CA3LE

Wow, that's a big difference on multithread for sure. Next time you do those comparisons try to keep the test size the same. Instead of that 6.8 MB and 42.1 MB ... maybe just do 50MB manual selection for both. In that instance the 6.6 Mbps result would have taken about 60 seconds for a 50 MB test. The 34.7 Mbps result would take about 12 seconds. Takes a little longer but it's a better comparison, apples to apples. Don't get me wrong, the results you show above are a great comparison. I'm just pointing out how to make them a little more definitive.

You should be able to stream without issues... except when your speed looks like this. https://testmy.net/db/3jtqGbupv At that time you may experience quality fluctuation but I would still bet it wouldn't cause major buffering issues. If you're able to hook directly into the modem and get better results we should look to improve your wifi. Could be as simple as better router placement or changing the channel in the router settings. For all of your testing in this instance, pick a size and stick with it throughout your testing. In your case, 50 MB should do the trick. You're already doing the right thing by focusing on the download test. Just navigate to the download page and manually select the size. * I know you already did but test directly to the modem again, reset the power to have the modem get the computer Mac address. Make sure we get a nice baseline from right now. Run the download test like 3 times. * Next plug the router in, reset the power on the modem, plug directly into the router this time and run 3 more tests. Keep the test variables the same for a more scientific comparison. The only thing that should change is the router being added. * Do 3 more tests on wifi about 10 feet from the router. Do another 3 tests further away in an area you often use the wifi. Bring your laptop near your TV and test the wifi to that location too. Compare those results. If the wifi results are terrible you may just have interference in the area. Sometimes going into the router settings (often found at http://192.168.1.1) and changing the wifi channel helps. Change it and test to see if there is an improvement. Also look at the placement of the router. You want it away from other electronics, even devices that don't emit wifi can and will interfere. The best placement is at the center of your home (away from any brick), the worst is in the corner. Most routers have omnidirectional antenna, so the radio waves go in 360 degrees. If you place it in the corner of your home you're only effectively getting 90 degrees of the total output where you need it, 270 degrees is going away from your home. Having said that, the router in my own home is placed in a way that wastes 180 degrees and I get great performance everywhere. You can actually take a reflective surface (in a pinch the inside of a Pringles can works) and place it behind the antenna in a way to make it reflect the signal back to the house. Just make sure it doesn't have lines, crinkles or wrinkles, nice smooth reflective surface.

I wouldn't be switching out that modem. In my experience Motorola Surfboard's are the best in the industry. It's all I ever run at my own house and it's all I ever recommend to friends... the Arris modem/router combo (Arris SB6900-AC) IS a Motorola surfboard by the way. If you already have a SB6180 you should put your money towards a standalone AC router, if anything. That modem has great specs, especially when you consider you're on a 50/5 package. I'm looking at the specs for the SB6180... DOCSIS 3.0 Supports IPv4 and IPv6 - the latest Internet standard Downstream 343.072 Mbps (8 channels) / 42.884 (single channel) Upstream 131.072 Mbps (4 channels) / 32.768 Mbps (single channel) Your modem specs are the same as mine (I have an SB6141 right now) and I often perform over 180 Mbps down and 25 Mbps up... with room to spare. Even if your modem only supported IPv4, they aren't going to leave you hanging when they start doing IPv6. They'd more likely support the existing IPv4 customers till they eventually dropped off. It's a moot point since your modem supports IPv6. If a tech told you otherwise... it calls everything else he said into question. Your router is faster than your connection. If you tested directly to the modem and got similar terrible results I'd be looking first at the modem. Since it's fairly safe to say that it's not the modem, being newer and all, you then must look beyond the modem. From the coax connection going to the back of the modem... back through your wall plate, on to the splitter, out to your drop coming into the house, then to the pedestal where your closest neighbors all hook in.... then beyond to Cox's network. --- it could always be the modem but IMO it's less likely in your case. I see that you've been testing with Firefox. Have you tried Chrome? I'd expect similar results but it's worth checking. Also, if you have other computers or devices you can test, their results will help make sure that it's not an issue with the computer itself. Similar results across your computers, tablets and phones for instance will show that it's not just an issue isolated to one device. Something else that helps diagnose connection issues is to test against a few more servers and compare multithread results. Don't go running out replacing perfectly good equipment. Let's be real here.. your router (CISCO WRT320N) isn't bad either. It's Dual-Band N. Wouldn't be my first choice today but it's still for sale on Amazon Prime so it's not that old. Unless you're concerned about the transfer speed of files within your home network, have a bunch of AC wireless devices or have a large number of devices in general.... save your money there too. Like I said, that router is much faster than your Internet, it's not the weak link.

Thank you for the suggestion. I added support for testing against port 8080. As you imagined, simply visit [no longer works] -- works with all servers, similar to the https (port 443) test at https://testmy.net. Let us know your findings.

What modem are you using? What router? Was your speed ever acceptable in the past? Have you tried testing with ethernet directly connected to the modem?

Hi Wayne, I wish more people questioned things like you do. If they did I might be able to catch up to ookla's traffic volume. First I'd like to show you this thread >> https://testmy.net/ipb/topic/28902-why-do-my-results-differ-from-speedtestnet-ookla-speed-tests/ There does need to be a movement. I've been trying to raise awareness of these issues for many years but I don't think it means as much coming from me. To have a tidal shift in TestMy.net's favor a major news outlet or someone with great influence needs to cover the story. Many people tell me what you're telling me but the majority doesn't question the largest player in the room. They assume they know what they're doing... and they do, they know how to make things appear better than they are in reality. Every aspect of their methodology is in place to render a more favorable result. Is it any wonder why so many ISPs use them? I focus my time on building the program and infrastructure that's TestMy.net. I rely 100% on word-of-mouth otherwise. My hope is that I'll continue to build my program and when the time is right more will know about it. It's grown organically to this point, it's a slow process but I'm one person. I don't have a PR team, programmers, designers and system admins. Just myself. There is a vision that's far from achieved. So I focus my time on building. Slowly more people are starting to understand the difference between my real Internet speed test and all the bs out there. Keep spreading the word. I appreciate your kind words and strong support.

You must be on a shared proxy. This is uncommon. Looking at https://testmy.net/stats/?&t=c&d=01172016&x=3&l=25&q=60619026442 You only need to be concerned about the tests that you've personally taken. Since your IP appears to be changing (and many other people are using the same IPs) I recommend that you make sure that you're logged in to properly keep track of your results. Sorry I haven't provided very good documentation other than answering questions in the forum. Need to work on that. Many things you may wonder about have tooltips. You need to hover your mouse for a few seconds to see them. Those icons are a quick visual aid to tell you that you're testing against San Jose, CA to your Android in TX US. If you change the identifier or server those icons will change. Some icons like android, iphone and ipad are automatically added depending on the detected device. The lightning bolt is the TMN logo and stands for a multithread result [you have to look at the result details to see the server(s) tested against]. The American flag stands for Washington D.C. -- To see more examples visit https://testmy.net/live and hover over the icons you're curious about. So don't go hunting anyone down. They aren't using your connection... they're just using the same proxy server as you are. Extra note since you're on an allowance. Once you have more tests taken you'll be able to see some additional information below your graph showing the amount of data transferred. Keep in mind that this only calculates based on the size of your final result, not the smaller tests leading up. If you're worried about data allowance you may want to use manual test size or express on the download test, upload test and auto test to better control the amount of data used in the test. Using those options will force the test to cycle less, using less data... which in turn will make the data calculation for your query closer to the actual amount of data used. My own as an example Below I change the query to reflect the last 30 days... then alter the results/page to show all the results on one graph. Whatever shows on the graph is what will be added up in this calculation. That 21.11GB could have be 40GB of actual transfer for all I know. Just keep that in mind when you see those numbers, the actual is always higher than what's displayed.

6 MB upload test... Activity Monitor, 1 second polling during the test. 4 kB/s to 43 kB/s fluctuation throughout.

What did you do, delete my post? Oh well. I was just saying that it ended up taking over two hours to upload that 158.5 MB to youtube. For fun I ran speedtest.net, also to Colorado Springs and got... ... if my upload were that fast the youtube upload would have only taken 11 minutes. Why so many people use that site, I have no idea. Maybe they like looking at a page full of lies and ads. There are so many ads I can't find the result! Visit us at ads.speedtest.ads. Why do people use that?!

Could it be a type of phased array radar? Cool! > BUILD A PHASED-ARRAY RADAR IN YOUR GARAGE THAT SEES THROUGH WALLS

That's a good explanation. When I've seen those in the past they look like this... But really, an array of any antennas would work. How certain are you that's what it is?

I keep seeing this at DIA and I can't figure out why they built it. Here's where it is... (lat: 39.8481 long: -104.6734) And here's what it looks like... Appears to be a huge array of some kind. Does anyone have any idea what it's used for?

Still running horrible. https://testmy.net/stats/?&t=c&d=01152016&x=3&y=u&l=25&q=4602014672148 Only 160 MB upload to youtube... That would normally be done in 60 seconds... ... even worse since I started this post.

https://testmy.net/stats/?&t=u&d=01132016&x=5&y=u&l=25&q=CA3LE Started yesterday, upload is almost nonexistent. Download is affected too but not as pronounced. Called Comcast and the automated phone system said that there are issues in my area and techs are working on it. I'll be testing to see how long it's like this and then call Comcast for a credit for those days. I appreciate that's it's still online but that degraded upload speed is affecting my daily use. No question, I just wanted to share what it looks like when your upload speed is severely affected by outside technical issues. ... and this is why you test regularly. You can be credited when it's not right. By the way, my Internet really feels fine. It's only when I upload something that it's an issue. Surfing around feels normal.

When you test a VPN service here you're still accurately testing your speed. Regardless of the method the data uses to get to you the principals for measuring the speed are the same. The basics. The test starts, a timer starts, data flows... when the data is done the timer stops. If if takes longer to send or receive the information it's reflected in the result. Keep in mind your speed with any VPN is going to be degraded. You're adding a middle man. The connection has to first route through them before it gets to you and then there's often encryption involved. For comparison, my best VPN tests only gave me about 60 Mbps out of a 150 Mbps connection. The information is being relayed back to you, it's to be expected. The time it takes to load speed tests at TMN directly correlates with the time you'd expect to see if you were downloading or uploading a large file. I might not be able to trace back to your IP but the principals for testing are the same.

That's a hell of a connection you have there! That TiP graph is also on your results page on all classic download tests. Keep in mind the single thread upload test here isn't going to show your full speed. I'm currently working on a new upload test for connections like yours.

Proof during the TMN test that you were topping out the connection. Funny the ookla tests read higher yet didn't cause your voip and rd streams to die out. The only way in my mind that happens is if there's funny business or it's just plain BS.

humm, well that's DOCSIS 3.0 and 16 channels down 4 channels up. I'd expect it to handle your 150 Mbps package but I've seen that exact same modem at a friends house... and their connection runs like sh*t despite being down the street from me and having the same provider. Spotty and will fluctuate from 1/2 their package speed to the full speed just by walking a few steps. I blame it on being a combo unit but what do I know. https://testmy.net/stats/?&t=u&d=01092016&x=3&l=250&z=61&q=spudler_t So, these are the best linear results you've seen? Try selecting Colorado or Dallas, pop off a couple of 100 MB download tests and make sure that it's not a routing issue to San Jose first. When you're comfortable with it's consistency select the server that gave you the most favorable results. Run another couple of larger tests, then directly connect the modem/router and run a couple more. If you get much better performance wired, then you know it's the wireless. Either bad router placement or interference are likely causes in that instance. If you want to go to great length to bypass that modem you could go to Best Buy or Target (Best Buy in my area didn't have what I needed, found out that Target has a good selection and the price was comparable to Amazon)... somewhere with a quick easy return policy in case there's no change... and pickup a surfboard. I have a SB6141 and more than max out Comcast 150, more expensive 16 channel modems didn't perform as well in my testing btw. If you pick up another modem and it performs better then you know that your lines coming in are fine and the Cisco unit is at fault for whatever reason. Then you'd need a router to pair with the new modem. If you get the new modem and it performs like the other one... then... You'll have to call to swap the modems around and it can be a little hassle but for me it's easiest to troubleshoot by eliminating variables and comparing to equipment that's known to perform well.

First, what modem are you using? (I rearranged the provider forums and moved your post btw)

Actually there was a major issue. Across all servers except the Master server and San Jose because they're compiled differently. Backend software updates were done yesterday across all servers and I forgot to reinstall php-mysql... MariaDB was installed but basically php didn't have instructions for how to talk to the database server. ... it will only cause an error in that instance if the test type selected uses the database algorithm. Thank you, thank you, thank you for bringing that to my attention quickly. I rarely use the combined auto test (I usually select one type or the other in my own personal use). So a mistake like that can persist for a while without awesome members like you. It started working for you all of a sudden because I noticed this thread and started working on it before I started posting. --- when @mudmanc4 tested on the other hand I hadn't started yet. So I assume that he must have been selecting the auto test that doesn't adjust size based on what it sees in the database. Combined vs Combined Express

I moved those forums to defunct a few weeks ago because of low participation in recent years. They'll be reorganized, updated and put back out front soon. As you've probably noticed by now AT&T Gigapower is detected as AT&T U Verse. So to research that you'd want to look into AT&T U Verse speed test log, extend the output to 1000/page, browse for the fastest results... then look in deeper detail at the users getting the fastest speeds. If you search for AT&T within TMN you'll find all the databases associated with AT&T. I think Google search is a nice quick way to search TMN. For instance, go to Google and search for site:testmy.net at&t ... when I want to find something here, I usually do a search like that in the address bar because it's just quicker.

I agree with @mudmanc4 -- most likely routing/peering with your ISP. This is why there are mirrors, so you can identify these types of issues by comparing the same test on different servers with different routes. Nothing out of the ordinary with my own testing against Dallas... https://testmy.net/db/mR0MlT-u1

That will take a little while to update but will update automatically in time.

Cool.. fixed it on my phone

I see one more instance I need to update... will do that in a bit.