SXT HG

RouterOS Wireless Networking
1

Tested the new HG and found it quite useful. Noticed the improved antenna. Has a L4 license and
GB/Ether. So it’s a real upgrade to the SXT Lite5 and can be used as sector for small spots further
away.

@MikroTik
What about the “special shielding” you did with the SXT SA? Not an issue with CPEs but an AP needs
shielding.

2

The licence 4 is a good thing

3

I 200% agree on the shielding! But what is the improvement you sensed in regard of the antenna? I mean, the gain is only 1dBi extra compared to an SXT 5HPnd or Lite but the tx is than actually 1dBm less.
I just bought myself 3 of these HG’s because I was expecting some improvement in directional beam. But now, after studying the documentations I found both the 5HPnD and the HG are 24º beam! In fact, all SXT’s are 24º units except the SA which is said to be 90º.

So, is the extra value (the HG costs approx 30% more than a 5HPnd, even 55%!!!) really paying off? To be honest I have serious doubts… did I buy myself a joke? :open_mouth:

I was really expecting that due the longer shape there would be better beam forming resulting in higher gains with narrower beam. But now I am wondering if the differences are really measurable. I think I have to break one open to see what really is done inside to get this ample 1dBi extra… :confused:

4

I actually found another tread with a link to an open unit; http://www.cdr.pl/p2008,mikrotik-router%20...%20tdoor.html
So, I see just 3 antenna micro prints one after the other. This creates only 1dBi extra gain? Very dissapointing.
It explains imho why the beam is the same as the other SXT’s. In fact, I wouldn’t be surprised that in a lab it might show the beam is actually wider. Due the 3 antennas in a row you create a slightly longer range for radiowaves to hit one of the microstrips.

So, unless I must obviously reading the available info completely wrong, and the antenna is really of a new design with really better caracteristics (or the given info is not correct) I think this antenna is not really worth the extra money.

Why can’t MT not just produce a 20-24dBi gain shielded narrow beam antenna with 31dBm radio? It would make it a perfect PtP or CPE antenna! Must be able for the same price as a HG and that would make it right than… :slight_smile:

I thing the are trying a bit in the right direction with the QRT 2 so maybe we will see a QRT 5 coming soon? (Same size 5Ghz antenna should give higher gain than 2,4Ghz device. So there is hope!)

5

Rudy,

In addition to the different antenna, the HG also has Gigabit and the new CPU. There isn’t another SXT with these features. You can either have SXT Lite with the new CPU but no Gigabit, or the now discontinued SXT G with the old CPU and Gigabit.

Even ignoring the antenna, this SXT is better anyway.

About the QRT, we did announce in the MUM USA that we will have both frequencies.

6

I replaced a SXT Lite with a HG and got more than this one db in improvement. It was a Near LOS shot.
So it does quite well. An Sextant with 7xx-Board performed worse. TX-Power is not interesting to me due to ETSI Restrictions. You have to consider the License cost.

Of course we get better results with 9xx-Board in a Mars 19db-Antenna which has a metal case for shielding.
But it is way more expensive.

7

hmm, well. Imho the antenna is the most important part of the device. Gb and high performance cpu is nice, but if we have a bad signal they are of no use. The mayor problem nowadays in crowded spectrum is the abiliti to sustain good to very good wireless links to try to get the high througputs over these antennas to make the better cpu and ethernet capacity worthwile…
So I really hope we will see more, and better, improvements on the antenna technology from you guys… 3 aims; 1). 20+ antenna gain. 2). narrow beam 3). shielding against unwanted signals.
If this all is done, than the cpu and gigabit becomes important. Because only than widescale high througputs are acheived by the antennas.
Now in general with SXT type of antennas I’ll bet 99% of users are not coming near to a 100Mb througput in production usage…

P.S.
I’m going to Venice, hope to see you guys in live for once! :smiley:

8

Hard to bend physics. To have more than 20dbi, bigger size is needed, so we have the QRT to give what you ask for. See you at the MUM!

9

True. But a ‘tube’ shaped antenna would give the option to have a more tubular shield around the antenna print which if it is good designed could improve the gain more than a lousy 1dBi we see now. At the same time the sheild would give better s/n ratio and lesser interferences. Just to charge 30-50% more for an antenna based only on the better cpu and gigabit makes this HG product one that will see eol very soon. Not enough sales…

10

But you already see in the posts above, that this SXT HG does perform much better. So there are other improvements that work

11

Well, by all means, and with all respect to ‘ste’, but his remark is the only that would point to some better performances. I’ll guess I have to try myself to see if the improvements in antenna pay out. Untill than the first impressions from documentations don’t impress me much… but I did pay 30€ extra for these boys…

12

Ok, today made a 2km P2P link between two SXT-G-5HPnd r2’s.

Well, I hope this is not a bad sign but the link has several problems.

  • signal is good, -47 to both ends

  • They have been ligned out very well

  • CCQ is around 20%! At best!

  • Tried to set the HT MCS value low to see if the ccq would improve. Actually found out that none of the MCS rates are used. I have 6Mb of the normal rates enabled and the antenna’s only connect to this. Do I set it to 48 or 52 it only connects to these.

  • Set the rates back to default and the best rate I see is somewhere in the 20’s.

  • I cannot alter the power output of the antenna. Not allowed; “tx-power-mode” not supported (6) error is produced… Same on client as AP.

The link is so bad I cannot even run a wireless scan on the other end. It takes a full minute for the station to come back online, thus winbox/terminal session is closed.
Running the scan on the AP does not show foreign signals nearbye. So no interference.

The client was first connected as a CPE to a mikrotik AP and worked fine. Haven tried the not supported settings here… So it looks like the AP function is the failing unit. But I can’t find any strange settings:

/interface wireless
set [ find default-name=wlan1 ] adaptive-noise-immunity=ap-and-client-mode band=5ghz-onlyn channel-width=20/40mhz-ht-below \
    country=spain default-authentication=no disabled=no frequency=5700 ht-basic-mcs=mcs-0,mcs-1 ht-supported-mcs=\
    mcs-0,mcs-1,mcs-2,mcs-3 hw-retries=15 l2mtu=2290 mode=ap-bridge nv2-cell-radius=10 nv2-preshared-key=************* \
    nv2-security=enabled radio-name=OP1-BH3 rate-set=configured scan-list=5600-5800 ssid="OP1<=>OP2" supported-rates-a/g=6Mbps \
    wireless-protocol=nv2



I hope anywone has a clue because if this the ‘better performing’ antenna, it soons ends in the bin (RMA). I am not happy! :frowning: :frowning:

13

You cant set country and antenna gain to reduce power to a sane level?
I guess the signal is to hot.
〵

14

? I’m afraid I don’t understand what you mean?

If my quest is answered by your question, you’re right. I can’t reduce antenna power. I have never experianced something like that in any router of mine…

My idea was also that the signals might be too strong.

So far I am clueless. But I’d had a break and going to throw me at it again…

15

Rudy “The link is so bad I cannot even run a wireless scan on the other end” - how are u doing that? If u do scan on the other side simple - link will disconnect!?

16

No, you can do a scan, but a short one.
Depending on the setting in the wireless config “Hw. Retries” part it takes some seconds before winbox considers a link is gone and shuts down.
So, you can disconnect the link for some seconds and still have the winbox (or telnet, or telnet in the winbox) session running, after the link is re-esthablished. Default setting is 7, but you can increase it to 15. Meaning the winbox session tries to stay aliver for 15 secons!

Open a telnet window in your winbox session to the remote (or do a telnet from the local) and type:

int wi

** this opens the wireless interfaces section **
** type:**

pr

(Which means “print”)
** Now a list of wireless interfaces pops up. *+

Look at the number of the interface for the one you do want to run the scan, on units with only one wireless this is “0”

Now, type;

scan 0 d=4

and hit enter

Now you’ll see no more activity on the winbox session and after a second or so the links towards this remote station drops…
BUT; after some (4) seconds the link comes back up and he; after another second or two the original winbox session (with its telnet window) comes back alive and see, the telnet session shows now the outcome of the scan! How marvallous is that! :smiley:


Some remarks:

  • If station has ‘default’ scan in the scan portion, the scan commands actually scans the whole band that is allowed for the country setting. So, this can be so much (special in 5Mhz bandwidth) that in 4 secs not the whole band gets scanned, or weak signals might not picked up.

  • For same reason it also can take several seconds before station finds its working frequency again to associate. This can be too long and the winbox session in your screen drops and thus you’ll not be able to see the terminal output anymore…

So another tip:

  • If you are looking for interference sources; set the scan for a 40Mhz off set to its AP frequency. For example; AP transmits at 5570Mhz. I would set the scan to 5530-5610. This means now the ‘scan’ command scans only in this part of the band. So it can scan longer and thus bigger change it also picks up weaker signals. Secondly, after the scan set time (4secs) is passed, the station tries to re-associate with the AP again. Since he only has to look in the scan set part of the band will find its working freq. much faster and thus much faster associate. Less change you might timeout the winbox session!

Sometimes it works to scan for 5 or even 6 upt 7 seconds. But that only works when the Station almost inmediately after the scan re-associates.
You can play a bit with the settings to get the best result.

I agree (and put it forward on this forum too) that a print dumb of the scan result would be very helpfull, you could also scan much longer and do the whole band for instance. But this procedure as given is a nice work around…

Good luck.

17

I meant frequency-mode=regulatory-domain country=xyz antenna-gain=17

This reduces tx-power on my test SXT-HG.

18

Found the problem.

I always used “card rates” to set different power mode. If I’d use it… probably during the days I used routerboards with radio cards in them…
But now I am using more and more the 9 series card with embedded radios. And hardly ever set an off set power mode.
But I’ll gues ROS is not considering these new unit’s radios as “card” so that’s why it not working.
Because via the “all rates fixed” option I can set different power values… :smiley:

So, you see a habit developped over years sometimes put you (me) on the wrong path with new hardware… shame on me…