[HamWAN PSDR] Why HamWAN is using MikroTik/RouterBoard devices

Bart Kus me at bartk.us
Fri Feb 8 10:59:57 PST 2013 

Alright, changing subject to address this one.  This really should be on 
the Radio Modem 
<https://www.hamwan.org/t/tiki-index.php?page=Radio+Modem&structure=HamWAN> 
component engineering page.  Shame on me.

The big problem with doing a wide area network is PtMP distribution.  
802.11a/b/g/n/ac is inadequate for the task since it relies on 
carrier-sense multiple-access (CSMA) multiplexing. You end up in a 
situation of nodes transmitting over top of each other and garbling 
packets at the remote receive site they're both trying to reach (a 
mountain cell site).  A good solution for this is to use time-division 
multiple-access (TDMA), where there is one time-slot controller node and 
all others only speak when they're instructed by the controller.  Think 
of this as your typical voice radio "net control" human being.

There are several systems which implement TDMA, the most open of which 
is the WiMax 802.16 standard.  In a perfect world HamWAN would be using 
WiMax and would have manufacturer independence. The problem with WiMax 
is that the equipment manufacturers always targeted telcos.  The 
equipment is incredibly expensive.

This expense provided a business opportunity for smaller manufacturers 
and gave rise to alternative TDMA implementations. Ubiquiti has their 
own, and MikroTik has their own.  There are others, but those are the 
prominent players.  So in the real world, we can't afford WiMax, and if 
the goal is to actually get something on the air and make it real, we've 
gotta settle on one of these alternative implementations.

Having studied both product lines, and having used both products over 
the years, I've come to the conclusion that MikroTik offers the better 
software story.  The capabilities of Ubiquiti devices are far behind 
what you can accomplish with a MikroTik device. Also, from what I've 
seen of the behavior of both companies over the years, MikroTik takes a 
far more serious and carrier-grade approach to things.  They're trying 
very hard to be Cisco, and that's a good thing.

Now I realize I have not mentioned specifics here about what makes the 
software better, but that would take pages of listing of features.  Best 
thing you can do if you're curious is get a couple 5SHPn 
<http://www.wifi-stock.com/details/metal_5shpn.html> devices and explore 
them. WinBox <http://download2.mikrotik.com/winbox.exe> makes this a 
pleasant adventure.  When you're done exploring you can use them to link 
up to HamWAN.  :)  I will mention one very important feature for HamWAN: 
MPLS-TE support.  This allows link bandwidth sharing across multiple 
nodes, not just on an aggregated link. It'll let the network scale 
gracefully.

If you think I'm missing something, please let me know.  Modem selection 
is a difficult decision with many variables, and I hope the decision is 
good.

--Bart


On 2/8/2013 10:12 AM, steve monsey wrote:
> I was thinking of the Ubiquiti system as a whole.
>
> Steve
>
> On Feb 8, 2013, at 9:59 AM, Bart Kus <me at bartk.us> wrote:
>
>> I'd like to identify a set of HamWAN-approved antennas we can use.  Ones which meet HamWAN's requirements and pass testing.  Any antenna which performs up to the requirements can be accepted.  So far, I have not found 1.  So we've got a few more candidates on the way.  The Laird is not quite out of the running yet, it needs to be outdoor-tested to measure the true backside performance.  The Laird is, however, incredibly expensive @ $180/ea.  And it has poor frequency response at the high end.  Unlikely to be selected.
>>
>> For the time being 5GHz is it.  There's plenty of spectrum to grow in there, and when that's all exhausted we can expand to other bands like 3GHz or 10GHz.  But I expect 5GHz is here for quite a while, a few years at least.
>>
>> Now, you mentioned Ubiquiti, so if your question was about modems, I'll have to reply in more detail.  :)
>>
>> --Bart
>>
>>
>> On 2/8/2013 8:08 AM, steve monsey wrote:
>>> Just curious, seems like you just want to settle on one manufacture ?
>>> Like Ubiquiti?
>>>
>>> Are you just using 5 ghz or 3 ghz as well?
>>>
>>> I agree that not making a line of equipment is the best idea for
>>> something that will last and replaceable. Sometimes you have to do
>>> one-off's for a special installation. Plus you can maybe get a price
>>> break for buying in bulk for the first go around.
>>>
>>> Steve N0FPF
>>>
>>>
>>>
>>> On Feb 7, 2013, at 10:21 PM, Bart Kus <me at bartk.us> wrote:
>>>
>>>> Attached are 2 pix of the ARC and 1 pic of the Laird sectors.  Excuse
>>>> the mess, these were taken before I re-organized and cleaned up the lab.  :)
>>>>
>>>> Excellent point about parasitic reflections.  I had not considered that.
>>>>
>>>> And yeah, I'd love to stick with off-the-shelf.  Really don't wanna get
>>>> into the business of manufacturing antennas.  Wanna build a network!
>>>>
>>>> --Bart
>>>>
>>>>
>>>> On 2/6/2013 10:33 PM, Rob Salsgiver wrote:
>>>>> Bart,
>>>>>
>>>>> Can you send me a photo or two of the internals of one of these sector
>>>>> antennas?  I've done some work with 2.4 but not 5.8. I'm curious to
>>>>> know if they are flat panels or if they are contoured (rounded or
>>>>> otherwise shaped).
>>>>>
>>>>> I'm thinking along the lines of your sheet metal shields. Depending on
>>>>> the internal design, I'm wondering if we can't apply some different
>>>>> approaches here.
>>>>>
>>>>> The potential downside of a sheet metal shield is parasitic
>>>>> reflections, unless they are spaced such that they enhance the signal
>>>>> (ala a reflector) as well as shield from adjacent RF.
>>>>>
>>>>> Using it as a reflector then gets me thinking of applying some other
>>>>> geometries -- how about a yagi or log periodic design for 5.8?  Given
>>>>> the frequencies involved, it likely would fit into a similar footprint.
>>>>>
>>>>> I need to do some calcs and brainstorming.  There's definitely
>>>>> advantages to having off the shelf equipment, and if possible it would
>>>>> be good to launch with standard items if they work, but there may be
>>>>> longer-term advantages to using some of the benefits of our "amateur"
>>>>> roots if we can get them to be stable and reliable.
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Rob
>>>>>
>>>>> *From:*PSDR [mailto:psdr-bounces at hamwan.org] *On Behalf Of *Bart Kus
>>>>> *Sent:* Wednesday, February 06, 2013 9:45 PM
>>>>> *To:* psdr at hamwan.org
>>>>> *Subject:* [HamWAN PSDR] Improved radiation pattern software and new
>>>>> antennas on the way
>>>>>
>>>>> Hello,
>>>>>
>>>>> I've taken some time tonight to improve the radiation pattern
>>>>> measurement software.  The algorithm for auto-centering the rotation
>>>>> of the pattern won't be fooled now by the presence of a non-centered
>>>>> peak lobe.  Full description of the software function and the software
>>>>> itself have both been published to the Antenna Analysis
>>>>> <https://www.hamwan.org/t/tiki-index.php?page=Antenna+Analysis&structure=HamWAN>
>>>>> labs page.  The radiation pattern for the Laid sector
>>>>> <https://www.hamwan.org/t/tiki-index.php?page=Laird+SAH58-120-16-WB&structure=HamWAN>
>>>>> has been updated and you can see the wonky alignment is now gone!
>>>>>
>>>>> Since neither of these antennas are thrilling performers in the HamWAN
>>>>> Labs tests, a couple more candidates have been added to the Sector
>>>>> Antenna - 120deg
>>>>> <https://www.hamwan.org/t/tiki-index.php?page=Sector+Antenna+-+120deg&structure=HamWAN>
>>>>> component engineering page.  A HamWAN community member who wishes to
>>>>> remain anonymous has generously purchased sample antennas of the
>>>>> Teletronics and RF Elements brands for lab testing.  They should
>>>>> arrive some time next week.
>>>>>
>>>>> In the meantime, I've had a thought about how to improve F/B ratios in
>>>>> these sector antennas.  Using a sheet of aluminum and my sheet metal
>>>>> brake <http://www.harborfreight.com/30-inch-bending-brake-67240.html>,
>>>>> I should be able to manufacture traditional RF shields, like this:
>>>>>
>>>>> VARIA 14-19dBi variable sector antenna 45-120°, 2.4-2.5GHz
>>>>>
>>>>> but without fancy holes or adjustable features.  It might be enough to
>>>>> keep that nasty high power adjacent RF out.
>>>>>
>>>>> --Bart
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> PSDR mailing list
>>>>> PSDR at hamwan.org
>>>>> http://mail.hamwan.org/mailman/listinfo/psdr_hamwan.org
>>>> <mime-attachment>
>>>> <LairdSector1.jpg>
>>>> <ARCSector2.jpg>
>>>> <ARCSector1.jpg>
>>>> <mime-attachment>
>>
>> _______________________________________________
>> PSDR mailing list
>> PSDR at hamwan.org
>> http://mail.hamwan.org/mailman/listinfo/psdr_hamwan.org
> _______________________________________________
> PSDR mailing list
> PSDR at hamwan.org
> http://mail.hamwan.org/mailman/listinfo/psdr_hamwan.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.hamwan.net/pipermail/psdr/attachments/20130208/09742ed7/attachment.html>

More information about the PSDR mailing list