Community discussions

MikroTik App
 
wirelesswaves
Member
Member
Topic Author
Posts: 311
Joined: Thu May 31, 2007 12:38 am

Unstable signal. !st Lobometrics Now Mikrotik

Fri Dec 21, 2007 3:17 am

I need some good advice. I am an RF Engineer, so sensible replies only please. I have been using Lobometrics 912N and Lobo 912 at 5.8G over a 1.5KM LOS link. When 1st installed signal levels at one end were -66 and the other -82 withh 20Mhz bandwidth. Eventually the signal decayed reported by the 912N to the point it was -91 and unreliable. I reduced the Bandwidth to 5Mhz and the circuit stabilised at -66 radio 1 and -88 radio 2.

Since then I have tried 2 further Lobo 912N's. Each with the same result.

After zero help from their supplier I now have switched to making my own and use a MT133 board with R52H card.
I replaced both ends, at first the signal levels were -46 radio 1 then -66 radio 2. After 2 days -46 and -66, after 3 days -46 and -82. If I turn up the bandwidth to 20Mhz then the signal drops to -88 and is too near the threshold of the unit.
Ping times are erratic.
The antennas were 22dB panels both ends, then 24dBi dishes both ends, then finally 22dB panel and 24dB dish.
Coaxials are about 1m long and have been changed both ends 4 times.

The radio that reports back that the signal is -83 (right now) is 1 of a dual radio. Ie this unit has 2 radio cards fitted. (the second lan points 90 degrees in other direction)

The antennas were spaced 2m apart, now 1.5m apart.

Strangely just before I ripped out the Lobometrics, I somehow managed to get the signal level reported by this radio to a sensibel lvel (-66) by changing channels and then changing back again. However if the radio was ever re-booted or any changes made which involved the radio card resetting, then the unit would go back to -66.

The datarates are set as configured (all except 54). Basic rates all unchecked except 1st rate in each row)

Radio card power set to default.
 
jcremin
Member
Member
Posts: 360
Joined: Fri May 25, 2007 7:57 am

Re: Unstable signal. !st Lobometrics Now Mikrotik

Fri Dec 21, 2007 3:45 am

Sounds like intereference to me. Especially on the radio2 side. How populated is the "wireless" spectrum in your area?
 
wirelesswaves
Member
Member
Topic Author
Posts: 311
Joined: Thu May 31, 2007 12:38 am

Re: Unstable signal. !st Lobometrics Now Mikrotik

Fri Dec 21, 2007 11:39 am

Thanks for response. I am in rural Spain and as far as I am aware I am the only operator of 5.8G. (the spanish are only just coming to terms with 2.4G)

What I didnt mention in the post is that I have a longer link of 6km, that is barely line of sight with the curve of a hill in the fresnel zone and that has been stable at -82dB for almost a year. This link is a quarter of the distance and its visable from both ends.

One transmitter is on the top of an apartment block at the furthest edge of town and the other end (1 half of the dual radio) is on a 20ft mast, 10ft up with ground sloping rapidly away infront of it.

The site at the apartment block is new and last week the transmitter was on top of a factory unit also on edge of town. The results are the same from both sites.

Also, last night before the posting I tried 5660, 5600 and 5580 Mhz, all frequencies react the same. I put the unit back into 20Mhz bandwidth and the signal went down do -91, thats too unstable with zero margin, so I quickly went back to 5Mhz bandwidth.

I remember seeing some other posts about a similar problem, but I dont think there was any satisfactory outcome.

The deteriation in signal has been by the same amounts each time, firstly -46 then -66 now -81.. Each time a drop by 20-25dB.

When I useed a Lobometrics 912/912N it was wrap2E based with atheros card MT OS, now I use MT 133 with R52H and the same things are happening.
 
bushy
Member Candidate
Member Candidate
Posts: 140
Joined: Thu Oct 20, 2005 11:56 pm
Location: Ireland

Re: Unstable signal. !st Lobometrics Now Mikrotik

Fri Dec 21, 2007 4:28 pm

If the signal was ok at one and and poor but stable at the other , it would have to be terrain/reflections from something nearby, but since the signal is fading over a few days , something is killing your radiocards.

We know that lightning nearby ( not in your case),static , powerful radio /radar nearby or a poor power supply to the board can kill them.

Which radio do you have to replace to bring the link "back to life " ?

At that end:

Is the same POE system still in use since you had the Lobometrics gear in there ?

Is there a nearby airport / radio transmitter ?

Static problems maybe ? ( do the antennas measure near zero ohms if you measure between the centre pin and outer of the connector ? )

Do a scan with a spectrum analyser at that location where teh cards are dying
 
wirelesswaves
Member
Member
Topic Author
Posts: 311
Joined: Thu May 31, 2007 12:38 am

Re: Unstable signal. !st Lobometrics Now Mikrotik

Fri Dec 21, 2007 4:55 pm

Hi, thanks for feedback. The nearest airport is 45km away and I am surrounded by Mountains anyway.

Yes I have tried both Lobo type POE's and MT, the result is same.

If it was reflections and thus 2 signals arriving at opposite phase then I would have expected that from the former site (which was 150M closer). The equipment was then mounted on a pole 20ft above a STEEL corrugated roof. But now the equipment is on a appartment block, high and in the clear.

With the naked eye you can even see the windows on the property clearly at the far end.

I am always careful to test the products before installed in the field and always use a dummy load especially when I use 350mW cards.

What is confusing is that when it was first installed it was -46 in both directions, then like I say the next day -46 (1 end) and -66 the other, then eventually ending up at -81to -83. (1end only)

Am I right in assuming that when in the window that says "registration" that the signal strength reported here, is the received signal strength of the transmission from the far end.

If that is the case it is the radio on top of the appartment block that has gone weird (tx only). This is a single radio too and not a dual unit. So there shouldnt be too much build up of heat.

Also the Lobo used a different card to the R52H.
 
bushy
Member Candidate
Member Candidate
Posts: 140
Joined: Thu Oct 20, 2005 11:56 pm
Location: Ireland

Re: Unstable signal. !st Lobometrics Now Mikrotik

Fri Dec 21, 2007 7:22 pm


Am I right in assuming that when in the window that says "registration" that the signal strength reported here, is the received signal strength of the transmission from the far end.
I know its pain but go to the easiest one to swap out a radiocard in and change it . It is the only way to know without doubt which one is actually causing problems.

Is there a very long POE run there ? Could you temporarily run a separate power lead to it ?

I think -46 is a bit high , try turning down the tx-power on both sides , you know more better than me but i think that most cards are very sad when you look at them , they work ,but really belong in laptops, same can be said for some antennas etc.

These : http://ubnt.com/products_xr5.php4 might be worth a try , were made for the job , guy from there hangs out on these forums
 
User avatar
tgrand
Long time Member
Long time Member
Posts: 671
Joined: Mon Aug 21, 2006 2:57 am
Location: Winnipeg, Manitoba, Canada

Re: Unstable signal. !st Lobometrics Now Mikrotik

Sat Dec 22, 2007 12:54 am

Problem is that radio 1 is seeing radio2 at too high power.
Thus, I beleive that your twin radios are going deaf.

Again swap out the radio cards to confirm this.

Try using dishes (less side lobing), and reduce the radio output power (~-60db is more than sufficient), and try seperating the frequencies as much as possible.

It would be nice if nstreme polling with csma disabled could impliment syncronous transmit, this way the input amp of the radios would never be over powered from other radios on the same support structure.
 
User avatar
ghmorris
Member Candidate
Member Candidate
Posts: 284
Joined: Fri May 28, 2004 12:14 pm
Location: Minden, Ontario, Canada
Contact:

Re: Unstable signal. !st Lobometrics Now Mikrotik

Sat Dec 22, 2007 3:04 am

We really like R52s and have a big bunch of them in service with great results.

Have never tried R52H, but also have never seen anyone on the forum that actually likes them.

Interesting. 8)

Has anyone here had good results with an R52H???

We really like Ubiquiti XR5s for long range work in a noisy environment. Great filtering, first class card.

The R52H is about half the price, so it would be nice to find out if it works well...

George
 
wirelesswaves
Member
Member
Topic Author
Posts: 311
Joined: Thu May 31, 2007 12:38 am

Re: Unstable signal. !st Lobometrics Now Mikrotik

Sat Dec 22, 2007 12:58 pm

OK TGRAND. There are issues in your post that I dont understand with regards Nstreme. I tried changing power on the card (in dual radio unit) from default to 17dbm but that didnt make any difference except to change the "signal level received" on the tower block radio from -46 to about -50 ish.

I suppose I should be changing the power of the Tower block unit, as the one at the dual radio site reports that the signal level received from the tower block unit is -83db.

Another words reducing the signal level received by the dual radio unit in a hope to clean up the signal quality.

The only problem there is that I could manage to cut myself off as I am at the far end of the link. It would then need a visit to the customer. Best I do this after the Xmas breaak.

So is there any published data about on what happens to these radio cards when their "front end" is overloaded. Seems to me that if it is the case that radio cards behave erratically when they receive too much signal then there are a lot of folk out there who are mistakenly trimming every last ounce of db out of a link when better advice for them might be to leave alone.

On the point of the Ubiquity, What results do folk have with the XR5 attached to a MT133 board. When I studied the board it seemed just possible for the supply rails on the MT133 to cope with the power of two R52H cards, but certainly not 3 of them.

Has anyone experience of using 3 XR5 cards with a MT133, cos it seems from the supply rail spec that it wont cope!

I am interested to know from anyone who talks the talk about modulation techniques, sensitivy, and dynamics and the heart of these cards who can explain what is going on with the card at high signal levels.

My experience is in Radio Broadcasting , transmitters and receivers at VHF and HF, I have bugger all experience at microwaves. But I guess some of the principals of overload may be the same.

The dish at the dual radio site is 24/25dB and the flat panel around 22/23dB, so the polar patterns are going to be similar, with side lobes well down in level.

The spacing in frequency of the dual radio unit is 5500 Mhz 20Mhz BW and 5600 5Mhz BW on the problem link. (formerly 5660)

Is someone suggesting that the frequencies arent spaced enough apart.
 
User avatar
ghmorris
Member Candidate
Member Candidate
Posts: 284
Joined: Fri May 28, 2004 12:14 pm
Location: Minden, Ontario, Canada
Contact:

Re: Unstable signal. !st Lobometrics Now Mikrotik

Sat Dec 22, 2007 4:34 pm

I can answer parts of your post.

XR5s will tolerate a receive signal in the -30s without obvious performance degredation, but its not good practice. -45 to -65 is a better range to be in.

Adjacent channel interference is one of the 800 pound gorillas in license-exempt. You simply can't get really good filtering on a mini-PCI card, its physically too small.

If you do a scan for unwanted signals from your installed radio, you should try to ensure that you see nothing higher than a -75. If you can keep to this rule, then you won't have any problems filtering unwanted signals.

If you have another card right next to you that you are seeing at -35 on a scan, there is a really good chance you are going to get some receive desensitization issues as a result.

If we use dual polarity antennas for example, we alway put up two separate boxes to get the radio cards as far from each other as possible. We only have about 30dB of seperation at the antenna feed horn, so we don't want to make that worse by leaving any other opportunity for cross-talk.

A 133 will happily take an XR plus another lower powered radio. We tend not to jam the max number of radios into any one enclosure these days, and since we've stopped doing that our performance has improved quite dramatically. If we do run multiple radios, they are usually on different frequency bands.

The price of a 133 now relative to the price of an XR5 means you can run one radio per box and still have a very ecomomical system.

George
 
wirelesswaves
Member
Member
Topic Author
Posts: 311
Joined: Thu May 31, 2007 12:38 am

Re: Unstable signal. !st Lobometrics Now Mikrotik

Sat Dec 22, 2007 6:18 pm

Thanks George, thats really good advice and is making sense. The common point in this scenario that has been relaced now 4 times is the dual radio.. After xmas I will knock up another and put in seperate enclosure. Aluminium this time and not polycarb. Should afford better shielding. Seems that 2.4/5.8 is ok but as you say, no 2 radios on the same band in the same box..

Perhaps I'll use the xr5 on the longer link (6.7km) and the r52h on the shorter


Thanks for advice, I knew there would be someone out there with experience of this issue.




Merry xmas
 
User avatar
ghmorris
Member Candidate
Member Candidate
Posts: 284
Joined: Fri May 28, 2004 12:14 pm
Location: Minden, Ontario, Canada
Contact:

Re: Unstable signal. !st Lobometrics Now Mikrotik

Sat Dec 22, 2007 9:15 pm

We use the PacWireless outdoor cases with good results. Made of aluminum. We buy the hinged version these days, easy to work on.

Please let me know how you get on with the R52H. We are using a ton of R52s with excellent results, but haven't used the high output version yet. I really like the XR5s.

Merry Christmas!!

George
 
bushy
Member Candidate
Member Candidate
Posts: 140
Joined: Thu Oct 20, 2005 11:56 pm
Location: Ireland

Re: Unstable signal. !st Lobometrics Now Mikrotik

Sun Dec 23, 2007 4:33 pm


The only problem there is that I could manage to cut myself off as I am at the far end of the link. It would then need a visit to the customer.
Write a small script to turn the power back up after say 5 minutes, then start it and turn down the TX power, if the link breaks , the script will save the day.

The easiest way i've found for boxes outdoors is to put the board into one of these type boxes :

http://www.aerial.net/shop/product_info ... cts_id=137
( they come with screws,standoffs etc)

then put that into an ordinary outdoor plastic enclosure.
Run the cables in through glands so the Nconnectors are INSIDE the box ( saves messing with waterproof tape)

If the radio dies , make up a spare at home and go out and swap out the metal box complete. No messing with small screws up on towers
, no danger of breaking pigtails

Who is online

Users browsing this forum: benavidezrudy and 24 guests