Upload drops out every few minutes for 10-20 seconds at a time

Perhaps you could post the modem stats (upstream/downstream & network logs).  Then we might spot something that isn't right.

We'd be looking for upstream modulation fluctuations or high power levels.

The Forum Team take a couple of days to get to a post and any additional evidence you post here will assist your case.

Hi joe4465

Thanks for your post, welcome back to the forums!

Sorry to hear you've been experiencing some issues with your connection. Your BQM looks spot on to be honest, are you noticing an intermittent connection across both wired and wireless devices?

We've tried to run further diagnostics but it looks like you may be in modem mode. If you could please pop back into router mode when you get the chance and let us know if the issues persist, and we can then run further tests to see what may be causing this.

Keep us posted 😊

I presume that they want to rule out your equipment as a drop out cause.   But here’s the catch: it’ll be just your luck that when they are checking (the couple of minutes allocated to that task), all will be well and we are no further forward. If you are not using a Hub 5X, Do pls take my advice and post your modem stats particularly if you do it during an outage.  The community is good at this and the elders know their stuff. 

Regarding your setup-please describe it and reconfirm that both WiF and Ethernet cut out fot those 30 or so seconds. The BQM should have reflected the dropout and indeed it is too goodto be true unless you are on Full fibre ( and the I wouldn’t expect modem mode to work). 

This exact thing has been happening to me. Most noticeable when on Teams calls or RDP where the upload channel drops out for about 10 pings and then comes back again. I can see/hear the other side but they can't see/hear me for that period. It seems to happen once or twice per hour. I tried logging a support call and they just fobbed me off.

Have you found any reason for this?

WiFi is a total crapshoot.  Drop-outs usually occur because of channel contention (or someone standing in front of the base WiFi router!).

The acid test for trying to locate the cause is to assess this against a wired device.  In such cases of drop-out, it's nearly always a circuit issue (when it's not the PC operating system suddenly getting busy).

The upstream is a delicate thing.  For a start, the channels operate in a noise-susceptible part of the RF spectrum (which is why you don't see high QAM numbers on the Docsis 3.0 channels).  Also, the spectrum is confined to a 5 to 65 MHz range into which all users in a locality must fit, the allocated range varying from segment to segment.

When you want to transmit on the upstream, you request what is known as a "mini-slot".  You'll get the first available mini-slot and that will allow a number of symbols onto the upstream, and so on until all your data has been sent.  You can imagine how congested that might get at peak times - notice might.  

It won't be your hardware, nor any firmware.

Mine is connected by ethernet via a UDM-PRO and that is connected from a SFP+ port to the 10GbE port on the VM Hub 5x. It is in Router Mode with DMZ to the port connected to the UDM-PRO.

There is no Wi-Fi involvement.
Every hour or so, it drops out 10 pings but only on the upload channel.

Here is an example:

---

64 bytes from 151.101.0.81: icmp_seq=7154 ttl=57 time=11.178 ms

64 bytes from 151.101.0.81: icmp_seq=7155 ttl=57 time=11.608 ms

64 bytes from 151.101.0.81: icmp_seq=7156 ttl=57 time=8.531 ms

Request timeout for icmp_seq 7157

Request timeout for icmp_seq 7158

Request timeout for icmp_seq 7159

Request timeout for icmp_seq 7160

Request timeout for icmp_seq 7161

Request timeout for icmp_seq 7162

Request timeout for icmp_seq 7163

Request timeout for icmp_seq 7164

Request timeout for icmp_seq 7165

64 bytes from 151.101.0.81: icmp_seq=7166 ttl=57 time=7.748 ms

64 bytes from 151.101.0.81: icmp_seq=7167 ttl=57 time=7.953 ms

64 bytes from 151.101.0.81: icmp_seq=7168 ttl=57 time=7.838 ms

---

Any ideas why it's only on upload channel?

@jim3108

Ah - Hub 5X.  My earlier remarks were applicable only to the DOCSIS network which you don't use.   In my experience, the usual reason for ping requests being timed out is to do with congestion.  Any forwarding device en-route to San Francisco (which I believe is the end-point) can discard a ping request because it is of lowest priority.  What happens if you ping BBC.CO.UK, which will have fewer hops?

ping VM DNS is a good test

Here is my BQM results: https://we.tl/t-q3KDotx9eJ

Please forgive me - it's not the XML we want (nor can conveniently use).  It's the simple link that Thinkbroadband provide to your BQM.