Slow Homeplug - Five Ways To Boost Powerline Network Speed

[memilanuk]

Hmmm... didn't think I'd have to go down this road so soon, but I need some help getting the speed up on my network.

At the one end... Buffalo Airstation WHR-HP-G54 router/gateway. It's connected to the local fiber provider via PPPOE via the WAN interface, provides 802.11G wifi w/ WPA2 security, and has an Ethernet cable running from one of the ports of the built-in switch to a Netgear XAVB2001 HomePlug AV 200 adapter.

At the other end... about 10-12 feet as the crow flies, but down a floor and over a bit... another Netgeare XAVB2001 adapter feeds a Linksys 5-port 10/100 switch via the 'uplink' port. Connected to the switch are two older PCs... one a Linux box for testing and the other is my FreeNAS box. The FreeNAS box has an Intel D945GTP mobo, 2.6GHz cpu, 2GB RAM, and three 500GB SATA drives in a RAID-Z1 configuration shared under /mnt/storage/.

Currently I have my Macbook plugged into a port on the Airstation under the premise that wired ethernet should be faster than 802.11g wifi. Using Filezilla I am copying over my iTunes library to the server via sftp. The files range from ~3-6MB (songs) to ~50MB (music videos) to ~500MB (tv episodes) to ~1.5GB (HD tv episodes).

According to the graph on the FreeNAS admin page, its currently swinging between 2-11MB/sec, and just eyeballing it... averaging about 5-6MB/sec. Looking at iStat on the Macbook... it's showing between 200-1200 KB/sec, hard to tell what 'average' is. When Filezilla is transferring one of the bigger HD files, it shows two sftp connections running at around 280-300kb/sec each. At this rate... transferring the ~60GB of music and videos from the Macbook to the server will take overnight, at least. Transferring the 260GB iTunes folder from my desktop PC... does not even bear thinking about.

So far as I can tell... both Homeplug adapters are plugged straight into the wall, not into any kind of surge suppressor. I don't see anything that looks like an AFCI bkr in the main panel downstairs, which would lead me to start wondering if the problem might be interference from some other source (such as the cell phone charger mentioned in the article). The question is... how far away should I be looking for potential sources of interference? Other outlets in the room? The next room? Across the house? In the garage?

Any other ideas would be most welcome. After seeing the results in the various HomePlug articles (looks like average of mid 40MB/sec for one stream) it seems like I should be getting a lot more speed than I am. Granted, it could be something else besides the HomePlug connections... but either way, I need to find it and get it under control!

TIA,

Monte

 

[memilanuk]

Just an update... at some point since my last post, the transfer speeds have jumped up to about 35Mb/sec from the previous ~5-6 Mb/sec. Not sure if something in the file transfer process was clogging things up...? Nothing has really changed here in the house that I'm aware of as far as things being plugged in, so I'm kind of at a loss as to what would cause the rather marked changed? I'd still like to get it a little higher if possible, even if the file transfer is moving markedly faster.

BTW, can anyone point me to a good explanation of the difference between bits/sec vs. bytes/sec in terms of real-world file transfer? Depending on the setting, the graph on the FreeNAS admin page shows either ~35Mb/sec or ~4.2MB/sec - which one is more useful for comparing LAN speeds?

Thanks,

Monte

 

[thiggins]

First, get a benchmark by doing transfers via Gigabit Ethernet. Then plug both HomePlug Adapters into the same outlet and see what that does.

Noise effects generally are pretty local, i.e. same outlet and perhaps same room.

You can use bits or bytes. There are 8 bits to a Byte, so you can do the math to get a rough conversion. Just use the same units consistently to make direct comparisons.

 

[memilanuk]

I have no gigabit ethernet... if I did, the house would be wired for it and I wouldn't be messing around with HomePlug AV I think one of the computers has a gigabit port... but it's upstairs with a dual-channel 802.11g/n wifi adapter. I could move one of the homeplug adapters up there, but nothing else has a gigabit port to connect to even for comparison.

Thanks,

Monte

 

[thiggins]

Fine, just use 10/100 Ethernet. You are just trying to get a baseline with the fastest and most reliable connect method first.

 

[memilanuk]

Well... here's what I've found. Not sure it all makes sense, but this is what I got.

For reference, just machine to machine, no HomePlug in between:

Transferring a dvd iso image of CentOS 5.5 from my Macbook to the desktop PC via cat5 ethernet (10/100 port on the Mac end, gigabit ethernet port on the PC end, Linksys 5-port 10/100 switch in between): network monitor gadget on PC says ~9.5-9.8 MB/s, iStat on Macbook says ~9.0 MB/s. Running LAN Speed Test from the PC to a SMB share on the Macbook reports 10.8 MB/s write, 8.6 MB/s read.

Plugging the Macbook into the other 5-port 10/100 switch downstairs (the one shared by the FreeNAS & Linux boxes) showed about the same transfer rate - ~9.8 MB/s.

Now for testing the HomePlug devices. Again, Netgear XAVB2001 adapters. Both plugged into the same outlet (granted, it has one of those 6-into-1 outlet adapters, which is the only reason both adapters fit), with one feeding the Macbook directly and the other one feeding the 5-port 10/100 switch that the PC is connected to. Transferring the same dvd iso image from the Macbook to the PC: about 4.0-4.1 MB/s per iStat, and about the same, maybe a bit more from the Network Monitor gadget. LAN Speed Test reported a little under 6 MB/s write, 3.5 MB/s read.

Repeating the LAN Speed Test from the desktop PC <---> upstairs 10/100 switch <---> HomePlugAV <---> basement 10/100 switch <---> FreeNAS box to a SMB share mapped as a network drive... 3.7 MB/s write, 3.0 MB/s read.

And just for the sake of comparison... LAN Speed Test from desktop PC via wifi (Atheros dual channel 802.11a/b/g/n) to the Macbook and the FreeNAS box, presumably with the Buffalo Airstation WHR-HP-G54 in between... about .33-.36MB/s... which seems a lot slower than I remember for some reason... but I never really tested it with this tool, it just was what it was.

TIA,

Monte

 

[thiggins]

Thanks for running the tests. The speeds you're seeing are somewhat slower than what I found in the HomePlug AV Adapter Roundup
but in the right ballpark. Note that the published test results are Mbits/sec, not MBytes. Just divide by 8 to convert bits to Bytes.

Your tests also show little throughput degradation from the house wiring (your upstairs / downstairs test).

 

[memilanuk]

Part of what concerns me, I guess, is the variability in real-world transfers day to day. Uploading one iTunes library the other day from the Macbook (ethernet) through HomeplugAV via sftp I got 4.0-4.2 MB/s, for the last two days doing the same thing (with a much larger library) from the desktop PC its running around 2.0-2.1, occasionally up to 2.5 MB/s. Granted its not the exact same location in the house, but its not *that* big of a house!

 

[HeyRadar]

Tim, for the line noise article, did you try plugging the devices that generate line noise, into a power strip that had built-in noise suppression?

 

[vicmdv]

Tim,

I understand that if the noisy device is plugged into the filtered outlet of the powerline adapter itselft the filter helps to keep the noise down and hence improves the bandwidth but there could plenty of noise generating devices (not plugged into the adapters) along the way between the 2 powerline adapters, how does the built-in filter help in this case?

I am looking into getting one of the Netgear powerline networking kits. They make 2 versions of the kit: with built-in filtered outlet and without it.
The new Netgear AV+ 500 kit is about to hit the stores. Are you saying the one with built-in filtered outlet (XAVB5501) is better than the one without it (XAVB5001)?
Is that what your contact at Netgear implied?

Thanks,
Victor

 

[thiggins]

I understand that if the noisy device is plugged into the filtered outlet of the powerline adapter itselft the filter helps to keep the noise down and hence improves the bandwidth but there could plenty of noise generating devices (not plugged into the adapters) along the way between the 2 powerline adapters, how does the built-in filter help in this case?

The only thing that matters is what the noise is at the receiver. Doesn't matter where along the path the noise comes from. High frequency noise tends to be attenuated very quickly with distance. So from a practical view, it's the noise sources near the receiver that matter most.

I am looking into getting one of the Netgear powerline networking kits. They make 2 versions of the kit: with built-in filtered outlet and without it.
The new Netgear AV+ 500 kit is about to hit the stores. Are you saying the one with built-in filtered outlet (XAVB5501) is better than the one without it (XAVB5001)?

The only difference between the product with and without filter is the filter itself. "Better" depends on whether you have noise sources to protect against.

 

[vicmdv]

Thanks, Tim.

Interesting... I did not realize that the noise only matters at the receiver.

Sounds like the built-in filter cleans the noise coming from both devices plugged into the powerline adapter and the ones along the way.
This feature makes it a "better" choice in my book. ;-)

 

[thiggins]

Sorry, but I misspoke. The filtering is only for the built in outlet, to prevent noise from whatever is plugged into it from interfering with the adapter.

 

[vicmdv]

OK. Then I guess either model is fine as I am not planning to plug any noise generators into it.

Tim. Are you planning to do a comparison of Netgear XAVB5001/5501 and Trendnet TPL-401E2K 500Mbps adapters any time soon?
It would be interesting to know how they compare to Netgear's WNHDB3004 wireless solution that you recently reviewed.

 

[thiggins]

Tim. Are you planning to do a comparison of Netgear XAVB5001/5501 and Trendnet TPL-401E2K 500Mbps adapters any time soon?
It would be interesting to know how they compare to Netgear's WNHDB3004 wireless solution that you recently reviewed.

The NETGEAR isn't shipping yet. Trendnet is on the way. See this thread for one reader's experience with the Trendnet. http://forums.smallnetbuilder.com/showthread.php?t=4261

 

[tannebil]

Hi Tim,

Did you ever swap out one of your AFCI breakers to see if a Square D or Eaton breaker really was the solution to the problem? I'm thinking about doing it (I have Siemens and see a huge performance hit) but I'd like to hear that it worked for someone before actually spending the money on an electrician.

Thanks, Bill

 

[thiggins]

Did you ever swap out one of your AFCI breakers to see if a Square D or Eaton breaker really was the solution to the problem? I'm thinking about doing it (I have Siemens and see a huge performance hit) but I'd like to hear that it worked for someone before actually spending the money on an electrician.

No, I haven't since my home is wired for Ethernet.

 

[Erik Pearson]

Couple more filters

Found this great site (thanks!) while researching the topic, and have another couple filters to add to the list:

Plug-in Line Noise Filter from Energy, Inc. (the makers of the TED monitoring devices.)
Only $4 for plugin, $11.95 for wired in-line
http://www.theenergydetective.com/plug-in-line-noise-filter

TV Power Line Filter - US Version from Gefen
$49
Single outlet box that plugs directly into wall outlet
http://www.gefen.com/kvm/dproduct.jsp?prod_id=5570

Have not tried either. We are probably going to order a few of the TED filters. TED uses powerline networking for their electricity usage monitoring detectors to their gateway, so they have a vested interest in making sure these things really work for the customers who use them!

 

[Unregistered]

Lakelady

Hi
Help!! I have a powerline home plug security camera system. I have one camera about 600 ft away from my house on 300 ft underground ethernet and about 300 ft of #10 electrical wire. The power adapter is in between the two. I have a lot of problems keeping the camera viewable on my laptop. It flicks in and out and sometime disappears completely. Would a long range outdoor repeater help or what will help increase the bandwidth and powerline readings? I know very little about this kind of thing so it has been a headache since Jan. I have buried electrical wire, pulled it back up buried it in pipe..don't know if burying it has anything to do with loss of signal or not. Not understanding how this stuff actually works I'm fighting a loosing battle. The camera manufacturer isn't much help. The Homeplug is a 200mbps and plugs into my router. I think a lot of my problems is router related. I would greatly appreciate any help.
Thanks

 

[thiggins]

Help!! I have a powerline home plug security camera system. I have one camera about 600 ft away from my house on 300 ft underground ethernet and about 300 ft of #10 electrical wire.

I don't think the problem is in your router. You are probably pushing the limits of powerline networking range.

You would be better off using a long-haul Ethernet extender. But you would need to run twisted pair (CAT5/5e is fine) all the way between the two locations.