120M : New upstream == pings over 500??!?

[pip08456]

Originally Posted by Rexz

New IP. Looking better.... NOT!

Looks like your issue was fixed at about 10.45 this morning.(if that is a live TBB graph)

[Rexz]

Indeed it was and thank you to everyone on here who helped. I was rather impressed at how quickly VM got this sorted! I was scared this would turn into long waiting game. pings seem better too :P

[pip08456]

Sometimes they do get it right!

[craigj2k12]

Originally Posted by pip08456

Sometimes they do get it right!

But mostly they dont...

[Sephiroth]

@ REXZ
Have a look here. VM say they've cracked it.

[qasdfdsaq]

Nice.

[Sephiroth]

But not so nice that it only affects the SH2 and the SH1 apparently is more tolerant of these "fringes".

[sollp]

Originally Posted by Sephiroth

But not so nice that it only affects the SH2 and the SH1 apparently is more tolerant of these "fringes".

The Intel chip within the SH2 is less tolerant of any deviation in Symbol rates and frequencies being slightly out on the UBR. But of course there is never anything wrong with the UBR it's the local network.(network engineers will understand this)

The SH1 isn't so specific and can handle a slight misplaced input on the config of the UBR

[qasdfdsaq]

Ooh it's an intel chip? I thought people said it was a Broadcom

[Sephiroth]

It's the Puma5 chip in the SH2- which was also used in the VMNG300.

Solly's got it right I'm sure. It'll have been off-centre frequency.

[Kushan]

To be fair, according to Intel the Puma5 is a series of chips, so it's not necessarily the exact same chip in both the Ambit 300 and the SH2.

[horseman]

Originally Posted by Kushan

…. so it's not necessarily the exact same chip in both the Ambit 300 and the SH2.

Particularly since Ambit VMNG300(aka uBee U10C035) apparently came with the Texas Puma5 chipset(c2008) and TI data sheets list 5 variants in their data sheets!

Intel only lists 3 Puma5 variants(c2011) in their family and two of those are eMTA versions which would circumstantially leave TNETC4830 as the likely (optimised DATAMODEM) candidate? or…. ?

[Sephiroth]

Originally Posted by horseman

Particularly since Ambit VMNG300(aka uBee U10C035) apparently came with the Texas Puma5 chipset(c2008) and TI data sheets list 5 variants in their data sheets!

Intel only lists 3 Puma5 variants(c2011) in their family and two of those are eMTA versions which would circumstantially leave TNETC4830 as the likely (optimised DATAMODEM) candidate? or…. ?

... which means that the Intel Puma5 is a minor evolution of the VMNG300 Puma 5.

[horseman]

Originally Posted by Sephiroth

... which means that the Intel Puma5 is a minor evolution of the VMNG300 Puma 5.

Different substrate possibly and fabricator but a TNETC4800/10/30 from Intel wet line must be same circuit design as TI TNETC4800/10/30?
Unfortunately as stated by others there's no parametric comparisons available in public domain for legacy Texas and current Intel manufacturers!

However the fact that Texas originally additionally produced TNETC4820/40(not currently listed in Intel P5 family) with some optimised Data Modem features begs the question as to what exact (TI)chipset variant was included in VMNG300 as opposed to latter (Intel)SH2?
Even considering VM/Netgear may have included the eMTA chipset but left the I/O redundant in VM's 2nd order product may explain some of the "dumbed" down feature code in firmware/eCOS?

[Sephiroth]

Quod erat demonstrandum.