$20 GPS+GLONASS receiver; $50 for RTK

3,035 views
Skip to first unread message

Matthew Mucker

unread,
Jan 24, 2014, 10:36:57 AM1/24/14
to diyr...@googlegroups.com
I found this on Hackaday here.

The article references NavSpark, an IndieGoGo campaign here. This is apparently an "Arduino-compatible" board with onboard GPS, GPS+GLONASS, or GPS+Beidou. Most intriguing for us is probably the $50 kit which includes two receivers and is supposedly RTK-Lib compatible, which should deliver sub-centimeter accuracy to our GPS navigation.

The processor is a 100MHz Venus 822 with 212K RAM and 1024K Flash ROM. The CPU core is a LEON3 architecture, which I have never heard of before but is a SPARC core. I/O looks a little anemic but still has UART, SPI, I2C, one PWM and 17 total digital I/O. The final product will be released open source with schematic, board, BOM, and Arduino IDE included. (Although the GPS/Beidou/GLONASS navigation kernel will be in binary format, whatever that means.)

The GPS core, if I'm reading this right, is also named Venus. (Am I getting this right or am I confused?) The Skytraq Venus 838, according to the manufacturer's product brief, supports a 50Hz update rate with 2.5m accuracy.

The ship date is currently set at March 24, but my experiences with Kickstarter tell me to discount that. It'll probably get to us too late for AVC 2014, but the idea of using RTK to navigate in AVC2015 is certainly tempting. (I believe that 0x27 uses RTK navigation with that super expensive GPS they have. 'Twould be nice to be able to do that with $50 of kit.)

I'd love to hear the group's input. I lack the knowledge and experience to fully evaluate this project.

Matthew Mucker

unread,
Jan 24, 2014, 11:21:52 AM1/24/14
to diyr...@googlegroups.com
With further digging, it appears that this project may be a Stytraq project. Skytraq is the manufacturer of the GPS chip.

A comment from the founder:

We expect to have 2 kinds of library for Arduino, one with GPS/GNSS, and one without. The new Venus822 datasheet will have detailed description on the hardware peripherals, enabling users interested to fully access the hardware can build things from ground up using the library without GPS/GNSS. The other library will have easy to use API provided. Description on LEON3 Sparc-v8 used in Venus822 can be found at: http://www.gaisler.com/index.php/products/processors/leon3?task=view&id=13

Here are some datasheets:

Venus822 = LEON3 Sparc-V8 + FPU + memory + peripherals. All LEON3 Sparc-V8 & FPU details can be found in section 69 of GRIP document. What we need to add from scratch for existing Venus822 datasheet are the memory map and peripheral register control descriptions, which isn’t difficult to describe clearly.

http://t.cn/8FbovJT
http://t.cn/8Fboh6t

The GPS/GNSS chip datasheet from other vendors are all like this, just enough information for making a GPS receiver. :-)

http://t.cn/8FboAEX



Another interesting comment by the founder:

Running RTKLIB on Venus822 has crossed our mind, but we aren’t sure there’s enough memory on Venus822. One interesting thing to note is that limiting factor is 212KByte on-chip SRAM. If SRAM is enough and only 1MByte Flash memory isn’t enough, we might take this into consideration and try to bring out quad-SPI Flash memory interface if we later package a new higher pin-count device, then it could run Flash program off chip, enable user choice of 2 / 4 / 8 / 16 / 32 / 64 MByte quad-SPI Flash.

Another comment:

RTKLIB has utility that converts our binary to RINEX. We also have program that converts to RINEX that could provide.

S1315F-RAW uses Venus 6 chipset. At first posting of NS-RAW, we were still having some issue with Venus8 port working with RTKLIB, so Venus 6 version was planned as backup. Last Friday we had Venus8 port working correctly with RTKLIB, so we expect to ship NS-RAW using Venus 8 solution.

And another:

The RTKLIB website, especially the RTKLIB document, has sufficient information on how to properly setup and run using low cost GPS receiver. NS-RAW is compatible to S1315F-RAW mentioned in the RTKLIB document. There is also a forum with much discussion on RTKLIB below for reference:

http://www.rtklib.com/

http://t.cn/8FG6RMr


I am a bit concerned that the founder seems all too willing to incorporate the ideas of the backers, suggesting the prototype hasn't been completely thought out yet. That could be good, in that the final product may have extra features, or it could suggest a roadblock or delay might come up.


Alex McClung

unread,
Jan 24, 2014, 3:46:21 PM1/24/14
to diyr...@googlegroups.com
As cool as it is, it's not allowed in the AVC. Quote from the 2013 rules: "The vehicle must be fully autonomous and self-contained. No transmitters or communication beacons (other than GPS) of any kind are allowing. You may NOT tether to a laptop or other device. Everything necessary for the vehicle’s navigation/processing/sensing must be attached and part of the vehicle itself."

Alex

Wayne Holder

unread,
Jan 24, 2014, 4:19:38 PM1/24/14
to diyr...@googlegroups.com
Actually, you can parse that sentence to read that "GPS communications beacons" are allowed, just not other kinds of beacons.  In fact, anyone who's used a GPS receiver that used a 3rd party DGPS correction service, such as Omnistar was using a GPS beacon to receive the differential correction data (from beacon transmitters hosted on satellites outside the GPS constellation.)  I would argue that bots using RTKLIB in just this same fashion are also within the rules.  Either that or some bots they've already allowed to run were in violation of them.

Wayne


--
You received this message because you are subscribed to the Google Groups "diyrovers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diyrovers+...@googlegroups.com.
To post to this group, send email to diyr...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/diyrovers/3a607ec7-e4ab-4938-a6d8-8a6c6b811d29%40googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

Michael Shimniok

unread,
Jan 24, 2014, 4:25:47 PM1/24/14
to diyr...@googlegroups.com
> Actually, you can parse that sentence to read that "GPS communications
> beacons" are allowed, just not other kinds of beacons. In fact,
> anyone who's used a GPS receiver that used a 3rd party DGPS correction
> service, such as Omnistar was using a GPS beacon to receive the
> differential correction data (from beacon transmitters hosted on
> satellites outside the GPS constellation.) I would argue that bots
> using RTKLIB in just this same fashion are also within the rules.
> Either that or some bots they've already allowed to run were in
> violation of them.
Ultimately we'd have to ask SFE. If I had to guess, my guess is that
they don't *intend* to allow folks to bring their own beacons, whether
dgps, RTK or whatever. :)

Michael

Wayne Holder

unread,
Jan 24, 2014, 4:35:06 PM1/24/14
to diyr...@googlegroups.com
Actually, you don't have to bring a beacon to use something like RTKLIB.  There USGS operates a network of systems from which you can receive differential correction data in realtime using a protocol called NTRIPS.  This is the scheme many portable survey receives use.  In fact, there's are several of these stations around Boulder.  For all intents and purposes, using DGPS via an NTRIP connection is functionally the same as using a service like Omnistar.  The difference is that Omnistar costs a lot to subscribe to the service and the receivers needed to use it are not cheap, either.  I would think that in the spirit of experimentation one could make a good case that it's not not fair to include systems using Omnistar and exclude homebrew systems that, in effect, work exactly the same way...

Wayne


--
You received this message because you are subscribed to the Google Groups "diyrovers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diyrovers+unsubscribe@googlegroups.com.

To post to this group, send email to diyr...@googlegroups.com.
Message has been deleted

Michael Shimniok

unread,
Jan 24, 2014, 4:48:45 PM1/24/14
to diyr...@googlegroups.com
On 01/24/2014 02:35 PM, Wayne Holder wrote:
> Actually, you don't have to bring a beacon to use something like
> RTKLIB. There USGS operates a network of systems from which you can
> receive differential correction data in realtime using a protocol
> called NTRIPS. This is the scheme many portable survey receives use.
> In fact, there's are several of these stations around Boulder. For
> all intents and purposes, using DGPS via an NTRIP connection is
> functionally the same as using a service like Omnistar. The
> difference is that Omnistar costs a lot to subscribe to the service
> and the receivers needed to use it are not cheap, either. I would
> think that in the spirit of experimentation one could make a good case
> that it's not not fair to include systems using Omnistar and exclude
> homebrew systems that, in effect, work exactly the same way...
Ah, well that's different then. I see your point for sure.

Michael

Ted Meyers

unread,
Jan 24, 2014, 4:51:25 PM1/24/14
to diyr...@googlegroups.com
Wow, that does look really great.  It seems to me (I don't have much knowledge of GPS or RTK) that for $50 it is well within the set of things that are too good to be true.  Aren't most RTK setups in the multi-thousand dollar range?  (A quick search on ebay shows Trimble RTK setups for surveying are all well over $1000).  I'm skeptical.

Matthew Mucker

unread,
Jan 24, 2014, 4:56:46 PM1/24/14
to diyr...@googlegroups.com
On Friday, January 24, 2014 3:48:07 PM UTC-6, Ted Meyers wrote:
Wow, that does look really great.  It seems to me (I don't have much knowledge of GPS or RTK) that for $50 it is well within the set of things that are too good to be true.  Aren't most RTK setups in the multi-thousand dollar range?  (A quick search on ebay shows Trimble RTK setups for surveying are all well over $1000).  I'm skeptical.

Which is one reason that I posted it here. On the other hand, the founder appears to be the company that actually manufactures the GPS chip, so if anyone can do it, it's likely they can. I suspect that those pricey surveying setups are pricey because A) it's a small market with a niche product selling to an audience that can justify the cost, and that B) you're paying for the software, not the hardware.

At any rate, I plan to watch this campaign closely for the next 14 days. I might be willing to risk losing $50 against the possibility that this really is what it claims to me.

Wayne Holder

unread,
Jan 24, 2014, 5:03:17 PM1/24/14
to diyr...@googlegroups.com
Actually, the protocol is NTRIP (I mistyped it in my prior message.)  You can read a but more about it here:

  http://en.wikipedia.org/wiki/Networked_Transport_of_RTCM_via_Internet_Protocol

RTKLIB is an open source software package designed to work with GPS receivers that can output RAW "pseudorange" data, which is then adjusted via differential corrections received via NTRIP and then these corrected pseudoranges are used to compute a more accurate GPS fix.  In theory, you can do all this with relatively low cost hardware.  There are several consumer-level GOS receivers, such as the u-Blox NEO-6T that can output RAW data.  And, I've read about people running RTKLIB on relatively low end ARM chips.  I've not tried it myself, but this has been something I've been actively investigating now for many months.  One project I've been following is the Piksi project on Kickstarter which, once it's finished, should be capable of cm level accuracy using NTRIP-fed RTCM corrections (or so they've told me.)  Still not cheap, but much less than a commercial solution.

The $20 receiver these guys are pitching is said to be able to output RAW reading (just quoting their writeup), but you'd still need to run RTKLIB on a host processor to get real time differential fixes.

Wayne




On Fri, Jan 24, 2014 at 1:48 PM, Ted Meyers <ted.m...@gmail.com> wrote:
Wow, that does look really great.  It seems to me (I don't have much knowledge of GPS or RTK) that for $50 it is well within the set of things that are too good to be true.  Aren't most RTK setups in the multi-thousand dollar range?  (A quick search on ebay shows Trimble RTK setups for surveying are all well over $1000).  I'm skeptical.

--
You received this message because you are subscribed to the Google Groups "diyrovers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diyrovers+...@googlegroups.com.

To post to this group, send email to diyr...@googlegroups.com.

Michael Shimniok

unread,
Jan 24, 2014, 5:07:24 PM1/24/14
to diyr...@googlegroups.com
On 01/24/2014 02:56 PM, Matthew Mucker wrote:
> Which is one reason that I posted it here. On the other hand, the
> founder appears to be the company that actually manufactures the GPS
> chip, so if anyone can do it, it's likely they can.
Yeah it's the folks behind the Venus chip. They're using the same Sparc
architecture found in the 638flpx.

I got the impression reading the descriptions and digging in that RTK
wasn't an out of the box thing, but more like a maybe we can do this
kind of thing?

> At any rate, I plan to watch this campaign closely for the next 14
> days. I might be willing to risk losing $50 against the possibility
> that this really is what it claims to me.
I'm also really curious. I don't think I'm ready to bet $50 on it. I'm
sure it'll work fine as a Venus GPS but RTK? Hmm...

They supposedly make the SDK available ... I had no response when I asked.

I wonder if would be possible to write GPS firmware that is IMU aware?
Has that been done? I wonder if it is doable on this.

Michael

Borna Emami

unread,
Jan 24, 2014, 5:12:10 PM1/24/14
to diyr...@googlegroups.com
The logistics of setting up an RTK base station at AVC even if someone has a commercial setup, out weights the benefits that you would get from it. The base station needs to be placed at some location for a few hours without any obstruction such as people walking nearby or the location of it changing at all. any problems with that causes the solution to diverge and be back to square one.

I do not think even omnistar is worth using in the AVC environment due to similar problems.



Regards,
Borna Emami



Chris

unread,
Feb 10, 2014, 2:04:58 PM2/10/14
to diyr...@googlegroups.com
I don't know Skytraq's Oliver Huang personally, he's just been kind enough to on several occasions answer questions I've sent him in email.  But that alone is unusual.

Also, the S1315F, is really a great GPS chip.

On Friday, January 24, 2014 11:21:52 AM UTC-5, Matthew Mucker wrote:
With further digging, it appears that this project may be a Stytraq project. Skytraq is the manufacturer of the GPS chip.
....

The RTKLIB website, especially the RTKLIB document, has sufficient information on how to properly setup and run using low cost GPS receiver. NS-RAW is compatible to S1315F-RAW mentioned in the RTKLIB document. There is also a forum with much discussion on RTKLIB below for reference:

http://www.rtklib.com/

http://t.cn/8FG6RMr


I am a bit concerned that the founder seems all too willing to incorporate the ideas of the backers, suggesting the prototype hasn't been completely thought out yet.

DON'T WORRY.
 
Skytraq has been working with open source hardware and software developers for a long time. They are in a completely different league than these other crowdfunding projects.

That could be good, in that the final product may have extra features, or it could suggest a roadblock or delay might come up.


I know for a fact that they have been thinking about this for a while because I was probably just one of many people to suggest something like this to Oliver Huang, in email several years ago.

He didn't dismiss it out of hand, he explained the economics of how the market is segmented - now its several years later, and my guess is that after probably lots of people suggesting that they do this, he just said, what the hell, lets do it. 

The Skyraq chips embed a 50 MHz Sparq CPU in them also. I don't know if the users will be able to access that for this price, probably not. But still, the opportunity exists to get the product into a lot of people's hands and potentially, a lot of new products, and really do something quite unique.

I just found this list while Googleing so I am not so familiar with what you are doing here- Is this list focused on plain GPS or RTK-GPS?

Rover is also a term used for RTK-gps where it refers to the mobile receivers used in RTK-GPS- See http://gpspp.sakura.ne.jp/indexe.html

Chris


Jon Watte

unread,
Feb 10, 2014, 3:15:49 PM2/10/14
to diyrovers
I got an update today from the indiegogo campaign.

They are marketing their 100 Mhz (!) SPARC CPU, with built-in FPU, as a general development board with additional GPS capability, which is an interesting take on it. And, yes, RTK support, too. Here's hoping it delivers, else I lost my money :-)

Sincerely,

jw
 




Sincerely,

Jon Watte


--
"I find that the harder I work, the more luck I seem to have." -- Thomas Jefferson


--
You received this message because you are subscribed to the Google Groups "diyrovers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diyrovers+...@googlegroups.com.
To post to this group, send email to diyr...@googlegroups.com.

C. Sirius

unread,
Feb 10, 2014, 5:29:27 PM2/10/14
to diyr...@googlegroups.com
I'm sure you will have a blast. Although I already have two, the additional ability to use the CPU is something Ive always wanted, since its there. I hope I can still get two of them. :)

Check out the FOSS-GPS mailing list for more info on how to use it!

http://lists.osgeo.org/pipermail/foss-gps/




C. Sirius

unread,
Feb 10, 2014, 5:33:51 PM2/10/14
to diyr...@googlegroups.com
You're going to need a good antenna to really take advantage of the Skytraq's accuracy.

If you use a standard patch antenna, they are usually designed to be mounted on a metal ground plane. Using one of the patch antennas you'll probably get okay but not the very best accuracy. There is an article here which is in Japanese but Google Translate translates it pretty well.

http://gpspp.sakura.ne.jp/anteva/anteva.htm


On Mon, Feb 10, 2014 at 3:15 PM, Jon Watte <jwa...@gmail.com> wrote:

Allen Crawford

unread,
Jan 19, 2015, 2:54:48 PM1/19/15
to diyr...@googlegroups.com, siri...@gmail.com
It is unfortunate that that article didn't include any antennas from Tallysman.

Tallysman's TW2410 is a high performing antenna with Tallysman's unique Accutenna technology which has proven to provide superior multi-path signal rejection and tight Phase Centre Variation (PCV).  http://www.tallysman.com/TW2410.php
Reply all
Reply to author
Forward
0 new messages