Initial issues with the R3

[Alias 1aB]

Got the R3 in today. I thought it would be kind of a simple swap out of the R2 but I was wrong. So far Xbox Series X is showing DoubleNaT (didnt have this issue with the R2),  ping optimizer configures itself at 99% download and 100% upload, built-in streaming apps were actually buffering on my TV, and my android IPTV box will not playback video when connected to the R3 as it just stays frozen. I had to connect the iptv box wirelessly to the fiber gateway to get it working again. The buffering stopped in the streaming apps once I manually changed the download and upload congestion to 70%. I still have not messed with the double Nat issue on the Xbox. I restarted both the gateway and the router, just have not turned the xbox back on to see if that solved that problem. I dont know if I will be able to turn DHCP on the gateway off if I cannot use my iptv box through the R3. I really dont think ping optimizer/ congestion control is working properly. I ran multiple bufferbloat tests and they were all Bs and Cs. I used to get A+ and As with the R2. I even had better results before when connected directly to the fiber gateway. I have congestion control set to ALWAYS ON right now. One of the positives so far is more accurate detection of actual internet speed. The R3 detected 914 down/ 883 up. My connection is 950 down & up. The R2 never came close. I miss the little connection benchmark tool the R2 had. The R3 is missing it.

 

Any suggestions to help resolve these issues would be very welcoming. Thanks.

[Alias 1aB]

update:  fixed the double NAT error. had to go the NAT/Gaming section in the ATT Fiber Gateway and had to choose Xbox Series X service for DumaOS device. I also turned on UPnP on the R3. I also changed the IP Passthrough Allocation mode from Passthrough to default server and assigned the default server internal address to the IP address the gateway gave the R3. 

[Netduma Fraser]

Glad to hear you got the NAT sorted. Regarding bufferbloat, speed test bypass will make those tests report your full speed and therefore not reflect your Congestion Control settings. You can disable this under Congestion Control

[Alias 1aB]

yeah, I disabled speed test bypass under congestion control. It makes no difference. when I manually adjust the congestion control sliders, I still get really poor Buffer bloat scores.  Almost every time I run ping optimizer, it gives me 100% for both upload and download. Also the ping test detects a much lower ping than what the ping optimizer test detects. The difference is from 15 ms to 33 ms. I assume that has more to do with the servers it is connecting to. Makes me think ping optimizer is not working as intended. Is there a way to force ping optimizer to use the same server it is doing the ping test from? Because my actual ping is anywhere from 12-15 ms to the fastest test server. 

 

Thanks

[Netduma Fraser]

Where are you testing bufferbloat? What bufferbloat results are you getting? How did you setup Congestion Control exactly?

For Ping Optimizer we're aware it's not 100% reliable at the moment, it's something we're working on to improve, there isn't a way to force the server it uses unfortunately but I can add that to the request list. I'd focus on setting CC manually for the moment.

[Alias 1aB]

waveform website  Bufferbloat and Internet Speed Test - Waveform

 

 

These are the settings on the R3:

The network speed was autodetected at 928 down / 866 up. ping optimizer autodetected 100% download / 59% upload. Congestion control always on. I disabled speed test bypass.

When I set congestion control manually, the Buffer bloat results are even worse.

DHCP is still active on the ATT Fiber Gateway (Humax BGW-320) but the only devices that are connected to it are a Ooma VoIP phone box (wired) and IPTV box (wireless - I had to leave 2.4 Ghz on just for that because that box will not work with the R3 and I don't use it that often). Neither device uses much bandwidth as is.

And like I said, ping optimizer uses a different server than the ping test does. I think it literally picks the geographically closest server to my location.

I am thinking I might try to factory reset the R3 and see if I can trick it like I used to the do with the R2 by telling it I live near the lowest ping server instead. 

 

[Netduma Fraser]

Start with a high value for Congestion Control, say 95% then do a bufferbloat test, decrease by 10%, test, decrease by 10% etc, until you get to a value that is pretty good and then try 5% either side of that value to see if it can be improved. It's important to note that Download & Upload on Congestion Control don't have to be the same value and that you may have a better experience with differing values.

[Blinker73]

@Alias 1aB, if you use a dhcp server other than the one included on the R3, it might not work.  See my post on the subject. 

 

[Alias 1aB]

Blinker,

I haven't had any issues with DHCP at all. I am able to use DHCP on both the fiber gateway and the R3. I am essentially using the same setup I had with the R2 except now I can use the Wi-Fi radio on the R3. The R2 gave major issues when wireless devices were connected. And I am now connecting all my other hardwired devices to the R3 also because it is at least detecting my network speeds a lot closer to what it is than the R2 ever did.

[Alias 1aB]

On 12/22/2023 at 11:08 AM, Netduma Fraser said:

Start with a high value for Congestion Control, say 95% then do a bufferbloat test, decrease by 10%, test, decrease by 10% etc, until you get to a value that is pretty good and then try 5% either side of that value to see if it can be improved. It's important to note that Download & Upload on Congestion Control don't have to be the same value and that you may have a better experience with differing values.

Ironically on the R2 the magic numbers were 78% download and 43% upload. Almost zero bufferbloat. But I manually set network speeds to 900/700. The R3 is going to take a little bit more trial and error. Does ping stabilizer affect the bufferbloat test? If so I have to take that into account.

[Netduma Fraser]

1 hour ago, Alias 1aB said:

Ironically on the R2 the magic numbers were 78% download and 43% upload. Almost zero bufferbloat. But I manually set network speeds to 900/700. The R3 is going to take a little bit more trial and error. Does ping stabilizer affect the bufferbloat test? If so I have to take that into account.

Do let us know how you get on with that! Steady Ping? In theory no it should just be applying to the Geo-Filter but did see some reports where it was affecting ping outside the Geo so perhaps just disable it before doing the tests to be sure.