Today I make some tests in good environment on those Routerboard products, to better understand how to choose the right of them.
Ap was a Basebox with 20dBi MIMO sector antenna, with power set @ 23dBm per channel (26dBm total), which is the max linear power of those devices (whithout clipping and poor modulation).
The distance was 4Km in open space and LOS.
The channel was 20Mhz and 802.11 standard with "rts cts" and threshold=1 .

Sextant G ($110.00)
-57/-57 stable.
Btest Rx UDP 90Mbps

SXT 5HPnD ($89.00)
-60/-60 or -59/-59.
Btest Rx UDP 90Mbps

SXT 5HPnD-SA r2 ($99.00)
-63/-62 stable.
Btest Rx UDP 90Mbps

SXT 5HPnD-HG ($95.00)
-59/-58 stable
Btest Rx UDP 90Mbps

So, Mikrotik, could you help me answer some obvious questions?
1) Why 2-3dB different hardware between SXT 5HPnD and SXT 5HPnD-SA r2 (and not only different license)?
2) Why SXT 5HPnD-HG (+4dB) costs 4$ less than SXT 5HPnD-SA r2 (same level 4 license)?
3) Why so different shape between SXT 5HPn (CPE) and SXT 5HPnD-HG (AP) for approximately the same antenna behavior?
4) How have I to choose between Sextant G and SXT 5HPnD-HG @ cpe side? 15 more $ (and bigger unit) for only 1 more dB (and only level 3)?
5) Why 4 different hardware to do approximately the same thing (+/-3dB), and none in the real middle between SXT and QRT5 (a 19-20dB antenna unit or so...)?
Thanks
Nicola Giampietro

P.S.: Very good units however!

interesting topic

Today I make some tests in good environment on those Routerboard products, to better understand how to choose the right of them.
Ap was a Basebox with 20dBi MIMO sector antenna, with power set @ 23dBm per channel (26dBm total), which is the max linear power of those devices (whithout clipping and poor modulation).

How have you tested that?

which ROS version and wireless driver did you use?

Sextant G ($110.00)
-57/-57 stable.
Btest Rx UDP 90Mbps

SXT 5HPnD ($89.00)
-60/-60 or -59/-59.
Btest Rx UDP 90Mbps

SXT 5HPnD-SA r2 ($99.00)
-63/-62 stable.
Btest Rx UDP 90Mbps

SXT 5HPnD-HG ($95.00)
-59/-58 stable
Btest Rx UDP 90Mbps

So, Mikrotik, could you help me answer some obvious questions?
1) Why 2-3dB different hardware between SXT 5HPnD and SXT 5HPnD-SA r2 (and not only different license)?
2) Why SXT 5HPnD-HG (+4dB) costs 4$ less than SXT 5HPnD-SA r2 (same level 4 license)?
3) Why so different shape between SXT 5HPn (CPE) and SXT 5HPnD-HG (AP) for approximately the same antenna behavior?
4) How have I to choose between Sextant G and SXT 5HPnD-HG @ cpe side? 15 more $ (and bigger unit) for only 1 more dB (and only level 3)?
5) Why 4 different hardware to do approximately the same thing (+/-3dB), and none in the real middle between SXT and QRT5 (a 19-20dB antenna unit or so...)?
Thanks
Nicola Giampietro

P.S.: Very good units however!

1) -SA is a sector (90 degree) antenna, the 5HPnD is a CPE...
3) More gain, I think not only in the antenna but the radio amplifier judging by its sensitivity.
4) In real life, bigger antennas = better signal. Also have a look at its sensitivity: for MCS7 at full power, SEXTANT has 6dB more sensitivity (quad!), and it's more directional (17 degrees vs 25)
5) wouldn't say +/-3dB is "approximately the same thing", its double or half the signal intensity. SEXTANT goes between SXT and QRT...

By real life experience, I'd use 5HPnD-HG for no more than 0-8km range; SEXTANT can perform great for 10-13km.

I was also puzzled why does the -HG exists for, and there's an updated model (AC) model coming soon, so it should have its application niche for sure, but have yet to find its application, maybe CPE also for 0-5km with more power to cut through trees, rainy weather or something.

For CPE, 0-5km distances my favourite is the SXT5 Lite, I also tested SXT5 HPnD, -HG and SEXTANT; I use SXT5 Lite, SEXTANT or QRT depending on distances.

SXT 5HPnD-SA r2 has 14dbi antenna.

Router Os for tests was 6.27.
Pukkita what do you see not correct in my results?
So, if you say that SEXTANT has 6dB more sensitivity than 5HPnD-HG, you would say that SEXTANT @ -71dBm and 5HPnD-HG @ -65dBm have same bandwidth performance?
Thanks.

Router Os for tests was 6.27.

Which driver did you use in the tests? wireless or wireless-fp?

Pukkita what do you see not correct in my results?

I didn't say they were incorrect

I was very interested in knowing which test (and equipment) did you to see the distortions and clipping, and asking to know the conditions. I'd test TCP results however, they're more revealing.

So, if you say that SEXTANT has 6dB more sensitivity than 5HPnD-HG, you would say that SEXTANT @ -71dBm and 5HPnD-HG @ -65dBm have same bandwidth performance?
Thanks.

Do you mean for that signal level? No, they won't.

What I'm saying is SEXTANT has -HG sensitivity x4, so its "hearing" capability is 4 times higher than -HG, so for a given AP signal, specially for distances > 5km SEXTANT will have much better signal.

Bandwidth capability is directly related to achievable MCS; for a Modulation Scheme to be stable, a certain signal level is needed (which is mostly the same for all these models as all have the same radio chipset), so yes, as SEXTANT will have a higher signal it will be able to "lock" in a higher MCS and thus get more bandwidth. But it will get better signal because it has higher "hearing" capabilities.

However real TCP bandwidth performance is affected by more factors than signal, and is more characteristic of each CPE implementation.

Blow some new life in this tread, never seen it before but raised the question once to MT like; Why buy a SXT-HG in comparison to a SXT-HPnD? Answer was; "Because it is just much better.....you'll notice the difference" Well, in real I don't......"
My return question would have been; But SEXTANT is also much better, better than SXT-HG, and only a few bugs more... so why is there any reason for existence of SXT-HG anyway?

This question actually stands still today.....

as SEXTANT will have a higher signal it will be able to "lock" in a higher MCS and thus get more bandwidth. But it will get better signal because it has higher "hearing" capabilities.

New question to raise;

If the absolute signal is not an issue but interferences from other sources is an possible issue, which one is better to use? SXT-HG or SEXTANT?

The link is presently made of one SXT-HG as ap (bridge mode) that accepts connection from QRT at abt 2km distance.
Signal seen from HG is -51/-51 and I use 10Mhz bandwidth.
We are upgrading the tower where the HG is located so all units are up for revision. Although this particular link runs at a set MCS10 and 11 and in general the CCQ is 100% I still see at times CCQ suddenly drop to 95-100%. I presume this is because the whole spectrum is full with signals anyway so at times some interferences occur.
I am in the option now to replace the SXT-HG for a SEXTANT and the latter will also be equipped with a metal shield.

Inmo the SEXTANT will have a slightly higher output (18dBi versus 17dBi) and a narrower beam.
My interest is more in the latter; narrower beam. So it is less susceptible in picking up unwanted signals from unwanted but close to distant QRT directions..
So I was hoping to have a higher s/n ratio which in return might give a more stable link. (The aim is 100%CCQ at all times)
But the fact the SEXTANT is actually 'hearing' better might also mean that distant signals coming from directions close to where distant QRT is located might also be 'heard better'? So the positive effect of a smaller beam might be lost against the higher level of noise entering the antenna?

Any comments on this?

WirelessRudy my original question for this thread was just the same of yours: "Why buy a SXT-HG in comparison to a SXT-HPnD?".
Infact until now the only reason tu buy SXT 5HPnD-HG for me is the L4 license in a similar-antenna device.
SXT 5HPnD-SA is L4 device, but its antenna has a -3db gain compared to SXT-HG, unexplainably because of the same device shape.
I install SXT 5HPnD-SA and SXT 5HPnD-HG, as BTS AP, to achieve different area coverages, in order to obtain a -60 to -65 dbm signal level @ CPEs.
After this, I choose the correct CPE (among SXT Lite5, SXT 5, SEXTANT G 5HPnD, QRT5 and finally Basebox + 30dbi dish), to reach the wanted -60 to -65 signal with the smallest CPE.
Each of those CPEs I appointed costs about 25-35% more than previous one (including mountig brackets), so my right choice is almost only a mathematical problem.

Speaking from pure theorization here, as I havent opened and traced it, or know real specs about this device, as "the proof is in the puddle" I simply did a PTP test changing one of the ends with different SXT available, SEXTANT and QRT.

The -HG performed worse than SEXTANT and QRT (this is to be expected).

It was a 8km link, so this kind of pushes this device, I'm not an expert by any means in antenna design, but usually you cannot cheat the laws of physics, and this device has the same diameter as a SXT (ok, its bulkier lenghtwise) but gain comparable to a SEXTANT? Hard to believe... and in fact it performed exactly the same as a regular SXT-5HPnD.

The -HG stands for High Gain, but is it really all in the Antenna gain, or has a High Gain signal amplifier?? that would explain the results, as High Gains in amplifier usually carry their own can of worms, like distortion, loss of SNR, etc.

I'm puzzled however as mikrotik is rather pragmatic, so there should be an application for this device, just cannot find it.

Would be interesting what Mikrotik has for comments. I doubt they will...

SXT 5HPnD-HG is simply the AP version of SXT 5HPn. It's better than SXT 5HPnD-SA r2 in some cases, and it's a bit cheaper (why?). SXT 5HPnD-HG shape is longer than SXT 5HPn, but I think that antenna gain is almost the same.

SXT 5HPnD-HG is simply the AP version of SXT 5HPn. It's better than SXT 5HPnD-SA r2 in some cases, and it's a bit cheaper (why?).

That makes no sense and is not true. The HG version is mend to be a P2P unit. It has a small beam, slightly smaller even than a normal SXT-5HPnD so that means its a backhaul unit, not an AP. The SXT 5HPnD-SAr2 on the other hand has a 90º wide sector which is typical for AP use....

SXT 5HPnD-HG shape is longer than SXT 5HPn, but I think that antenna gain is almost the same.

No need to 'think' the specs are telling you its 1dB better than the normal SXT.

Off course you can use any antenna as AP or CPE, the software (if higher than level 3) allows both modes. But only in very specific situations a narrow beam antenna will be used as AP (to reach only some clients in small sector) where the use of a wide beam (SA = Sector Antenna => 60 or 90 degrees sector) antenna makes no sense to use in P2P or as CPE. It CAN but is not recommended. It will pick up and spread signal from and to direction where it has no use....

WirelessRudy I simply tend to minimize the number of devices on the same base station, to minimize the number of channel required, and choose a device according to the area coverage I expect. My nework spreads over a complex mountainous territory, with big distances and limited population. I have a ratio of about 100 users spread over 20 sites and 50-60APs, so I use also SXT 5HPnD-HGs and QRT5s as AP for 3-4 users. At the same time the hyperlan use on this territory (Italy) is
rather uncontrolled, so there are lot of private illegal links and channels are very busy. When you put 2 or more hyperlan devices on the same site, you have to observe distance among them and avoid situations where device1 hears device2 with a signal stronger than -40dBm, and this isn't a simple goal. I noticed that I can achieve better results using one single sector nstreme AP on a tower, to connect two or more sites among them (source backhaul and 3-4 clients), than using two devices (sector and P2P client) close together. Obviusly if you cannot use a TDMA protocol, this situation became poor instead.

@solelunauno: Your statements in your previous post were incorrect. The way you use the available devices in the actuall situation is what I do myself. I also use high gain narrow beam P2P antenna's just to reach some distant clients (or 2-5 client + one backhaul). Sometimes you have to find inventive usage of devices and frequencies. I have in a range of 20km in circle more than 60 (!) radio's working in the 5Ghz band. Probably my network and other providers have to serve some 1000-1500 clients.
I have AP's that have to serve a client a 100meter and clients at 12km.

So yes, its a big and complicated puzzle to find a working situation. Were we were OK to use -80dB (or better) signal level but today I don't accept a client that can't reach -70 and preferrably we look for -60 or better. Backhauls have to be better than -60 or are useless to transport too much traffic.

So we use were we can shielded or metal housing antennas and radios and high gain (=narrow beam) antennas where possible.
On crowded AP's the battle is between more clients per radio (=lesser channel needed) versus more AP's that can server lesser clients (=better time ration and higher signal due smaller beam AP) but now we need more channel to reach same region.....
4 x 90 = 360 degrees => 4 x 20Mhz = 80Mhz at for instance average -60dB signal,
Or 8 x 45 = 360 degrees => 8 x 20 = 160Mhz at for instance at same clients -57dB signal.

Sometimes 4 x 90 is better, but if there are 400 clients to serve 8 x 45 might be better for the radio.

i think you both Right.
but sometimes technicians - had to make use of things that "not SO optimal for".
and sometimes budget (both both ISP and client sides)was SO small so you Had to find compromises.