Getting read for SmartLink -- test your upload performance

  • 7
  • Idea
  • Updated 2 years ago
SmartLink is coming and a number of you are going to want to give a try from Starbucks, your Mother in Laws, etc.  :)

This is going to bring a new phase of 'how good is my internet'.  We have always focused on download numbers since we were streaming or downloading.  Now, your upload is going to be part that is going to affect just how good your Remote operating is going to be.  

It is more than speed and latency is just as important.  In fact, latency is going to be critical.

For over 11 years, I have been operating remote using the internet.  At my remote base, my best speed is still about 14mb/sec down and about 1200kb/sec upload.     For years, I used Remote Rig and a TS480, and that worked well enough.

Once the Maestro came out, it was a game changer for me.  For over the past year, I have been running Softether VPN and the Maestro that traveled with me.   But, that required more upload and instead of using about 110kb/sec, I was needed about 400kb/sec with Softether.  That continues through to today.

As more upload was required, I had to work to tune my upload.  I learned about QOS and Bufferbloat.  Not to go into details here, but Bufferbloat is bad for streaming.   Streaming is what we are going to want when remote operating your Flex 6000.   

At the moment, there is only one good website for testing your Bufferbloat, and that is   You can do it manually by doing the steps outlined here:

This is my test on a Sunday night (I should have been focusing on the hockey game).  BTW, this is considered somewhat higher speed for rural Canada.  High speed internet is not found in places where you want to install an HF remote station.

In the diagram above, my Bufferbloat number is a C.  At times it is an A.  For the most part, I drop about 0.01% packets, which is almost non-noticeable.  

So, how do you fix your BufferBloat?  Your ISP can't/won't/doesn't care.  That leaves it up to you or your friends.  If you have a gamer in the house, he or she might have already fixed this for you.

First, you'll need a router that you can manually tune your network connections.  Usually, even doing some QOS work will help, but it will be trial and error.   You'll be making changes and then running a test to see if it gets better or worse.  Just like tuning an antenna.

Routers that you can load DD-WRT, OperWRT, pfSense, etc. are your best bet.  

There is a new commercial router that is plug and play, sold by Evanroute.  It takes a week have the IQrouter automatically tune your network.  Personally, I have had great success with them on the networks I have installed them on.  Your mileage may very.  Some of the Alpha Team members have been playing with the IQRouter for a few weeks.  They are available on Amazon.  They can be configured for use on DSL, Cable and even all in one modem/routers.  

The IQrouter website is worth the read if and only you want an easy to read description of the problem.   BTW, the reading above was done with a pfSense router I am testing and not an IQrouter.  When I tested the IQrouter, it was almost always an A or B for my bufferbloat, however, I had to take that router and install it at a customers site.
The internet (or pipe as it is called) is something that FRS can't fix or control.   In the diagram above, my Bufferbloat number is a C.  At times it is an A.  For the most part, I drop about 0.01% packets, which is almost non-noticeable.  

The next phase of your HF operating is going to make a huge change.  Have fun with it and brag when you got to do it differently.  After all, that is what Amateur Radio is all about.

Mike va3mw

p.s.  I will be in Dayton, so if you see me, feel free to say hi!  I look forward to meeting a bunch of you.
Photo of Mike va3mw

Mike va3mw

  • 824 Posts
  • 199 Reply Likes

Posted 2 years ago

  • 7
Photo of Larry Benoit

Larry Benoit

  • 91 Posts
  • 23 Reply Likes

Thanks for opening this thread on remote ops, Internet service issues and bufferbloat.  I wanted to share my experience with Spectrum/TWC service and a Netgear R7800 X4S gaming router. 

First, the router was originally configured with Quality of Service (QoS) enabled with automatic default bandwidth settings (about 40 mbps download and 6 mbps upload). I was typically getting grade "F" or "D" for bufferfloat by the speed test.  Average speeds were just a little less than the default QoS bandwidth settings.  

Second, I had read several reports on router forums that manually lowering the automatic bandwidth setting for QoS would reduce bufferfloat.  After some experimentation with these settings, I found that manually reducing the default settings by about 25% resulted in a good balance between bufferfloat and average speed. The test reported bufferfloat had improved to grade "B" on average with improvement particularly large with downloads.  This helped mitigate Spectrum's evil dynamic buffering and throttling, cutting peak latency by more than half. 

Despite the previously poor bufferbloat condition, I've still had excellent results operating remote over VPN (OpenVPN and SoftEther), seldom experiencing dropouts, including a full month of daily sessions from California to Maine. However, I hope this improvement in bufferfloat latency will help in the future when network conditions are marginal. 

Thanks again for your post -- I would have otherwise not investigated this issue. 

Larry KB1VFU
Photo of Mike - W8MM

Mike - W8MM

  • 220 Posts
  • 50 Reply Likes
Looks like I'm good to go at home:

Now, I have to play with my work LAN connection to keep all green bars.  I occasionally get 2 or 3 yellow bars when operating my Maestro from there.
Photo of Mike va3mw

Mike va3mw

  • 824 Posts
  • 199 Reply Likes
Excellent.  Don't worry about the Maestro Bars.  I routinely have 1 bar and never hear a difference.