What's new

TP-Link PA9020P kit giving REDUCED internet speeds after I upgraded from Vivid 200 to Vivid 350!

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

The test I would try is an iperf between just the powerlines with and without the virgin equipment plugged in. If you can get your old speeds without it plugged in and it only drops after the virgin equipment is plugged in, the newer speeds may require more ampage and power draw changing the scenario on the power circuit which the powerlines use.

Finally got round to doing it; testing from one Mac to the other using iperf3 just testing network speeds across the wired powerline.

server set using iperf3 -s
client set using iperf3 -c 192.168.0.60 -P4 -t 30

First result with VIrgin Hub 3 plugged in switched on and connected to network as usual:

[SUM] 0.00-30.00 sec 1.06 GBytes 305 Mbits/sec sender
[SUM] 0.00-30.00 sec 1.06 GBytes 304 Mbits/sec receiver

Then tested with the Virgin Hub switched off and unplugged from power and disconnected from powerline or switch by pulling out the ethernet cables

[SUM] 0.00-30.00 sec 1.10 GBytes 314 Mbits/sec sender
[SUM] 0.00-30.00 sec 1.09 GBytes 313 Mbits/sec receiver

Then finally just for good measure I pulled out the coaxial cable from the back of the hub too

[SUM] 0.00-30.00 sec 1.11 GBytes 317 Mbits/sec sender
[SUM] 0.00-30.00 sec 1.10 GBytes 316 Mbits/sec receiver

Looks to me like has very little impact on the issue.

Unless TP Link aren't really talking about interference leaking from the hub or coaxial lines as such but that there is an actual issue with the way the data transmits at higher speeds from coaxial/hub into the powerline, as that always halves the speed to 160-170Mbps now with the 350 service, whereas used to be 200+ with the 200 service.

Any thoughts?
 
Slight drop as you've mentioned, but not significant enough to be the cause.

Is there a duplex setting on the virgin hub for each of the lan ports that you can manually set? I've seen mis-negotiated half/full duplex cause problems like this too.
 
If you're getting half-speed on the far end, check the ethernet cable attached - you could be half-duplex for whatever reason (yes, this happens).
@phillipwj Did you check this on your Mac Pro Desktop network adapter?

I may have missed it but in all your testing I can't see what combination of equipment is giving you the 170Mbps speed. Presumably you're doing the download test on the Mac Pro Desktop plugged into powerline #2. Have you tried testing from the laptop plugged into powerline #2 (using a known good cable)?

I think we can rule out any RF interference as the cause.
 
I haven't seen any way to set duplex settings on the Virgin Hub 3.0 LAN ports.

On the Mac Pro I checked the network adapter and it's set to: Automatic - 1000BaseT - full duplex, flow control - MTU Standard 1500.

Yes the only traffic with a speed of 170 is internet traffic downloads coming across the powerline, all other network traffic across the powerline is evidently running at 300+ and gigabit speed when not crossing powerline:

ISP(Virgin Coaxial Cable)---Virgin Hub3---TL-PA9020P(No1.)~powerline~TL-PA9020P(No2.) —PC2 (Mac Pro Desktop)

I just plugged in my laptop to powerline #2 the with the same cable used to plug in the Mac Pro and another known good cable, all the same speed.

Interestingly, I just discovered the laptop can achieve a higher speed via wifi from this same location, 210Mbps. Must be especially low interference this evening as I'm sure it's rarely beaten the wired poweline before. The laptop is only N wifi, not AC.

Alas, the N wifi in my Mac Pro barely scrapes 60Mbps. Both N, such a difference in speed. I guess the Mac Pro metal box never was the best for a good wifi signal, hence I've always preferred wired ethernet or wired powerline.
 
I would try setting the flow control to off and see what happens. Also, try setting full duplex to half, but I don't think this is a valid setting in gigabit ethernet.
 
I'll give the flow control idea a try. Cant set half duplex in gigabit. How about Jumbo or Custom MTU?

I'm giving strong consideration to the idea of asking Virgin to return me to Vivid 200 for a week, just to see whether this whole palaver is as a result of going to 350 and powerline simply not playing well with it.

BUT I do still have my Attenuator to arrive this week to try that also.
 
I've once--only once--have helped diagnose a problem that was similar to yours that was due to some unexplainable reason, but that a dumb switch fixed. Put a dumb switch in between the virgin router and the powerline and see what that does.
 
I've once--only once--have helped diagnose a problem that was similar to yours that was due to some unexplainable reason, but that a dumb switch fixed. Put a dumb switch in between the virgin router and the powerline and see what that does.

Alas, I've already tried.
If you see in my first post: "In an effort to reduce interference I separated the Hub3 further from the Powerline by adding in a gigabit switch, so it is now:
ISP(Virgin)---Virgin Hub3---Netgear Switch(Gigabit)---TL-PA9020P(No1.)~~~TL-PA9020P(No2.) —Mac Pro Desktop (PC2)"
I did this at the same time as trying to put as much distance between the powerline #1 and the Virgin Hub as I could, based on the coaxial cable length I have. It's not much, still only about 2-3 feet, but they were slap bang next to each other before whereas now the Hub 3 is 3 feet above the powerline adapter and away from the power socket. Of course the coaxial cables are still lounging around in close proximity to the ethernet cables and the powerliine etc. Only way I could change that would be to get some much longer coaxial cables and some much longer ethernet cables and try to spread everything out a bit.
 
I would try setting the flow control to off and see what happens. Also, try setting full duplex to half, but I don't think this is a valid setting in gigabit ethernet.

Cheers for this one Samir, appears cancelling the Flow Control on the Mac Pro ethernet adapter (so just setting it to Full Duplex, but no flow control) has managed to notch up the speed a tad. It's gone from c160-170 to more often 175-185 even 190 occasionally, so at least I'm getting back to where I used to be with the Vivid 200 service.

Still baffling me tho why I just cant get beyond this level and pull a faster download speed from the Vivid 350 service through the powerline when all the other network traffic is whizzing through the powerline no problem at 300-380Mbps and gigabit everywhere else.
 
I've once--only once--have helped diagnose a problem that was similar to yours that was due to some unexplainable reason, but that a dumb switch fixed. Put a dumb switch in between the virgin router and the powerline and see what that does.

Actually, thinking on this again, you dont happen to remember what make of switch worked do you? I'm currently using a Netgear GS105 unmanaged switch which are pretty efficient at passing the signal through. However I also happen to have spare Belkin 5 port and 8 port unmanaged switches which I'm now wondering might possibly help as I know they're reputedly not quite as efficient as the Netgear, but that maybe a positive attribute in this instance, acting as a kind of signal filter between the coaxial hub and the powerline network. Who knows, I maybe spouting complete clap trap, but based on what you just mentioned maybe worth a try and nobody from TP-Link, Virgin, or anywhere else has so far been able to fully explain the situation based on all the evidence presented.
 
I forgot that you tried that already, but what's interesting is that in the one time I read where the switch was the issue, it was a netgear gs105/108 I believe, so trying another brand switch might be worth trying for sure.

I'm glad the flow control helped something and I bet some of the more knowledgeable members here will be able to use this as a clue to possible solve the issue completely perhaps by messing with the jumbo frames or mtu settings.
 
Ah right. The switch definitely isn't the issue in this case though as I've tested both with and without it. I ended up adding it as a way of trying to increase separation (both distance and electronic) between Hub 3 and Powerline adapter and indeed to negate the likelihood of any potential faulty LAN ports in the Hub (which I've seen reports of also). Anyway, still worth a try replacing with the Belkin at some point when I have a moment.
 
It won't hurt to try different ethernet lan ports on the virgin router if there's more than one. Probably won't make a difference, but would eliminate any potential 'bad port' issues from being the source of the problem.
 
Hmm.. just a thought, and a passing thought of something that @System Error Message mentioned in passing about the chipset and ports on the PA9020P TP-Link stuff...

The ports on the PA9020P are actually interfaces into the Broadcom chipset, not dedicated switch ports like the QCA based AV1200 PA8030P kit...

The PA8030 kit does not support Jumbo frames, so the MTU is 1500 bytes at most... I'm not so sure about the PA9020 and Broadcom chipset - it might try to do jumbo frames there, and this might correlate to packet fragmentation and lower IP thruput... it's an odd angle to look at, but still a consideration...

So, try disabling Jumbo frames on one of the client PC's used for testing and see what happens, or turn down the MTU if you can't...
 
Last edited:
Hmm.. just a thought, and a passing thought of something that @System Error Message mentioned in passing about the chipset and ports on the PA9020P TP-Link stuff...

The ports on the PA9020P are actually interfaces into the Broadcom chipset, not dedicated switch ports like the QCA based AV1200 PA8030P kit...

The PA8030 kit does not support Jumbo frames, so the MTU is 1500 bytes at most... I'm not so sure about the PA9020 and Broadcom chipset - it might try to do jumbo frames there, and this might correlate to packet fragmentation and lower IP thruput... it's an odd angle to look at, but still a consideration...

So, try disabling Jumbo frames on one of the client PC's used for testing and see what happens, or turn down the MTU if you can't...
also different devices that support jumbo frames may support up to a specific size. CPU connected ports will usually support the maximum an ethernet frame can while switched ports may not. If the port is CPU connected, then it will support jumbo frames if the firmware does, and seeing that this is tp link with its degrading AV firmware (each update the software gets worse and worse for configuring it). The other factor is the buffer, smaller chips cant handle a large buffer so they cant handle large ethernet frames.

As an example, a mikrotik router with CPU ports can support 10K or higher ethernet frames. Their dedicated switches like the CRS support up 8-9K while the older routerboards with switches may only support up to 4K frames. However since it is a chipset with RAM it doesnt need to fragmant packets but the netgear prosafe, even if it says supports jumbo frames performs poorly with jumbo frames leading to high latencies and packet losses because as the jumbo frame enters the switch it gets fragmented as theres no huge RAM, only onchip memory and it uses the store-and-forward method which while reliable is slow for jumbo frames as it has to process the entire frame or fragment it internally.

When it comes to AV, it is actually better to use smaller frames than jumbo frames because the AV transmission itself is not a reliable mechanism despite using copper as it rides along the AC wave so a large frame will be harder to encode than a smaller frame.
 
Just for future reference for anyone coming across similar issues, in the end I decided to get myself an Archer C9 router, put the Virgin SH into Modem mode and connected the Powerline to the Archer with a Belkin Gigabit Switch in between the Archer and the Powerline.

Hey presto, my download speeds suddenly jumped back up from 150-160Mbps to around 250Mbps. Maybe the Virgin SH just has crap LAN ports/hardware or somehow there was interference between the coaxial being in close proximity to the Powerline, I really still dont know, but adding the Archer into the mix appeared to do the trick one way or another.

More recently I've ditched Virgin and switched to Hyperoptic FTTP 1Gbps Up/Down service. Now I get full 1Gbps when I attach direct the Hyperoptic router (Tilgin HG2381), but via Powerline it tops out around 300Mbps and via a/c wifi I can get up to around 500Mbps when in close proximity to the router.

Haven't tried fiddling around with the Archer C9 too much yet, but initial tests seemed to suggest the Tilgin was the better speed performer. Now wondering whether to install DD-WRT onto the Archer C9 and see whether I can push that to perform better than the Tilgin. Any recommendations much appreciated! :)
 
Similar threads
Thread starter Title Forum Replies Date
intArray MoCa Packet Loss with TP-Link Equipment MoCA, HomePlug, HPNA 0

Similar threads

Latest threads

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top