Help for Newbie with QoS/VOIP

[shelleyevans]

@FreshJR, thanks for the answers to my questions, and this script. Thanks to everyone for the generous help. I am noticing something interesting-- not urgent, but since I am learning...

My bufferbloat is always "B" when I do speed tests on my ethernet connections, even though it's always A (since applying changes) on WIFI 5G. I have older CAT5 cabling in my house, so I run the ethernet at 100 (fast ethernet) speeds. I am assuming this is why--? My ATA's are on the LAN-- is a "B" score good enough for solid VOIP?

Thanks in advance.

 

[JDB]

If you are running older cables and getting only 100mbps it may be a reason for decreased rating. Have you tried a 1 metre new cable direct from a computer to the router? You should be getting 1000mbps assuming your Ethernet adaptor supports it (even with a few year old cable).

Either way a B rating should be fine for VoIP


As a side note, can I return your compliments and thanks and say it’s great to have a new member who is willing to listen, try, learn and be appreciative of the forum support!


Sent from my iPhone using Tapatalk

 

[shelleyevans]

Stumped:

After hours of troubleshooting, I finally realized I can't test a direct cable connection yet, because my Mac adapter is 100 mbps connection (head slap). But, according to the router, I definitely have 1GB WAN, which I can see from my speedtest results when I am on WIFI. I have ordered a 1000mbps adapter, but I am pretty certain that the old CAT5 cabling (+15 years) in my house is the rate limiter at my desktop, which has a 1000mbps card but consistently gets ~95mbps speeds. Way too complicated to move my desktop setup three floors down to the router, to confirm.

I have confirmed (see above, hours of troubleshooting), that when I run sfq with manual bandwidth settings, I get very low latency (no higher than 20) and scores of A+/A/A+ on Dslreports when I use wifi, which obviously allows the right sized pipeline. When I use my desktop (or the 100mbps adapter), my scores drop to A/B/A (where "B" is blufferbloat). More concerning, the bufferbloat numbers are running pretty high (?), around 70ms on download. This suggests to me that I am essentially creating my own bottleneck, with download bandwidth limits (175) that are higher than my wired connections maximum bandwidth.

I see a few possible solutions:

1) stop here, stop worrying about it, and accept the occasional glitch in VOIP that might occur because of bottlenecks. (My ATA's are all LAN.) (And I have had a couple glitches, but nothing like before.)

2) put my ATA's on the wireless network. Which makes me go yikes. Perhaps things have changed, but I have moca adapters all over my house to ensure stable connections where they are essential.

3) set different bandwidth download limits for LAN than WIFI. (don't know if this is even possible)

4) something else that I haven't thought of, because I basically have no idea how this works...

I welcome opinions, including "stop here, stop worrying about it..." Over the past couple days, I have been reading the WHOLE script thread, just for fun, and I was struck by something @FreshJR said-- when his latency (bufferbloat?) scores went from 200 to 20, he stopped fussing, and moved on to better problems.

 

[FreshJR]

I welcome opinions, including "stop here, stop worrying about it..." Over the past couple days, I have been reading the WHOLE script thread, just for fun, and I was struck by something @FreshJR said-- when his latency (bufferbloat?) scores went from 200 to 20, he stopped fussing, and moved on to better problems.

That thread is super old. Just the first two posts should be enough!

But onto your issue, wouldn't each ethernet cable act as a 100mbps link.

So even if your desktop is getting bufferbloat due to bottlenecking itself on the cable, wouldn't the neighboring VOIP ethernet devices still be fine?

(Do two simultaneous speedtests on two ethernet devices)

 

[shelleyevans]

the first two posts should be enough!

True, but every now and then something comes along that helps me understand better how this works. Extremely inefficient, but it's the best way I have found, so far, of gathering information-- it keeps me from posting questions fifty times a day!

even if your desktop is getting bufferbloat due to bottlenecking itself on the cable, wouldn't the neighboring VOIP ethernet devices still be fine?

I'm not completely sure what you mean, but I THINK you mean wouldn't other cables, with other devices, perform better? As it turns out, you are right— I did a speed check on a different desktop computer, that has a better cable, and I have results identical to my wifi. So I conclude that I have an old cable feeding my desktop (and also one of my ATA’s), and possibly another ATA. That cable is giving me slightly less impressive scores on Dslreports (A/B/A).

I had hoped I might be able to improve those scores by making special rules for the slower part of the LAN, that would prevent those devices from getting flooded. But perhaps that’s not possible, and/or I should leave good enough alone….?

EDIT: I answered my own question about putting those ATA's on WIFI-- not enough signal to that part of the house.

 

[Datalink]

@shelleyevans, out of curiosity:

1. Do you have a MoCA Point of Entry filter installed on the incoming cable to prevent any other MoCA system from accessing your MoCA network? The follow on to that is, do you have the MoCA adapter's privacy settings enabled and password protected?

2. You indicated that you have Cat 5 cabling in the home. Is that from inspecting the cables to determine that they are marked as Cat 5 cables? There are probably a good number of homes where there is Cat-5e cabling installed, but, the cabling is connected to the wallplate ports in such a manner that only 100 Mb/s is supported. Cat 5, 5e & 6 cabling has fours pairs of wires.

https://planetechusa.com/wp-content/uploads/2016/02/category-cable-wiring.png

Only two pairs have to be connected end to end to support 100 Mb/s. To support gigabit connect rates, device port to device port, all four wire pairs have to be connected end to end. It wasn't very long ago that 100 Mb/s cabling installation was standard. So, to cut corners and save time, the cabling was installed to support 100 Mb/s connect rates instead of gigabit connect rates. So, its up to the homeowner to check out the cabling and connectors to see how their connected and determine whether or not the other two wire pairs have been cut, or simply wrapped around the ethernet cabling where the cable connects to the wallplate connector or keystone.

Edit: doing a quick search, Cat 5 cable is built to support 100 Mb/s out to 100 metres. I haven't found a chart that shows the degredation beyone 100 Mb/s, but I'm sure there's one out there somewhere. Have a look at the following page, specifically the chart near the top that shows the cable data rates.

https://customcable.ca/cat5-vs-cat6/

Assuming that your cabling has four wire pairs and that all four wire pairs are connected end to end, I would expect you to see data rates higher than 100 Mb/s but, each cable data rate would be dependant on the cable length. I suspect that the performance wouldn't fall off of a cliff, but that there would be a degradation in the data rate as the connect rate was increased from 100 Mb/s to 1 Gigabit/s.

Looking back at your comments above, you indicate that your house is +15 years old. Our home is 14 years old with structured cabling which uses Cat-5e ethernet cabling. If your home isn't much older than 15 years old, I would suspect that it might have Cat-5e cabling installed. Your comment that you see +95 Mb/s really says to me that the cabling and connectors are only connected with two wire pairs to support 100 Mb/s, so seeing 95 Mb/s thru that cable and 100 Mb/s port is pretty reasonable.

On an ethernet connected device, select the internet connect symbol on the lower right hand side of the task bar. Right click and select "Open Network and Internet Settings" (Windows 10). Select "View your network properties". The link speed that is shown on that page is the port to port link rate. For cabling and connectors that only use two wire pairs, that should show 100/100 Mb/s (Receive/Transmit). For cabling and connectors that use all four wire pairs, that should show 1000/1000 Mb/s (Receive/Transmit). If you have a device (desktop or laptop) with a gigabit port, that link rate will confirm what the wallplate cabling/connector will support 100 Mb/s or 1 gigabit/s.

 

[FreshJR]

I'm not completely sure what you mean, but I THINK you mean wouldn't other cables, with other devices, perform better? .

So imagine you have your gigabit router connected to 4 clients with 100mbps cables.

Desktop1 = 100mbps
Voip1 = 100 mbps
Voip2 = 100 mbps
Desktop2 = 100mbps

So even if desktop1 is pulling 100mbps of data and bottlenecking itself, Voip1, Voip2, and Desktop2 still have the remainder of 60mbps to share internet traffic, or 100mbps each to share for intra-LAN traffic.

So the bufferbloat numbers on your desktop shouldn't introduce bufferbloat on the VOIP lines (which is the critical part of the network).

 

[shelleyevans]

@Datalink

Do you have a MoCA Point of Entry filter installed on the incoming cable to prevent any other MoCA system from accessing your MoCA network? The follow on to that is, do you have the MoCA adapter's privacy settings enabled and password protected?

Um, whoops. Nope. At least not that I know of. I imagine that you are suggesting I do a forum search, and figure out how to to do those things.....?

its up to the homeowner to check out the cabling and connectors to see how their connected and determine whether or not the other two wire pairs have been cut, or simply wrapped around the ethernet cabling where the cable connects to the wallplate connector or keystone.

I am going to do that! Maybe I will have a sudden increase in speeds. What a miracle that would be.

If you have a device (desktop or laptop) with a gigabit port, that link rate will confirm what the wallplate cabling/connector will support 100 Mb/s or 1 gigabit/s.

Well, I have OSX, but when I check the advanced settings, I see 1000baseT which I assume refers to the internal ethernet card, not the wall plate. True?

 

[Datalink]

Here are a few images of a MoCA Point of Entry filter:

https://www.google.ca/search?rlz=1C...c.1.64.psy-ab..6.2.178...0i13k1.0.7Wc5_7_w1Yo

This is installed on the incoming cable, prior to the splitter. Also note, to run MoCA adapters, you should also have a MoCA 2.0 qualified splitter installed, which is designed to allow port to port frequency ranges for MoCA operation on the house side of the splitter. That MoCA splitter should have as many ports as you have cables, ie, a one to one match so that you don't have any unoccupied ports on the splitter. Here is one example from Holland Electronics:

https://www.amazon.com/dp/B00OTO99VY/?tag=snbforums-20

Holland also makes other MoCA 2.0 qualified splitters:

http://www.hollandelectronics.com/catalog/catalog.php?product_id=catv-moca-splitter

Spec Sheet:

http://www.hollandelectronics.com/catalog/upload_file/GHS-PRO-M.pdf

For now, until you obtain a MoCA filter, I would advise shutting down the MoCA adapters as you would be feeding MoCA signals onto the ISP network and potentially causing interference for anyone else running an unprotected MoCA network. Depending on which frequency band your MoCA adapters are set to run, you could be causing interference with the ISP data range which is 0 to 1002 Mhz. Your MoCA adapters should hopefully be MoCA 2.0 qualified and set to run in the Extended D Band. Coupled with the MoCA filter and MoCA 2.0 splitter, you should see reasonable performance from that network.

Having said that, hopefully if you inspect your wallplate connectors and the structured wiring cabinet in the basement, you might find that only two of four wire pairs have been used in the connection between the ethernet cabling and the connectors. If so, then its a matter of replacing the connectors at all locations with Cat-5e keystones with additional short jumper cables in the structured wiring cabinet to connect to a gigabit switch. That's the easier way to do it. You can install RJ-11 connectors on the cable so that you can connect the cables directly to a switch or router, but, thats another tool to buy and its a little tougher to install those correctly. Those RJ-11 connectors are also installed on ethernet cables that you buy to connect various devices to a wallport or switch, etc.

My guess would be that you might be able to install keystones on the cables, with all four wire pairs connected and at the end of the day, that might allow you to remove the MoCA adapters and use ethernet throughout the house.

If you look behind the wallplates that have ethernet ports installed, it should be fairly evident that there are only two wire pairs out of the four that are actually connected. If in fact all four wire pairs are connected, that would imply that two wire pairs are installed incorrectly, and as a result, the devices at both ends negotiate a data rate with the remaining usable wire pairs that both devices are satisfied with. As a result, you end up with 100 Mb/s.

If you look at the cable ends in the basement, at the structured wiring cabinet, you should also see the same situation, two wire pairs connected, possibly four pairs connected incorrectly.

For the mac, have a look at the following link to show the link rate with the connected router, modem or switch:

https://www.mactip.net/how-to-check-ethernet-connection-speed/

The link rate should show 1000 Mb/s. If you're on a 100 Mb/s download plan from the ISP, then a link rate of 1000 Mb/s with a switch or router, with a download rate of 95 Mb/s would make sense.

 

[shelleyevans]

@FreshJR

So even if desktop1 is pulling 100mbps of data and bottlenecking itself, Voip1, Voip2, and Desktop2 still have the remainder of 60mbps to share internet traffic, or 100mbps each to share for intra-LAN traffic.

Well, regrettably, math is not my strong suit. Assuming you are not tired of this, let's say I have:

Desktop 1 = 1000mbps
Voip1 = 100 mbps (limited by my crummy Cat5 cable)
Desktop 2 = 100 mbps (downstream of Voip1, on passthrough)
Voip2 = 100 mbps (on a second crummy cable)

And let's say I have your script installed, with 20% allocated for VOIP, VOIP at the highest priority in the GUI, and also the ATA's at highest priority in Bandwidth monitor. My bandwidth maximum download is 175. That would reserve .20 of 175, or 35mbps for VOIP, well below my 100mbps bad cable limits. Your script would allocate and distribute the remaining 140mbps according to the priority defined in GUI, and percentages defined in your script. This might mean that desktop2 would suffer, if the router decided that it should get 70mbps of downstream data, which would be more than the total (100mbps) that the cable can support. But VOIP, on all the devices, good cable or bad, should be fine. (I am inferring that ALL the LAN devices share the same 175mbps download stream defined in the GUI; I am also inferring that the WIFI operate independently...) Is any of this correct?

Feel free to ignore my questions. I realize that I am way behind the curve in understanding, and others need your help with coding!

 

[FreshJR]

@FreshJR


Well, regrettably, math is not my strong suit. Assuming you are not tired of this, let's say I have:

Desktop 1 = 1000mbps
Voip1 = 100 mbps (limited by my crummy Cat5 cable)
Desktop 2 = 100 mbps (downstream of Voip1, on passthrough)
Voip2 = 100 mbps (on a second crummy cable)

And let's say I have your script installed, with 20% allocated for VOIP, VOIP at the highest priority in the GUI, and also the ATA's at highest priority in Bandwidth monitor. My bandwidth maximum download is 175. That would reserve .20 of 175, or 35mbps for VOIP, well below my 100mbps bad cable limits. Your script would allocate and distribute the remaining 140mbps according to the priority defined in GUI, and percentages defined in your script. This might mean that desktop2 would suffer, if the router decided that it should get 70mbps of downstream data, which would be more than the total (100mbps) that the cable can support. But VOIP, on all the devices, good cable or bad, should be fine. (I am inferring that ALL the LAN devices share the same 175mbps download stream defined in the GUI; I am also inferring that the WIFI operate independently...) Is any of this correct?

Feel free to ignore my questions. I realize that I am way behind the curve in understanding, and others need your help with coding!

The only issues are that desktop2 might pull data at the link rate and introduce bufferbloat on all of its traffic .

Desktop1 and the VoIP devices should be fine .

 

[shelleyevans]

The only issues are that desktop2 might pull data at the link rate and introduce bufferbloat on all of its traffic .

Which would explain why my cloud backups are running ALL the time now. I guess I better see if I can fix those cables, as @Datalink suggested.

You are very generous with your time-- both in coding, and also your clear explanations (which I have followed in that long thread) to people like me who don't know what we're doing but install your script anyway.

 

[shelleyevans]

If you look at the cable ends in the basement, at the structured wiring cabinet, you should also see the same situation, two wire pairs connected, possibly four pairs connected incorrectly.

You imagine a much tidier set-up than I actually have. No basement, no structured wiring cabinet. I think whoever upgraded this 1926 house basically ran very long stretches of cable from four different wall plates, snaked them through the studs, and dropped them into an old linen closet, where they just dangle, like Spanish moss, with RJ11 (whoops-- RJ45!) connectors, until I plug them into the router. I am going to bring in a friend to check the RJ 45 plug and the wall plate on the offending cable. I hope very much you are right, that they are only partly connected. And I really thank you for the thoughtful, thorough consideration of all my troubles.

 

[Vexira]

Information about Asuswrt-Merlin version and Adaptive QoS

Testing with factory reset, without backup on both versions.

- With Asuswrt-Merlin v380.70 + FreshJR QoS, when I play NEVER have ping spike, lag or lost packages, when I'm using all the bandwidth.
- But with Asuswrt-Merlin v384.4_2/384.5 + FreshJR QoS, when I play I GET ping spike, lag and lost packages, when I'm using all the bandwidth.

Adaptive QoS is broken in Asuswrt-Merlin v384.4_2/384.5 Firmwares, regardless of whether the script is running.

The same thing happens to me that he with Asuswrt-Merlin v384.4_2/384.5:
- Need help on flashing back older version of Merlin

@Find the Door @FreshJR

Which router are you running?

I have no issues on my RT-AC88U, with the latest test build and freshjr script have you tried the bandiwth % commands?

 

[shelleyevans]

Hello all! I am busy updating my network with some of the recommendations that I received in this thread-- a Comcast tech is showing up tomorrow to add a POE MOCA filter outside of my house (beyond my technical skills and courage), and is going to check my LAN connections in my office, to see if the wall plate might have only two pairs connected instead of four. BUT: it turns out I have an unused coax hiding behind another wall plate, and so I bought some MOCA 2.0 adapters to see if they would work on my old Moca 1.1 network. I tested the adapters on an existing coax outlet. The speeds are actually identical to my ethernet speeds (the gigabit ones). Furthermore, when I run the DSLreports speed tests, when I use the Moca connection DSLreports test always reports 16 streams down 16 streams up; when I run the tests over ethernet, I get 16 streams down 6 streams up; when I run the tests over 5g wifi, I get 16 streams down 12 streams up.

So my hopefully last question is this: what are these "streams" DSL is keeping track of, and is it better to have 16 up and 16 down than 16 down and 6 up? If so, perhaps it would make sense to just have the guy hook up the coax cable to the wall plate so I can install the moca adapter, and not worry about fixing the misbehaving Cat5 cable. Thoughts?

One piece of background information: NO cable TV, not even basic, in this house, just internet. So the COAX lines are basically empty of traffic, except, at the moment, the Moca 1.1 network which only handles two rokus and one ATA.

PS everything is working so well, you guys. Clear phone, speedy internet, and great bufferbloat scores. THANK you!!!

 

[HuskyHerder]

@shelleyevans

OS X user here as well, you have done very well with your endeavors.

Re: DSL Reports, Here lately I have had results from their Speedtest all over the place. Not always 16 Streams, often not returning grades at all.
Streams : They are running tests to multiple servers aka Stream and using the results to compile your score. In my understanding.

Re: ISP, My speeds from Spectrum are 300/20 advertised achieving 340/22, NO TV either, only internet. Some of my issues have been related to new work in the neighborhood and not inside my home. Sometimes noise bleeding in on the Coax etc. Bad connections at the street.

Re : Home cabling, I installed cabeling once upon a time. Lots of Cat3, 5, 5E, I have just started using Cat 6 in my home for new pulls. I have very few pulls of 5e remaining. No need for Cat6 mind you, its just that I had some left over from work at the office. 2/3 of a free 1000' box.

Also the misbehaving Cat cable may simply be a broken end or wire in the plate or even mis wired.

Re : You notes, Well done. I was thinking about something similar for macOS / OS X users, but you already documented it.

@FreshJR

Just a suggestion maybe link to the text or include it on you topic for those Mac users who may see putty etc and go nah I’ll pass. Just my 2¢

 

[Datalink]

Hi @shelleyevens, it sounds like the your internet service is improving bit by bit. Just to add my 2¢, its definitely worth looking at the Ethernet connectors to determine if the cable is connected with all 4 wire pairs (8 wires). That alone should greatly improve the data rate thru the Ethernet cables, regardless of whether they are Cat-5 or Cat-5e. Normally those cables should provide a faster data rate compared to a MoCA adapter, but, it depends on whether or not they are Cat-5 or Cat-5e and also on the cable length, but, I can’t imagine that those cables are anywhere near the max length of 100 metres that the cable specs are good for. Getting those cables working to their full capability should be part of your overall plan of improving your internet service.

The additional MoCA adapters as you have indicated are MoCA 2.0 qualified. They are interoperable with the MoCA 1.1 adapters but have additional frequency ranges compared to the 1.1 adapters. The end result, when you mix them together is that you won’t be able to take advantage of the additional frequency range of the 2.0 adapters. You have to be careful at this point to ensure that all of the adapters are set to operate in the same frequency range. Here is the frequency band selection for the Actiontec ECB6200:

http://communityforums.rogers.com/t5/Internet/Rogers-Hitron-CODA-4582-Hardware/m-p/419670#M51438

Note the selection of D High and D Extended. The 1.1 adapters won’t have those selections, the 2.0 adapter will. If you look at the frequency chart on page 7 of 34 of the MoCA 2.0 spec, you will see the various frequency ranges above 1 Ghz. The ISP uses 0 to 1002 Mhz for cable operations. Typically, if a home user was running some some type of whole home pvr system, that system would use the D Low frequency range to communicate box to box and provide program data to any other pvr/receiver in the home. The 2.0 adapters easily accommodate this by increasing the frequency range of the adapter and include a selection to use the D High frequency range.

http://www.mocalliance.org/MoCA2/specification/MoCA_2_Device_RF_Characteristics-150406d.pdf

From page 7, the Extended D range has a frequency span of 550 Mhz.

From page 7, the D range has a frequency range of 400 Mhz

From page 9 of the spec, the E range has a frequency range of 300 Mhz.

From page 10 of that spec, the F range has a frequency range of 225 Mhz.

So, what does this all boil down to? Ideally, all of your MoCA adapters would be MoCA 2.0 qualified. That would give you the ability to select the Extended D range for MoCA operations, using the widest bandwidth possible and as a result, the highest data rate possible with the MoCA adapters, given the cable losses and signal losses that you would see with a splitter.

Thinking aloud here, say for the sake of the argument that you only needed the cable to the modem to be connected to the external ISP cable. If you didn’t need that same cable to backfeed MoCA data to the splitter, you wouldn’t need the splitter for the modem cable feed. There would simply be a barrel connector to join the external and internal cable together to feed the modem. That connector looks like this:

https://www.lowes.com/pd/IDEAL-4-Pack-Brass-Screw-On-Adapters/3363004

So, you wouldn’t have any splitter signal loss to the modem, which is inherent with a splitter installation.

At that point, if you have access to an Ethernet cable or other co-ax cable, you would use either one, or both to feed their respective systems. The Ethernet would be feed from the router, same for the MoCA adapter. In this case, if you’re mixing the MoCA 1.1 and MoCA 2.0 adapters together, your first choice for frequency range would be the D Band, followed by the E and F band. The E and F band would be useable if in fact you only had the modem fed from the external cable. The remaining MoCA system would be entirely internal and there would be no chance of ever feeding MoCA data to the external ISP cable system. You wouldn’t need the MoCa POE filter at all, but, if it’s installed, it shouldn’t cause any issues.

So, this is where you can start thinking out of the box, if in fact the Ethernet cabling situation is all figured out and you have the cable situation figured out as well. My usual advice to anyone running a combination of a modem and router is to park them near each other in a location where the router’s wifi will do the most for you. Usually that’s somewhere that you usually use your wifi devices. To do that you would need a cable outlet, and an Ethernet outlet as well. The modem doesn’t have to go in a closet somewhere, same for the router, but you really need an available cable outlet, and an Ethernet outlet as well to do that. If you can do that, you can install an unmanaged gigabit switch in the closet and feed that switch with the Ethernet cable from the router. The switch can then feed the rest of the Ethernet cables throughout the home. The tricky part in your case is the cross-over from Ethernet to MoCA. You could probably park the one cross-over MoCA adapter anywhere that you have both Ethernet and cable outlet. The MoCA adapter then feeds the rest of the MoCA system.

At the end of the day, you could end up with much higher wifi data rates as the router is closer to your working location and you have Ethernet and MoCA data access at their respective locations. It all depends on the location of the closet where the modem and router are currently located and how much of a wifi signal loss you experience as a result of the homes construction. I imagine that with a house that old, you probably have chicken wire covered with lath and plaster for the walls. Not really conducive to wifi operations these days, but, depending on the locations of the cable and Ethernet wallplates, you might be able to substantially improve on the overall Ethernet and wifi performance.

Here’s what this might look like:

External ISP cable -->> MoCA filter -->> Cable run -->> Barrel connector -->> Cable run -->> Modem


Modem Ethernet port -->> Router -->> Ethernet port -->> Cable -->> wallplate -->> Cable run -->>

gigabit switch -->> other switch port -->> device

-->> other switch port -->> device

-->> other switch port -->> device


Modem Ethernet port -->> Router -->> Ethernet port -->> Cable -->>MoCA Adaper -->> cable -->>

wallplate -->> Cable run -->>MoCA Splitter -->> Cable Run -->> MoCA Adapter

MoCA Splitter -->> Cable Run -->> MoCA Adapter

MoCA Splitter -->> Cable Run -->> MoCA Adapter


So, there are probably several variations on a theme that could be applied. You could cross connect the Ethernet and MoCA systems somewhere else in the home where both Ethernet and Cable outlets are available. Just have to be willing to think outside of the box and make use of the cabling that you have available. Don't be afraid to try different configurations to determine what provides the best overall performance for your requirements.

Only last point is to log into the MoCA adapters and ensure that they are all set to use the same frequency band for the cable configuration that you happen to be using.

 

[shelleyevans]

@HuskyHerder Thanks for the compliment, but honestly, everything that I wrote in that document is a direct copy of advice that I got from several people on this thread. I wrote it all down for myself, so that when I have to troubleshoot later I will remember what I did! Then thought, hey, maybe others could benefit.

@Datalink Thank you for this thoughtful reply, which came in as I was outside (in the snow! in April!) playing with our cable box (for hours... ). After the Comcast guy left, having installed the POE filter (which allowed me to learn how to get into the box, and also where the POE filter goes) I discovered, to my disappointment, that the cable jack he installed by my desk (a remote corner of the house with no wifi and poor LAN) will not talk to the Moca network at all. By this I mean, when I plug the adapter in, the green coax light doesn't go on. I tried both my new Moca 2.0 adapter, and my old adapter. No joy. I also used those same adapters successfully at other outlets, so I know the adapters are not faulty.

Thinking there might be a problem with splitters, I went to the outside box. We had a rats nest of cables, along with two splitters-- one three-way leading to one five-way. I tidied it up by disconnecting everything, and then, as I reconnected the cables, figured out which cable fed each outlet in the house. By using a Moca adapter at every outlet I discovered that four of my jacks can't get Moca of any kind at all (no green light), and four work perfectly. This leads me to assume that there are splitters in the wall, and/or that the cable run from those jacks terminates somewhere else or possibly was never activated or the person who did the wiring initially never bothered to finish the run. Or.... ?

At the end of the long cold day, I think I might have to give up on the high speed Moca idea all together. It will be too expensive and time consuming to figure out where those dead cable runs go, or why they don't work.

I tried to tidy up my old Moca 1.1 network, and the cable runs into the house, according to your advice.

My external cable box setup is now: ISP cable>>POE filter>>Holland Two Way Moca Splitter. One leg of the splitter feeds my cable modem inside the house; the other leg goes to one of the Comcast splitters, which feeds three other coax cables in the house. I use those three outlets for the two Rokus and the ATA on Moca 1.1. Since my modem is at the first split, I am hopefully getting a stronger signal. And the POE filter keeps me safe from outside interference.

My friend will come by on Thursday to see if my Cat 5 cabling has been improperly terminated, which I really hope, and to assess the possibility of using the old wire to snake a new one. If that doesn't work, I'll live with 95mbps at my desk. I need to go back to my day job.

 

[shelleyevans]

@Datalink It turns out the cable wasn't properly terminated! With a few quick snips and twists, I am up and running, taking advantage of my gigabit connection at my desktop for the first time ever! THANK you.

 

[Datalink]

Woohoo ...... nice to see success at the end of it all Was that a problem with the RG-6 cable inbound to the modem, or a problem with the Cat 5/5e ethernet cables not being properly terminated to support gigabit data rates?