Link to home
Start Free TrialLog in
Avatar of ArtG2521
ArtG2521Flag for United States of America

asked on

Powerline adapter slow Mbps?

I recently upped my internet speed from 75 Mbps to 100Mbps.  At my main computer connected directly via ethernet cable to the modem, I get 99 to 101Mbps on two different internet speed test sites.  This is all good.

Next, I am using Netgear Powerline adapters model AV500.  Capable of handling up to 500 Mbps.  I have 3 computers that can access the internet via ethernet/via 2 different powerline adapter locations.  These 3 computers (one old desktop and two new laptops) can only pull 37 to 44 Mbps.  I don't expect the powerline adapters to pull 100 Mbps, but I was expecting somewhere around 80 to 90 Mbps.  These adapters should pull equal to or close to what my direct connection pulls, but they do NOT.

I also tested the 2 laptops on wireless via my D-Link router (model DIR-619L-ES) which is connected directly to my modem.  The laptops get 53 to 57 Mbps (should this be better as well?). Better then the powerline adapters!

So what gives on these powerline adapters?  Why so slow?  I have tried re-booting modems and the adapters and it does not seem to matter.  Does this mean my only choice is to wire my house with ethernet connections?
Avatar of John
John
Flag of Canada image

Are your routers capable of the throughput needed. My main router (Cisco) has 900 Mbits/second throughput and readily handles the higher speeds.

Are the powerline connections via the slow router?
Avatar of giltjr
Your D-Link router only has 100 Mbps Ethernet ports, so the best through-put you will get through any of the wired ports is 100 Mbps.

Wireless connections are typically half-duplex and wireless is shared between all wireless devices, so that will slow wireless connections down.  Do you know what speed your wireless devices are connecting at?  You may want to run a few tests with just one device connected wirelessly at a time.

It's been awhile since I looked at power line adapters, but one issue they used to  suffer from is drastic performance drop when  the two ends were on different electrical circuits.  Having to  go through a breaker panel would drastically reduce through-put.

Are the two end points on the same circuits or different circuits?
Avatar of weifai
weifai

Are your power lines copper-based or aluminum-based? They are not popular in our country precisely because our lines are mostly aluminum, and that type of adapters perform really bad on aluminum...
Your D-Link router only has 100 Mbps Ethernet ports   <-- I find that kind of low throughput hampers most modern networks.
Testing throughput using your internet connection is a bad idea for a lot of reasons.
You are much better off doing the tests *within* the network.
You can do some large file transfers and use a stopwatch.  Not perfect or elegant but it generally does the job.
You surely want to avoid any Fast Ethernet switches and make sure the computers all have Gbit NICs.  Otherwise, you will be limited to 100Mbps +/-.  Then you may be able to actually measure the "500Mbps" powerline extender link performance.
I think knowing via one computer what the incoming internet can do is a decent thing to do. I do the same.

But the low internal speeds (100 mbits/sec) is surely hampering things.
SOLUTION
Avatar of JesterToo
JesterToo
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
I didn't understand that you needed to measure the internet speed as you have already done that.
The issue would be the computers and the links (e.g. the powerline extenders).  Right?

You said:
These adapters should pull equal to or close to what my direct connection pulls, but they do NOT.
But at least one of the computers is "an old computer".  Do they have Gigabit or Fast Ethernet NICs?  If not Gigabit NICs then forgeddaboudit - the NICs could well be the bottleneck.  Something under 100Mbps is reasonable for Fast Ethernet transfers.  Also, you can't have the computers competing for bandwidth during the tests.
Could you maybe tell if the sockets, with the adapters, are grounded.
The newer Powerline adapters (with MIMO) are using the ground wire and can only reach higher speeds when the ground wire is available.
This will be completely down to your electrical system.

To test, put the powerline adapters on the same double-socket and see what you get through them.  You'll probably see a massive improvement.
The specs are up to 500Mb. ie, that is the maximum speed under absolutly optimal conditions. If you used two outlets that were on opposite sides of a common wall, connected by one foot of wiring, you might get close to that. If two outlets are on opposite side of the same wall but connnected to different phases, you might get 2Mb or possibly nothing.
Powerline is just for "emergency" cases where its not possible or feasible to put lan cables. I've never seen power line adapters getting the trated speed, not even on multiplug on the same wall outlet. If you want your full speed you may need to install lan cables.

i'm using powerlan myself, it really depends on the electrical installation in your house. some of my links are sufficiently fast others are as slow as hell. For powerlan in a viruit which uses all the phases from the power grid (L1,L2,L3) you need  a phase coupler installed in your circuit breaker box. This device allows to pass the power lan signals from one phase circuit to another. Without it it only cann pass to other phases by cables or different phases running close together, but its not as good as a phase coupling device.
Given the situation you might be better off installing a few Wireless NICs into the machines in question. Powerline adapters should be a LAST RESORT. While cabling might be troublesome (or even impossible considering the situation), you want to avoid "funky" network adapters when able. Something that you might look into it is running an ethernet over coaxial setup. They're relatively cheap, can be run outside, and will get cables to practically wherever you need them to go.
Avatar of ArtG2521

ASKER

Wow! Comments from 10 different gurus!  Thanks everyone.

I have read through all suggestions/comments.  The wireless router is hooked into the modem and your explanations for the speed of the dlink wireless makes total sense.

On the powerline adapters, from reading all your thoughts, I think I have come to the decision that I will do fine for the moment with the adapters, but I will switch to running ethernet cables throughout the house.  My whole area is now gigabit fiber and they are offering speeds from my present 100Mbps to 1000Mbps.  The company gives you a modem and the signal is carried to my modem via coax cable. IF I WISH to go higher than 100Mbps, they would need to give me a new modem to handle the newer speeds up to the 1000Mbps.  Keeping this in mind, how should I approach installing what I need? What should I look for? What types of cables/equipment should I consider?  I don't want to go cheap junk, but I cannot afford high end either.
Realistically you only need to insure that your machines can handle the new speed. Make sure that all NICs are gigabit capable. The problem you're going to run into is transmission rate over a cat5e connection. While theoretically it can support gigabit speeds, you're probably going to need to run cat7 to all machines with a hardwire, while insuring for any WiFi connections you have a router that supports such speeds as well. Which, over gigabit you should rarely want WiFi.
I have a fairly high end gaming rig, not even a year old:

Integrated Qualcomm Atheros Killer E2201 LAN chip (10/100/1000 MBps)

Asus Maximus Hero VIII Intel Z170 based chipset, ATX Motherboard
Intel Core i7-6700K Processor, 4-core @ 4.0GHz with HT, 8MB L3 Cache
6GB MSI NVIDIA GeForce GTX980 Ti GDDR5
512GB Samsung 950 Pro PCIe 3.0 Solid-State Drive

My LAN/NIC card is quite capable if I go for the higher speeds.  Is the price drastically different between cat5e to cat7?  The modem they provide will have at least 4 ports out.  So I would feed one port out into the wall, but then in that wall would have to be a 4 port switch which would then lead to 2 or 3 other rooms in the house. Or a 4 port panel on the wall with each separate port wired to the room destination. Does that seem right or do you have other suggestions?
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Last question, as I addressed above about port switch or port panel in the wall leading to the other rooms, is that correct on my part?  And if the cable runs are as far as 40 or 50 feet from the modem to destination, will it suffer any dramatic speed or quality loss?
Cat5E will do you fine.  It'll be cheaper and easier to install than Cat6 and you won't ever need Cat7 unless 40Gbps over copper becomes affordable.

Installing a patch panel is overkill.  Just terminate the cables with RJ45s and plug straight into your switch.
ASKER CERTIFIED SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
As a general rule: Installing a patch panel at a central site (where all the cables come together) has some advantages:
1) it keeps the "building" cables in a static physical situation so they aren't flexed around and potentially broken / made intermittent.
2) the terminations should be easier to make correctly than many people would do with RJ45 plugs.  So, you eliminate that big variable.\
3) It provides for using patch cables at the ends that are sacrificial.
Same applies for wall ports - the exact same idea.

That said, it could be overkill if there aren't many cables in the first place.
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Thanks to everyone!  Had to make hard decisions on awarding points because of so many comments.
I'm glad we could help
I disagree with the answer chosen, but no bother.  I used to have a pair of these and could easily achieve over 100Mbps.  I used to average about 130Mbps.
I agree you could get more when you're sitting beside it. But if its in the basement and providing you put it on the least used outlet to get the through put through and electrical system, and electrical interfere with the operation, then you may get more than that, listen I HAVE GIGABIT ETHERNET N I COULDN'T GET MORE THAN THAT UNLESS I RUN WIRED.
So your statement about being spot on with 53-57mbps is incorrect, like I said.
If it helps any, I chose Natty's answer as the best because I am experiencing exactly what he did.  I believe because my house was built in the mid 1980's, it's circuit wiring is NOT modern or "normal" and this plus other factors will not get me the real speeds I would like.  The assisted solutions I chose because as I said, I've decided to go with a wired network.
Appreciate why you chose that answer, ArtG2521, but the chosen answer implied that the product itself would only ever yield around 53-57Mbps.  That's not correct.

I highlighted this just in case anyone else searches the forum for answers and comes across this.  I don't think it is correct to allow others to think their equipment is defective or sub-standard when it isn't.  If this wasn't highlighted people could end up chucking money at other brands of powerline adapters for nothing.
Just to add to the information:
I would put network extensions within a building in the following order:
1) Always use Ethernet cabling when you can.
2) Use powerline extenders next.
3) use wireless extenders last.
This is not to say that one can't extend *WiFi* but that the best way to do it is to hard wire the access points.

My experience with powerline extenders has been good.  When Ethernet is out of the question, it's much preferable to adding wireless extenders in my opinion.  Between-building wireless links are another matter.
I've not noticed slowness with them so much and have the impression that they do better than legacy wireless.

I have heard of the issue with going through the breaker panel from one circuit or "leg" to the other and can perfectly understand the concern.  It's usually hard to know if this is going to happen or not when one selects an outlet.  I suppose it's for things like this why the extenders have indicator lights on them.  But, I've not seen much in the way of confirmation that this is an issue.  Maybe that's just me.

A comment on speed:  Most home applications work well with Fast Ethernet; i.e. 100Mbps even though I would highly recommend Gigabit connections these days - even if the ISP connection *capability* remains with Fast Ethernet.  That will change.
You are running wired through electricity, you and I know that any powerline close by an unshielded Ethernet causes interference, that is why companies used twisted shielded pairs, correct. So if electricity is annoyance to Ethernet, imagine running the signal through the electricity it self converting pulse to data and back. What you think is going to happen?

Next is congestion on the channels there are so many wireless devices in a home that clearly shared the same frequency another interruption.

Then there is the signal strength itself from the device.

So if you hook up 100mb to powerline in the basement, then plug the wifi up stairs on the main floor you're telling me when you connect through your laptop upstairs that you will get 100mb.

If your answer is yes. Then networking should be your field of choice.
Natty Greg:  It appears your last was in response to my post.  
FM radio works very well for a reason.  The parallel seems evident.  

It's common practice to not run Ethernet cables for any distance in close parallel layout to electrical power lines.
But powerline extender signals aren't Ethernet signals - they only emulate the channel.
And, AM radio and FM radio and TV and... and... all run through the ether together don't they?  How can that be different?
Heard of frequency multiplexing?

One does not run signal through electricity but rather through copper wires.  All of the signals are "electricity" after all.

Congestion is always an issue independent of the transmission medium.  Nothing new there.  The internet connection is the most common example of this.  So might be cascaded or daisy-chained switches that are common in office buildings.

I didn't say that there would be no speed hit with powerline extenders.  I said that my experience with it was good - meaning acceptable.

I didn't propose to tell you anything but rather the author ArtG2521.  This isn't a chat room - but this post is my own break in practice.  We should all strive to keep that objective.  I apologize in advance.
Me too, I apologize
Natty Greg, the two are different.  Powerline isn't Ethernet, as Fred said.  It is designed to be passed through the electrical system without interference.  Noise and attenuation is different, and that's usually where the loss in throughput comes from.

It's quite feasible that if you connect your adapters between top of the house and basement, you could still achieve 100Mbps over wireless.  Let's assume that your adapters are running at 140Mbps with a good signal and your internet circuit is 100Mbps.  The adapters have a gigabit Ethernet port and the wifi gives you a link at 433Mbps.  I'd expect the 100Mbps to be close.  No problem there.
Assuming we have good signal.
That's obvious.  I don't see why you felt the need to add that comment at all though.

I said...

It's quite feasible that if you connect your adapters between top of the house and basement, you could still achieve 100Mbps over wireless.

...as opposed to...

The Av500 will not give you the 300mbs it advertises, however the 53 -57mbs is what I was able to replicate when i was using it so you're spot on.

Notice my use of the words quite feasible and could, which implied that it's possible to achieve faster speeds than the 53-57Mbps that you mentioned.  You categorically stated that the product won't give the advertised speeds and implied that 53-57Mbps is the top-line.

That is an assumption based on your limited testing and is incorrect, owing to my real-world use of the product tells me that it is possible to achieve more than the 53-57Mbps, that you stated was (by implication) the top line.

In ideal conditions these units will give speeds far greater than the figures you stated.
So argumentative, and as a network professional when as it ever been an ideal situation where you get exactly what you pay for.

You attack me as if we ever get close to posted speed where their R n D person did the research. Gas mileage 7.5L city however test in a lab didn't account for passenger weight, speed, and wind resistance among other things like climate so real world test with all variables involved fuel mileage is 13.7L city.

Real world example number 2 I have D-company Gigabit internet directly plugged in via Ethernet with laptop in hand nothing between me n it. I got 749mbps on their wifi standing right beside the router I barely went over 300mbps so assuming houses have no walls/electrical interference and no microwave or cordless phones or Neighbours with all these things mentioned and soaking up the same channel you're using then absolutely,
better yet lets lock the consumer in a faraday cage so absolutely no interference, including cell phone signals then we'd achieve absolute.
@Natty - the only thing I'm arguing about is your statement which was incorrect.  I'm not attacking you.  I'm merely providing some clarity to the chosen answer; an answer which may confuse people and lead them to incorrect conclusions.

I know full well that we never achieve ideal conditions, and I've never, not once, stated that I have ever achieved the best speeds possible using the kit.  All I've said is that you are incorrect in stating that the product can only achieve between 53-57Mbps.

Your examples are interesting to say the least, but I honestly don't know why you continue to try to argue.  The simple fact is that you stated something which was incorrect.  I merely highlighted that to help anyone searching the forum in the future.