Re: [DMV-HSMM] Re: Fwd: Baltimore Radio ATV Society meeting last night

[David Bern]

Friends:

The 50 mile long range test at Shenandoah National Park was a publicity stunt to see if it could be done.  Chris, KB3CS pointed out that we were outside the operational parameters of the Atheros chip and he is actually surprised that it worked at all.  The test was with the Xagyl miniPCI cards.  We ran out of time to do a test with the Doodle Lab cards.

More realistic numbers are found on page 25 and page 26 of my presentation at 2012 DCC Experimenting with High-Speed Wireless Networking in the 420 MHz Band

     http://www.tapr.org/pdf/DCC2012-High-SpeedWirelessNetworking_W2LNX_slides.pdf

The distance for these two tests were about 13 miles.  The Doodle Lab cards consistently gave us better throughput.  What I would like to find out is how the Doodle Labs and Xagyl cards work at 25 miles apart.  Note that all these tests used TCP/IP and not UDP/IP.

You can see the video of the presentation at

     http://arvideonews.com/hrn/HRN_Episode_0041.html

Thank you,
David, W2LNX


On 04/17/2014 07:26 AM, LLEA...@aol.com wrote:

 

Jesse,

Regarding:

"Although we were able to associate and get an IP address, we couldn't get any traffic across - there was too much noise, interference, and collisions on 2.4."

http://www.youtube.com/watch?v=j2BMLoJpeAA

Actually that may not be the case. At 33:00 in the presentation given by David Bern, W2LNX at the 2012 ARRL/TAPR DCC, Atlanta notes: ACK [Frame] Timing! But he was 50 miles (80,400m) apart, 2.3 miles (3700m) should be within the working parameters of most default dettings.
http://www.dd-wrt.com/wiki/index.php/WebInterfaceWirelessBasicSettings#Sensitivity_Range
Not all devices allow you to set Acknowledgement Frame Timing, but I'd surmise that could be the issue. We'll defnately have to account for it in links >=5 mi.

-KB3VWG

---In DMV-...@yahoogroups.com, <wb2ifs@...> wrote :

Great information from the BARTS group!

---------- Forwarded message ----------
From: Kevin Sherwood <kshrwood@...>
Date: Wed, Apr 16, 2014 at 9:55 AM
Subject: RE: Baltimore Radio ATV Society meeting last night
To: jesse Alexander <wb2ifs@...>

To give you a little background, we started playing with WRT54Gs and mesh in 2004.  When we started looking at network for ham radio, we thought 2.4 might work better because of path loss, so we got out some 2.4 GHz gear and a pair of Yagis and went to our sites 2.3 miles apart with a clear line of sight between them.

Although we were able to associate and get an IP address, we couldn't get any traffic across - there was too much noise, interference, and collisions on 2.4.

We talked with the Stateline group in PA and made a small investment in a pair of Ubiquiti 25 dB dishes (about $85 a piece) and were able to make a 300 mbps link across that path with no trouble.

Aside from that, though, mesh means every additional hop cuts your throughput in half and doubles your latency, and with ad hoc mode collisions are frequent. 

I'm attaching a few pages from the presentation on mesh vs a point to point network.  The important pages are that mesh provides 20% packet loss and 500 kbps throughput across two hops in a small area, whereas point to point backbone/point to multipoint sectors provides full bandwidth (they use 10 MHz channels in Seattle for 5-10 mbps, Stateline in PA uses 40 MHz channels and gets true 95 MBps throughput on its links) with low latency - less than 10 ms from end to end.

Obviously you've made the investment in 2.4 so you'll want to experiment with it, but I don't think you'll even see the access point from 5 miles away with omnidirectional antennas.  

Kevin

---

Date: Wed, 16 Apr 2014 08:32:08 -0400
Subject: Re: Baltimore Radio ATV Society meeting last night
From: wb2ifs@...
To: kshrwood@...

We're just getting started at 2.5GHz. Still in the "chalk board" phase.

A few of us have installed hsmm-mesh on our routers  and we're experimenting with mesh protocol, antennas, and powering ideas.

Lynwood Leech, KB3VWG, has acquired a swath of IP addresses from AMPRNet , and is experimenting on using 440MHz. As a IT guy working for the county, he's in communications about acquiring county sites and towers.

We're also working with Jim, WI3N, about sites and uses.

I've created a proof-of-concept cloud system described here: https://sites.google.com/site/ecsmesh/ My objective is to reach MedStar Southern Maryland Hospital Center from my house--about 5 miles away. I'm interested in the Ubiquity equipment and hacking WRT54Gs.

I'm also playing around with dd-wrt, and I was able to create a "part 15" mesh--also at 2.5GHz.

-Jesse

On Tue, Apr 15, 2014 at 6:15 PM, Kevin Sherwood <kshrwood@...> wrote:

Hi Jesse,

We're working on reformatting it.  Most of the presentation itself is just pretty pictures, the notes on what we said are the important parts.

Can you tell me where you guys are in the process?  I know that's a very broad question, but do you have any sites picked out?  Have you tried any gear?  Do you have any planned users/uses?

Depending on where you are I can tailor my information for you better.

Kevin
KB3PLX

President, BRATS

> To: brats@...
> Subject: Baltimore Radio ATV Society meeting last night
> Date: Tue, 15 Apr 2014 13:02:34 +0000
> From: wb2ifs@...
>
> From: Jesse Alexander WB2IFS/3 <wb2ifs@...>
> Subject: Baltimore Radio ATV Society meeting last night
>
> Message Body:
> Sorry I missed this meeting. Were there any notes or slides? I'm interested in the BRATS network because we're attempting a similar thing in Prince George's. I'd apprieciate any and all information on the network.
> -Jesse
>
> TNX ES 73s DE WB2IFS/3
>
> --
> This mail is sent via contact form on Baltimore Radio Amateur Television Society http://bratsatv.org
>

--
Jesse Alexander, WB2IFS/3
Prince George's County ARES/RACES
http://www.pgares.org/
@pgcares
wb2ifs@...
Cell:201-247-0831

--
Jesse Alexander, WB2IFS/3
Prince George's County ARES/RACES
http://www.pgares.org/
@pgcares
wb2ifs@...
Cell:201-247-0831

__._,_.___

Reply via web post

Reply to sender

Reply to group

Start a New Topic

Messages in this topic (2)

Visit Your Group

• Privacy • Unsubscribe • Terms of Use

.

__,_._,___

[KB3PLX,Kevin]

The point was, using commercial 2.4 GHz gear (Ubiquiti bullets) we had a strong signal at 2.3 miles but were not able to do more than associate.  Using 5 GHz gear we get 100 mbps easily. 

Yes, across mountainpeaks, where there is no noise, cobbled together 420 MHz gear can work over 50 miles for 500 kbps.

In real world environments, true commercial off the shelf 5 GHz gear is working today for links of 50 miles, giving 10-100 mbps speeds.  This is with a $150 dish and $80 radio, designed for mounting outdoors and continuous use in a commercial environment.  The Xagyl gear costs more than that for much less performance.

http://bratsatv.org/wp-content/uploads/2014/04/Mesh-vs-PtP.png

These are maps taken from a mesh group and a 5 GHz group.  The Austin group has 17 nodes within 2.8 miles.  

The HamWAN group has 19 nodes within 100 miles.  HamWAN uses 1/4 width channels and they get 5-10 mbps sustained through their links, pretty much from any two endpoints.  Latency is also low - on the order of 10 ms between any two endpoints.

Mesh cuts throughput with every hop, and doubles latency, plus with collisions packet loss is huge.

You can see current numbers from HamWAN by going to www.hamwan.org and clicking around on the map.  The sectors are represented by the arcs, and the dots represent nodes connected to the sectors.

If Baltimore-DC goes with this approach, this is the area we could cover with 4 sector sites at a reasonable height to a 16" dish mounted on the roof of a house:

http://bratsatv.org/wp-content/uploads/2014/04/Regional-coverage.png

We're deploying sectors now and should be able to provide this coverage within a month:

http://bratsatv.org/wp-content/uploads/2014/04/House-coverage.png

This is modelling an $80, 16" dish (more discreet than most TV dishes) connecting to a sector antenna on one of our sites.  The modelling criteria shows areas where 25 dB dish mounted 8m (26', not unreasonable for a roof) can get a -70 dB signal into our 16 dB sectors.  We have a link going through a hill at -86 dB where we still get 26 Mbps actual transfer speed, not link speed, so the green regions should provide coverage even if there are trees not modelled accurately.

[David Bern]

Friends:

What will mostly likely emerge as we build out the network is that 3.4 GHz and 5.7 GHz microwave equipment will be used for dedicated point-to-point backbone links running at many megabit/s broadband speeds.  End users will probably use 70 cm data narrowband radios at speeds anywhere from 9.6 kbits/s to 56 kilobits to connect to the backbone network.  70 cm is an attractive band since it much more forgiving in terms of line of sight and higher power is possible.  An example of a data radio is the UDRX-440

http://nwdigitalradio.com/

Also, there are many older VHF/UHF radios with data ports in the back for connecting older 9600 bps TNCs such as a KPC-9612+ Packet Communicator

http://www.kantronics.com/products/kpc9612.html

What can we do with that?  End user user stations will come and go and they should easily access the backbone network.  The end user station will access the network via a end user network to backbone network access point router.  This access point router will be typically located at a high point not unlike today's FM repeaters.  End user stations unable to directly access the access point should be able to access it via a peer end user station.  This suggests that the end user network topology will probably be some sort of ad-hoc network such a mesh topology.  Anyway, I think it is important to focus on the end user network in addition to building the backbone since, in the long run, an high speed broadband backbone without end users will not be successful.

I am going to Dayton in May and the first booth I plan to visit is NW Digital Radio at booth 515.  I want to know when they are shipping their UDRX-440 data radio they had announced two years at Dayton.  Who is going to Dayton?

Thank you,
David, W2LNX

On 04/19/2014 08:32 AM, KB3PLX,Kevin wrote:

[KB3PLX,Kevin]

Why should end users be limited to 56 kbps for $400?  Again, this is the area we expect to be able to provide 15+ mbps to a house - an end user - with an $85 dish within a few weeks using just three sites:

http://bratsatv.org/wp-content/uploads/2014/04/House-coverage.png

We're setting up sectors to encourage experimentation.  Providing the sector coverage I listed above and dishes for three houses will cost less than a pair of UDRx-440s, and provide high bandwidth, low latency connections.  

A single 56 kbps radio at a high site like a repeater would be worse than APRS, b/c APRS has a network of fill-in digipeaters that are able to hear units nearby.  Once more than one end user gets attached to the same UDRx at a repeater site though, the channel would be unusable.  

Kevin

[David Bern]

Kevin:

I my previous comment, the assumption I made is that line of sight is needed for microwaves.  So the question I have is how well at a end user's location does a 3.4 MHz or 5.7 MHz microwave dish work with a building obstructing the path to an access point on on the backbone network at a high site?  Maybe with commercial gear such as Ubiquiti airMAX 5.7 GHz equipment this is not so much an issue.  I don't know.  We need to find out by experimenting.   Please announce when you are deployed so we can come to Baltimore to experiment.  What equipment should we get?  The Ubiquiti NanoBridge M5?  Which version?  The NanoBridge M5 dishes are lightweight and can be mounted on a portable tripod and pole.

I worry about the end user line of sight issue at 5.7 GHz so I look at 70 cm.  What can we do with the Doodle Labs 420 MHz DL435-30 1/2 watt data radio cards?  How can we attach a broadband 70 cm amplifier to it on a  ham radio budget?  The UDRX-440, even though is narrowband, it is 25 watts.  How poorly does it really work with several UDRX-440s accessing the same access point using TCP/IP?  What can we do with UDP/IP?  multicast UDP/IP?  Only one UDRX-440 is transmitting while other UDRX-440s are listening.  For example, multicasting ham Web pages in non-real time can be quite useful.  It may be that the UDRX-440 is not practical.  I want to find out by experimenting. 

Thank you,
David, W2LNX

On 04/20/2014 10:37 AM, KB3PLX,Kevin wrote:

[Kevin Sherwood]

The UDRx-440 doesn't support TCP/IP, but AX25.  Latency for TCP/IP across that would probably be measured in seconds.

The bigger question to ask is what do we gain by encouraging houses to get on the air?  I'm all for experimentation, as that's how we'll get the next generation of applications developed, but how far do we go to support experimentation?

If we setup a backbone that clubs can use to link repeaters and EOCs can use to share data, and sectors that houses in good locations can use to reach in. we have immediate use, and room for experimentation.  Total cost for the radios to provide that coverage for this area is $2600:

http://bratsatv.org/wp-content/uploads/2014/04/Regional-coverage.png

That would allow a substantial portion of central Maryland hams to access a high speed network for $85.  Not that I expect many to do so, but at least they'd have the option, and we'd have built a network that links clubs and repeaters today.

Adding UDRs (which currently only support 9600, and aren't yet released) adds $1600 to that $2600 for a single 56k channel that doesn't support TCP/IP.  

"Dennis Rosenauer built a complete 56 kbps modem and RF deck back in the 80s. He had a 70 mile RF link fully operational for many years, but eventually gave it up because there wasn’t anyone to play with."

If you want to pay $800 for a pair of 56k  packet modems that don't offer much more than the current APRS backbone to see what kind of coverage you get, have a blast, but 56k packet is 1980s technology and a huge step backwards.  

I agree we need something between packet and wifi for endpoint access, but it isn't 56k packet for $400/radio.  If their 170 kbps 100 KHz channel device that will support TCP/IP ever comes to fruition, that would be worth investigating.  But the UDRx, while fascinating, is designed for a specific use, and this isn't it.

---

Date: Sun, 20 Apr 2014 11:58:31 -0400
From: David...@Engineer.com
To: boar...@googlegroups.com
CC: DMV-...@yahoogroups.com
Subject: Re: [DMV-HSMM] Re: Fwd: Baltimore Radio ATV Society meeting last night

[Keith]

Kevin,

One reason might be that you have voiced a reluctance to let other use then BRATS network.
When I asked you about connecting other networks to the BRATS network, you said you were only interested in connecting repeaters or D-Star.

I hope i misunderstood you, would your please clarify you position on connecting individual  Radio Amateur locations and club station to the BRATS network.

Will you allow regular Amateur Radio compatible Digital IP tragic over the BRATS network or if not what are your restrictions?

Keith

--
 
The next generation of communications technology in the Amateur Radio service!
 
Broadband Over Amateur Radio Networks
 
---
You received this message because you are subscribed to the Google Groups "BOAR-Net" group.
To unsubscribe from this group and stop receiving emails from it, send an email to boar-net+u...@googlegroups.com.
To post to this group, send email to boar...@googlegroups.com.
Visit this group at http://groups.google.com/group/boar-net.
For more options, visit https://groups.google.com/d/optout.

[Chris KB3CS]

On Sunday, April 20, 2014 12:20:23 PM UTC-4, KB3PLX,Kevin wrote:

The UDRx-440 doesn't support TCP/IP, but AX25.

[...]

Incorrect.  The first deliverable (firmware) will "support AX.25" but that is not all it can do. With reference to their published specs and plans, the initial capability will be a 25kHz IF bandwidth and "9600 MSK and 4800 GMSK Modems". A later capability will be a 100kHz IF and "4FSK or QPSK Modulation (2bits/symbol) up to 170kbps, Forward Error Correction (FEC), Automatic Repeat reQuest (ARQ), IPv6 Friendly, Support for TCP, UDP and RTP, Authentication, Designed to work with Ad-Hoc Networking. In addition we will scale this down to 42.5kbps in a 25kHz Channel."  The 170kbps capability is contingent on successful petition before the FCC to remove the "speed limit" from 70cm operation in favor of a simple 100kHz maximum occupied spectrum requirement.

You pipe your bitstream into the modem and you then "support" whatever data communication you like.  They specifically say they have no intention of implementing "legacy AX.25 56kbaud" (my paraphrasing).

[KB3PLX,Kevin]

Our restrictions are that amateur traffic won't be using our internet gateways and will be segregated from our protected network.  Basically we'll be using QoS to prioritize our repeater traffic, and not allowing hams access to our repeater controllers.  Since repeater traffic is < 50 Kbps, and the network is > 50 Mbps,that shouldn't be an issue.

My point was that our priority will be to establish dedicated links to clubs with traffic to send - clubs with repeaters able to be linked across IP is a good place to start.  We'll work with clubs with known uses for IP links to setup dishes to establish these links.  

We're deploying sectors to encourage experimentation among individual hams, but we have no idea what regular ham users will do with it.  So far we have two sectors deployed.  Our experimentation with the network includes additional receive sites for our ATV repeater, ham software development, and more, but again that will be isolated from someone accessing the public network.  Hams wishing to work with us on these projects are welcome to, of course.

If there are other clubs that want to link in but don't have a use case for it right now, we can evaluate if they can link through a sector or if it's worth setting up a dish.  All of our sites already have 3-4 dishes and sectors on them or planned, so we need to be careful with density.

The BRATS are not running a Wireless ISP - nor should any amateur groups, so don't expect to use this for regular internet traffic.  If someone does have a reason to have a device on the network accessible via the internet, we can certainly work with that ham to provide access through our gateways, but that would be like a remote weather station or webcam with no other way to get the data out.

So again, two sectors are deployed now providing the northern and eastern regions, two more and we'll have coverage for ham households in this area.

http://bratsatv.org/wp-content/uploads/2014/04/House-coverage.png

Kevin

[KB3PLX,Kevin]

I misread, I thought the 170 kbps link was for a new design and not a software update for the existing plans.

That doesn't change the fact that this product doesn't yet exist and still hopes to ship with 9600 bps packet.  170 kbps TCP/IP is a planned software development.

I hope they succeed, as it will be neat to play with, but performance will still be awful as soon as you get more than two devices on the same channel.  

[David Bern]

The UDRX-440 is an interesting device since it has an embedded Linux system, a UHF transceiver and a 25 W brick amplifier all for about $400.  It is narrowband compared to 820.11 stuff we've been playing with.  Right now, there is only the KPC-9612 PLUS 9600 bps TNC ($399) that is available off the self as far as I know.  So, I think, the UDRX-440 worth evaluating it.  Anyway, I look forward going to Dayton next month and find out what the story is.

David, W2LNX

[Keith]

Just thought I update you periodically on our progress.

We made several links with Frederick from Leesburg, Virginia today.
39.166216, -77.588928

We did a 28 mile link between the Leesburg Tennis and WN3R, 39.125207,
-77.567111

The signal strength ranged from 73 to 85.

Testing to the Leesburg hospital is next.

http://www.MAIPN.com

It is too early to draw any definite conclusions but I am willing to say
that a 28 mile link is pushing the limits for
a backbone link unless it is between to perfect high points. I'd think
that about 12 miles is a good distance under average conditions.
It seems you want a signal around 78 dbi or better for an on/off ramp
and a signal in the 60-70 dbi range for the backbone.

Keith