Author Topic: XRange with sx1272.  (Read 21352 times)

0 Members and 1 Guest are viewing this topic.

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
XRange with sx1272.
« on: June 20, 2014, 08:57:57 am »
Hi everybody,
I have run a crowd funded project (XRange) which involves the new semtech's LoRa sx1272 transceiver.

https://www.indiegogo.com/projects/xrange-extreme-long-range-wireless-module-for-868-or-915-mhz

Additional information on http://www.netblocks.eu/
http://www.netblocks.eu/xrange-sx1272-lora-datasheet/

What do you think about the project?
What do you expect from this type of projects?

Any feedback is welcome (positive or negative).

Thanks,
Milen
« Last Edit: April 06, 2015, 10:04:37 am by mpavlov »
 

Offline fake-name

  • Regular Contributor
  • *
  • Posts: 75
Re: XRange with sx1272.
« Reply #1 on: June 20, 2014, 10:50:59 am »
Flexible Funding

What the hell are you thinking?
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #2 on: June 20, 2014, 11:17:00 am »
Hi fake-name,

Unfortunately I have chosen the “flexible funding”, because I can use bank account and PayPal to receive the funds. If I use “fixed funding” Indiegogo will transfer funds only by PayPal. But  withdrawing form PayPal for non US account can be pain. because I can use only Credit /Debit Card to transfer funds. I believe that the project can collect the asked funds and the “flexible funding” will not be a problem.

Thanks
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #3 on: June 20, 2014, 11:46:21 am »
Unfortunately I can not change the campaign's funding type any more.
Anyway, even the campaign will not reach the target funds everyone who make donation will receive the claimed perks.
 

Offline Legit-Design

  • Frequent Contributor
  • **
  • Posts: 562
Re: XRange with sx1272.
« Reply #4 on: June 20, 2014, 02:05:19 pm »
Lack of information is disturbing. Can you provide some form of proof of anything. Flexible funding on indiegogo is just scammer heaven.  :palm:
I'm not saying you are scammer, just that you do everything like a scammer would do. Website that has nothing on it, same pictures as indiegogo page.

Since you are going to deliver the perks even if you don't finish the campaign, what is the point? Don't use scammer heaven if you want to become known and sell that product which seems kinda ok. Is this going to be open source or something? Use open source as your selling point? Will create heck of a more buzz that scamgogo.
 

Offline yramgu

  • Contributor
  • Posts: 44
  • Country: fr
Re: XRange with sx1272.
« Reply #5 on: June 20, 2014, 02:36:41 pm »
What does "+20 dBm at 100 mW constant RF output" mean? (see indiegogo page). That it always outputs 20dBm when transmitting?

Nevermind, got it :

http://www.semtech.com/wireless-rf/rf-transceivers/sx1272/

You just copied/paste Semtech's page. Not very honnest of you, as they might not reflect the performances of your device... any measurements to support your claims?

What kind of instruments did you use to design this? -137dBm sensitivity is really hard to reach.
« Last Edit: June 20, 2014, 02:55:24 pm by yramgu »
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #6 on: June 20, 2014, 03:25:05 pm »
Hi Legit-Design

It is my fault, I have not realized that “flexible funding” is used for scam. I can not use kickstarter because it does not accept projects form Bulgaria.

I accept that there is too fake projects on Indiegogo. But I spent last 6 months to develop the prototype you can see on the photos that the some components are hand soldered. I can produce small batch of product under 50 but it is not profitable.

The www.netblocks.eu web site is not finished yet. I will update it soon.
The firmware will be open source it should be downloadable from web site. I have working code for Keil MDK and I'm working to implement such for Mbed it should be quite easy because Semtech has a driver of sx1272 for Mbed
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #7 on: June 20, 2014, 04:21:51 pm »
Hi yramgu,

I am sorry that I provided only information from data-sheet. I have attached some pictures of actual measurement. This test is with LoRa modulation.
 

Offline yramgu

  • Contributor
  • Posts: 44
  • Country: fr
Re: XRange with sx1272.
« Reply #8 on: June 20, 2014, 10:27:43 pm »
Fair enough :)

Are you just using the sx1272 as a simple transceiver (ie using only LoRa modulation) or are you compliant with the LoRaMac as well?

I will insist on sensitivity: how do you plan to test it / have you tested it? Killer sensitivity (-137dBm!!!) is the key aspect of LoRa: it has no reason to exist otherwise.
If you can't guarantee it there's no reason to choose LoRa over another (cheaper?) "standard"
 

Offline daveshah

  • Supporter
  • ****
  • Posts: 356
  • Country: at
    • Projects
Re: XRange with sx1272.
« Reply #9 on: June 21, 2014, 07:47:32 am »
What distances have you tested it over so far? You claim over 12km.
 

Offline fake-name

  • Regular Contributor
  • *
  • Posts: 75
Re: XRange with sx1272.
« Reply #10 on: June 21, 2014, 11:24:03 am »
Hi Legit-Design

It is my fault, I have not realized that “flexible funding” is used for scam. I can not use kickstarter because it does not accept projects form Bulgaria.

I accept that there is too fake projects on Indiegogo. But I spent last 6 months to develop the prototype you can see on the photos that the some components are hand soldered. I can produce small batch of product under 50 but it is not profitable.

The www.netblocks.eu web site is not finished yet. I will update it soon.
The firmware will be open source it should be downloadable from web site. I have working code for Keil MDK and I'm working to implement such for Mbed it should be quite easy because Semtech has a driver of sx1272 for Mbed


Not to be rude or anything, but did you do ANY research about crowd funding before sticking your widget up?

Seriously, 2 out of every 3 threads just in this sub-forum alone are probably making fun of some scammer and their latest flex-funded boondoggle. Read ANYTHING critical about indiegogo AT ALL, and it'll probably have MANY paragraphs about scams, flex funding, and their commonly close relationship.
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #11 on: June 21, 2014, 11:30:34 am »
Hi yramgu,

I use the sx1272 as a simple transceiver with LoRa modulation. My first goal is to clear the hardware design.
About LoRaMac I have not port it to XRange but it should not be a problem because Semtech  provides source code with BSD license.
About sensitivity, I had the same question when I stared work with sx1272. I came up with idea to compare the distance of possible communication between two modules with LoRa modulation and FSK modulation (sx1272 support FSK).
My fist tests had not been very promising, I tested FSK compare to LoRa with Spreading Factor - 8  the LoRa was only about 2x better than FSK but far from what the Semtech promising. But real magic comes when I managed to set  Spreading Factor -12, LoRa is over 5x better than FSK in my tests (dense urban environment).
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #12 on: June 21, 2014, 01:24:54 pm »
What distances have you tested it over so far? You claim over 12km.

Well here is one point
https://www.google.com/maps/@42.590226,23.368248,3a,75y,19.3h,87.82t/data=!3m4!1e1!3m2!1ssQmvcHeJu2LHSbpATi03vw!2e0

other point is 14 - 21 km far from this point, but reliability of communication depends from terrain.?f course if there is a building between two transmitters this brakes the communications.
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #13 on: June 21, 2014, 01:39:57 pm »
What distances have you tested it over so far? You claim over 12km.

Well here is one point
https://www.google.com/maps/@42.590226,23.368248,3a,75y,19.3h,87.82t/data=!3m4!1e1!3m2!1ssQmvcHeJu2LHSbpATi03vw!2e0

other point is 14 - 21 km far from this point, but reliability of communication depends from terrain.?f course if there is a building between two transmitters this brakes the communications.

Other point on googlemaps:
https://www.google.com/maps/place/%D0%A1%D0%BE%D1%84%D0%B8%D1%8F/@42.691231,23.452451,3a,75y,336.64h,90t/data=!3m4!1e1!3m2!1sDzHmujbo1wD-15MljB-7Tg!2e0!4m2!3m1!1s0x40aa8682cb317bf5:0x400a01269bf5e60
 

Offline yramgu

  • Contributor
  • Posts: 44
  • Country: fr
Re: XRange with sx1272.
« Reply #14 on: June 21, 2014, 01:45:47 pm »
Hi yramgu,

I use the sx1272 as a simple transceiver with LoRa modulation. My first goal is to clear the hardware design.
About LoRaMac I have not port it to XRange but it should not be a problem because Semtech  provides source code with BSD license.


Don't be so sure, the LoRaMac behaves a bit "weirdly" in my opinion


About sensitivity, I had the same question when I stared work with sx1272. I came up with idea to compare the distance of possible communication between two modules with LoRa modulation and FSK modulation (sx1272 support FSK).
My fist tests had not been very promising, I tested FSK compare to LoRa with Spreading Factor - 8  the LoRa was only about 2x better than FSK but far from what the Semtech promising. But real magic comes when I managed to set  Spreading Factor -12, LoRa is over 5x better than FSK in my tests (dense urban environment).


I get your point. It's normal that you get better sensitivity with SF12, as this decreases the radio datarate. But the problem is that with this method you have no absolute reference for the sensitivity measurement. A better (a proper) way of doing this would be using an RF signal generator to inject a LoRa frame direclty in your module. The RF Gen would allow you decrease the power gradually until you reach your target FER. Then you would get the actual sensitivity of your product. Ask Semtech for help regarding LoRa format, as they didn't publish details about the modulation.
As those Gens are expensive, I think you could just rent one from R&S / Anritsu... for the time needed to do your tests.

And (for now) I don't think your project's a scam
 

Offline Legit-Design

  • Frequent Contributor
  • **
  • Posts: 562
Re: XRange with sx1272.
« Reply #15 on: June 21, 2014, 02:02:45 pm »
How about more data than just one connection?
Leave one point in your headquarters. Go around town with the antenna sticking from your backpack. Have gps running and time correlate the measurements and location. Or doesn't have to use gps just choose some data points and log them. I really want to see that >3km dense urban range in action.
Can you provide some basic real world measurements how much the maximum speed decreases with range? Doesn't have to be all real world, just correct them based on your measurements.
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #16 on: June 21, 2014, 02:34:46 pm »
Hi yramgu,

I use the sx1272 as a simple transceiver with LoRa modulation. My first goal is to clear the hardware design.
About LoRaMac I have not port it to XRange but it should not be a problem because Semtech  provides source code with BSD license.


Don't be so sure, the LoRaMac behaves a bit "weirdly" in my opinion


About sensitivity, I had the same question when I stared work with sx1272. I came up with idea to compare the distance of possible communication between two modules with LoRa modulation and FSK modulation (sx1272 support FSK).
My fist tests had not been very promising, I tested FSK compare to LoRa with Spreading Factor - 8  the LoRa was only about 2x better than FSK but far from what the Semtech promising. But real magic comes when I managed to set  Spreading Factor -12, LoRa is over 5x better than FSK in my tests (dense urban environment).


I get your point. It's normal that you get better sensitivity with SF12, as this decreases the radio datarate. But the problem is that with this method you have no absolute reference for the sensitivity measurement. A better (a proper) way of doing this would be using an RF signal generator to inject a LoRa frame direclty in your module. The RF Gen would allow you decrease the power gradually until you reach your target FER. Then you would get the actual sensitivity of your product. Ask Semtech for help regarding LoRa format, as they didn't publish details about the modulation.
As those Gens are expensive, I think you could just rent one from R&S / Anritsu... for the time needed to do your tests.

And (for now) I don't think your project's a scam

yramgu,  thanks for your feedback
Unfortunately  I do not have such resources to do this tests.
The RF part of XRange is exact replication of Semtech's
kit http://www.digikey.com/product-detail/en/SX1272DVK1BAS/SX1272DVK1BAS-ND/4285674
and this will be acceptable for such kind of projects like XRange.
 

Offline yramgu

  • Contributor
  • Posts: 44
  • Country: fr
Re: XRange with sx1272.
« Reply #17 on: June 21, 2014, 05:01:42 pm »

yramgu,  thanks for your feedback


You're welcome ;)

How about more data than just one connection?
Leave one point in your headquarters. Go around town with the antenna sticking from your backpack. Have gps running and time correlate the measurements and location. Or doesn't have to use gps just choose some data points and log them. I really want to see that >3km dense urban range in action.
Can you provide some basic real world measurements how much the maximum speed decreases with range? Doesn't have to be all real world, just correct them based on your measurements.


I second that. You're gonna have one major potential interferer: LTE. But as you said, you've copied Semtech's design and the rest of the board is nothing fancy so you should be OK :)

 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #18 on: June 21, 2014, 07:19:38 pm »

yramgu,  thanks for your feedback


You're welcome ;)

How about more data than just one connection?
Leave one point in your headquarters. Go around town with the antenna sticking from your backpack. Have gps running and time correlate the measurements and location. Or doesn't have to use gps just choose some data points and log them. I really want to see that >3km dense urban range in action.
Can you provide some basic real world measurements how much the maximum speed decreases with range? Doesn't have to be all real world, just correct them based on your measurements.


I second that. You're gonna have one major potential interferer: LTE. But as you said, you've copied Semtech's design and the rest of the board is nothing fancy so you should be OK :)

About test for 3km dense urban range, unfortunately I can not arrange this test soon. But from my experience if two transmitters are  near to ground and you are surrounded  by many buildings you can not achieve  3km distance. If you can put one transmitter one of the highest building you could have 3km range.
Last thing I can say is that there is very good reception  inside of the  buildings.
 

Offline Legit-Design

  • Frequent Contributor
  • **
  • Posts: 562
Re: XRange with sx1272.
« Reply #19 on: June 21, 2014, 09:13:03 pm »
Even half of that with those small stick antennas would be amazing accomplishment, taking into account that it's using some wireless technology.
 

Offline Legit-Design

  • Frequent Contributor
  • **
  • Posts: 562
Re: XRange with sx1272.
« Reply #20 on: July 03, 2014, 12:25:41 pm »
Still no news on this one? Any chance of just doing really low cost transceiver only for maybe arduino, raspberry pi or something? Does the LoRa part actually require more connections than just SPI to control it externally?
http://www.digikey.com/product-detail/en/SX1272RF1BAS/SX1272RF1BAS-ND/4490407 like this one, but maybe with some working examples for the popular prototype platforms?
I don't think most people want the mbed thing, except people who are already familiar with using mbed.

https://www.kickstarter.com/projects/flutterwireless/flutter-20-wireless-arduino-with-half-mile-1km-ran Similar to this one, but keep it as cheap as possible, if possible.
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #21 on: July 03, 2014, 08:52:16 pm »
Still no news on this one? Any chance of just doing really low cost transceiver only for maybe arduino, raspberry pi or something? Does the LoRa part actually require more connections than just SPI to control it externally?
http://www.digikey.com/product-detail/en/SX1272RF1BAS/SX1272RF1BAS-ND/4490407 like this one, but maybe with some working examples for the popular prototype platforms?
I don't think most people want the mbed thing, except people who are already familiar with using mbed.

https://www.kickstarter.com/projects/flutterwireless/flutter-20-wireless-arduino-with-half-mile-1km-ran Similar to this one, but keep it as cheap as possible, if possible.

I am working to clear the minor hardware bugs on prototype and preparing to produce final prototype. Also I am porting LoRaMac firmware to XRange hardware. I hope that I will have working  LoRaMac next week. 
I am considering how to lower the cost of XRange, because the price can be obstacle for the hobbyists and makers to use this technology.
It is possible to use sx1272 with Arduino but I think Arduino is not the best platform because the LoRa firmware if fairly complicated is takes about 20K (  LoRaMac about 26K for STM32L152).
To control LoRa you need SPI, one pin to control TX/RX of RF switch, MCU need to monitor DIO0- DIO5 because SX1272 toggle some of this pins on TX or RX event (depends of regime).
I have started with this boards http://www.digikey.com/product-detail/en/SX1272RF1BAS/SX1272RF1BAS-ND/4490407   (I had to manufacture them because they had not been on stock at that time) but it was very frustrating until I have managed to run semtech's source code on my hardware (I attached the photo of my first hardware). Actually you need two of this board total $140 + shipping. XRange is better value  for money, you can take two XRange boars with antennas for $99 , free shipping, source code of firmware and documentation.
Mbed is interesting platform platform but it is optional, firmware can be compiled with any ARM C compiler (I have project for Keil MDK if there is interest I can provide project for GCC).
Flutter is really inexpensive but they can achieve this price because of mass production. They have very successful project on Kickstarter. Did they provide source code and design files of Flutter?
 

Offline Legit-Design

  • Frequent Contributor
  • **
  • Posts: 562
Re: XRange with sx1272.
« Reply #22 on: July 04, 2014, 10:31:38 am »
They have very successful project on Kickstarter. Did they provide source code and design files of Flutter?
As far as I understood they haven't even shipped anything yet. So nothing except promises and some pictures are provided. Also in their updates/comments they seem to have hit a wall with the regulations. Their product can't be certified by FCC, as it was originally promised to be fully open source. "FCC basically won't certify a product that allows illegal frequency transmissions" And yes their product is open source so anything could be possible. I think they need to lock some things down and this won't be fully open source.

So like many other hardware crowdfunding campaigns, they are delivering late. People should be more sceptical about crowdfunding, nice words don't mean anything. We can look at crowdfunding campaigns that are years late and still haven't delivered anything or even close to delivering. That is the reality.

Are you planning on releasing some code and wiring for that prototype? For people who want to diy the whole thing? Since you are crowdfunding the polished easy to use product, diy version for people to test out might be a nice boost? That transceiver board (http://www.semtech.com/wireless-rf/rf-transceivers/sx1272/) provided gerbers. And looks like they offer maximum of 5 free sample chips.
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #23 on: July 04, 2014, 07:54:04 pm »
Are you planning on releasing some code and wiring for that prototype? For people who want to diy the whole thing? Since you are crowdfunding the polished easy to use product, diy version for people to test out might be a nice boost? That transceiver board (http://www.semtech.com/wireless-rf/rf-transceivers/sx1272/) provided gerbers. And looks like they offer maximum of 5 free sample chips.

I have not thought that there will be interest of this prototype. I will write article how to build it and I will provide source code also. I have 6 not populated PCB for sx1272. I can put then like a perks in my crowdfunding project.
 

Offline mpavlov

  • Contributor
  • Posts: 25
  • Country: bg
Re: XRange with sx1272.
« Reply #24 on: July 07, 2014, 07:11:01 pm »
Do you have any data on power consumption? I mean actual real-world measurements of the unit in operation, not just stuff from datasheets?
This is rough measurement of XRange board: 
Only CPU  - 7.5 mA
CPU + sx1272 (TX is set to max power +20dBm) - 80 mA
CPU + sx1272 (RX) - 12 mA
------
Note: the measurement is done before the switching regulator with 6V power supply so you have to calculate the real current!
--------------
You are entering a very crowded, competitive market. There are lots of cheap transceivers with on-board microcontrollers using the same or similar chipsets, and costing the same or less. What makes yours special?

The target of this product is to satisfy the need of long range add robust  communication solution. The current solutions can not compete whit LoRa Modulation (at same power level). At the moment there are a few products with sx1272 and their price is a bit higher. What do you expect from this type of product  about price , futures, firmware , support  ... ?  Please share your opinion. I am looking to refine the product to fit the needs of the people who will use the product.
« Last Edit: July 08, 2014, 12:22:48 pm by mpavlov »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf