Stumped by simple MoCa setup

[amallah]

I have this very simple MoCA setup with 2 goCoax adapters that has been working fine for about a year.

Recently it stopped working. I was able to trace the issue down to the splitter in the line. When I replace the splitter with a coupler (taking cable out of the equation), the two MoCA adapters can see each other (screenshots of it working)

So I replaced the splitter with this:
https://www.amazon.com/dp/B0054EILOE/?tag=snbforums-20

The moment I replaced the splitter, I got the MoCA connection light without even power cycling anything, so I thought I figured it out! But then when I went to clean everything up and put all the adapters back out of sight in their various rooms in the house, after some unplug/replugging of cables and power, it stopped working and I have not been able to get it to work again. The only thing I can think of is that there is a specific channel where everything works together and it has nothing to do with the splitter but just which device boots in what order? It seems like I have some control over channels, but I don't know enough about MoCA to know if this theory even makes sense.

Thanks for any help in solving this puzzle.

 

[degrub]

Do you have any DVRs or Cable TV on the coax ? Or is it just internet service ? Is their a MOCA light on the DOCCIS modem ? is it on ?

What you describe (boot order) has been found when other MOCA devices are present on the coax. That may also be a partial explanation for the low bandwidth you are seeing.

Otherwise, the Out to Out internal signal loss is probably too high (isolation) by design.. That looks like a straight splitter for signal distribution.
Get a Holland MOCA2 rated splitter, put a MOCA POE filter in the inlet to keep your MOCA signal inside ( and to improve the signal quality )

BTW, ~1700 is roughly half what it should be for MOCA2.5 between the modems. Do you have some cable/connector issues or other moca devices present ?

Usually, what i have seen with DOCCIS modems, is that you have to place a MOCA splitter ahead of the DOCCIS modem and the MOCA modem rather than as you show. You will still need the head in splitter (assuming you don't have another physical way to do the layout that would isolate the DOCCIS modem from the rest of the coax) and MOCA POE. If it is still working with the DOCCIS 3.0 modem, then you probably don't have any overlap of bands.



Look through the many threads here about dealing with DOCCIS modems. It gets more difficult if the ISP switches to DOCCIS 3.1 as you will loose bands.

 

[amallah]

No DVRS, no CableTV - it is just internet service and there is no MOCA light on the modem.

MOCA POE and MOCA rated splitter, I will definitely try that. I never even noticed the 1700 before today after reading other threads and seeing how much higher everyone else has. I'm starting to suspect a bad line outside my house that may just finally have reached the point where a signal can't be established.

 

[krkaufman]

When I replace the splitter with a coupler (taking cable out of the equation), the two MoCA adapters can see each other (screenshots of it working)

If that’s working, it’s not working very well for a MoCA 2.5 connection. 5 channels bonded should have those 1700 figures upwards of 3500 (up to 700 Mbps PHY rate per channel x5).

It might be interesting to see what the adapters report if you directly connect them using a short coax cable. If they don’t report nearly 3500 Mbps in that table then there’s something wrong with one or both of the adapters or their power adapters. (Still, wouldn’t mean there aren’t improvements needed for your setup, starting with using actual MoCA 2.x-rated splitters — not just splitters spec’d above 2Ghz — and getting a “PoE” MoCA filter properly installed.)

 

[degrub]

Log into your doccis modem and post the downlink and up link bands signal strength table.

what does the physical cable layout actually look like - from the demarcation box outside all the way to the doccis modem and the moca modems ?

 

[amallah]

If that’s working, it’s not working very well for a MoCA 2.5 connection. 5 channels bonded should have those 1700 figures upwards of 3500 (up to 700 Mbps PHY rate per channel x5).

It might be interesting to see what the adapters report if you directly connect them using a short coax cable. If they don’t report nearly 3500 Mbps in that table then there’s something wrong with one or both of the adapters or their power adapters. (Still, wouldn’t mean there aren’t improvements needed for your setup, starting with using actual MoCA 2.x-rated splitters — not just splitters spec’d above 2Ghz — and getting a “PoE” MoCA filter properly installed.)

Short cable showed the rates I expect:

Which is unfortunate, because if I get half the rate using a coupler, that means there is definitely a problem between the two MoCa adapters, which is probably the most inaccessible piece of this system.

 

[amallah]

Log into your doccis modem and post the downlink and up link bands signal strength table.

what does the physical cable layout actually look like - from the demarcation box outside all the way to the doccis modem and the moca modems ?

This is my rough drawing of the physical layout. I included the WiFi router, because there are a few ethernet devices which need to be in the location they are. I thought I might need to replace Coax B in the drawing, but now that I am seeing the speed difference, maybe it's Coax A?

If I can't find the issue with Coax A (or I find the issue and it's in the wall), I could move the DOCSIS modem, add a hub for the extra Ethernet needed at that part of the house and some additional coax cables to get here if you think this may be an improvement:

Although with that amount of effort, maybe Powerline Ethernet would be easier.

 

[amallah]

Log into your doccis modem and post the downlink and up link bands signal strength table.

what does the physical cable layout actually look like - from the demarcation box outside all the way to the doccis modem and the moca modems ?

I am not sure if this is good or bad:

 

[degrub]

What you have should work, you just need to replace that splitter with a Holland MOCA2 splitter and a POE cap on the inlet side of the splitter at the connection to COAX B. However, that may reduce the signal level to the doccis modem too much. You might have to use a moca2 splitter with "zero insertion loss" on all ports. This would require DC power injection to the splitter either direct plug in or across a specific coax port.

Here is what i would do since you mentioned you can move the equipment.
1) move the doccis modem and the wireless router to the downstairs location moca modem. connect the doccis modem directly to the cable, COAX B, from the demarcation box. Connect the moca modem coax port to the COAX A cable and its ethernet to a wireless router lan port. Connect the wireless router WAN port to the doccis modem lan port.

2) if the wireless signal upstairs is not good enough you have two choices. A) if you don't need wireless downstairs, turn it off on the wireless router and add an AP upstairs ( either another wireless router in AP mode or just an AC or AX wireless Access Point) or B) add an AP upstairs to cover that floor.

What this does is give the ISP modem the maximum signal strength and, more importantly isolates your internal coax layout from the ISP. You do not need any splitters.

If you use a router in AP mode upstairs you will have both the AP and a 4 port switch to use in the same location as the current position of the wireless router. If you need lan ports in the lower floor, you have at least 3 available on your current wireless router.

If there is a coax or connector termination issue, it is on coax A.

 

[krkaufman]

I thought I might need to replace Coax B in the drawing, but now that I am seeing the speed difference, maybe it's Coax A?

Evolving the short coax direct-connect test, you could connect the two adapters over each of the coax runs in question to see if one reports as more or less problematic than the others.

Aside from ensuring compatible splitters and getting a "PoE" MoCA filter installed, you'd also want to consider reterminating the coax connectors to eliminate bad connectors as a source of your issues.

---
edit: p.s. Re:

I thought I might need to replace Coax B in the drawing, but now that I am seeing the speed difference, maybe it's Coax A?

With a "PoE" MoCA filter properly installed (on the input port of the first splitter encountered by the incoming cable signal), "Coax B" wouldn't have any effect on your MoCA setup, since it would be on the upstream side of the "PoE" MoCA filter. (Coax B *would* be relevant to the signal quality delivered to the cable modem, of course.)

 

[krkaufman]

Short on time, here's a diagram of what *should* work ... w/ the "PoE" MoCA filter added, assuming splitters are upgraded to recommended MoCA 2.x-compatible versions, and coax connections are tested and reterminated as needed:

But I was also going to suggest the same thing as @degrub in their just-prior post -- ASSUMING 2 coax runs are possible between that outside 2-way splitter location and the inside MoCA adapter location. (Using a 3 GHz F-81 coupler to link Coax B into the room; and a second 3 GHz F-81 coupler to link Coax A into the room.)

Relocating the modem ensures maximum DOCSIS signal strength to the modem and future-proofs the setup for DOCSIS 3.1 (equipment or actual implementation), avoiding the frequency overlap issues between MoCA and DOCSIS 3.1+. The resulting setup, detailing degrub's suggestion, would be similar to the following:

​

NOTE the absence of a "PoE" MoCA filter in this later setup, as it's no longer needed due to the physical isolation of the DOCSIS signals from MoCA.​

​

And the AP in the upstairs could just be a Gigabit Ethernet switch if the additional wireless coverage is not needed.​

 

[krkaufman]

Short cable showed the rates I expect:

These still aren't as high as I would expect for a short direct-connect test. Here's an example from a parallel thread, with per-channel rates much closer to 700 Mbps:

​

I guess what I'm saying is ... don't exclude the possibility of MoCA hardware failure. Got a friend nearby with a couple spare adapters you could try out?

 

[kelllogg9]

Short cable showed the rates I expect:
View attachment 37900
Which is unfortunate, because if I get half the rate using a coupler, that means there is definitely a problem between the two MoCa adapters, which is probably the most inaccessible piece of this system.

In case it matters, just now i decided to connect both my Moca 2.5 units together via a "short cable". Technically, I'm using a male-male F-type coupler to reduce the coax length to its shortest possible distance that i can get (1 inch). Here is what you SHOULD be getting at their peak performance...

I'm pretty confident these values would hold if i used a 3 foot coax too. So if ur values are what u showed for short cable, then something seems wrong with one or both of your units. Granted, those values might just be normal for GoCoax devices?!? I'm using ActionTech ECB7250 units (has true 2.5Gbe ethernet ports) not GoCoax. Interestingly, our firmware are identical in appearance, but i am @ version MXL371x.1.18.2 while yours is at MXL371x.1.15.7. I can give you my firmware file but it may not be compatible since ur units are 1Gbe ethernet ports vs my 2.5 ports. Here is my Device Status page, if it also matters...