Which access point to use

Hello,

I would like to use two access points to connect a house with an internet connection to another house. I have two directional antennas pointed at each other and am currently using two older TP-Link Access Points (TL-WA801ND Version 3.0) with the original manufacturer firmware. The distance to be bridged is about 300 meters, which isn't a problem given the signal strength and the two directional antennas.

In principle, the whole setup works, but the connection keeps hanging. For example, when I watch TV over the internet in the house receiving the Wi-Fi connection, the picture freezes frequently and regularly.

However, if I watch TV over the internet directly at the source internet connection, there are no problems.

I am now looking for two good, high-performance access points that simply offer more power and also support external antennas (i.e., have an external antenna connection or detachable antennas).

And, of course, they need to be flashable with OpenWRT.

Suggestions or tips are welcome.

My location is Germany, so it would be best if I could buy the access points here in Germany.

Regards, Per

There is a problem with using directional antennas of your own added to an access point, the cables have MASSIVE attenuation.
This is bad enough when transmitting, but when receiving can be a show-stopper.

There is a reason that professional point to point equipment has the actual access point mounted directly behind the antenna.

At 300 metres, the signal could easily be enough before these cable losses, but after even 2 metres of cable, losses will result in many packet drops.
This can easily become unusable, particularly when streaming video.

But outdoor grade access points with full mainstream OpenWrt are very few and far between and mostly obsolete eg Ubuiquiti Nano.

There is another problem, particularly in ETSI regions eg Germany, in that 5GHz outdoor use is ether prohibited or restricted to very low power depending on channel.

The answer is a DIY weatherproof case for a pair of wifi6 access points running on 2.4GHz HE40, giving a 600Mb/s link for 2 stream/2 antennas, or 1.2Gb/s for 4 stream/4 antennas.
Channel 13 is usually best and 100% legal/unlicenced in ETSI regions.

I have used pairs of GL-MT3000s (two stream/two antennas ie 600Mb/s) in suitable ABS weatherproof instrument case, running OpenWrt with mesh11sd for the link, successfully over a ~1.2 Km link in some installations, in direct line of sight, so 300 metres should be quite doable.
I don't have one of the boxes to take a photo, but here is an MT3000 as it would be mounted in the box. As it is configured as a mesh11sd node, it only needs the power cable and the cable gland is for that. A 3 or 4 metre usb lead for power is fine. Pencil for scale.

The two antennas give a fairly wide active beamforming function, so all you need to do is roughly align the boxes at each end ie the lids roughly facing each other.

Direct line of sight is important.

60GHz mikrotik bridges?

Hi everyone,

As a ham radio operator (dc0hs), I already know my way around RF and antennas a bit.

The antennas and the cabling are not the issue. The signal is definitely coming in very strong and going out just as strong.

It's also not like the connection isn't working at all. When I run a speed test, I get over 50 Mbps in both directions. So that part works fine.

Because of that, my main question here was really just whether there is a solid Access Point that allows me to connect an external antenna.

I'm pretty sure I won't be able to bridge this distance without an external directional antenna.

Sure, there are off-the-shelf wireless bridges specifically designed for this kind of challenge, but I'd prefer to avoid total overkill if the problem can be solved in a more elegant and budget-friendly way.

There has to be an access point out there that can be flashed with OpenWrt and has connectors for an external antenna.

Any further suggestions are highly welcome!

Cheers,
Per

Aio have this advantage that you need to configure them and mount, no worries about antennas, cables, proper sealing etc. But if you want diy then mikrotik have nice parts that you can easily play with. For this distance you dont need high power cards and high gain antennas. For bridging you need transparent wds bridge.

Mikrotik sxt 5nd lite can do 97Mbps and in my place can be bought used for cheap.

Okay
Maybe I really should pay some attention to these devices.
The
Mikrotik SXTsq Lite5, 5 GHz, 10/100 Mbit/s
I can get brand new for under 40 euros.
If I understand correctly, you plug the network cable directly into the device?
Do you still have to configure something, or is it just plug and play?
I really have to think about it.

Obviously, a lot of money could have been saved if I had gone with this device right from the start.

Regards, Per

https://www.kaufland.de/product/346097656/?kwd=&source=pla&id_unit=386119519797&utm_source=google&utm_medium=cpc&utm_id=23214772797&gad_source=1&gad_campaignid=23214783807&gbraid=0AAAAADprSmFKSdykpNzdfqlsk3FTe0f26&gclid=CjwKCAjwmJjSBhB-EiwAkZgxiy2Opcc2ykymNOqYCWZnrdD0ZfMARDQEcwz9adnrMAZj_9fnIVpuARoCogAQAvD_BwE

maybe spent some more bugs for

Mikrotik RBSXTsqG-5acD

https://www.kaufland.de/product/346098205/?search_value=Mikrotik

Good for you.

It is the "coming in" that really matters most.
So what is the received signal strength? Sounds like you actually measured it....

As you know your way around antennas a bit, you will of course have done some measurements. What is the actual loss in the cables?

Your TL-WA801ND Version 3.0 has two antenna connections.
You will of course know why it has two antenna connections.
What are the two directional antennas you have at each end?

Hello,

I just took a look for some old pictures.
The setup is located in Sweden and I live in Germany, so I can't just quickly check the signal strength, but I checked what I documented back then and I recorded a signal strength of 50 dB.

The exact same access point is used in both places.
I installed a directional antenna on each of the access points and left one standard antenna attached (the factory vertical antenna).

The TL-WA801ND has two antennas, presumably so I can point them in different directions.
Therefore, an outdoor grid antenna TL-ANT2424B with 24dBi 2.4GHz is connected to the access point.

And the exact same setup is installed at the other house.

Regards, Per

The standard unit in both networking and RF engineering is the "dBm", where dBm stands for decibels relative to one milliwatt (dB(mW)).
The formula is:
dBm = 10 × log₁₀(P / 1 mW)
where P is the power in milliwatts.

Radiated power in "WiFi" is VERY low and received signal strengths will always be negative numbers.

  • Example: -50 dBm = very strong/good signal (0.00001 mW).
  • -70 dBm = good/usable signal.
  • -80 dBm = weak but often still functional.
  • -90 dBm or lower = very poor/marginal (barely usable or disconnected).

For the benefit of those reading this that are not radio engineers, this is a Logarithmic scale.
A 10 dB increase means the power is 10 times stronger. A 3 dB increase is roughly double the power.

FYI: From the TL-WA801ND specs, a minimum RSS of -68 dBm is recommended to achieve the best data rates.

This will defeat the built in 2x2 MIMO configuration, at best significantly reducing sensitivity, at worst, resulting in a high level of dropped or missed packets.
(MIMO - Multiple Input Multiple Output)
As you are seeing stuttering and freezing when streaming video, my guess is you are suffering from the latter case ie a high level of dropped packets.

No. It is designed for two identical omnidirectional antennas, both orientated vertically for best coverage in a domestic situation.

These, with the correct industrial grade radios are good for ~30 Km point to point links, so somewhat overkill for 300m.
But to work with the TL-WA801ND, you need another two of them....

Hello bluewavenet,

Just out of interest, where are you located? I am located in southern Germany, as you can also tell from my amateur radio callsign. Near Heidelberg.

My name is Per, as you can see from the posts.

I am not sure if you are interpreting the signal strength reading correctly.
I do think you are right about the decibel-milliwatts, but the router doesn't display it with a negative sign, and the weak signals have a lower value. My remote station has the highest value.
The signal should definitely be strong enough if, as you say, the antenna is designed for several kilometers.
It is obvious that something is going wrong in the transmission and that's why dropouts are occurring.
However, I don't think it's like you say, that you would have to connect two directional antennas to each access point just because the router has two antenna ports.

I know that the manual for another router of mine (Fritzbox) with 3 antennas states that you should point the antennas in different directions to cover different areas of the house.

Why shouldn't I use the small vertical antenna for the immediate surroundings and the other antenna for the bridge to the other house?
That would mean, if I understand you correctly, that if I had a router with only one antenna, this problem wouldn't exist.

I could also just disconnect the one antenna or connect a dummy load to it.

I always assumed that the TL-WA801ND always transmits on both antennas simultaneously and also receives on both simultaneously, like a "dumb" network hub and not like an intelligent switch that knows on which network port the partner can be reached, but perhaps my assumption is wrong there.

But apart from that, what do you think of the suggestion M10 made to use Mikrotik bridges?

Regards, Per

Mmmh, and here I thought that for multiple spatial streams (with 2 antennas at most 2 IIRC) it can help to orient the antennas differently so different spatial paths are more likely. Is that completely wrong?

I would tray that. I would naively? assume that cheap routers likely expect all antennas to be quite similar and might struggle when the actual antennas have noticeably different characteristics as your seem to have, no?

I don't have a purchase recommendation nor am I an RF/WiFi expert but here are a few thoughts on the problem.

You checked the signal strength. If you haven't monitored the level while the drop-outs occur, that may be worth doing. Using good test equipment would be better than relying on what the proprietary software in the cheap consumer grade devices shows.

Consider sources of interference. Other WiFi devices in range. Noisy electrical equipment. RF on the mains power.

You mention using one antenna connection for the inter-building link and the other for local use. Transmissions on the "local" use could be cutting into the available bandwidth needed for the streaming data. Is the tv data stream bursty like Netflix or does it keep packets flowing every few milliseconds?

Analyzing packet captures on the wired side could reveal more details about the nature of the problem

Cheers

The OP's router is quite old tech, 801.11n and 2x2 MIMO.
It relies on both antennas for the MIMO to function (Multiple Input, Multiple Output), meaning it can multiplex Tx and Rx, but it requires both antennas to have the same characteristics, defaulting to two 5 dBm gain omnidirectional in a plane at 90 degrees to the antenna.
They are little stubby antennas so have a quite broad angular beam pattern, typically around 120 degrees between 3db points if you looked at a polar plot.
Yes, in a building, orienting the antennas 10 or 20 degrees off vertical can give a better signal on other floors of the building, but this will have a performance penalty, albeit often negligible.

Once you move up to 802.11ax or 802.11be the wireless chipset has the speed and capability to do phased array beamforming and keeping track of where signals are coming from and being sent to. Two streams with 2 antennas with beamforming is a big improvement on 802.11n, but 4 streams / 4 antennas is gobsmacking. 8 stream / 8 antennas is... well absolutely f*cking amazing akin to alien technology.

Anyway, a device similar to the mt3000, or maybe an mt3600be is 2 stream / 2 antennas, fairly cheap, small enough to fit in a small ABS box and powered by usb type C connector. Mounted outside can eat up 300 metres without trying when connecting to a mirror image of itself as a point to point, using 802.11s frames for the backhaul with optimised rts threshhold and ddr scheduler quantum settings, the performance will be rock solid. No fancy antennas required.

Because this router is 802.11n 2x2 MIMO. It multiplexes both antennas for each wireless link giving a much increased throughput.

Either way, as there is no way to tell the wireless chipset you are doing this, and it is hard wired to expect both antennas to be of the same characteristics, you will be breaking the 2x2 MIMO, resulting in significant packest dropping.

No, it does a kind of full duplex operation - Multiple Input Multiple Output (MIMO).
The antennas are not independently assigned permanent "Tx-only" or "Rx-only" roles — the chipset coordinates them dynamically for optimal performance in AP/Client/Repeater/Mesh modes.
Effectively the router can transmit and receive at the same time (not quite, but a very good approximation) This is the 802.11n standard, very outdated these days and superseded with better things eg 802.11ax, 802.11be and more in the development pipeline.

A Real-World Analogy - Think of it like two people talking in a noisy, echoey room:

  • They take turns speaking (half-duplex), but listen at the same time as speaking.
  • But each has two ears positioned slightly apart, allowing them to focus on the desired voice while ignoring echoes and their own voice (diversity + processing).