysports4evr
Premium
join:2010-01-23
kudos:1
reply to JeepMatt

Same here, 8 DS chans, 2 US chans.



EG
The wings of love
Premium
join:2006-11-18
Union, NJ
kudos:9
reply to telcodad

said by telcodad:

said by cypherstream:

But can you get more than 27.5 dB upstream SNR on Older 750 MHz Antec nodes?

I'm assuming you're talking about, in reference to that article, the greater uSNRs that are needed to support higher-order QAM modulations (than the current 16- and 64-QAM) on the upstream, such as 256-QAM.

From this previous thread: »QAM Questions

said by DrDrew:

said by pclover:

What is the SNR required for 128 QAM?

It's a 3dB increase for every doubling of the QAM number.

Min. Good Modulation
24dB 27dB 64QAM
27dB 30dB 128QAM
30dB 33dB 256QAM
33dB 36dB 512QAM
36dB 39dB 1024QAM

I believe that is for the forward path ?


telcodad

join:2011-09-16
Lincroft, NJ
kudos:3

Theoretically, it shouldn't matter which direction, but in the real world, I suppose there could be certain types of signal distortions that may be unique to the upstream path that may cause a difference (and require a slightly better SNR to achieve the same error rate performance).

As gar187er said in that same thread:

said by gar187er:

anything is possible, but will it ever happen? ehhhhh i wouldnt say never, but highly unlikely. The snr needed for 128/256 really isnt possible in the upstream. i would think you will see mid/high split systems before seeing anything more robust then a 64qam upstream.

as for 512 QAM on the down, in an all digital world, it isnt really needed, how many channels does one person need??? i really dont see 512 happening either due to plant design limitations....if every node was designed n+3 then sure...but 6,7,8,9 actives deep, i couldnt see 512 being a viable modulation scheme.

and scdma is in trials in asia last i read anything on it. if and when this comes to the US it will make my job a lot easier!

said by pclover:

said by gar187er:

The snr needed for 128/256 really isnt possible in the upstream.

Seems maybe possible to me. If 256 QAM needs 30 dB SNR in the downstream direction wouldn't it be the same for the return or no? But that seems really borderline tho. Maybe possible if you had a very clean plant?

Comcast says my uSNR right now is 31 dB for QAM64

said by gar187er:

again if the plant is 3-6 actives deep, its not that hard to acheive...i work in a node+9 system...average SNR for a decent size node is 30-31....there is no room for error if you push 256 qams up as 32 is BARE minimum.....not to mention if you bond 256 qams...lollololololloololololerzzzz



EG
The wings of love
Premium
join:2006-11-18
Union, NJ
kudos:9

Isn't the lower frequency band of the return path subject more to noise than the higher forward band ?



telcodad

join:2011-09-16
Lincroft, NJ
kudos:3

Exactly, that's why the the upstream SNRs are lower than the downstream SNRs, and why only 16- and 64-QAM carriers are currently used there, rather than the 256-QAM carriers that are commonly used on the downstream.



EG
The wings of love
Premium
join:2006-11-18
Union, NJ
kudos:9

O/k. I must admit that these are rhetorical questions bud !



telcodad

join:2011-09-16
Lincroft, NJ
kudos:3

OK sorry, I just thought you had questioned whether those needed SNRs that DrDrew had provided for the various QAM modulations pertained to a particular path.

As cypherstream and gar187er indicated, it can be tough getting the 32 to 33dB uSNR needed to support 256-QAMs on the reverse path in many of the older/longer systems.