I’ve been trying to setup a single/dual channel gateway, but more and more I’ve come to realise that unless it’s a full 8 channel concentrator gateway there’s little to no point.
I’ve seen the ic880 based concentrator and also the Hope RF boards, both of which look to come in around £200. Are there any cheaper alternatives as that’s a bit beyond the budget.
I would really like to give it a go as a coverage predictor website shows that from my location I could potentially provide a gateway for a large part of the south east of England.
I did see a thread where someone suggest a bulk buy of the Hope RF boards, I can’t find it now, but if that’s still on the cards I might be interested.
I did mention a group buy here a while back when a client of mine ordered 50 boards (I though well maybe we can get to 100 and get an even better price) but in the end I didn’t find a simple way to organise such a thing without incurring a large risk.
So to me it could be interesting as a ‘real’ business but far too much to take on as a one man organising a group buy thing.
For reference the ‘bulk buy’ prices I have from Rising HF are:
plus shipping and handling fees on top of course.
We could also make a “go between” board for it to make it easy to connect to a Pi which should add no more than $5. I have a design of this for the NanoPI M1 but if there’s interest it’s easy to change to the Raspberry Pi instead (or have both options)
If anyone is interested and has an idea how to approach this let me know.
@jmarcelino I looked at that Nano Pi and it looks great but it doesn’t have eMMC memory so probably not a good idea for a gateway. Currently my top choice is the OrangePi Plus. I plan on doing an adapter board for it so I wonder if your board was done in KiCAD and if you might have it on github?
I’d like to understand your concerns around needing eMMC because there are a lot of Raspberry Pi based gateways out there, all without eMMC.
The adapter board design I have is very specific to a project I worked on so I’d have to change it and remove some parts before releasing it, also it’s done in Altium Circuit Studio (but should open with the free Citrcuit Maker too)
@jmarcelino From what I understand, SD cards are not tuned to the host system and tend to have poor interface designs which over time result in what people call bitrot. The result is that over months of use there will be data loss and eventually file system corruption. Onboard eMMC is designed with the host controller platform in mind besides having it’s own NAND controller. Not only does this make them generally faster but also able to handle bit errors better. I think there might be better wear leveling too but not sure there.
Originally my concern was mostly having to do with the physical tension based contacts of an SD card in a system which would be mounted outside. The slightest moisture on the contacts would lead to corrosion while eMMC are soldered on. No mechanical issues either.
So I’m even skipping over the low end Odroid because the eMMC is an addon-card. The OrangePi and Beaglebone Black eMMC are soldered on.
Yeah but would come at a cost for their own gateway sales, which is the product they actually want to develop and support.
Let’s be fair here TTN is far more than the gateway, they’ve put a ton of work into the backend - open for everyone to use, dealing with the LoRa Alliance and having all these great resources like the Website/Forum/Slack we all love. They wouldn’t be able to do that on razor-thin gateway margins, that’s for sure.
The only downside is this was probably also the reason they went all-in with a PIC32/Microchip gateway solution, to me far from ideal compared to an open and hardware independent Linux one - a solution also proving troublesome to them considering the Kickstarter delays.
I was the one who suggested a “crowdbuy”, and I’m still game, but it was kinda put off for a while to see what the pricing would be on gateways with the new and cheaper SX1308 chip. Have yet to see a gateway with this though.
“[Multitech] Conduit AP supports eight uplink and one downlink LoRa channels capable of supporting thousands of LoRaWAN end points” sounds like a Sx13… chip to me
Sorry if I seemed to be knocking the hard work of the people who have created TTN, that wasn’t my intention. It is a fantastic achievement to have developed TTN to the point where it is today.
Before I discovered The Things Industries, I did wondered how they “keep the lights on”.
However, if TTN is going to be a network that is actually worth using for anything other than seeing how far you can send a signal, it has to have enough gateways. There are Zero active gateways showing on the map within 100 miles of my location in the South East of Britain!
Without a company paying to install gateways we enthusiasts will have to bear the costs.
I’m prepared to put some into this, but I can’t justify £250 on the official gateway.
The commercial side of TTN is/will benefit from our investment, reducing the barrier to entry helps everyone.
Bulk buying, and selling at a small profit, would reduce the cost to us to get a gateway up and going.
Those people who want a ready made solution will still buy the existing gateway, those of us who just want to get going on the network will have built our own anyway.
@Epyon I was under the impression that the SX1308 was a lower power/range chip, or am I wrong?
Currently, if a crowdbuy is organised would be interested. I’ll probably limp along on my single channel gateway for a while until I get frustrated with OTAA not working correctly.
[quote=“SgtWilko, post:14, topic:5571, full:true”] @Epyon I was under the impression that the SX1308 was a lower power/range chip, or am I wrong?[/quote]
I actually don’t know. I was under the impression it was a production optimized version of the SX1301, but haven’t followed up on it.
The sx1308 is functionnaly the same as the sx1301. The only difference is the temperature range wich are the ‘commercial’ one +0°C +70°C for indoor use. While the sx1301 is in the ‘industrial’ range -40°c +85°C suitable for outdoor use.
