Preparing to buy a set of MoCA 2.5 adapters...
- Thread starter OzarkEdge
- Start date
FloorPizza
Occasional Visitor
First off, thank you so much for sharing all of your info on this product.
I'm just about to pull the trigger on the ScreenBeam ECB7250K02 kit over on Amazon, but...
Whenever I look at a product on Amazon, I skip the good reviews and go right to the one and two star reviews. There are a lot of one star reviews panning the product for poor longevity, especially regarding the ethernet ports. I won't be constantly plugging/unplugging stuff into the MOCA ports, so I'm not really *too* concerned about this, but you can't help but feel a bit of a spidey sense when there are a lot of reviews saying the product doesn't last and the company's customer support is severely lacking.
Good to hear that yours are still in service for a year and a half or so.
I think I'll just jump in with both feet and do it. I gave my daughter WiFi based body trackers for VR as a Christmas gift, and she's not getting a good enough 2.4 ghz signal in her room upstairs to use them. The plan is to MoCA our LAN upstairs and place a switch and wireless AP in her room.
Can anyone else that has purchased the ScreenBeam ECB7250K02 comment with their experiences regarding product longevity and/or customer support?
OzarkEdge
Part of the Furniture
The Screenbeams appear and behave like a solid product... I would consider them again. As for the reviews... a few good ones is enough for me... a lot of reviews on commerce sites now are uninformed consumers ranting about their troubles in life... or the competition dissing their competitor's product. Trust reviews that make sense... trust yourself. Ignore the reviews that just complain and whine.
I bought a second kit for another media center, but have not started that project yet. The Amazon pricing was jumping all around so I jumped when they were priced a bit lower than usual.
OE
FWIW, I've been running a pair of ECB7250s since about July with no issues. One thing I have noted is that while they seem to be able to get pretty close to the rated 2.5Gbps throughput, they add a bit under 4ms to round-trip ping time. Maybe if you're a twitch gamer you'd be unhappy with that, but it's not causing a problem for me.
kelllogg9
Occasional Visitor
Been using my ECB7250K02 kit for hmm maybe 2-3 years now non-stop. Although i posted here way back about not being able to achieve the advertised 2.5Gbps in one direction (other direction worked perfectly), i discovered it seems to be windows randomly throttling when network activity is slow..when it picks up then both directions go back to 2.5Gbps. So i left it at that.
Other than that issue, i have never had issues regarding longevity. It's a well built product. Even their 1Gbps models were good (i upgraded from that way back).
Although u mentioned only getting a pair, just keep in mind if u add another to your network (so u would now have 3+) i believe the speeds would now drop lower. Cant be sure, but i believe that is what i heard way back. Just a thing with Moca iirc. But thankfully i only needed a pair.
By far superior to Powerline units. I used to have those too (still do, but never use anymore) and they are inconsistent and total bs in their claimed speeds. Moca all the way.
So after about six months, I've convinced myself that there is something wrong with the pair of ECB7250s I have. What I am seeing is that I can get full-rate 2.35Gbps iperf3 speeds in one direction, but in the other direction there is some sort of throttling or something going on: iperf3 between two 2.5G-capable machines maxes out at not much better than 200Mbps. Much though not all of the performance can be recovered by telling iperf3 to use half a dozen streams, but it's flaky: in a series of test runs with parallel streams, some will be fast and some slow. To add to the mystery, things look fine when measuring between 1Gbps-capable clients. This is all sufficiently weird that I didn't believe it for a long time, but having now acquired some additional 2.5G-capable machines to test with, it's impossible to conclude anything except that it's the MoCA link at fault. (@kelllogg9 mentions having seen something similar, but blames it on Windows which is not the case here ... this is a Windows-free household. I'm seeing this with Linux and macOS machines.)
One idea for how it might not be all the ScreenBeams' fault is if the coax line is asymmetrical --- I think it's just an isolated run with no splitters or filters anywhere, but I didn't install it so I can't be sure. Has anyone heard of performance like this being due to splitters/filters? How could I test the cable to be sure it's OK?
Failing that, does anyone have recommendations for other 2.5G MoCA adapters to try? I see from a bit of googling that competing units are available from TrendNET, Motorola, and ASUS for instance.
One idea for how it might not be all the ScreenBeams' fault is if the coax line is asymmetrical --- I think it's just an isolated run with no splitters or filters anywhere, but I didn't install it so I can't be sure. Has anyone heard of performance like this being due to splitters/filters? How could I test the cable to be sure it's OK?
Failing that, does anyone have recommendations for other 2.5G MoCA adapters to try? I see from a bit of googling that competing units are available from TrendNET, Motorola, and ASUS for instance.
Tried it with a brand new, never-used, 2-foot coax patch cable (one of the ones that came with the 7250s). Same results: wire speed in one direction, variable-but-200-ish-Mbps in the other. Makes no sense at all, particularly since earlier in the evening I ruled out swapping the adapters (no change in which direction was fast) and the boot order of the two adapters. After the swap test I thought maybe it was indeed something weird about the in-the-wall cable, but this test seems to refute that too.
kelllogg9
Occasional Visitor
Yup all your tests is what i also did a year+ ago and came out with the exact same disappointing results Wow, so finally after a year+ i am now seeing other people finally show up experiencing the same thing. As mentioned in the other thread, i spoke to their official tech support. We spent a week straight trying all kinds of tests and he giving me an upgraded (still in beta? at the time) firmware. After a week, i concluded his help was not helpful at all. Had to do my own endless tests for several continuous months afterward. I even started tweaking the TCP window size and the method it used (i forget all the terminology since it was a long time ago), etc. None of it made a difference.
I did notice a pattern that the speed would spike back up to both directions being at their max speeds (280MB/s+) when i was doing more activity on the network like streaming...if i stopped streaming the speeds would continue to operate at their ideal peaks. But if i stopped being active for some time and rechecked then the same one direction dropped back down to 112-180MB/s (i forget what my one direction was...too lazy to re-read my posts here). I concluded i think it is Windows 11 (for me) throttling speeds for whatever reason at specific times, only allowed higher rates when it decided it needed to ie, seeing higher activity on the network so it increased the tcp window size or something allowing the pc's transfer to flow at its maximum.
Last edited:
I actually saw this when I first got the ECB7250s back in July, but I didn't believe it --- how could two identical units not behave symmetrically? I blamed it first on the then-new Ugreen 2.5G USB Ethernet adapter I was using at one end of the test, and later on some other components in my network, but now I've done enough testing to say that none of those are at fault.
No Windows here boss
. I've spent time fooling with ethernet flow control on my switches and suchlike, to no avail. I cannot see a reason for this behavior, nor can I figure out why the bits preferentially flow from west to east not vice versa.
kelllogg9
Occasional Visitor
Much appreciated on taking Windows out of the equation. Helps narrow it down even more. Oh and thanks for even trying a brand new unit. I was always meant toNo Windows here boss
tbc btw, my speeds do return to full speed in both directions only under the condition i become active across the network for a period. but once i stop and leave it relatively low for some time (an hour or two) then the speed in the 1 direction drops to about half and stays there.
kelllogg9
Occasional Visitor
Yea that is what i thought years back and even the ScreenBeam tech guy thought the same. So we DID disable "Energy-Efficient Ethernet" and "Green Ethernet" on all nic cards. As i check, they are still disabled to this day. But either enabled or disabled it didnt result int he problem permanently disappearing. However, it is possible maybe that might explain why at times my speed does reach its peak max. Maybe the other guy consistently never gets faster speeds because those two options are enabled on his pc.
Interesting theory, but I don't buy it for my case. One ScreenBeam is plugged into a Cisco switch, which reports that EEE is disabled because the remote (i.e., the ScreenBeam) doesn't support it. The other one is plugged into a very minimally-featured Zyxel switch that doesn't appear to have any EEE features at all.
More info about the weirdness I'm seeing:
I went so far as to buy another pair of 2.5G MoCA adapters (these are ASUS MA-25 units), and was disappointed to find that they exhibit the same odd behavior as the ScreenBeams. After quite a lot of trial-and-error testing, what I believe I've proven is that the weirdness is associated with having specific client NICs plugged into a MoCA adapter. With "good" 2.5G clients, I can get solid, repeatable 2.35Gbps iperf3 rates in both directions, with one or several iperf3 streams. A "bad" client can transmit at 2.35Gbps to a good client, but it struggles with inbound data. Typical results are 200-250Mbps with one iperf3 stream. If I use six or more streams I can still get a total throughput of 2.35Gbps. This misbehavior generally isn't visible with 1Gbps data streams, only when the source and sink are capable of more. (Also, to be clear, the "bad" NICs can definitely handle 2.5G ... just not when plugged into a MoCA adapter.)
I don't have a huge range of more-than-1G clients to test. I've identified these NICs as "good": 2.5G ports on Cisco CBS350 switches and the 2.5G port on a fairly new Intel NUC (driven by an Intel i226-V controller). The "bad" ones are a Ugreen 2.5G ethernet-to-USB-C adapter and the high-speed ports on a Zyxel XGS1250-12 switch. Unfortunately the place where I need a MoCA link is between a Cisco switch and the Zyxel. I'm not buying another Cisco switch just to fix this, and I'm hesitant to start buying cheaper 2.5G-capable switches at random. So I'm just going to live with it for now.
I would like to understand exactly what is the cause of this sort-of-incompatibility, but I can't think of a way to investigate in any greater detail. Given that two different sets of MoCA adapters behave much the same, it seems like it must have something to do with data transmission patterns of the MoCA spec itself, but what?
I went so far as to buy another pair of 2.5G MoCA adapters (these are ASUS MA-25 units), and was disappointed to find that they exhibit the same odd behavior as the ScreenBeams. After quite a lot of trial-and-error testing, what I believe I've proven is that the weirdness is associated with having specific client NICs plugged into a MoCA adapter. With "good" 2.5G clients, I can get solid, repeatable 2.35Gbps iperf3 rates in both directions, with one or several iperf3 streams. A "bad" client can transmit at 2.35Gbps to a good client, but it struggles with inbound data. Typical results are 200-250Mbps with one iperf3 stream. If I use six or more streams I can still get a total throughput of 2.35Gbps. This misbehavior generally isn't visible with 1Gbps data streams, only when the source and sink are capable of more. (Also, to be clear, the "bad" NICs can definitely handle 2.5G ... just not when plugged into a MoCA adapter.)
I don't have a huge range of more-than-1G clients to test. I've identified these NICs as "good": 2.5G ports on Cisco CBS350 switches and the 2.5G port on a fairly new Intel NUC (driven by an Intel i226-V controller). The "bad" ones are a Ugreen 2.5G ethernet-to-USB-C adapter and the high-speed ports on a Zyxel XGS1250-12 switch. Unfortunately the place where I need a MoCA link is between a Cisco switch and the Zyxel. I'm not buying another Cisco switch just to fix this, and I'm hesitant to start buying cheaper 2.5G-capable switches at random. So I'm just going to live with it for now.
I would like to understand exactly what is the cause of this sort-of-incompatibility, but I can't think of a way to investigate in any greater detail. Given that two different sets of MoCA adapters behave much the same, it seems like it must have something to do with data transmission patterns of the MoCA spec itself, but what?
Yeah, it's possible. I observe a very large difference in their boot-up times, though: about 80s for the ScreenBeams and 135s for the ASUSs. So I'm inclined to guess they're not the same, but I don't have real hard evidence for it.
krkaufman
Very Senior Member
Didn't your post determine that the difference is seen between different NICs? So the issue would seem to be with the network cards or their drivers/settings, something that makes them less compatible with MoCA.
That said, I've seen a few reports where network performance improved after running "TCPOptimizer" on the computer. (example)
Similar threads
Similar threads
| Thread starter | Title | Forum | Replies | Date |
|---|---|---|---|---|
| W | 'Simple' MoCa Set up not working, need help | MoCA, HomePlug, HPNA | 10 |