Speed Bottleneck on Aimesh

[The Lorax]

I currently run an ASUS Aimesh LAN using RT-AX88U (main) with two nodes ZenWiFi and one RT-AX1800S. Backhaul is ethernet. The 1800S is remote connected by UeeVii wireless bridge CPE-820. ISP speed is >1 GbE, modem is 1 GbE and all nodes connect at 1 GbE except the bridge which connects at 100 Mbps. The bridge is 1 GbE port selected (has both 100 and 1 G). All cabling is Cat 6 except for bridge which is Cat 7. I understand that the NICs set the speed so older ones only go to 100 Mbps but the bridge is 1 Gbps. Where is the "bottleneck" causing the bridge to connect only at 100 Mbps while all over nodes connect at 1 Gbps? Would like to recover some of the 450 Mbps from ISP on the 1800S (P2P bridge connection). Any suggestions? Can find an option to force a particular port on the main router to force 1 Gbps instead of fall back to older 100 Mbps.

Oh, I have five ethernt LAN connections on the 88U, three at 1 Gbps, one at 100 Mbps (older device, Zigbee hub) and the CPE bridge at 100 Mbps.

 

[bbunge]

I currently run an ASUS Aimesh LAN using RT-AX88U (main) with two nodes ZenWiFi and one RT-AX1800S. Backhaul is ethernet. The 1800S is remote connected by UeeVii wireless bridge CPE-820. ISP speed is >1 GbE, modem is 1 GbE and all nodes connect at 1 GbE except the bridge which connects at 100 Mbps. The bridge is 1 GbE port selected (has both 100 and 1 G). All cabling is Cat 6 except for bridge which is Cat 7. I understand that the NICs set the speed so older ones only go to 100 Mbps but the bridge is 1 Gbps. Where is the "bottleneck" causing the bridge to connect only at 100 Mbps while all over nodes connect at 1 Gbps? Would like to recover some of the 450 Mbps from ISP on the 1800S (P2P bridge connection). Any suggestions? Can find an option to force a particular port on the main router to force 1 Gbps instead of fall back to older 100 Mbps.

Oh, I have five ethernt LAN connections on the 88U, three at 1 Gbps, one at 100 Mbps (older device, Zigbee hub) and the CPE bridge at 100 Mbps.

I noticed that the CPE-820 is PoE capable. If you are powering either device with PoE is there a chance that the PoE switch or power adapter is only a 100MB device? I looked for a manual on the CPE-820 but could not find one quickly.

 

[The Lorax]

Indeed. I'm way ahead of you. Since UeeVii always shows the setup topology with LAN connected to bridge rather than data over the POE connection, I assumed they supplied 100 Mbps injectors. So I looked it up and they claim the injectors are 1 GbE with fall back if using the 100 Mbps port. Unfortunately, I can't confirm because all the POE speed detectors cost $1300 or more. Any other suggestion?

 

[Samir]

I would try powering the device without the POE injector and see if it will connect at gigabit. If so, you found the culprit--the injector. Otherwise, it could be a very rare incompatibility that I've seen just a handful of times in many decades. The way to test for this is to get a small gigabit unmanaged switch and put it in between the asus and the ueevii and if it connects at full gigabit, it's that rare incompatiblity. If it does not connect at gigabit, it's a 100Mb port on there no matter what they're saying or a bad/marginal cable.

 

[The Lorax]

Cables are fine when connected to other 1G ports and devices. Also swapped cables between 5E, 6 and 7, all show same result. Unfortunately the UeeVii bridge does not have power cord/module capability, just POE. And others who tried skipping the POE injectors for their POE switches/routers could not get the bridges to work. UeeVii claims no power cord available although they have a power cord slot. Very difficult to get power cord 30 ft up on a roof and through the wall so POE is basically needed. I thought of purchasing 3rd party POEs rated at 1 G but it is very, very difficult to match bridge input requirements. I will try another run at the Chinese manufacture for Tech support to verify both the POE injectors AND 1 gigabyte port on antenna. Although there is a huge language understanding problem with English requests. Chinese companies that produce consumer electronics are usually very good. Just wish they would hire English speakers to translate user instructions to be understandable (better quality).

 

[The Lorax]

Another piece of the puzzle. A customer on Amazon responded with this about the CPE-820:

"We are using cat 5e cables with a 1 gbps router and we are seeing transfer rates of 450-500 mbps."​

So it must work as described.

 

[The Lorax]

I just read in another two forums that people claim Cat 7 cable has to be grounded or it only connects at 100 Mbps. (Odd since Cat 6 UTP, Cat 6 STP and 6a S/FTP all connect at gigabit speed just not 10G Base T.) Anyway, my UeeVii bridge connected with Cat 7 SFTP has no ground connection to POE or router. The Cat 7 RJ45 connectors are crimped onto cable collar and mesh wire jacket with copper foil tape. The POE injector has no single pin ground and the router has no third pin on its' plug. How does one ground either device? I can install an ethernet surge protector in-line with the POE which carries a ground loop. Those surge protectors often degrade connection speeds themselves so how would I know?

 

[degrub]

The shield is what has to be properly grounded. Usually, that will be to the case earthing at the device on one end. The important thing is that it is only grounded at one end. Otherwise you can get a ground current loop and cause normal mode noise or bias on the signal conductors.
One reason CAT5E and CAT 6 are more useful.

 

[The Lorax]

Thanks for the ground explanation. However, I'm connecting a wireless bridge of 400 ft LoS. Grounding the Cat 7 cables that supply POE on each end should not connect a ground current loop. I will try inserting a POE+ pass through grounded surge protector (lightning) between the bridge antennas and the POEs before going to router port. It may work but at least I have some insurance against near lightning over current damaging equipment.

 

[Samir]

Cables are fine when connected to other 1G ports and devices. Also swapped cables between 5E, 6 and 7, all show same result. Unfortunately the UeeVii bridge does not have power cord/module capability, just POE. And others who tried skipping the POE injectors for their POE switches/routers could not get the bridges to work. UeeVii claims no power cord available although they have a power cord slot. Very difficult to get power cord 30 ft up on a roof and through the wall so POE is basically needed. I thought of purchasing 3rd party POEs rated at 1 G but it is very, very difficult to match bridge input requirements. I will try another run at the Chinese manufacture for Tech support to verify both the POE injectors AND 1 gigabyte port on antenna. Although there is a huge language understanding problem with English requests. Chinese companies that produce consumer electronics are usually very good. Just wish they would hire English speakers to translate user instructions to be understandable (better quality).

While it may be difficult to get that unit down and find a proper power unit for it, that's the only real way to test it. If there's already a language barrier, I would be surprised if you're being told the truth--usually the chinese want to dump crap and get their money and then repeat, and any support pretty much is to get rid of you vs solve your problem.

 

[The Lorax]

Certainly you have a good suggestion but there is a great risk involved. If the manufacturer does not have an AC/DC power unit for the antennas they also will not provide specific power needs for the units. Finding a 3rd party power adatpter that matches those unknowns is risky due to possible damage if not found. Already I am having difficulty locating 3rd party POE injectors that match their specs as most supply far more power, voltage, frequency. I'm betting on the issue being in the Cat 7 connections since that is the only difference compared to all other connected devices. Bypassing shielding, grounding and wire gauge I should see if the gigabit ports are actually correct. Since no one has publicly noted a similar problem with the CPE bridge I assume the ports and POEs are correct. Lots of networks use these bridges without issue. Otherwise they would not sell so well if misrepresented. Of course, I could have gotten the odd unit with a defective port but unlikely since I tested four bridge units all with the same 100 Mbps limit. Thanks.

 

[degrub]

so there are two lan ports on each module
and the bit rates listed are :

it is not clear to me if the POE device is gigabit or if one of the ports is gigabit. You could have a situation where they provide a universal POE injector with capability of gigbit input/output and only 100mbit input lan port on the bridge.

check the sync lights on the bridge ( should have them, maybe not though) or POE injector output to bridge. No need for tester.

Not sure why they would need two lan ports on the device unless one is power injection only and the other is data only.

 

[degrub]

so there are two lan ports on each module
and the bit rates listed are :

it is not clear to me if the POE device is gigabit or if one of the ports is gigabit. You could have a situation where they provide a universal POE injector with capability of gigbit input/output and only 100mbit input lan port on the bridge.

check the sync lights on the bridge ( should have them, maybe not though) or POE injector output to bridge. No need for tester.

Not sure why they would need two lan ports on the device unless one is power injection only and the other is data only.

So from the user manual the two lan ports are different - any chance the lan cable carrying poe is plugged into the 100mbit port instead of the gigbit lan port circled ?

 

[The Lorax]

No chance. Verified gigabit port on both bridges and POEs are gigabit. I did test the 100 Mbps ports on bridges to see if there was a mix up or defect, no. I think I may have narrowed the issue but testing will determine. The Cat 7 cable/connectors to the bridge units are rated 10 G and triple shielded, grounded. However I re-checked and the cables are 26 AWG stranded copper. So it is possible that the RJ45 gold connectors only contacted a few strands which would account for positive continuity but have greater impedance in smaller wire. (Guess) To test, I am replacing with 23 AWG solid copper, shielded CAT 6 rated at 10 G and 550 MHz. All other connections service gigabit on this cable. Need extra shielding because cable passes through common conduit with several MoCA, ethernet cables. The reason I chose the Cat 7 was triple shielding and weatherproofing to withstand temps of -23 F to 100 F, ice, rain, snow, critters and lightning strikes nearby.

 

[Samir]

Certainly you have a good suggestion but there is a great risk involved. If the manufacturer does not have an AC/DC power unit for the antennas they also will not provide specific power needs for the units. Finding a 3rd party power adatpter that matches those unknowns is risky due to possible damage if not found. Already I am having difficulty locating 3rd party POE injectors that match their specs as most supply far more power, voltage, frequency. I'm betting on the issue being in the Cat 7 connections since that is the only difference compared to all other connected devices. Bypassing shielding, grounding and wire gauge I should see if the gigabit ports are actually correct. Since no one has publicly noted a similar problem with the CPE bridge I assume the ports and POEs are correct. Lots of networks use these bridges without issue. Otherwise they would not sell so well if misrepresented. Of course, I could have gotten the odd unit with a defective port but unlikely since I tested four bridge units all with the same 100 Mbps limit. Thanks.

Wait, there's no power specs on the port? That is a problem. If you can get the specs, there are universal power supplies that will work as I've had to do this before.

Most cat7 or other such rated cables are usually some sort of fake crap. I would get a long cat5e cable and test that first if you suspect a cable is the issue. I'm also leaning on it being a cable issue as I've seen ethernet cables fail to 100Mbs if they're marginal.

This is the first time I've even heard of the company, and if their primary market is asia where people are used to buying stuff and it not work, getting ripped off, etc., then there wouldn't be any evidence of the systemic issue--and that's just the way business works over there.

 

[Samir]

No chance. Verified gigabit port on both bridges and POEs are gigabit. I did test the 100 Mbps ports on bridges to see if there was a mix up or defect, no. I think I may have narrowed the issue but testing will determine. The Cat 7 cable/connectors to the bridge units are rated 10 G and triple shielded, grounded. However I re-checked and the cables are 26 AWG stranded copper. So it is possible that the RJ45 gold connectors only contacted a few strands which would account for positive continuity but have greater impedance in smaller wire. (Guess) To test, I am replacing with 23 AWG solid copper, shielded CAT 6 rated at 10 G and 550 MHz. All other connections service gigabit on this cable. Need extra shielding because cable passes through common conduit with several MoCA, ethernet cables. The reason I chose the Cat 7 was triple shielding and weatherproofing to withstand temps of -23 F to 100 F, ice, rain, snow, critters and lightning strikes nearby.

I would not use shielded cabling at this point--just some proper quality branded ethernet cable to test that gigabit is possible. Then you can get cable that is more suited to the environment. A lot of the higher rated cabling is fake unless being custom made by someone who uses branded cabling and terminations.