HH3 - Infinity - Local WiFi Channel 'fade'?

[GRAEME1873]

Update from me. Long story but after initially saying they would send me a replacement HH3 I have actually been sent a new HH5. Installed it last Friday and have had no issues since. Speed is actually slightly slower than I was getting with HH3 (28 mbs vs 36 mbs) but is totally solid.

[keefo]

I've had to reset my channel every morning this week.. and one evening.  I've sent this to BT

I'm having the exact wifi "fade" problem described in this forum thread: https://community.bt.com/t5/Other-Broadband-Queries/HH3-Infinity-Local-WiFi-Channel-fade/td-p/144629...

This is clearly a general problem with HH3.

I've spoken to India and been advised to change the channel, this only provides a temporary solution, have to do this at least once a day.

I would be grateful for a replacement Hub, as offered to others.. I don't want to spend further time on the phone to india when this is a known issue with HomeHub 3
Thank you, Keith

[fentigolf]

Like you Keith, I also have had to restart all but one day in the last week since I first notified BT of this.

In fact, on Monday evening when I returned home from work, the speed was less than 1mbps. A restart sorted that out and got the fastest speed I've had since the beginning of the year at 8mbps (normal BB).

No response from the Hub Team seems to suggest they are at a loss as to what is wrong and I also think it's some sort of change at BT's system which is causing us all grief.

[Gneiss]

"I also think it's some sort of change at BT's system which is causing us all grief."

Absolutely this fentigolf!

It is in my opinion absurd to suggest that all these HH3s have simultaniously developed hardware faults after working perfectly in some cases for years. This has to be a change BT have pushed that is incompatable with the HH3 and has caused this instability.

We know the firmware hasn't been updated recently so it has to be some sort of configuration change they have pushed through....

I think it's more than coincuidence that this occured at the same time as the disastrous HH5 update that caused HH5 owners so many issues.

[Keith_Beddoe]

A change in the Infinity network characteristics, is something that is being looked into, as this issue only seems to affect HH3s on Infinity, not ADSL.

Changes in packet size or other TCP parameters seem most likely, possibly filling up the wireless buffer.

[keefo]

" this issue only seems to affect HH3s on Infinity, not ADSL."

Not so Keith_Beddoe  I'm not on Infinity and it is definitely affecting me!

[jac_95]

Could possible be the change to the MTU on the network side which would mean the IP datagrams would need to be fragmented by the IP protocol if the original datagram had too many bytes to be transmitted to the next router.

However I think the IEEE 802.11 protocol has a MTU size of 7981 bytes.

When we were using a HH3 (on ADSL though) every now and again would see entries in the log saying the TR-069 had changed some values/perameters.

Is the HH3 MTU still set to 1500?

[Keith_Beddoe]

---

@keefo wrote:

" this issue only seems to affect HH3s on Infinity, not ADSL."

Not so Keith_Beddoe  I'm not on Infinity and it is definitely affecting me!

---

The majority of complaints do appear to be from Infinity users, but there may be other factors affecting ADSL.

I believe the BT hub team are correlating the results.

The problem is that this forum only represents a very tiny number of HH3 users, compared to the total number of HH3s out there.

[Keith_Beddoe]

---

@jac_95 wrote:

Is the HH3 MTU still set to 1500?

---

I do not have a HH3, so I cannot check, but whatever it is, seems to be affecting the way the wireless module handshakes with the network. Perhaps a build up of unacknowleged packets filling up a buffer store somewhere? It a bit outside of my scope, having mainly being involved in X25, well before TCPIP

Its an interesting problem, which should keep the hub team busy, assuming that it isn`t a bad batch of HH3s.

[keefo]

Maybe there are more reports from Infinity users because they have an expectation of higher speeds anyway?

There doesn't seem to be much urgency in resolving this - collating results sounds a bit useless. Why not get hold of an affected HH3 and try to reproduce the problem.