Hello,
I'm Guillaume F4HDK from France.
I would like to share my last project with the AREDN community; it's called "New Packet Radio".
This protocol allows to transport up to 500kbps IP trafic over 70cm amateur radio links, in a point-to-multipoint topology.
With an "off the shelf" DMR amplifier, you can get 20W RF power or more.
Everything is open source : software, PCB, protocol specification, documentation.
All details are here :
https://hackaday.io/project/164092-npr-new-packet-radio
I think that this could be a solution to access AREDN network from areas where "line of sight" links are not feasible.
Here in France, we plan to use it as an "access" method for an existing 5.6GHz IP backbone.
If you have questions, do not hesitate to ask!
73,
Guillaume F4HDK
Andre, K6AH
I'd love to have 2 of those right now. I even have the 2 DMR UHF amplifiers (they're sold by BTECH here).
Andre was right on when he said most of us just don't have the time to send out the board, order the individual parts, and assemble it from scratch. If there was already a product, or even a kit, I would surely buy 4 of them right now as well, so would half my ham buddies depending on price.
This sounds like a great project for a young entrepreneur to take up over this summer.
-Damon K9CQB
Hello,
Many thanks for your feedback, and your interest about the project.
I will organize the sale of kits, but at short term only for Europe. I don't know if I will ship to US later; I will keep you informed.
It would be much better if someone in the US could organize the sale of kits. If someone does it, I can mention his contribution and his contact e-mail inside the project web-page.
You say that you would like a kit... but it's really not that hard getting all the parts, and receiving the equivalent of kits. It only takes several dozens of minutes. Only 4 or 5 different orders to different suppliers, including the PCBs (Gerber files for PCB manufacturing are already provided, you don't need a PCB software).
The BOM (Bill Of Materials) is short and easy to provision.
Of course, soldering and cutting the enclosure takes more time. But that's what many people also like to do : hacking, making, and then using what they have build. And local fablabs-hackerspaces could help if needed, if you are not confident.
73,
Guillaume F4HDK
Regards,
Kent, KL5T
I'll talk to some of my folks here in the US to see if they'd be interested in cranking out the NPR, either as a kit or some type of finished device.
I may even have 2-3 of these boards made and assemble/solder the boards and load the FW to test how well these work with each other.
I'll also poke around and see if we are bound by some regulations that prohibit hams from using it's full bandwidth/throughput. I know we have amateur digital TV in this same band that is wider than this signal. So I don't see immediately any impediments. Fortunately in the US we are blessed to have a robust group of hams whose favorite hobby is telling other hams what they cannot do, so hopefully I will come across a highlighted FCC regulation that shows without a doubt to what extent these things are regulated. ;-)
Thank you,
-Damon K9CQB
If you are able to accomplish huge data rates using a set BW and power without interfering with anybody, nobody should be able to tell you to be less efficient in that given band. That's the opposite of progressive. That sounds very antiquated.
Just throwing out my 2 cents.
By the way, I would still buy these because they are still 5-6 times faster than the fastest packet radio nets we have out there.
-Damon K9CQB
"Fortunately in the US we are blessed to have a robust group of hams whose favorite hobby is telling other hams what they cannot do" <<< Loved this!!
Guillaume - can this radio be programed to run at 56 kB symbol rate and 100 kHz band width?
56kSps, this would need lots of modification on the software, and the protocol itselfs. And the speed would be very slow.
In France, we already manage to put New Packet Radio in our bands... and we only have 10MHz in the 70cm band, whereas you have 30MHz in the US!
I thing the best way you can support the project is :
1) Test it, even if it does not comply with FCC rules.
2) If you find it interesting, then negociate the removal of the very outdated FCC rule about 56kbps and 100kHz. We are in the 2019, not in 1999...
73,
Guillaume F4HDK
Guillaume,
What kind of collision avoidance technology is built into the on-the-air protocol for point-to-multipoint use?
73, Mark
It's detailed in the documentation. "New Packet Radio" is not CSMA like the old packet radio. It uses "managed TDMA" instead. Therefore all modem have their precise time slots, allocated dynamically by the "Master Modem" according to the needs expressed by each modem. It's like 2G mobile phones.
Therefore, there is no collision at all, even if "client modems" don't hear each others!
The "hidden terminal problem" is 100% solved by such protocols, without introducing additional delays.
73,
Guillaume F4HDK
What if you make this a full duplex system on 70cm. 5MHz shift for the US or 7.6MHz for Europe. Makes the system a lot faster.
It's what we did in the 90's with packet radio..
73 Ruud
Hello Ruud,
I plan to add this feature. Frequency split feature. One frequency for uplink and one for downlink. At Master-Reapeater side, it will be "full radio duplex" with 2 radios and a duplexer (FDD, frequency division duplex), and at client side only 1 radio+amplifier which changes its frequency rapidly between uplink phases and downlink phases. Exactly like 2G phones.
The main advantage of this feature (more than the gain in datarate) is that you can put a NPR Master-Repeater in a tower where there is already an FM or DMR or D-Star repeater, on the 70cm band.
73,
Guillaume F4HDK
Hello.
I have updated the firmware and the documentation. There are 2 new features which were expected by the US community:
https://hackaday.io/project/164092-npr-new-packet-radio
73,
Guillaume F4HDK
I didn't know you were even working on these 2 new additions to the firmware.
I'll be getting a couple of these built soon. When I get back from traveling I'll get with a my crew and we'll order all the components/boards we need and knock this out. We already have a few of the amplifiers. We even thought about modifying the case and amplifier heatsink to put the modem inside the amp's case.
Thank you so much for your efforts.
-Damon K9CQB
I can see different use cases to deploy this with AREDN, if not multiple options at the same time. I've scanned the documentation, so I know enough to be dangerous at the point :) .
LAN use case:
The 70cm link is on the LAN subnet of a mesh node. A mesh node on one end (master) and LAN devices on the other end of the 70cm channel -- laptop, etc.
DtdLINK use case:
The 70cm link is on the DTDLink network of a mesh node. The devices at endpoints of the 70cm channel mesh nodes. OLSR traffic on the 70cm may consume significant capacity of the channel if a large mesh network.
WAN use case:
A mesh RF island is remote in a canyon. The group wants ability to get to winlink CMS servers and http://arednmesh.org site to update firmware, etc. One of the mesh nodes is a gateway using ther 70cm link to get to the internet.
F4HDK, will this layer 2 bridge link pass vlan tags? What use cases have been tested? For the P2MP case, do clients receive the signals from other clients ( (through master, direct, or not at all). Any limitations?
Joe AE6XE
Hello Joe.
I'm not familiar with AREDN topologies and use cases. I will read documentation about it and try to answer to you.
Warning, currently NPR only has "IPv4 access" mode. It does not carry Ethernet header. Its is very limited in terms of topologies, but if you have read the "advanced user guide", you are aware of that.
I will try to implement a "pure ethernet" mode, in a few months, in order to add flexibility, but it will be (very) limited in term of number of MAC managed by the network (~100 or 200 maxi). I will try to be compatible/transparent with VLAN, like a standard (non-VLAN) ethernet switch.
The current use cases :
* "Access point" to a larger IPv4 network. Here its our "european Hamnet" : https://hamnetdb.net/map.cgi
* point to point but for a single remote site with static IPv4 address, and no IPv4 routing possible at remote site.
We are currently testing VoIP, and other "voice" technologies. It's promising, but it will be even better if I can add QoS prioritization.
Yes, of course, 2 clients can communicate together via the Master (not directly, otherwise it would be a multipoint to multipoint topology).
73,
Guillaume F4HDK
OK, then the 3 defined use cases are how it could be deployed with AREDN. An 802.11q switch would be needed to split out the desired network, if vlan tagged trunked out of the AREDN device.
"Yes, of course, 2 clients can communicate together via the Master (not directly, otherwise it would be a multipoint to multipoint topology)."
My understand is the following behavior:
A) a client sends a broadcast frame, the master will retransmit for all clients to receive.
B) a client sends a unicast frame to the IP address of another (may be hidden node, may not be) client, the master will retransmit.
C) a client sends a unicast to another subnet, the master does not retransmit.
Is this correct?
Joe AE6XE
"implement a "pure ethernet" mode". Probably no urgency to do this for integrating with AREDN. This would enable more networks (vlans) on top of 70cm and further stress the available capacity. But more importantly, mixing these AREDN use cases together, over a long distance 70cm RF link, probably isn't a realistic or necessary combination -- doing only for the sake that one could.
Joe AE6XE
I had spoken to my hardware development folks at work and we were originally thinking of modifying the hardware to make it smaller and more capable by redesigning the board and modules, but using the same base chipsets so that F4HDK's software will always work on them. HOWEVER... we came to the conclusion that the hardware was fine in it's current state to do what we needed (juice not worth the squeeze).
If we ever wanted to actually change this design we would likely create a daughterboard (although larger) for the MikroTik USB150 dongle to replace the processor and give it extra capabilities. IF we did this we would likely tell F4HDK what we wanted to do and get him involved, but we would still do the hardware building and software porting for him. This does not seem likely right now as my hardware guys are rather busy doing stuff that earns a profit ;-)
USE CASES - We intend to use these for 2 different use cases. 1) We have a few local mesh islands in rural/mountainous/wooded areas that we'd like to connect to and these would be perfect for those situations. 2) We may see if we can use these to go over long distances to link NoVa Mesh to Hampton Mesh and Culpepper Mesh by placing the master NPR node at our southernmost AREDN node and shooting several slaves nodes out toward the southeast and southwest. We'll see if this whole NPR thing becomes a viable option once we start messing around with it. I think Ken (KE2N) was going to put together 3 of these boards for experimentation as well - so we'll have a decent trial going here.
-Damon K9CQB
Hi Gulliaume, nice work. Some guys from Austria have same ideas. I can see a chat usecase for this.
I talked to Lars DL4APT.
He brings some kits to Friedrichshafen HamFair this weekend. We will test it with AREDN.
73 Timm
Here are some news about NPR - New Packet Radio.
I have made a big update of the firmware. Check it here:
https://hackaday.io/project/164092-npr-new-packet-radio
The main evolution is a great improvement of the radio sensitivity. If you have noticed lack of sensitivity, it will help, for sure!
In fact, I have switched to GFSK instead of GMSK, with double frequency deviation compared to GMSK.
After studying it in details, we have concluded that the radio chip that I use (SI4463) is quite bad at GMSK demodulation.
This evolution makes it (of course) not compatible with previous version! You have to update all your modems at the same time.
The bit per hertz ratio decreases, and I managed it like this:
* for low symbol rates with bandwidth constraints, I remain in existing constraints (100kHz for modulation 20, 200kHz for modulations 11 and 21)
* for higher symbol rates, the SR is the same than before, and the radio bandwidth is increased.
Check all new version of documentation, release notes, etc...
Other improvements:
* I now publish a firmware for the 144-148MHz band (you need adequate radio modules of course).
* Frequency setting resolution is improved to 1kHz (instead of previously 40kHz).
* RSSI is now displayed in "dBm"
The next step of the development will be frequency shift and FDD (Frequency Division Duplex) feature. The uplink and downlink frequency could be different, and you will be able to install 2 modems at Master side (one for RX the other for TX) plus a duplexer. The main goal is to be able to locate a NPR Master (repeater) in places (towers) where shift repeaters are already present.
Finaly, about the kits sale, a new world-wide sale will probably start before end of october. I will keep you informed.
73,
Guillaume F4HDK