Problems with poor Ethernet Over Mains performance - Develo 1200Mb/s

I wonder if anybody on this forum can throw any light on why I'm not getting particularly good results. (low speeds, perhaps complete dropout) with 4 Devolo units within our home. One on a DIR 882 router and 3 remote units. We've got 4 ring mains and it happens these units are spread over 3 of them. I've upgraded from 500 Mb/s units with little gain. Speed are in the 100 - 220 Mbit/s range. But I've seen less. I'm technical (electronics eng. by training) but limited experience outside of our own home in using these devices.. The units are plugged directly into wall sockets. We've a Victorian property professionally re wired approx 24 years ago. But lots of devices with 4 adult adults. Our consumer unit had a more recent upgrade to unique per ring RCD's . I've seen reference to these causing problems but not sure what type?

I've submitted a support request to Devolo with the generated files supplied. But doubt they will be able to locate my specific issues.

Possible culprits include , some very appliance that's running all the time, RCD's , some strange impedance issues on the wiring ?

Thanks in advance for any thoughts or advice.
 

neilball

Well-known Member
It could easily be one or more devices you’ve got connected that may be generating noise in the frequency ranges used by the homeplugs, and there is not really any practical way to measure this or easily identify the culprits. When they work homeplugs can be great, but just like with wifi, when they do not perform as expected it can be almost impossible to determine the exact causes for the poor performance or to do anything about it.
 

Sloppy Bob

Distinguished Member
Homeplugs are hit and miss and pretty much dependant upon your house wiring/layout.

I have similar ones to you and they work great in 3 rooms streaming full UHD files with no issues and next to useless in one (struggles with 1080p). A friend of mine was interested in buying some so I took a couple of mine to his house and you might as well have connected 2 cups with a bit of string.

Unfortunately, if you can't find a device that's causing interference, you're back to wifi, perhaps looking at some kind of MESH system or hard wiring everything.
 
Hello, thanks for the comments. So far I've not managed to find the 'source' of interference if there is a single device. I've attempted to isolate some computers and other devices with no success. Indeed it could just be down to the scale of wiring . Our house is large ish. 3/4 bedroom Victorian on 3 floors. So wiring runs will be significant. Hence my question on RCD's or any other known common problems. I worry (but haven't tried) if Mesh will work with doors closed . Walls are brick construction of course. Thanks again.
 

ChuckMountain

Distinguished Member
Powerlines are a pet hate of mine as I believe that they are falsely advertised.

For a start, the advertised speed is always full-duplex so both up and down speeds combined. So a 1200Mbps would be 600Mbps at best, the link speed will state significantly higher than the actual connection speed. The 500/600Mbps variants typically have 100Mbps Ethernet connectors so won't ever go faster than that regardless. Then you have the actual impact of the housing, wiring and so you end up with connection speeds around the 150-200Mbps speed for the higher end units in most of the reviews. Yes, some people will see higher but it's hardly reliable.

Having said that 100Mbps+ should be more than enough for a couple of UHD streams from Netflix and co. We are seeing powerlines become the bottleneck though particularly with FTTP and VM faster broadband offerings.

Are you routing your traffic over these or do you have WiFi AP plugged into them too? Some (all?) powerlines are like WiFi and only one thing can talk at once. If you are all connected to both these and WiFi, you could end up with a situation where your network is getting "stuck" between handling requests.

I assume there are no options to hardwire?
 
Hello,

Thanks for all the responses so far.

My issues are not just speed but what appears to be 'drop outs' . I've seen the CRC errors in the log files and they look very high. 20-30% between pairs. Indeed wiring direct might have to be the way. But runs are 20m+ from the router into at least 3 rooms. Also difficult to hide due to existing flooring / decoration. Mesh might have to be the way forward, but another set of uncertainty to try out.
 

mickevh

Distinguished Member
Ethernet is good for 100m over UTP cabling (strictly speaking, 90m plus a 5m each end for the patchcords) - 20m is easily within standards. It also works at the same rate ("speed") whatever the length of the cable lobe, so it doesn't fall off with distance as other technologies do. (It either works "full speed" or not at all - there's nothing in between.) Of course, routing and hiding cables is often a challenge in locales not designed for it.

A lot of IT uses CRC - essentially it's a checksum used to test the integrity of some "thing" - ie. blocks of data retrieved from an HDD or data networking transmissions. Ethernet uses CRC, Wi-Fi uses something similar called a "FCS" (frame check sequence) and if I ever knew what HomePlug uses, I've forgotten, but I can happily believe its a CRC.

In data networking the terms "reliable" and "unreliable" have specific meaning. Technologies such as ethernet, Wi-Fi and HomePlug are "unreliable" in that if errors are detected in transmission the offending packets are silently discarded without notifying the sender. I don't know whether HomePlug does this or whether is has any facility to request re-transmissions.

However, CRC errors are indicating corrupt or incomplete transmissions and the more of them you have, it suggests the worse the link is. How "normal" this is for HomePlug I couldn't say.
 
UPDATE - I've been carrying out a few more tests and can pretty much confirm that the performance of my units is dramatically improved if I choose to have all devices on the same electrical ring! (In practice not what I actually want) I've now seen speeds of 700Mb/s reported with very small error rates. (previously less than 200. Error rates can be observed for Devolo by generating their support information file). I'm also pretty clear that my Residual Circuit Breakers in the consumer unit are the cause of the problem. They have a small transformer in the circuit which will act as a filter for the comms signals. I've got to now do some rather convoluted cabling to achieve a minimum workable solution in the house. Obviously complex wiring of adaptors starts to encroach on the very need to have them in the first place. Thanks for all of the replies which have spurned me to do some more research of my own.
 
Run the Develo cockpit app. I assume this is the raw channel rate , not an actual communication rate. Forgive my ignorance . See below. Run the Support function by 'selecting' the speedometer symbol with the mouse. You then get a report and you can see errors between adaptors.
Snap 2020-06-25 at 10.59.38.jpg
rance on the standards.
 

ChuckMountain

Distinguished Member
Ah yes the cockpit view, which bears little resemblance to the real-world throughput. To get a proper test you need to run something like iperf which runs on two computers at either and then sends test information between them. If you can do that that would give you a much better idea of the throughput you are really achieving.
 
In principle I could do that as I have a desktop and a laptop . I don't know how to run iperf though. My informed guess is that it will be a lot better than it was, just looking at the BERates via Devolos app. I've a little bit of experience working on DAB radio development (manufacturing side) but often the BER is used as a GO/NOGO verification of a good product. So looks better to me as is. If there's a simple guide to iperf I might give it a try.
 

Rodders53

Distinguished Member
I tried to use Powerlines in my home but I have, instead run some CT5 cables where possible.

Even then I've found that a cheap WiFi mesh system has helped some devices (Chromecast TV and audios) in a location that has Gigabit ethernet capability and done a general improvement overall to wifi connected devices.

Like you, I have found the powerlines are fine on the same ring circuit, but not good on two separate ones.

Some say they work. I don't believe them ;)
 

Monster900

Active Member
UPDATE - I've been carrying out a few more tests and can pretty much confirm that the performance of my units is dramatically improved if I choose to have all devices on the same electrical ring! (In practice not what I actually want) I've now seen speeds of 700Mb/s reported with very small error rates. (previously less than 200. Error rates can be observed for Devolo by generating their support information file). I'm also pretty clear that my Residual Circuit Breakers in the consumer unit are the cause of the problem. They have a small transformer in the circuit which will act as a filter for the comms signals. I've got to now do some rather convoluted cabling to achieve a minimum workable solution in the house. Obviously complex wiring of adaptors starts to encroach on the very need to have them in the first place. Thanks for all of the replies which have spurned me to do some more research of my own.
Doesn't surprise me.

I've found that some consumer unit circuit breakers can play havoc with powerline performance. I also understand that it is the RCD breakers that give the biggest reduction in performance. Unfortunately, that is about the limit of my knowledge so I can't offer much help on how to improve matters.
 
As a quick and dirty test I ran the Netflix networking speed test on my TIVo box. Develo connected. Now seeing 100mb/s , previously only around 20 i.e. about a 5 to 1 improvement which in fairness does align with the cockpit reported speeds (ratio metrically). The right hand Devolo above unit showing only 190Mb/s is still on another ring to the other 3 units.
 

jonoro

Active Member
Fixing the "other ring" effect is easy, as long as you are comfortable around consumer unit inards. Just a few pence on some coupling capacitors. Alternatively you could buy some "phase couplers" (i.e. very expensive capacitors).
I posted the results of my investigation here:
 
Hello and thanks for the link. Interesting that somebody has tried to 'bridge' the rings electrically with some apparent success if I've read this correctly? I've definitely seen a difference here but moving rings. But my resultant solution needs more home Cat cabling to achieve. Thanks for all of the support and replies to date.
 

jonoro

Active Member
Yes bridging the rings was just as good and often much better than being on the same ring - it all depended on cable run lengths I believe.
On different rings I had Devolo reporting 700Mb/s but as ChuckMountain said a test with iperf revealed nothing like that but still the ring bridging still showed with iperf the circa 505 increase.
 
Not sure I'm going to try the experiment though. Not over keen on having capacitors between the rings. Seems like a bit of an unnecessary risk if nothing else. But interesting nevertheless. I think I'm going to stick with my new condensed ring arrangement. I'm seeing speeds of 500+ Mb/s now on 2 of my Devolo 1200 devices (200+ on the other but connected to another ring.) Although the house was professionally rewired under my supervision I still having a nagging doubt about a break say on a ring. But too many other issues today. The hardest bit now is getting this extra cable into the living room :( without making a mess.
 

grahamlthompson

Distinguished Member
Risk? What risk?. There is no voltage across them. Phase couplers are openly for sale.

Pinch a capacitor from an isolated TV output plate.

Twisting the live ends of the two ring mains before the MCB terminations may produce enough capacitance to pass the home plug rf.
 
I'm an Electronics Eng. not an Electrician. The risk is to me fiddling with my consumer unit! A mains capacitor I'd guess is an additional risk when coupling mains lines. Capacitors as you know can fail. I'm pretty risk adverse and try to minimise additional issues. Didn't realise phase couplers exist. Will have to investigate. Thanks for the 'Heads up'.
 

Similar threads

The latest video from AVForums

Podcast: New Denon AV Products, HDMI 2.1, 8K, DTS:X Pro Sound Q&A with Denon Product Specialists

Trending threads

Latest News

Movies preferred to TV shows during lockdown in UK
  • By Andy Bassett
  • Published
Ruark announces R3 all-in-one music system
  • By Andy Bassett
  • Published
Disney Plus announces Star Wars: The Bad Batch spin-off
  • By Andy Bassett
  • Published
LG CX and GX OLED TVs gain AMD FreeSync Premium
  • By Andy Bassett
  • Published
Samsung launches HW-Q950T and HW-Q900T Dolby Atmos soundbars
  • By Andy Bassett
  • Published
Top Bottom