50M : 16QAM & 64QAM Interference

[chambohambo]

Is it correct that 64QAM is more prone to interference than 16QAM?

[heero_yuy]

Originally Posted by chambohambo

Is it correct that 64QAM is more prone to interference than 16QAM?

In a nutshell, yes. As the modulation goes to a higher degree the guardband between identifiable states gets reduced.

Here is a phaser diagram for QAM16:



Each state is well defined with a large guard band, each point can move around quite a bit due to niose, propgation effects etc and still be easy to distinguish.

Contrast to QAM64:



HTH.

[Kushan]

Originally Posted by chambohambo

Is it correct that 64QAM is more prone to interference than 16QAM?

Yes, in the sense that if you try to cram more information into the same space, it takes less for that information to get corrupted. It's not quite right to say that it's more prone to interference, rather it's better to say it's more sensitive to interference.

However, the extra bandwidth you get from it makes it worthwhile if the network is capable of it. The systems Virgin use mean that if interference rises, it'll drop from QAM64 to QAM16 and even QPSK, reducing overall bandwidth but increasing stability until the issue is fixed.

[General Maximus]

although it seems like we have been waiting forever the upstream upgrades and channel bonding, I can see VM have been doing some work on the qt. For as long as I can remember upstream maintenance requests have always been qpsk and when I looked at my power level last week because of the heat, I noticed everything on upstream is now qam16.

Just checked again and very interestingly although I am still on 2 upstreams, one of them is qam64 and it defo wasn't last week. Things are looking up.

If they can get 3 bonded upstream channels on qam64 sorted before xmas then we will be in an excellent place for 300/20.

[Kushan]

Originally Posted by General Maximus

although it seems like we have been waiting forever the upstream upgrades and channel bonding, I can see VM have been doing some work on the qt. For as long as I can remember upstream maintenance requests have always been qpsk and when I looked at my power level last week because of the heat, I noticed everything on upstream is now qam16.

Just checked again and very interestingly although I am still on 2 upstreams, one of them is qam64 and it defo wasn't last week. Things are looking up.

If they can get 3 bonded upstream channels on qam64 sorted before xmas then we will be in an excellent place for 300/20.

Interesting, I thought QAM16 was standard for quite some time now and qpsk was only used as a fallback.

My 2 upstreams are still both QAM16, sadly

[General Maximus]

the data has been qam16 for yonks but the maintenance requests have always been qpsk. I have just had a quick flick through old threads and I don't think I have ever posted a dump from the upstream burst tab so I cant show you what it used to be like. This is the new one:

[Ignitionnet]

Could be nothing, could indicate you've been moved from a profile using 16QAM for data and QPSK for maintenance to one using 64QAM for data and 16QAM for maintenance.

There are a few 'interesting' software caveats on one of the CMTS platforms. If you've been resegmented onto one of the new CMTS that'd explain it.

---------- Post added at 13:43 ---------- Previous post was at 13:42 ----------

Originally Posted by Kushan

Interesting, I thought QAM16 was standard for quite some time now and qpsk was only used as a fallback.

QPSK is used for the station maintenance and requests as that's data you really, really want to get through untouched, so if the CMTS allows it you tend to use a lower order modulation for those.

---------- Post added at 13:44 ---------- Previous post was at 13:43 ----------

Originally Posted by chambohambo

Is it correct that 64QAM is more prone to interference than 16QAM?

Yes.

[General Maximus]

cpc24-linc11-2-0-custxxx.12-1.cable.virginm.net

I would have to trawl through previous posts but the linc11-2-0 is defo the same as it was years ago. I'll need to check the cpc24 bit.

Edit: just checked and it has been the same since 2012. I found a post for 2010 where I was cpc2 but it hasn't changed since 2012.

[Ignitionnet]

Cisco 10k.

Must be a profile change.

[Martin_D]

It was QPSK now QAM16 for for maintenance too

vm.PNG

cpc67132-uddi22-2-0-cust***.20-3.cable.virginm.net

[Ignitionnet]

Cheers Martin, must be a network-wide configuration change on the 10ks, probably part of the upgrade to more channels

[Jon22]

Just out of interest, what make is this CMTS?

cpc65166-telf11-2-0-custxxx.16-1.cable.virginm.net

Still using QPSK for the maintenance bits.

[Ignitionnet]

Okay yes it's an upstream profile change to allow the fallback mechanisms to work properly.

---------- Post added at 12:36 ---------- Previous post was at 12:35 ----------

Originally Posted by Jon22

Just out of interest, what make is this CMTS?

cpc65166-telf11-2-0-custxxx.16-1.cable.virginm.net

Motorola/Arris BSR 64000.

[Jon22]

Originally Posted by Ignitionnet

Motorola/Arris BSR 64000.

Ok, thanks. Do they require more work to move to 64QAM on the upstream or even replacing?

[Ignitionnet]

Originally Posted by Jon22

Ok, thanks. Do they require more work to move to 64QAM on the upstream or even replacing?

Nope. They're good to go; all of the CMTS must support 4 x 64QAM bonding out of the box.