Hi @Vincent-TP 

 

I have already referred to the link you shared last week and have done my level best in managing the bands and width given the clients environment and circumstances. Hence this was why I disabled 2.4Ghz and DFS. 

 

Btw, could you help explain about the EAP683's statistical rules of Rx dropped packets. I would like to understand it better.

 

So far the client feedback today that speeds are fine just the occasional disconnections for some of the staff.

 

I have received the ticket ref by email. I shall follow up via email then.

 

Much thanks

Long time customer first time poster.

 

I have a EAP-683UR with latest firmware EAP683 UR(US) v1.0.  Omada control software latest version 5.14.26.

 

I am getting extremely high RX packet loss.   I'm talking like 300k per day? Pic related. Doesn't seem to affect any of my clients, but something doesn't seem right with this.

I've tried reducing the antenna power and doesn't seem to have an effect even at low settings. My neighborhood is pretty interference free (using wifi analyzer app) so its not from rival networks. I have just one AP too.    I like the AX6000 infrastructure and why I went with this AP. Not sure if I should return it.  Tried DFS channels and not to no effect. Disable /  Enable LLDP doesnt do anything.  All the clients are updated android iphone mainstream products (ipad pro, iphone pro, pixel 7 pro, samsung s23 ultra)

 

 

Is this a hardware issue or an app reporting issue?  Seems like the latter, but I don't want to miss my return window if the device is defective.  Apparently a lof of people with the european version had issues with this one.  

 

Let me know.

 Hi @MarktheRed

 

I'm afraid I have had no further improvement since reporting about this anomaly weeks ago.

 

Even in an empty office environment, with no users around, all 4 EAP683 URs I've deployed still report Rx dropped packets, even on wired LAN. However probably not as much as what you are experiencing. Thankfully users haven't reported any unsual behaviour in day to day use.

 

I've asked @Vincent-TP to explain about the EAP683's statistical rules of Rx dropped packets but havent got a reply from him. Perhaps if this could be tweaked maybe things could change.

 

My previous deployments using lower spec EAP620HD don't produce these Rx errors, so why the higher end EAP683 UR does bewilders me.

 

@MarktheRed, perhaps you could try replacing your current unit with another new unit and see if it reproduces the same results. If yes, then you know its a TP Link issue (hardware or software). If it doesn't, then you know you had a faulty unit.

 

I experienced this before with the EAP620 HD - had 2 new units out of 10 that became faulty after initial deployment as I detected unsually slow and low throughput. Once replaced everything was fine.

 

Good luck!

 

  @WiFi_Done_Right 

 

Thanks for the response.  I hate being the return guy at amazon, but I think I will do that.  I am looking at comparable Unifi Enterprise APs.  I really like the 4x4 antenna, as while I may not have a lot of active users of the system, I do need quite a bit of range.   I would always prefer to have one "big" AP as opposed to many small ones, just because its much easier to tune.   The thing is I really drank the Koolaid on the AX6000.  Merlin over on smallnetbuilders (this guy knows his stuff) raves about it and considers it the ne plus ultra of wifi for the next 5-10 years.  I also really would prefer to stay on the TP-link ecosystem (pic related) over Unifi because of better customer service.    Minor quibbles though, the Unifi AP mounts a lot cleaner and hides the cat6 cable in my upper floor ceiling.  I'll wait a week before "swapping" this on Amazon.  Is there a TP-link AP you would recommend?   Seems like the 650HD is next best, although it doesnt have a 4x4 antenna.

 

 

  @MarktheRed 

Since I am going to be the return guy.  I am going to try the EAP772 to compare coverage.  I've always read that 4x4 antenna's are better than 2x2, but maybe TP-Link is putting all their juice into the newer wifi7 models.    The basement of my house (my office) pretty much only gets 2.4GHZ so this will be the make or break decision.   

 

(1) Step 1: Use the 683 UR (replacement model - will it stop dropping packets)

(2) Step 2:  Use the EAP772 and see if it works.  My switch can handle 10gb ETH but its a 100' Cat6 run and this will require a LOT more juice to amplify hence me going with the 2.5GB POE version (772 vs 773)

 

Wish me luck!  Will post findings this weekend.

  @MarktheRed 

 

Still waiting on the replacement 683ur and the Ubiquiti Enterprise U6 to come in for the showdown.   In playing around, I did manage to tame the 5GHZ errors somewhat.   Its still a high number on an absolute scale but its 9%.   I could probably live with this, but its disappointing considering this is new tech and my eap 225v3's were nigh perfect for years without issue.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

THe 2.4 GHz band is still a mess regardless of channel I put it on.   I read this helpful tuning strategy from the enemy here and made 2.4 GHZ only 20hz channel and tried eithe 1,6,11.   All have the same error rate despite channel 6 being significantly more open in my neighborhood.  I made the power setting -7dbi to the 5ghz power setting to no avail.  In fact the errors seemed to be inversely proportional to each other. (i.e, if 5ghz goes down, 2.4ghz goes up)

 

I think this is a hardware issue implementation for this AP by TP-Link.  The AX6000 is a very proven design and used by some of the most popular routers in the world (Asus ROG, etc.)

 

Ill swap for the new one and see if that resolves this, but I think I might just go with the Unifi 6 enterprise as it (1) looks nicer mounted in my ceiling (2) has wifi 6e and (3) the software runs MUCH easier than Omada on my home Truenas server.   THe (3) point is not TP-Link's fault, more of a fact the Omada controller cannot detect / adopt AP's inside a Kubernetes cluster whereas the Unifi one does.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hi @MarktheRed 

 

Looks like you're gonna have a busy, but hopefully, an interesting weekend comparing the EAP772 and U6 Enterprise.

 

Hope you do have devices that can ake advantage of the 802.11be standard as well as MLO.on the 772.

 

Re: recommended replacement AP, if you're looking for something closely similar, the 660HD would be the one as it has identical specs to the 683 UR but with slightly lower power ratings I think. The 660HD is also design for higher density environments but if all you need is just 1 unit, you should do just fine.

 

Looking back as your earlier postings, I realize you did not share your topology. What gateway and switch models are you using? 100m LAN cable length is quite long too, I've often preferred to use SFP+ over fiber for longer runs which requires an additional switch but this brings the AP closer to switch resulting in shorter Cat6/Cat6a cable runs, which I prefer.

 

Interestingly enough the customer I deployed the 683UR for was formerly using Ubiquiti U6 lite APs and a Ubiquiti 24 PoE switch.

 

Personally - I did not like how the U6 lite mounts as the mounting brackets and screws were so tedious to put on compared with the 683UR, but that's a personal thing. The cabling is indeed better looking on the U6 lites however I did make them just as neat on he 683UR though they show a little.

 

Do share your findings once you've compared the 2 APs, it would be interesting to know. 

 

Good luck🤞🏼 

 

  @WiFi_Done_Right 

I guess I need to clear things up; I may have changed my mind.  I am not going to bother with the EAP772; the lead time was too long (+2 weeks).  I am getting a replacement new 683ur and a Ubiquiti U6 enterprise.   One (per your advice)  to see if I have a defective 683ur.  The other to see how the u6 enterprise performs in the same relative environment.

 

Its really unfair to TP-Link if their hardware is working great but their controlling software is jank (?) and reporting these things incorrectly.  That said my eap225v3 worked great, I'm beginning to regret getting rid of them considering the hassle these new and improved ones are supposed to be. 

 

Running a single AP on a high speed highly modern network with ANY 2.4ghz and 5GHZ channel combination concurrently in a one acre property with no outside interference should NOT be this hard and should not generate hundreds of thousands of dropped packets / failures..   Sorry if Ubiquiti can  do it and TP Link can't, well sayanora TP-link (after 10+ years as a customer).  

 

Finally I don't have a 100m run.  I have a 100ft run of Cat6, which is actually reasonable considering its running from my basement rack to the center of my 2nd floor hallway.

 

 

  @MarktheRed 

 

Well tried the replacement EAP-683UR and didn't get any better results.  The results pretty much flatlined with the 5GHZ slightly upticking to my last one and the 2.4ghz stayed the same.   I used the exact same channel width, power, channel, etc. as the last one.   

 

The unit that came in the "New" box was acually a returned unit.  I noticed this when the radios fired up it had "Madi Manor" broadcasting.  This is more of an endictment of Amazon than TP-Link as they as sending new items as clearly used.   Maybe Mr. Madi had the same problems and got fed up with it and swapped for a new one.   This if anything makes me think I made the right choice in going to the U6 Enterprise which I will summarize here. 

 

Here is quick totally last minute benchmark I did of the three units:

 

 

Both of these devices were in the basment of the house in my office, as here is where I noticed the most problems with random disconnects during the first install of the 683.  You'll note the speed of the devices were largely similar, but I will say the laptop STAYED in 5ghz mode with the Unifi.  Gave zero f's at all.  The 683 constantly switched me to the 2.4ghz band down there and the only way I could run the test was create a standalone ssid for 5ghz.

 

The Unifi gave far better coverage outside my house.  The treehouse, front yard, driveway all had coverage whereas the 683 dropped /vanished in those areas.

 

The Unifi software works better on Truenas Kubernetes and is much more intuitive to use (I still hate how it asks me 2FA every time I log in).   The topology map should be the industry standard in presenting wifi networks IMO.  Every device is clearly identified (i.e., Apple, Ecobee, Pixel, etc.) whereas the Omada one doesn't even recognize iphones and ipads as such and you have to cross reference router lease tables if you want to know what is what.

 

Lastly the U6 just looks nicer mounted on low ceilings.  Cant see the cable and looks professional and clean.  The mounting kit that comes with it is night and day more professional and cleaner (the silicone plug wrap is great).

 

The truth is I hate Ubiquiti as a company.  Their tail covering during known security breaches is shameful.  They shipped my ER-4 in a gimped state whereby you have to do fancy CLI commands just to disable traffic flow restrictions and actually get 1gps throughputs.   I wanted TP to work, but the U6 enterprise is night and day better than the 683ur and worth the extra money.

 

 

 

 

 Hi @MarktheRed 

 

Apologies if I misread the length of cable run to your basement. 

 

Thanks for sharing your findings. I guess the decision is quite clear for your setup with the U6 enterprise. Hope it works out well for you going forward.

 

There are quite a number of settings within the Omada controller that requires experimenting but the whole process can be time consuming and worst still - not fruitful.

 

Hopefully TP Link figures out a way to further optimize the 683UR as in my case - the customer had paid for and is deployed.