You are here

Finding a way to improve AREDN hardware

13 posts / 0 new
Last post
KD9MOQ's picture
Finding a way to improve AREDN hardware

Hi, everyone. I'm a senior electrical engineering undergraduate at Case Western Reserve University. I'm doing a senior design project with two other electrical engineering students. We have ~14 weeks to take an idea from start to finish and we would like to work on wireless mesh nets. We are looking for issues the community faces to turn into a problem statement. Our best idea so far is to make a router/AP built around an OpenWRT compatible SoC specifically for AREDN. Any input on the problems you face with your AREDN node hardware that we can try to solve with a custom box (or something else entirely) is appreciated.


w6bi's picture
10 GHz

Don't know how much RF engineering you're willing to go into, but we really could use a transverter to move the output of a 5 GHz node (Ubiquiti Rocket M5 or Mikrotik BaseBox 5) up to the 10 GHz ham band.  We'd use two of them, one for each of the two transceivers in the nodes.

Contact me for more info if you're interested in pursuing this.


Orv W6BI

K9CQB's picture
I concur. We need 2x2 MIMO in 10ghz Ham band.

I agree with Orv. Due to the changing landscape of FCC/Telecoms squeezing amateur radio. We could really use a board/box that converts a 2x2 MIMO AREDN node (2.4 or 3.4 or 5.9 GHz) and convert one of those devices to the 3cm ham band (10.0-10.5GHz). This is a pretty big challenge because both chains of the 2x2 MIMO are receive and transmit (TDD, so not RX-TX at the same time). This would solve a growing problem in our community to have an affordable solution for our Point to Point (PtP) links.

-Damon K9CQB

KD9MOQ's picture
Thank you both for the input.

Thank you both for the input.
What is the benefit to operating on the 10.0-10.5 GHz HAM band instead of the 2.4, 3.4, and 5.9 GHz bands? Is it simply underutilized?
There are number of 10GHz transverter designs out there already such as this and this. If I understand the needs of an AREDN node correctly the issues with these designs are the size, lack of compatibility with PoE, probably the price, and lack of a second channel. Does that sound about right?

-Jason KD9MOQ

K9CQB's picture
These transverters are for >10W PTT radios not <1W MIMO radios

That's one of the issues. All of these ham 10GHz transverters are for fairly high power (multi-Watt), narrow bandwidth (15-25KHz), single antenna (SISO) ham radios to usually communicate voice from 50MHz, 144MHz, or 440MHz, etc. A similar system would work if doubled and combined, but you'd have to keep in mind that our AREDN Mesh nodes are 2x2 MIMO, low power (1W or less), wide bandwidth (5-20MHz), and may be easier to LO/mix if the frequency is closer to 10GHz, but much harder to get any TX gain.
A great way to get RX gain may be to use readily available 10GHz LNB (for satellite stuff) to drop it to 1100MHz then upconvert to match 2.4 or 5.9GHz, but that may complicate the directional dish's feed. It would be like having a 2x2 MIMO TX feed-point and 2 LNBs clamped together (one for horizontal polarization chain, one for vertical polarization chain) making a crowded mess at the feed-point of the dish.
I see the biggest benefit of using 10GHz in our area is that it would replace 3.4GHz links (not many), but our 5.9GHz links are getting crowded out here in Northern Virginia.

-Damon K9CQB 

We1btv's picture
3cm transverter?

Just my 2cts to add to this 3cm transverter idea..
The wifi bands are getting crowded indeed. Moving to the 3cm band would be a logical step. The existing equipment is costly though. To use an up-converter could be a possible solution.
The use of a modern LNB suggested by Damon would give an RX device readily available and easy to modify. Replace the internal reference by an external free programmable PLL and you can use any IF you prefer between 400 and 2000 MHz. Ask the QO100 satellite community and you will find out it’s done by many of them. On the TX side you could use half the W1GHZ design suggested by Jason. Looking at the mixer stage on the GHZ pcb  there are no objections against a higher IF frequency like 900MHz
If an 900MHz rocket is used you have an existing device you could use as an IF. When used with two separate PLL’s (RX-TX) in the transverter circuits you could set up a system using two separate (RX-TX) frequencies. Not really MIMO but close. Looking in to the Rocket design is needed to find out if rx and tx can be separated to individual sma connector for each.
There are some pro’s and con’s.
Pro’s, quiet band compared to the regular wifi bands. Almost no others users to share the desired frequencies with.. High(er) power tx devices can be used. Higher antenna gain. No bandwidth limits :-) resulting in higher throughput.. (why go for MIMO if you can get the same or more with a higher bandwidth?) And… there is 500MHz to play with.
Con’s. you need some skills in microwave building, although the W1GHZ design is not to complex. There is no cheap out of the box solution… If you want MIMO a new feed for the system needs to be designed.. Scaling down the 5.8GHz feed? If the tx and rx cannot be separated indide the rocket there will be the need to do some switching on 900MHz (and on 10GHz if a single frequency is used..)
The cost of a transverter like this… No idea but less then $100 is possible, depending on the used materials, needed output power or/and personal wishes..

Ruud, WE1BTV

K9CQB's picture
That's right, let's not forget about our 900MHz band.

Ruud, these are great ideas.
In my area we use the Rocket M900 (0.8 Watt) and NanoStation M900 (1 Watt) for wooded areas usually in rural areas. The pager base-stations in our area are too powerful in populated areas to find a decent channel to use 900MHz, but it's pretty quiet in the rural area.
Sorry about the tangent, but you've reminded me that a Rocket M900 is closer to the L-Band for a 10GHz LNB and not so hard to up-convert the TX side to 10GHz using a modified version of the existing transverters. Yes, the hard part would be splitting RX/TX from their MIMO chains. 
I'm still hoping that someone will get some inspiration and design a fully integrated 2x2 MIMO (2TX/2RX) transverter from scratch. I will even help design the built-in 10GHz dual-polarity antenna for them. 

-Damon K9CQB

K6CCC's picture
What is the benefit to

What is the benefit to operating on the 10.0-10.5 GHz HAM band instead of the 2.4, 3.4, and 5.9 GHz bands? Is it simply underutilized?

2.4 is massively overloaded in many areas, 3.4 we're losing in a few years, and there is more and more sharing with non-amateur users on 5 Gig.  At least around here, I believe the intent for 10 Gig is to replace the long haul links that are mostly operating on 3 Gig.

KE6GYD's picture
Our RACES organization in

Our RACES organization in Irvine, CA has created a citywide network that uses three relay node sites in the surrounding hills that use the 3 GHz band for the links to the PD.  With that band going away for secondary amateur use, unless the ARRL can change the FCC's mind, the 10 GHz band would really be a big help. 

I believe the 3 GHz Rockets already had a transverter in them to place them in the 3 GHz band. But don't quote me on that. But that may help in this quest.


10 GHz Equipment

I have never touched one in person, but according to the spec sheet, Mimosa Networks has a radio that covers the entire 10 GHz band.

I know they are not currently supported buy the AREDN software, but for true backhaul, meshing is less important. In fact, I seem to remember reading several posts about using the Ubiquiti AirOS for backhaul links here in SoCal.

If these radios can be tuned to work in the Part 97 spectrum, it seems to me that they are exactly what is being asked for.

nc8q's picture
Mimosa B11 MIMO 4X4:4ac Backhaul Point-to-Point, 10,000-11,700 M

Mimosa B11 MIMO 4X4:4ac Backhaul Point-to-Point, 10,000-11,700 MHz, high speed up to 1.5 Gbps, Connectorized Brand: Mimosa Networks
Does not include antenna

K6CCC's picture
Damn!  And I thought 3 GHz

Damn!  And I thought 3 GHz radios were expensive!

KD9MOQ's picture
QAM tolerance to distortion

Thank you everyone for your input.
We're going to try developing a preliminary design based on what we've learned.
We want to support 2.4 GHz, 3.5GHz, and 5.8 GHz hardware with an adjustable LO if possible.
Our primary concern is with the distortion introduced by the W1GHZ design.
The passband of the pipe caps seems like it would introduce amplitude distortion in the transmitted and received signal.
Is there a way to quantify how tolerant QAM would be to distortion?


Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer