Bought the wireless wire kit. Bridging between a shop building and my home. Distance is 50 meters. Home unit is about 12 feet up on the side of the house. Shop antenna is about 15 feet up. Clear line of sight. Running 7.14.

Using the black, plastic aiming device I got both antennas lined up really well. Then I tried to follow the arrow LEDs on the back for fine tuning but no matter how hard I tried I just could not get the arrows to extinguish and the green light in the middle to turn on. The closest I could get would be for a pair of arrows to appear, say up and left, and eventually I would move the antenna enough so that they would switch to down and right. I could never get down to a single arrow, much less no arrows. And following the arrows often drove the pointing of the antenna well away from pointing at the other antenna. Finally I gave up and lined them up with the black, plastic optical sights again and left it at that.

I'm not complaining too loudly because testing with iperf3 shows the link providing approx. 950 Mbit/s (TCP) in both directions. WinBox says signal strength 95 and MCS 8 (what is MCS, anyway?) But I'd sure like to figure out how to fine tune them. Maybe there's some odd multipath confusing the internal algorithm?

50 meters is too close to get any exact alignment without the most minute changes. Looks like you have achieved maximum throughput as is.

MCS stands for Modulation Coding Scheme. Has to due with calculating maximum throughput using the number of spatial streams and channel widths, I believe.

Don't worry about the alignment arrows, its not the correct way to align the 60ghz products anyway, but at 50 metres it doesn't matter. You can spin the dish 70 degrees to the side and still get over a gigabit

Don't worry about the alignment arrows, its not the correct way to align the 60ghz products anyway...

I'm curious, what is the correct way to do this?

Start with a rough alignment by eye
Log into both radios, open terminal and manually set tx-sector=36 in order to disable beamforming. And if the range is >400m the use 64800 for the frequency. And you'll want the solidmount
Then type in 'align 0' to enter alignment mode and give a real-time readout of various stats, do on both sides at same time
Run a UDP bandwidth test in both directions between the 2 radios.

Now fine tune your radio alignment, looking at RSSI values but paying more attention to the error rate %
Your main focus is on getting the error rate as consistently close to 0% with a few single digit spikes. If you're seeing over 10% then forget the RSSI and start making small tweaks until that's as close to 0% as possible
RSSI should end up approx equal on both ends as a consequence of that but again error rate is more important. Very carefully lock the adjusters down

Now disable alignment mode and you can set tx-sector=auto again. They should read as being on center or very very close to it. Don't worry too much as the beamforming may be compensating. The most.important thing is it should be as solid and reliable as it'll ever be, and beamforming can deal with small issues like the wind flexing the dish

For anything under 200m you don't need to bother with this

Start with a rough alignment by eye
Log into both radios, open terminal and manually set tx-sector=36 in order to disable beamforming. And if the range is >400m the use 64800 for the frequency. And you'll want the solidmount
Then type in 'align 0' to enter alignment mode and give a real-time readout of various stats, do on both sides at same time
Run a UDP bandwidth test in both directions between the 2 radios.

Now fine tune your radio alignment, looking at RSSI values but paying more attention to the error rate %
Your main focus is on getting the error rate as consistently close to 0% with a few single digit spikes. If you're seeing over 10% then forget the RSSI and start making small tweaks until that's as close to 0% as possible
RSSI should end up approx equal on both ends as a consequence of that but again error rate is more important. Very carefully lock the adjusters down

Now disable alignment mode and you can set tx-sector=auto again. They should read as being on center or very very close to it. Don't worry too much as the beamforming may be compensating. The most.important thing is it should be as solid and reliable as it'll ever be, and beamforming can deal with small issues like the wind flexing the dish

For anything under 200m you don't need to bother with this

nice

did not know tx-sector 36 disables BF

Good info. I searched this forum for aiming info and the Microtik setup instructions are pretty short. I'd this stuff documented somewhere?

Good info. I searched this forum for aiming info and the Microtik setup instructions are pretty short. I'd this stuff documented somewhere?

No, much to my frustration
I voiced this on this forum several times, even after figuring out exactly how to work with the 60ghz devices and posting detailed information on these forums it's still completely absent

There's still a lot of mystery around the 60ghz as evident by an extremely simplified configuration. My guess is even mikrotik themselves have NFI and the 60ghz chipsets are predominantly a white label product with a simplified driver written by a third party. But I have enough of an understanding to know how to make them work

This is a very interesting and useful discussion. I had some time this morning to drill down deeper with WinBox. Looking at these screens was somewhat enlightening. The bridge (master station) is on the left, the station-bridge (slave) on the right. I also finally found the correct search term is "alignment", not "aiming" and found this very useful post from way back when.


It would be nice to have a sector map to know how the sectors are numbered. Is there any such thing?

As you can see, the 50M range estimate was a bit off. At 30M range it's easy to see how it would be very difficult to move the unit only a few tenths of a degree with the massive backlash that exists in the mount design. Heck, just tightening the screws afterwards might move it that much. And it's also easy to see why the TX sector has to be set manually, otherwise it'll just start scanning sectors again continuously and there's no way to reliably converge the aiming like that.

There doesn't seem to be a way to set the TX sector manually in WinBox, but it would be nice if there was a way. I'm totally comfortable doing it in the CLI, just wondering if there is a way in WinBox?

Also, selecting "Align" and then pressing "Start": what exactly does that WinBox function provide over and above just looking at the Status tab?

ETA: the nRAY wireless wire kit ships with the "solidmounts", so I have those. Further research with the search term "alignment" seems to imply that anything under 0.8 degrees is good to go, and you can see above that has been achieved.

The sector map should be this one:
viewtopic.php?t=199599

To further add to my 'even mikrotik doesn't know' statement, i've very successfully used a LHG60G as a multipoint that covers a 90 degree spread with 8x other LHG 60G dishes, yet according to them it only has a 3 degree beam width and must be aligned to within 1 degree. It's just blatantly wrong but in a positive way at least. I've no doubt that the 'optimal' transmission power is within 1 degree of perfect center (and you definitely need to get close to it once you start pushing >1km and want any sort of reliability) but the beamforming capabilities are quite impressive for such a low cost device. Even with beamforming disabled it will still quite happily maintain a link 10+ degrees off center out to quite a high distance

Completely disregard the numbers shown to you in the radio, they're just flat out wrong. Try it for yourself if you want just spin the dish 60 degrees to one side and it'll say something like "3.4 degrees", it'll also quite happily say that when you're almost on center due to it fluctuating in the phase array transmission pattern (which is not perfect) and I suspect it also has dynamic power which you have no control of

This video is good as giving a simple visual example of how beamforming works https://www.youtube.com/watch?v=vtPPAnvJS6c
If you understand that, as well as the software and hardware are definitely not perfect you can both appreciate the inaccuracy in the readouts but also the absolute brilliance of whoever designed the 60ghz chipset used in these radio's, it's very simple with no real control of them but it does link up and operate where it absolutely should not. I've used other 60/80ghz products that truly do need to be within 1 degree and they are a giant pain in the ass. If a piece of fabric briefly flaps in front of the radio it dies
MikroTik 60ghz if you're within 200m pretty much just throw it up, tighten down the bolts and as long as you're kinda vaguely sorta pointing in the right direction you'll have a link. I've even stood in front of one and still maintained link for a time, this is due to the beamforming making it take a different transmission path and use reflections intentionally. You don't intentionally want to do this and you get the cleanest and strongest signal right in the center but it does work. RSSI is not a valid indicator due to dynamic power adjustment, as i've seen the same RSSI numbers for a 100m link as a 800m one. This is why error rate is a better metric, and you need to transmit as much data as possible over the link whilst checking it (thus the UDP bandwidth test). And keep the tx-sector at 36 so it's always shooting the signal forward and not steering it to the sides whilst doing your alignment

40 years in RF engineering, I'm GTG with respect beamforming theory

I did another round of fussing over the pointing with both units locked in sector (beam) 36 and it did not go any better. So, as you say, I just pointed it through the plastic aiming device and locked it down.

Here's a mystery: in WinBox, when I got to Wireless and double click on wlan60-1 to get the Interface <wlan60-1> window, on the Traffic tab it shows absolutely nothing on the bridge (master) unit, but on the station-bridge (slave) unit it shows traffic just fine.

As a side note/question/doubt (possibly indirectly confirming the "even them don't really know"), how can the aiming tool here:
https://mikrotik.com/product/nrayaim_dh1
actually work?

How are you supposed to verify the alignment of your (rifle) scope with the device?

And which kind of "random" base misalignment can one expect due to tolerances in manufacturing and/or in deformations of the plastic support?

I find the best alignment is the same as what I do with any other radio
Lay your phone flat against the top/bottom surface of the dish and use the camera to align it (use an app that gives you a center cross/mark). This'll do your left/right alignment
Then rotate your phone around 180 degrees and do it again, split the difference in where its actually pointing. This will compensate for any misalignment of the phones camera. It helps to have optical zoom or a very good sensor so you can zoom in as much as possible
Same again with it being flat against the left/right side of the dish, this'll do up/down alignment
Very simple, easy and fast. And much more accurate than a rifle scope as it has no fixed point of reference and as you say can be tilted left/right/up/down from center. Though a scope helps you with very long distance alignments - way beyond what a 60ghz is going to do anyway

This obviously assumes the dish is not malformed and the radio is indeed centered in the dish (and you actually have a flat surface to hold against), but in the real world it's going to be close enough to getting you within a bees dick of perfect alignment. At that point thats when you can do your micro adjustment and look at radio stats. But honestly it's so incredibly close 95% of the time that its not required, you're essentially playing with slop and backlash in the mounting hardware at that point

If it's high humidity, dark clouds or rainfall you've got no hope of accurately aligning 60ghz by the radio readouts, but you can at least use the camera method to get you extremely close with a visual alignment

Yep, the method you use/recommend seems very sound, simple and practical, once you know that in most cases pointing the device in the general direction of the other is enough to later be able to use the radio locked on sector 36 to do the fine tuning.
Without this piece of info one would be led to believe that a high precision (rifle scope) optical device is needed (though as said above it makes very little sense to couple a very accurate scope with a plastic support without any way to verify the alignment).
The less young of us may remember the original day of the jackal movie when the rifle scope alignment is tested shooting at the watermelon...

Here's a mystery: in WinBox, when I got to Wireless and double click on wlan60-1 to get the Interface <wlan60-1> window, on the Traffic tab it shows absolutely nothing on the bridge (master) unit, but on the station-bridge (slave) unit it shows traffic just fine.

Anyone have an idea about this?