Bandwidth loss over distance

For a hardwired connection, what's a good guideline for bandwidth loss?

I have a 30 foot run of cat 6 cable going from my modem to my router. At the modem I see 200. At my the router, hardwired, I see 50. 

Comcrap are putting the difference as not using their speed test on my end of line. Funny though, every single speed test EXCEPT their specific one shows my speed at 50/10. Their special test shows me at 200/10.

Those are both E36 M3ty speeds as far as I'm concerned. Is this just their way of using a dynojet instead of a mustang dyno?

They sent a tech out today because I've only been sending in 25% of the bill payment. Hey, if I'm only get 25% of what I'm paying for, then why should I pay 100% of the fee?

Citizens fiber can't get their parallel gigabit service here fast enough. Not only is it a parallel gig up and down so 5 times faster down and 100 times faster up, but it's only 60% of the price as comcrap. Only 3 more miles and 4 dozen lawsuits to go, they have had to sue for every single pole they run fiber too. I could have sworn antitrust laws existed to prevent just such an occurrence. 

6 foot Ethernet at the modem you see 200/10?  30 foot Ethernet into the router and then Ethernet to your computer you see 50/10?  What do you get when you use the 30 foot Ethernet to your computer?  Having troubleshooting thousands of these type issues I would have to tell you this is on your end as you already say you get 200/10 at the modem. Anything past that is yours not the ISP. 

In reply to Stampie :

Only when I use their specific test do I see advertised speeds. Every single other test I can find to run, hardwired and wireless spit back about the same, which is about 25% of what I'm paying for and what their specific test gives.

I feel like it's a rigged test. I don't know how, but I can certainly understand the why. Especially with this group of corksoakers. 

I can tell you that they run their speed test servers on their network because that's all they can control. They can't be responsible for speeds on the open internet. 

So I'm still not understanding your statement that you see 200 at the modem but only 50 at the router. Using Joe Blow speed test at the modem what do you get?  Run the same test at your router and what do you get?

First question, if you're using their test, do you see 200mbps at the router?

Second one, normally you're getting "up to" speeds, not guaranteed speeds and bandwidth unless you're leasing a line.

As long as the 30 ft cable is in good shape you should not really see any drop over that distance.  Is it a home made cable or store bought?

Now plugging my laptop in tells me no internet, but plugging the router into the modem gives me internet. Power cycle the modem and it recognizes the laptop again, reads me out 50/10. When not an hour ago comcrap was here pulling 200/10 on the same cable, in the same port, on their special test. 

I have the "newest and best"  wireless gateway garbage modem, with the Wi-Fi shut off.

Speedtest.net has been the standard speed test, local servers, low ping, and  repeatable results.

I know, I really shouldn't expect anything but frustration with comvajajay, but seriously, 25% of advertised speeds? 

Unless they're intentionally confusing megabytes with megabits on the bill and on the website. 

Store bought cable, about a year old. I can't get all the wires to stay put with my homemade cables so the tool just gathers dust. 

Ok I'm out. I suggest that when the tech shows up you present your argument in a logical calm pleasant manor. They see these complains a lot and 99% of them are either the customers equipment or just missinformed customer. Hopefully you are the 1% but make the tech want to help you. 

Try a different speed test (I like speedof.me) to see if you can duplicate the results?

Advertised speeds I could quickly find on Comcast suggest their speeds are in Mbps, not MBps. The numbers you mentioned seem to back up that both of you are using the same units of measurement.

One other thing that might well be worth doing (especially if you have a Linux box around) is running a ping test for 10-20 minutes to something like www.google.com, stop the ping and see if it shows any packet loss. Depending on how Comcast measured the line speed, they might be seeing 200Mbps but a surprisingly small amount of packet less can bring this down to 25% or less of actual throughput.

You should have zero bandwidth loss over a distance like that. Consistency is key in performing the test; it's possible your router is costing you throughput. If you ever get a consistent 200/10 result with a given configuration then it changes based on you moving stuff around then it's your equipment causing the slowdown, be it a cable, router, interference from outgassing limestone, etc. OK probably not that last one.

Paying only a portion of your bill, even if due to a service problem, is not going to end well for your credit or continued ability to connect to the internet, unless you've come to an agreement with Comcast.

Stampie said:

Ok I'm out. I suggest that when the tech shows up you present your argument in a logical calm pleasant manor. They see these complains a lot and 99% of them are either the customers equipment or just missinformed customer. Hopefully you are the 1% but make the tech want to help you. 

He's been here and gone, I'm not even going down that road again.  Left right before I made my first post. There was noise on the line that he eliminated by getting rid of an old splitter, but that's made little difference inside the house as far as I can tell. 

I was pleasant, for me, because I was taken by surprise, until the end where my router magically quit working and he began "you saw what I did I didn't change anything, it must be your router" and he left saying everything was good and the router should just come back on because he's seeing internet. After he left I power cycled the modem and got everything back up again and started doing speed tests. Something has changed a little, the hardwired Xbox is showing an average increase of 15 Mbps, Wi-Fi and the hardwired tv show no changes. 

Caught me by surprise, with a 12-2 window he pulled in at 1205. 

I suspected the cable itself needed replaced. It's not the new, fat RJ11, and has been here a good 15 winters since the last replacement. They came out way back then to replace the line from the road to the house because we had terrible cable and bad internet.

Well, from this past December to February when I finally cancelled the tv service, I went through 7 cable boxes and still had channels that would cut out for no reason, on top of the internet issues. Again, made me think main cable. 

I still have nightly service losses which I've had at this house for 5 years. Every night between 130-2am, I lose everything for 20-30 minutes. That's just my imagination according to the low level reps I've spoken with at the store and on the phone. I've given up that fight. 

Just so sick of dealing with them and their bullE36 M3. Needing to call and fight every 6 months because the bill starts going up even though I'm in a contract, getting disconnected from half the customer service calls, junk hardware, it just seems like now if the only way I can see the numbers I should be seeing is to go to their special place to see them, then maybe they're not the right numbers. 

They're just the only game in town, and fighting the competition I will happily leave for every inch of the way. 

So you get good speeds at the modem (and yes they lock onto a specific MAC address, so when you switch between the router and your laptop, you need to force it to release it by rebooting, etc.) but not at the end of your 30 foot store bought cable?

Try another cable, especially if its routed outside or prone to being stepped on, etc. they don't take well to abuse and quite frankly store bought cables can be pretty crappy.

Generally 300 feet of CAT5 is the limit before speed loss is significant.

Windows is terrible at networking, combine that with Comcast's typical management of their network (which is send it.  All.) you get a lot of crap packets being sent back and forth and generally sucking up bandwidth and router/modem CPU cycles.  Turn off all of the network options you don't need, filter everything you can at the router/modem to cut down on the excess "noise"

I second the 20 minute ping test to determine if there is a lot of packet loss, sometimes just tightening the cable connections can help reduce packet loss.

As for only paying a portion of your bill, you're being a dick and you're going to lose that battle because their lawyers made sure that they used the "up to xxx speed" in their advertisements.  You don't have a leg to stand on and they could easily kill your service and send you to collections.  If you want action, call them every day until they get it right.  Having people talk to you on the phone and sending techs out will cost them more in the long run and may actually satisfy your needs.

There is no ISP that's the only game in town. Just some that have no good competitors. Most companies like this have systems in place for repeat issues. I suggest that if you truly feel that it's the ISP's problem you call back. For example in my business if we're at a house 3 times in a month it really starts getting looked at. 1st tech screws up sure it happens. But that 2nd tech sure doesn't want the hey you missed this call so the 2nd tech cyas really well. The 3rd tech knows that sups/managers will be coming to look at the house so at that point it better get fixed. Squeaky wheel gets the grease. 

@RevRico, those reoccurring problems remind me of an issue I had with Charter a few years ago.

It looks like most cable monsters are happy to send out the domestic troubleshooters, but when the problem is outside the house, they need to send out a proper line crew at much higher expense and they really don't like that. I eventually ended up nagging my way up to a line crew, who found that the drop from the pole had been chewed on by wild rabbits...

Took them a couple of hours to fix, but it took me months of nagging to get to that point. And in my case, Charter was the only ISP servicing that location. Couldn't even get microwave service there.

1000baseTX is spec'd for 100 meter runs over cat6/cat7 equipment.  Note that that doesn't just mean the cables, technically all of the connectors need to be cat6/cat7 as well.

Odds are that the difference between their speed test and the other ones you're running is related to the "distance" (really the amount of intervening network gear and utilization percentages) between the comcast network and the server that they're using.  Their test is likely using a server that's sitting in their own data center, directly connected to the cable modem network.  (Hell, it may even be a built-in DOCSIS test that's just checking speed between your modem and the head end).

Speedtest.net, OTOH, requires that packets cross their cable modem network, make their way through the routers inside the Comcast network to the internet gateway, and then get to whatever public server they're using.  The more networks in between, the more latency there is, and the more chances that you'll have to traverse a network that's got enough congestion on it to hurt performance.  Arguably this is a more realistic test, but one could also argue that if the congestion is occuring in the public internet, then it's not Comcast's fault.