Worse performance in newer adapters

This thread has been locked for further replies. You can start a new thread to share your ideas or ask questions.

Worse performance in newer adapters

This thread has been locked for further replies. You can start a new thread to share your ideas or ask questions.
Worse performance in newer adapters
Worse performance in newer adapters
2022-09-14 13:41:32 - last edited 2022-09-20 22:23:37
Model: TL-PA8010P KIT  
Hardware Version:
Firmware Version:

Hey,

 

So before, I was using 4 TL-PA4010P powerline adapters. They worked fine.

After upgrading my ISP plan from 500/100 Mbps to 1000/200 Mbps, I decided to upgrade my devices because my previous adapters were limited by 600 Mbps PLC and 100 Mbps connection to my devices because the Ethernet port on TL-PA4010P only supports up to 100 Mbps.

I ended up buying a kit of TL-PA8010P which support 1200 Mbps PLC and 1 Gbps connection to my devices, in theory.

 

At the moment I have one of the new adapters connected to my router and the other one to my desktop. And I also have one of the old adapters connected to another device on the same room of my desktop. However, the speed between adapters that I'm seeing on TP-Link software is worse between the new adapters. 

 

 

Considering the fact that the two devices are on the same room, why am I getting worse performance on the most recent adapter? On my desktop, I had 100/100 Mbps in SpeedTest before. Now, with a better device, I get worse results (half of that).

Any tips of what I can do?

 

  0      
  0      
#1
Options
1 Accepted Solution
Re:Worse performance in newer adapters-Solution
2022-09-20 21:39:28 - last edited 2022-09-20 22:23:37

  @calves 

 

OK so lot to digest there.. Ill try to answer each one individually.

 

Basically we have 3 values:   In short, Yes

 

A - Theoretical maximum speed: 1200 Mbps   -  This is what the physical chipset on the device is capable of doing, you will never see this figure in real usage.   This is a bit like a car manufacturer saying it has a new engine that produces 1200 horse power.  That is the PERFECT scenario with clean oil, best fuel and sitting on a test bench in a lab designed to maximise that engine output.
 

B - Speed between powerlines (on TP-Link software): 425 Mbps  This is what your plugs are detecting for your specific setup for the WHOLE NETWORK at maximum.  It may appear that Plug1 is connected to Plug2 at 200 and Plug 3 at 400 etc etc.. but in reality, its not, that is more a guide to interference.  

 

If you actually think about how your house is wired, all plug sockets are connected to breakers and splitters, its basically one large cable in reality.  If we send a signal from one side of your house to the other, every socket in your house sees that signal.   Now have 6 plugs sending at the same time, its going to collide and force retries to happen, and performance will drop as a result.   The 425 you are seeing is the best result that Plug 1 can get going to Plug2, that however may not be direct and could be via Plug 3 as that part of your house has better cabling so the signal is a tad faster getting there.   The numbers really are just a finger in the air scale

 

The speed drop happens as you add more plugs to the mix.   Without going into too much detail on collision detection, the plugs will look to see if there is no signal on the powerline before sending.  This wait-and-see reduces the speed as its latency after all, add more plugs and more traffic.. the wait becomes longer.  Should it think its safe to send but its not, a collision happens.  All plugs then have to discard that data, wait a random period of time and try again, this add more latency and reduces speeds further.   In short these things do not scale well!  

 

So 425 in your case is from Plug1 to Plug2 with nothing else going on, more plugs will slow this, more data will slow this and should the dryer be pulling 3kw down the cable also its going to massively kill it.  

 

C - Speed on end device (SpeedTest): 100 Mbps  -  This is your actual throughput after collisions, retries, errors, contention, your breaker box and Dryer, etc is all taken into account.    As mentioned in a burst this will be higher, if you start to saturate the plugs, each fights for that 425mbps it will drop massively.     There is little you can do about this sadly as you are fundamentally sharing one large cable

 

 

I understand there will always be a huge difference between A and C. But I don't understand the difference between B-C. If the upstairs's powerline is actually receiving 425 Mbps from the one downstairs, why isn't this speed reaching my end device?

Basically forget A, its BS...   the real numbers are B and C difference and thats ~33/40% in most setups.    Going back to the single cable thing, take 2 plugs talking... Plug1 asks plug 2 for data, plug 2 cant reply until Plug 1 is done.. this is known as single direction communication, straight off the bat you are at 50% speed already.    Plug2 wont class the data sending as OK until Plug1 says it is.  Plug1 has to wait to say its ok as Plug2 is sending.. its 50% at best.  Add a 3rd plug, thats now 33%, 4x 25%..   obviously not all plugs are going to be used at once, but you get the idea..

 

It could also be the way you are test this.  Im guessing you are using something like Speed Test or Fast  ?    Thats not for testing LAN stuff, you need to look at something like iperf3, it will give you a far more accurate reading od only the plugs. 

 

There should be no power interference there, it should be only dependent on the Ethernet ports (of the powerline and the end device) and cable used for the connection

between the two. In my mind, it would be logical that the second powerline and the end device received the same speed.

 

LAN cables will always run at 10 / 100 or 1000mbps, the speed issue likely isnt getting from your PC to the plugs, its getting from Plug to Plug..   You are only as quick as your slowest part (which is the powerline).

 

The one exception to this would be your plugs are connected to your PC at 100mbps port speed, windows should tell you what they are connected at however.

 

 

In summary.. homeplugs.. good for up to 200mbps (roughly).  If you want faster splash out for a fancy mesh WiFi system, or run cables.    Dont take my word for it, google   Real World Speeds with Powerline Networking    others get the same as you, myself included

 

 

 

Recommended Solution
  0  
  0  
#9
Options
9 Reply
Re:Worse performance in newer adapters
2022-09-14 19:49:16

  @calves 

 

Hey and welcome

 

It could well be the mix of the older and newer adaptors.  The older adaptor is AC1 standard so uses a single data connection, the AC2 uses dual data connections

 

However I have seen that the newer adaptor un-performs when mixed with the older adaptors as its required to change its connection to suit that device, the old "as fast as your slowest device" idea.  You may well find that removing the 600mbps adaptor would speed up the AC2 models.

 

Also static on the wire could cause this. Homeplugs do cause a level or static after a while on the mains connection and from experience disconnecting them overnight once in a while helps massively.  This allows the static to clear and the route tends to work better, i done this every 3-4 months with my plugs.

 

However.. I have to be honest and say if you are just upgraded to 1000/200 WAN speeds, this may be the wrong choice for you.    In short homeplugs are SLOW, even compared to WiFi.   Their main advantage is stability, something WiFi doesnt do well at all.  

As the plugs are rated 1200mbps, thats the theoretical maximum and real speeds will be a LOT less, I would expect around 33% of what the connection speed reads at.    The connection you have reading 427, in real speeds that is likely around 220/250mbps.  With such fast internet that is a MASSIVE bottleneck for you

 

If you are after faster speeds you may find that the Deco range (something like S4) would be a better option for you, seen these getting 400mbps solid are generally very good.   

 

In short for the internet speeds you are throwing, power lines are not the best choice... fast WiFi would be easier and arguably faster for you.

 

 

  0  
  0  
#2
Options
Re:Worse performance in newer adapters
2022-09-14 20:49:36

Hi @Philbert 

Thanks for such a detailed reply.

 

I tried removing the old adapter but it didn't improve.

 

 

To be fair, when I first paired the new adapters, I plugged everything in two rooms with one of them being where the router is and the speed between adapters was pretty good at that moment. 

 

 

Obviously I wasn't expecting to have 1 Gbps through powerline but I was hoping to have at least 200-300 Mbps. I guess the problem here is range. Because my normal setup has one adapter in the living room, connected to the router and the other one in my bedroom which is in another floor. But looks like the old adapter has better range or doesn't get as much interference? I think that's kinda weird, but I'll try your suggestion to unplug overnight since I've been running powerline adapters 24/7 for like 6 years. Should I unplug both adapters or just one?

 

Regarding mesh systems, how reliable would be something like Deco S4? Those systems are signifcantly more expensive so they would need to perform very well.

At the moment I have a WiFi 6 router downstairs and another ONT on the first floor which also supports WiFi 6 (Huawei OptiXstar K562e) but even though I can get 900 Mbps on my iPhone downstars near the router, I can't even get 200 Mbps upstairs near the ONT because the WiFi signal that reaches the first floor and is being extended is already weak, compared to the one downstairs. Wouldn't I have the same problem using Deco? I know those systems are good to eliminate dead zones but I don't have a coverage problem, my problem is speed and I cannot use Ethernet cables through floors/rooms so that's why I was looking at powerlines.

  0  
  0  
#3
Options
Re:Worse performance in newer adapters
2022-09-14 21:38:31 - last edited 2022-09-14 21:39:12

  @calves 

 

Hey

 

It does sound like you are in the same boat i was a couple years ago.   Like yourself I had homeplugs (1200 also) and got the same issues that you are seeing, while homeplugs are great they dont scale well and performance is.. well meh sometimes, however they are reliable

 

You do have a quite a few plugs now which isnt ideal if honest, usually performance does suffer after 4+ plugs due to the volume of data on the cables.   Also the mix of AC1 and AC2 isnt ideal either, homeplugs will only work at your slowest speeds so at moment thats 600mbps regardless of what plugs you use.  Ideally you should be moving everything to AC2 plugs (1000-2000), running any AC1 plugs at all will slow you down.   Sadly however that is $$ to do.. i know first hand

 

If I was in your scenario and its at all possible, disconnect all the plugs tonight.. leave overnight to let the circuit discharge any static clear then replug only the AC2 (1200) tomorrow and run tests to see if the speed improved, then add the older plugs back in.   Least that way you have a baseline on clear wires with new plugs...

 

 

In terms of Deco, they are quite reasonable really.   The S4 model is 100 euro / 120 USD for a 3x pack.  If you place these striaght above eachother on the floors it might be a good option for you. 

Check something like WiFi Analyser on Android and see what signal you get straight above your current router, if its -65db or lower then WiFi could be an option for you.   Never use Apple devices for this, they lie something bad in terms of speed and signal, android or windows all the way for testing. 

 

Ultimately with 1gbps coming in you really should be considering spending some money on hardware for this.  As said I had the exact same setup as you, homeplugs all the way and got 1gb internet which i could not use.   In the end I had no choice but to run cables externally up trunking to the roofspace then stick in APs. 

 

In most cases you will find WiFi to be faster than the homeplugs are, however its far from a simple answer.  

 

Check out this site, it might help you with some tips.  Just take the 3 underscores out of the URL.. stupid external link blocking..

ww_w._makeuseof._com/tag/6-tips-improve-speed-powerline-network/

 

 

 

  0  
  0  
#4
Options
Re:Worse performance in newer adapters
2022-09-16 08:36:02

  @Philbert 

 

Keep in mind that I only mixed all adapters in the beginning, after that, I removed all my oid adapters, even before creating this post so there should be no problems related to incompatibility since I'm only using 2 adapters of the same kit.

 

I unplugged both adapters overnight and I think it helped a little bit. Now the speed between adapters is 300 Mbps and in SpeedTest I get 95-115 Mbps which is a little bit better than before.

 

Regarding mesh systems, I did try one of those apps a couple days ago and I noticed that upstairs I get around -80 dB. That's why I get "slow" speeds on my WiFi extender, because the signal that gets upstairs is already weak. I have other extenders but there are no suitable places to install them between the router and the extender so I don't think there will be a better option here. With mesh, I would face the same problems. And since Ethernet cables are not an option because I don't want to drill or see cables or pvc concealers around the house, I guess I'll be stuck with these speeds.

  0  
  0  
#5
Options
Re:Worse performance in newer adapters
2022-09-20 10:57:24
@Philbert Is there a reason why the speed I see on the software does not match the speed I get on speedtest? I asked a friend to test my new set of powerlines in his house in a much more controlled scenario. He has a 500 Mbps connection, with both powerlines connected super close to each other, the software shows 920 Mbps between the powerlines but the speedtest only indicates 220 Mbps. Both cables and computers support Gigabit so if the internet speed is 500 Mbps and the speed between powerlines is 920 Mbps, I would expect the computer to be able to use 500 Mbps or similar to that, while it's achieving half.
  0  
  0  
#6
Options
Re:Worse performance in newer adapters
2022-09-20 17:39:46 - last edited 2022-09-20 17:40:20

  @calves 

 

Hey

 

Yeah this is not totally unexpected, I had mentioned in the post earlier I expected to see this result

 

As the plugs are rated 1200mbps, thats the theoretical maximum and real speeds will be a LOT less, I would expect around 33% of what the connection speed reads at.    The connection you have reading 427, in real speeds that is likely around 220/250mbps.  With such fast internet that is a MASSIVE bottleneck for you....

 

Homeplug work in theory the same way as WiFi does, it will give you a theoretical maximum connection speed but as you are going through a contended medium (literally power going down the same lines) this will massively affect the performance.  In general from experience I find homeplugs to be about 33% of what they claim on average.  Some newer homes with better wiring and lower power LEDs etc a tad faster, older cabling a tad slower..   but 33/40% is the ball park range

 

The main advantage that home plugs have is stability, you will get 220 or whatever it is now and that will be consistent...   WiFi connection would be faster for you, perhaps even the full 500, but it will be erratic and at times drop massively.   Basically its speed vs stability here

 

The issue you are going to have is the same that I did, your internet as it passes 500+ will greatly outstrip the capability of the homeplugs, you simply wont get near those speeds and thats you bottleneck.   WiFi will get you closer to the full speeds, possibly even a tad more but getting good coverage is expensive!      Something like the Deco range is good, but littering the place in nodes can quickly become expensive and the more nodes / hops you have the slower it gets.

 

To be honest, if you want the full 500+ you will need to run cables.   I ended up running up black trunking on the outside wall, coming in at each floor and placing an AP there.   3x WiFi5 APs all connected by a cable backhaul, solid 500mbps all around the house.  I have 1gb coming in so upgrading to WiFi6 would get me closer to the 1gb speeds, but I dont have a need for the full 1gb over WiFi at present.  

 

With cable you have a un-contended connection and it will deliver ~960mbps out of 1gb all day long.    Its maybe not the answer you want, but it may be your only choice if you really want the fast speeds all around.

 

 

 

 

  0  
  0  
#7
Options
Re:Worse performance in newer adapters
2022-09-20 18:27:04 - last edited 2022-09-20 18:28:58

Basically we have 3 values:

  • A - Theoretical maximum speed: 1200 Mbps
  • B - Speed between powerlines (on TP-Link software): 425 Mbps
  • C - Speed on end device (SpeedTest): 100 Mbps

I understand there will always be a huge difference between A and C. But I don't understand the difference between B-C. If the upstairs's powerline is actually receiving 425 Mbps from the one downstairs, why isn't this speed reaching my end device? There should be no power interference there, it should be only dependent on the Ethernet ports (of the powerline and the end device) and cable used for the connection between the two. In my mind, it would be logical that the second powerline and the end device received the same speed.

  0  
  0  
#8
Options
Re:Worse performance in newer adapters-Solution
2022-09-20 21:39:28 - last edited 2022-09-20 22:23:37

  @calves 

 

OK so lot to digest there.. Ill try to answer each one individually.

 

Basically we have 3 values:   In short, Yes

 

A - Theoretical maximum speed: 1200 Mbps   -  This is what the physical chipset on the device is capable of doing, you will never see this figure in real usage.   This is a bit like a car manufacturer saying it has a new engine that produces 1200 horse power.  That is the PERFECT scenario with clean oil, best fuel and sitting on a test bench in a lab designed to maximise that engine output.
 

B - Speed between powerlines (on TP-Link software): 425 Mbps  This is what your plugs are detecting for your specific setup for the WHOLE NETWORK at maximum.  It may appear that Plug1 is connected to Plug2 at 200 and Plug 3 at 400 etc etc.. but in reality, its not, that is more a guide to interference.  

 

If you actually think about how your house is wired, all plug sockets are connected to breakers and splitters, its basically one large cable in reality.  If we send a signal from one side of your house to the other, every socket in your house sees that signal.   Now have 6 plugs sending at the same time, its going to collide and force retries to happen, and performance will drop as a result.   The 425 you are seeing is the best result that Plug 1 can get going to Plug2, that however may not be direct and could be via Plug 3 as that part of your house has better cabling so the signal is a tad faster getting there.   The numbers really are just a finger in the air scale

 

The speed drop happens as you add more plugs to the mix.   Without going into too much detail on collision detection, the plugs will look to see if there is no signal on the powerline before sending.  This wait-and-see reduces the speed as its latency after all, add more plugs and more traffic.. the wait becomes longer.  Should it think its safe to send but its not, a collision happens.  All plugs then have to discard that data, wait a random period of time and try again, this add more latency and reduces speeds further.   In short these things do not scale well!  

 

So 425 in your case is from Plug1 to Plug2 with nothing else going on, more plugs will slow this, more data will slow this and should the dryer be pulling 3kw down the cable also its going to massively kill it.  

 

C - Speed on end device (SpeedTest): 100 Mbps  -  This is your actual throughput after collisions, retries, errors, contention, your breaker box and Dryer, etc is all taken into account.    As mentioned in a burst this will be higher, if you start to saturate the plugs, each fights for that 425mbps it will drop massively.     There is little you can do about this sadly as you are fundamentally sharing one large cable

 

 

I understand there will always be a huge difference between A and C. But I don't understand the difference between B-C. If the upstairs's powerline is actually receiving 425 Mbps from the one downstairs, why isn't this speed reaching my end device?

Basically forget A, its BS...   the real numbers are B and C difference and thats ~33/40% in most setups.    Going back to the single cable thing, take 2 plugs talking... Plug1 asks plug 2 for data, plug 2 cant reply until Plug 1 is done.. this is known as single direction communication, straight off the bat you are at 50% speed already.    Plug2 wont class the data sending as OK until Plug1 says it is.  Plug1 has to wait to say its ok as Plug2 is sending.. its 50% at best.  Add a 3rd plug, thats now 33%, 4x 25%..   obviously not all plugs are going to be used at once, but you get the idea..

 

It could also be the way you are test this.  Im guessing you are using something like Speed Test or Fast  ?    Thats not for testing LAN stuff, you need to look at something like iperf3, it will give you a far more accurate reading od only the plugs. 

 

There should be no power interference there, it should be only dependent on the Ethernet ports (of the powerline and the end device) and cable used for the connection

between the two. In my mind, it would be logical that the second powerline and the end device received the same speed.

 

LAN cables will always run at 10 / 100 or 1000mbps, the speed issue likely isnt getting from your PC to the plugs, its getting from Plug to Plug..   You are only as quick as your slowest part (which is the powerline).

 

The one exception to this would be your plugs are connected to your PC at 100mbps port speed, windows should tell you what they are connected at however.

 

 

In summary.. homeplugs.. good for up to 200mbps (roughly).  If you want faster splash out for a fancy mesh WiFi system, or run cables.    Dont take my word for it, google   Real World Speeds with Powerline Networking    others get the same as you, myself included

 

 

 

Recommended Solution
  0  
  0  
#9
Options
Re:Worse performance in newer adapters
2022-09-20 22:26:08

  @Philbert I just want to thank you for such a detailed explanation! I'll stay with this, I don't really need a better internet upstairs, where I actually need good speeds (Apple TV 4K) I have a Ethernet connection directly to the router which results in 920 Mbps in Speedtest. If I could have better speed everywhere, it would be better, but can't justify the investment in a mesh system, for example, at the time being. Thanks again.

  2  
  2  
#10
Options

Information

Helpful: 0

Views: 526

Replies: 9