What Could Be The Problem ? Uplink Interference ? Or What ? Dirty Channel ?
?
What problems are you actually experiencing with the link? From everything you’ve shown, this link doesn’t actually look too bad.
There’s really only two things you can do to reduce retransmissions, latency/jitter… reduce channel width and/or find a less noisy channel. You might also be able to get a bit better latency/jitter by locking the maximum MCS rate on both radios to whatever the highest % MCS rate is (e.g, looks like MCS7 on downlink side, and MCS8 on uplink side). You could also try changing to a long or short guard interval on the AP.
Changing channel and channel width can only be done by ISP so I have limited options.
I am experiencing high ping and ping spikes while gaming more especially from 15:00 to 18:00
the downlink locks itself to DS MCS 2 - QPSK 3/4 during that period.
The Link Quality Flactuates Every 5 Seconds And Drops Down To 49% Or Even Worse 20%
Then you will need to get your ISP involved, or find a new ISP.
1 Like
Seems Like The Only Way To Stop This Is Limiting Downlink To DS MCS 5 - 64-QAM 2/3 And Uplink To DS MCS 3 - 16-QAM 1/2 That Way Jitter Is Around 5ms And Ping Is 18ms In Most Tests
Ive written about this before, and youre right. Do not listen to those telling you tonreduce channel width or lock mcs to highest rate - this is the worst thing to do. What you’ve done is exactly right, you need to lock the mcs rate to the lowest rate that has packets passing, accept the lowest rate of course. I tune every single coe radio this way and my latency is 10-20ms with zero packet loss and zero retransmits. At 80mhz you can still get over 100mbps at mcs2, and with uplink set at mcs3 you can get more than 25mbps.
You worked it out yourself, but its crazy ppl still think reducing channel width is helpful. I tell people with noise issues to increase width to 80mhz and fix max mcs lower and lower until theres zero retransmits. Reducing channel width is old thinking that is wrong with these radios, no matter how big Eric’s response is 
1 Like
Exactly. This thing of locking mcs to the one which passes the most packets doesn’t work. Better to keep reducing mcs till 99% or 100% of packets pass at chosen mcs while also monitoring retransmissions and capacity drop packets.
Incase someone is looking for a fix. This is how I did it. “Aim for 99% of traffic to pass at a single mcs value”
That is precisely what i said 
1 Like
2026 and we’re still having to lock in max MCS rates. Ridiculous. Needs a more pessimistic MCS rate set in software based on retransmit rates. We don’t have time mess around setting max MCS rates on thousands of radios
2 Likes
A different brand has an adjustment for Exactly that: “Pessimistic, Normal, Optimistic” still adjust automatically as conditions change, but just erring on the side of caution If that works the better in your situation
Yeah, like how in ePTP mode it gives you a low latency, balanced, and high throughput options.
2 Likes
Yes, exactly, that same idea. In most scenarios, they are the essentially the same, but when the link gets full and things start to get stressed… what do you want it to err on the side of. 