Surf Soho, Intermittent, occasional dropouts

Problem statement:
I am experiencing occasional, intermittent Internet connection dropouts from all LAN client nodes through my WAN connection. They usually last less than a minute, but occur at what appear to be random times of the day. I cannot say if they are concurrent with Internet access events because I have a mail server running on my LAN, so there are spontaneous incoming connections as well as outgoing connections. Drop-outs are evident on multiple clients when they occur, and when they are experienced directly are matched by health check reports from the router emailed to admin (me).

System architecture:
I have a business-class cable Internet connection with fixed IP, and a mixed topography LAN with a MoCA system to distribute Ethernet to various places in the house, and four connections direct-wired to the router, all with no intervening switches: one to the cable modem, one to my mail server, one to the MoCA transceiver, and one to a MAC client.
The main system components are in my basement. There is a 15 foot section of coax between the cable splitter at the internal cable distribution point in the basement and the cable modem, which carries both cable TV and MoCA signal. All splitters in the system are MoCA compliant. The rest of the house cable wiring carries TV and MoCA Internet.
All wired connections are fixed IP connections on a private subnet, and DHCP running on the Surf SOHO serves wireless devices such as mobile phones and iPads. while wireless is enabled on the Surf SOHO, there is a second WAP serving first floor users.The fixed IP range does not over lap the DHCP pool. The Surf SOHO provides DNS to the LAN.

Diagnostic efforts to date:
I have a ping-type health test running to my IP’s DNS servers, and the router reports failures, consistent with the (extremely agitated) reports from my two Internet-game-playing sons.
I have had multiple visits from the cable company: we have checked signal levels, replaced wiring inside and outside, replaced the modem, and are now completely at a loss for a cause.
During the outages I*** believe ***that clients can connect to each other on my LAN through hard wiring and through the MoCA system, as I have not witnessed any LAN connectivity issues. I am disinclined to suspect the latter specifically, or LAN integrity in general. I have yet to confirm that I have LAN connections ***through ***the router during dropouts, but I believe that I do (my brain is not in diagnostic mode all the time, regretfully). I have observed WAN disconnected notification on the status screen, but only during downtime that was storm induced, not during any of the Ninja-style failures which are the subject of this post. I will endeavor to capture such an event to assert that I can still connect to the router during an outage. That will demonstrate unequivocally that there are no MoCA system issues (I’m thinking unlikely but possible disturbance on the MoCA line that would interfere with the cable internet signal).
Except for storm related activity (this is summer in Minnesota, so there is plenty of it), there is no cable TV outage concurrent with the Internet outage.

I hope I’ve described the situation completely.

What steps can I take to isolate the Surf Soho as a possible cause for the problem?


WAN health check is just a method to detect or checking the WAN condition (Internet link’s logical connectivity) in real-time base on the defined settings. For more information, please refer to the URL below:

Do you try before to fine tune the WAN health check settings ?

Health Check Parameters that can be fine tune:

  1. Reliable health check servers.
  2. Health check methods
  3. Health Check Interval
  4. Health Check Retries
  5. Recovery Retries

Last but not least, you can disable the health check and see whether user still experience for the internet disruption (from your two Internet-game-playing sons :o). If disruption still happen, this shown Internet WAN link issue.

Thank You

Thanks for that, Sitloongs.

Yes, I am familiar with the behavior of the healthcheck parameters. As stated above, I am pinging my provider’s DNS servers. These should be persistently available, so I am unlikely to get false failure indications, which might happen were I pinging some independent, commercial organization.

I’m not sure why you are suggesting fine tuning the healthcheck parameters. I have it at the default values of 5, 5, 3, and 3, which seem reasonable to me, and the failures which are detected coincide on my occasions with failures detected by my uses. Failures which aren’t detected occur at times when the Internet connection isn’t being used. I am not trying to obscure drop outs, so I don’t know why I would do anything differently.
You suggest disabling the health check… why? Are you suggesting that doing so my prevent disconnects? Are you saying that the router might be causing them by the health check process itself???

A lot of ISP’s DNS servers can be a little flaky and cause this type of behavior. I recommend to change the heath check to DNS and then use Google’s DNS servers and

I’m already with you on that, Tim S.; My first DNS server is the ISP’s, the second is Google’s at

But I’m not sure if I’m conveying this problem correctly. I’m not just saying that the health check is failing: I’m saying the WAN connection really appears to be going down, and that the health check is detecting the event. My users and I really are losing our connections. But the connection usually goes down when there is no need for DNS activity: my gamers are already connected to their sites, and I’m connected to my employer’s LAN through a VPN, RDP-ing into my office workstation.

To reiterate, I don’t actually KNOW that the router is blameless in this, so I am looking for a way to isolate it as a possible cause of the problem.

If I could set up my file server on my LAN as a DNS server and have the router ping that just to make sure the router isn’t failing it would give me a bit more confidence, but the router health test doesn’t say WHICH of the DNS pings fails, so the available information doesn’t have the requisite granularity. I have had a router in the past (a Cisco RVS4000) lose WAN side but keep LAN side functionality, so I know it can happen.

Again, any ideas on how I can determine that it is my WAN that is going down intermittently, not the router itself?

Are you seeing any WAN up/down events in the Event Log? I do recommend not to use the ISP’s DNS server for the health check but instead use and

Yes, Tim, the router sends me reports every time it sees the WAN down and up events - as I said in my opening post :slight_smile:
I will replace the ISP’s DNS server with the other Google number as you suggest.

What is the purpose of this change? Are you anticipating that I will stop getting reports of my WAN interruptions? If the WAN goes down because of ISP problems I’m going to lose access to Google’s servers too, so I’m not sure where you’re going with this.

Additional info:
yesterday I set up a mobile phone as a WiFi WAN source. The router successfully switched over to WiFi WAN during a broadband WAN outage.

The purpose of the change is to see if the WAN health checks keep failing when using Google’s DNS servers instead of the ISP’s. We have seen this many times before and if the ISP DNS server is a little flaky it will cause these false positives…

Tim, would you kindly permit me to say this one more time? It’s not just health checks that are failing, I am actually losing connection.

Unless you’re telling me that if a DNS health check fails your router actually kills the connection until the health check passes again? That would be silly…

As mentioned, health check just a method to detect or checking the WAN condition. There are 2 reasons will cause the health check failed.

  1. WAN link was down

  2. Unreliable health check target

Tim was suggested to change the health target to Google’s DNS. This is to eliminate the problem 2 above. Please help to change it to confirm whether the problem persists.

Thank you.


Did you ever get this issue resolved. I am having the same problem.

Not resolved, really, but I’m virtually certain it isn’t the router itself but my provider.They keep fiddling around overnight and dropping out the connection. The Soho appears to be just doing its job and barking like a good guard dog :slight_smile: