That is my experience now, but I suspect it may have been a quick workaround and now the systems have been updated so that g.inp are supported up & down according to the capability of the modem. To have that adaptive capability may have taken some changes to the configuration software.

So you concede what you posted is wrong

It was right at the time it was added to the article. That was the first fix so that ECI modems on Huawei cabs didn't lose 20Mbps or so.

Hope I haven't lost track here, but are you saying the ECI Modems have now been updated to support G.INP in both direction's?

No, I'm not.

I'm pointing out the date of the linked article and the fact that the problem was diagnosed as the upstream G.INP on Huawei cabinets knocked ECI modems for six. So upstream was disabled - as per the "Latest" addition to the article.

That was why I said "It was right at the time it was added to the article". I didn't and don't say anything about the position now.

Edited by RobertoS (Wed 03-Feb-16 13:31:56)

Right, thanks for clearing that up..

Upstream was never disabled. It was changed so it came on a couple of weeks after downstream.

In reply to a post by professor973:

Fourth post here says there is firmware in the wild.
http://forums.digitalspy.co.uk/showthread.php?t=2066551

That post is nearly a year old and was very likely referring to 4.7.5.1.83.8.204.1.11

Oh, and my HH5 reboots every 14 days as well.

No, it changes nothing. What I linked to is a report that upstream g.inp had been disabled due to compatibility issues. That was some time ago, and it seems to me that upstream g.inop has been enabled where the customer's modem is capable of supporting it. That would fit with my experience that the HH5A syncs about 10% slower than an HH5B in the upstream direction on the same line whilst the downstream on both are pretty well the same. Somebody else has also reported that the combination of a separate OR modem & HH5A gives about 10% more upstream throughput against an HH5A on its own but similar downstream.

Not definitive of course, but fairly strong evidence. Your HH5A seems to exhibit unusual symptoms, must notably not rebooting every 14 days (approximately) which everybody else seems to have noted.

So I would say there's decent evidence that the HH5A has some g.inp support, but it has some limitations too.