Point-to-point links are usually single band. Unless you're getting 6 dB or more in gain over a typical patch antenna, you're unlikely to "feel" the gain at all. mk24's suggestion of at least 10-12 dBi seems like a reasonable threshold. I personally would be going for 15-20 dBi, but I have tooling, connectors, and LMR-series cables to make my own feed lines.
You may need to combine a unit intended for a point-to-point link with general-purpose AP for your local clients.
Like the R6220 with a outdoor antenna or outdoor bridge?
My idea was to now get a router dual band with dbi antenna enough to grab the wireless in front of me and in the future when I need to login at more far away WiFi then get a high gain antenna in that time.
Having free WiFi here its quite easy, besides many coffee shops share them passwords, the city have a portal made to tourists have free web. For my bad luck, that web portal reach me in the other side of the apartment from where I catch this one and I just grab signal if im almost outside of the window.
If i could find a something that could give me the high gain and extend in my budget then great, if not then just the router now and then in the future the high antenna if this 2 WiFi go no available or the router sucks grabbing the signal of them.
@Phoenix1:
You don't have to compile your own image in case your hardware has no bad eraseblocks in the NAND memory in front of the firmware. You only have to change the offset if you experience issues with reading the firmware. After installing the firmware provided by OpenWrt login to the router using the ssh console and run the command:
dmesg | grep eraseblock
You will see a list like this in case that your NAND has errors (applies only to one of my three R6220):
[ 2.935827] Bad eraseblock 181 at 0x0000016a0000
[ 3.011489] Bad eraseblock 435 at 0x000003660000
If you don't have any bad eraseblocks before the firmware offset you can just use the stock version from OpenWrt. My first bad eraseblock is located before the regular factory partition offset (0x2e00000). I had to shift the offset to 0x2de0000 as a workaround. Please find my Github repository at:
The structure of the files has changed with 18.06.2 but you can still use the information on this page to identify the affected code blocks in the current repository. I didn't update my Github repository anymore as I am waiting for the final fix and my workaround got rejected by the developers with good reason as it is only a workaround and they don't allow workarounds.
I bought the R6220 in Germany on amazon.de - it was quite cheap 1.5 years ago. I suppose amazon just wanted to reduce the stock as this model is getting old. Right now it is priced at 54.99EUR.
Thanks for the info, pity the release with the fix its taking so long. Will take that in mind when I decide. Anyway its a very good hardware.
Buy it and test my PCIe hardware mod. 
Kidding aside, we have the fix for the bug (2nd post) and I can confirm it does indeed work flawlessly. However the need for custom built firmware is quite annoying, you basically have to have your own repo for packages. That aside I personally am very satisfied with the purchase. The router provides great signal strength, stability and a solderable extra miniPCI-e port to fill with whatever (USB 3.0). But if you don't want to build the firmware, this device may not be for you. Until the fix gets incorporated into the release that is.
I hope that this will be mainline:
http://lists.infradead.org/pipermail/openwrt-devel/2019-November/020043.html
Has nothing to do with mainline... The R6220 is a Mediatek (mt7621) device, the patch you're linking to is for an Atheros (ar71xx/ath79) device.
This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.