Hello there. It’s my first time setting up a Mikrotik Hex Refresh as a replacement for my Synology Router and it was a very steep learning curve in terms of configuring all the basic set up. After spending hours tinkering this router, I think I got it to the point where it is functioning the way I wanted except for one thing. My WiFi download speeds are considerably slow compared to my previous setup. The download speed appears to max out at 300mbps.
I have a 1Gbit connection from the ISP and all wired connections to the RouterBoard are downloading at 950mbps.
My WiFi access points are Google WiFi AC nodes configured as bridge so all DHCP leases are coming from the RouterBoard. I connected a computer directly to this node and can confirm that I can download at 950mbps no problem.
I already have fasttrack enabled as initially I was having issues with hardwired devices maxing out at only 600mbps. After enabling, I’m now getting the correct speed but this only works for wired connections.
I also tried connecting my old Synology AC router directly to the RouterBoard and configured it as a basic router with WiFi. I get the same thing—->hardwired devices connected to the Synology AC Router are getting a steady 950mbps download but wireless devices are maxing out at 300mbps.
Is there any configuration that I need to change in order to get the old speeds I’m getting without the RouterBoard?
Yeah, same question as above. The description clearly says, that over the cable you get full speed, but your wifi AP is not from MikroTik, so I don’t see how the hEX could have any influence on the problem
Some interference in form of timing jitter affecting TCP window scaling?
Experience with official test results says that figure listed under “Ethernet test results → Routing → 25 ip filter rules → 512 bytes [packet size]” resemble real-life performance quite well. And hEX refresh has 498.1 [Mbps], which means that 950Mbps @OP sees when using wired computers is quite an achievement. And any slight disturbance can mean a very noticeable drop in performance.
To isolate if this is a problem with the Google Wifi AP, i switched to my old Synology router and set it as an AP behind the Mikrotik router and I got the same exact results.
Getting max throughput when PC is connected directly to the Synology AP but maxes at 300 when connected via WiFi to the Synology AP
Already several times it has been hinted to show your config although the info you provide, does not seem to indicate an issue with the wired part.
Even if you add a non-MT AP into the mix, it is still a wired device for your Hex.
However …if from wired to wired you “only” get around 820Mbps, then something is possibly not ok. You should get 950-ish.
I also have Hex Refresh and using ports 2-5 it can reach 950Mbps (= 1Gb connection) without even sweating if set properly.
Unless you also added ether1 into the mix ?? That’s a special case on this device since it does not pass switch chip, only direct connection to CPU.
But then we need to see your config AND small diagram how everything is connected.
(moved post with config into this thread since you replied in another thread, from 2023)
Please also add drawing how you test your iperf setup.
From which device to which port on Hex, which port from Hex to which other device.
Same with wireless AP, what port is it being connected to ?
Quick comment: this is not a standard config you show here and may very well be the reason why everything slows down …
As an example: using 2 bridges makes sure everything has to pass CPU once you pass bridge boundaries. Bye bye throughput.
Thanks. The 2nd bridge I have was just configured for future expansion in case all goes well in my configuration. This is currently not being used.
The config I have is mainly to utilize Netwatch to handle the multi-wan failover.
There came a time when both WAN1 and WAN2 failed, and we were without internet for a week, so I had to add WAN3 hence the need for a multiwan router.
WAN1 is active for most connections. Im using WAN2 and WAN3 as failovers and not load balancing. I have other IPs configured to the other WANs (different docker containers running speedtest trackers for each WAN connection—they work fine. I get the correct speeds advertised by my ISPs on each connection). Other than that its just WAN1 for everything.
The iPerf3 server is on PC2 and PC1 as the client for the Wireless to Wired test
PC2 as server and PC3 as the client for the wired to wired test
If you are testing from PC1 wireless to PC2, you are not really passing Hex.
It stays on that switch.
What sort of switch is it ?
Simple test since both ether4 and ether5 are on the same bridge:
remove that switch from the mix.
remove/disable bridge2 (to be sure HW offloading is fully active on the remaining bridge) - reboot might be needed, I am not sure here so better be safe and do it.
Add AP1 to ether4
Add PC2 to ether5.
There are months apart between the results from the two screenshots, over half a year. When is the most recent “good” result (obtained with the old router)? Your wireless environment might have changed in between, maybe the neighbor bought more powerful equipment and now there is more interference? Maybe the Google nodes got some firmware upgrades, or the MacBook some driver upgrades?
