First flight test with a ublox M8 GPS

[Peter Plischka]

Hi there,

attached a log file from a flight with a ublox M8 GPS and a 3DR ublox 6 GPS.

Between the houses a maximum of 13 Sats were achieved with the M8.
The ublox 6 had reached a maximum of 8 Sats.

The copter was very good in the air in hybrid mode.

The values ​​of ofsl ofsQ magl magQ look completely different than a ublox 6 

@ Paul
It was the version e924da4fec871ec5f0df4c786d9ef2d124f8e0b6  


regards Peter

[Julien Dubois]

Sounds great Peter!

From your log, the flight seems not very agressive and it would be interesting to see how the GPS module works in hard conditions
cf. Andrew's post here: http://diydrones.com/profiles/blogs/ublox-m8-series-gps-glonass-receiver-test

Still from you log, HDop was very bad at the beginning even with 8 sats.

Finally, Hybrid should be usefull to test GPS in hard condition as it's the more agressive flight mode that relies on GPS imo... so your choice for this test was the right one!

Where have you got the M8 module? Is it something like http://www.csgshop.com/product.php?id_product=170

Julien

[Peter Plischka]

Hello Julien.

It is this one:
http://www.csgshop.com/product.php?id_product=171

Unfortunately, here in front of the house is not enough space for a longer flight.
I'll try later on a larger area.

With enough Sats the accuracy is very good. Yesterday it was 15 Sats and HDOP of 1.16 

regards Peter

[john...@gmail.com]

A high number of satellites found is usually a indication of good GPS conditions, but does not automatically translate to a low HDOP value. It is the position of the locked satellites in relation to your GPS receiver that decided how good the horizontal resolution will be (triangulation). A high satellite count means that you have a strong lock and there is less chance you will lose the GPS solution during flight. And in practice it usually results in low HDOP, but there is no guarantee.

- JAB

[Cornel Fudulu]

I'm playing with an M8N myself, and I also saw more sattelites but worse HDOP around the house.

I'm waiting for a NEO-6 GPS with external LNA and SAW to swap the M8N module in, because I can't find a config option for the internal SAW filter.

[Paul Riseborough]

HDOP is supposed to represent the ratio of horizontal position error to range measurement error

http://en.wikipedia.org/wiki/Dilution_of_precision_%28GPS%29

and is not an absolute accuracy (for example two receivers could have the same HDOP, but one could have better range accuracy and therefore better positional accuracy). HDOP as reported by UBlox does not appear to conform to this convention.

There isn't much value comparing HDOP values between different models of receiver. A better test would be to put two receivers side by side in a static test (with the same level of  motion filtering selected) and compare the movement in their solutions over time.

[Randy Mackay]

 

     I wonder if it might help if MichaelO added the ability to graph the 2^nd GPS’s positions in the KMZ/KML files that the mission planner produces or when you click on the “show map” in the dataflash log viewer.  I’m not totally sure if it’s necessary or if it’s only something that developers really care about and it would just add complexity and confusion for normal users.

-Randy

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Peter Plischka]

Hi Randy,

I think it's a good idea if we could graph information in the dataflash log viewer for both GPS Modules .

So then everyone has the opportunity two different GPS modules, or two identical modules in another location, easy to compare.
This function needs only in the advanced view available.  

 

regards Peter

[Peter Plischka]

Here are two logs with the M8 and a ublox-6 module.
Once a fast flight and then a longer loiter.

With Loiter 19 Sats were obtained with a HDOP of 1.19.
In fast flight there were 18 Sats with a HDOP of 1.06

The video shows that the hexa is very good over a period of 11 min on the spot.

 

https://www.youtube.com/watch?v=bN8JwNQk_a8

Compared to the flights with the ublox-6 module I am very satisfied.


regards Peter

[Julien Dubois]

Very good Peter, thanks for the test!

I've never got so many Sats netheir so low HDOP with my ublox-6. That sounds very promising as I had  compared the position hold on the APM + Ublox-6 VS Naza + DJI GPS and APM was >= even with strong wind.

So, now, if GPS is again better, that will push APM on top for sure!

Can't wait to try it and see how it behaves in the worst cases

--
You received this message because you are subscribed to a topic in the Google Groups "drones-discuss" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/drones-discuss/r7gREVd3Ilw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to drones-discus...@googlegroups.com.

[Peter Plischka]

Julien,

I am looking forward to your test results.

 

The delivery from csgshop.com to Germany took only 4 days.
I'm going to order a few more for my other Copters.

regards Peter

[Tom Coyle]

Hi Peter,

Did you have to load this new uBlox GPS with the 3DR configuration firmware to get it to communicate correctly with your APM?

Regards,

Tom C ArduRover2 Developer 

[Peter Plischka]

Hi Tom,
No I have not. It is much easier to adjust the few values ​​manually.
I have:

PRT (ports)	UART 1
	Protocol out	UBX
	Baudrate	38400
RATE (Rates)	Measurement Period	200 ms
	Measurement Frequency	5 Hz
NAV5 (Navigation 5)	Dynamic Model	Airborn 4G
	Fix Mode	3 - Auto 2D/3D
	Min SV Elevation	10 Grad
	DGPS Timeout	120 sec
SBAS	Subsystem	Enabled
	PRN Codes	EGNOS (EUROPE)
		120, 124, 126, 131

 

regards Peter

[Tom Coyle]

Thanks for the response Peter, much appreciated.

Regards,

Tom C ArduRover2 Developer

[Peter Plischka]

No Problem Tom,

I am very happy with the M8. I get even in difficult situations (between trees and houses) more than 11 Sats. HDOP is between 100 and 200 cm.

With the ublox-6 I had only 6 Sats and was not as happy with the Loiter and Hybrid.
In an open area there are about 20 Sats.

Meanwhile, I have loaded all my copter with the M8, as the backup there is still a ublox-6 installed.

 

regards Peter

[Tom Coyle]

Hi Peter,

I am presently running a two GPS configuration (#DR uBlox LEA-6s) on my Pixhawk equipped Traxxas Slash rover.

I live in southern Florida, USA and typically can get 11 to 12 sats with an HDOP as low as 1.2 when I am at my parking lot test course.

I can get up to 10 sats and a HDOP of less than 2.0 inside my house in the afternoons:-)

We are using two GPSs mainly for redundancy since one GPS will usually maintain a lock if the other one momentarily goes out of lock.

I am hoping that the M8 will get the sat count up even higher and get the HDOP close to or less than 1.0.

Regards,

Tom C ArduRover2 Developer

[Peter Plischka]

Then yes you live in a good GPS area.
With the ublox-6 I've never had more than 9 Sats in the wild.
I usually fly in an area with houses and get to 5-6 Sats.

For me, the M8 is a great help.

I have ordered one with the helix antenna. and look forward to the comparison with the patch antenna.

Believed to be the helical antenna will not be as well suited for Copter or airplanes because of stronger directivity.

Could be interesting for your Rover.

 

regards Peter

[Julien Dubois]

100% OK with you Peter, helix has too narrow beam for copter/plane imo.

I had ordered the M8 with patch antenna you linked me but, Lucky as I am, it's been damaged during shipping (that's what answered CSGShop to explain why I've still received nothing).

So, I wonder if they are serious seller and if I'll get the item one day...

BTW, I get up to 12-13 sats with my Neo-6m single GPS with HDOP between 1,7 - 1,9. worst cases, 7 sats.

 

regards Peter

--

[Peter Plischka]

Hi Julien,

but that is a shame. The shop sents unfortunately only in a plain paper envelope.

Whether they are serious, I do not know. There is no Impressum in the shop and mails are not answered.

Another shop for the M8 I have not found it yet.

 

regards Peter

[Łukasz Wasik]

Peter how you configure that GPS? With 3DR configuration file?

[Peter Plischka]

Hello Lukasz, 

for the few values ​​you do not need a configuration file. 
You just have to change the following values: 

[Łukasz Wasik]

Thanks!

[Paul Riseborough]

With the extra satellites available with this receiver, it may be worth experimenting with a higher mask angle. There is an interesting paper which looks at the optimum mask angle and concludes that with the addition of Gallleo and GLONASS, the optimum angle is higher than it would be for just GPS.

http://gracegao.ae.illinois.edu/publications/conference/2013_ION%20GNSS_Liang_RFT%20mask%20angle.pdf

[Ben Nizette]

Huh, interesting. I've been doing some experiments with some
GPS/Beidou receivers and came to a similar (but far less rigorous!)
conclusion. In our case, dynamic maneuvers were putting low-elevation
satellites out of the FoV of the antenna, causing jumps in the
position solution. By increasing the software mask angle we can
simply ignore the SVs that are at highest risk of this and exclude
them from the solution entirely. Essentially we're sacrificing
accuracy for consistency but with enough sats, the trade-off seems
acceptable.

Ben.

> --
> You received this message because you are subscribed to the Google Groups
> "drones-discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an

[Peter Plischka]

Thank you Paul and Ben,

The difference of 5 degrees to 10 degrees in practice, be very difficult to detect.
I'm about to read the Document from Paul.

regards Peter

[Andrew Tridgell]

Hi Paul,

> With the extra satellites available with this receiver, it may be worth
> experimenting with a higher mask angle.

makes sense, yes

We should add a GPS_ELEV_MASK parameter I think, with some reasonable
default (5 or 10 degrees maybe?). That would allow users to tune for
terrain too.

Cheers, Tridge

[Julien Dubois]

Hi Andrew,

and what about a real time autotune? the controller adjusts this mask value to get a decent accuracy with the highest mask value possible.
In worst case, the mask is limited to 5 (no change from current perf)
In other cases, we increase consistency while keeping the same accuracy we currently have.

Julien

[Peter Plischka]

That's fine, Andrew.

 

Due to the frequent plugging the GPS units, the DF-13 connectors are on PIXHAWK gotten loose.


regards Peter

[Łukasz Wasik]

Someone said that Airborn mode is better for fast planes. Pedestrian mode is better for copters. Is it true?

Setting NAV, NAV2 and NAV5 to Pedestrian mode cause better accuracy for position hold.

W dniu wtorek, 3 czerwca 2014 21:19:49 UTC+2 użytkownik Peter Plischka napisał:

[Peter Plischka]

It will be hard to see a difference.

 

You can even try out different settings and show us the difference. 

regards Peter

[Łukasz Wasik]

OK this is the best config. Max HDOP 1.3 on the 2nd of the 4-floor building, 3m from window!

[Ben Nizette]

On 6 June 2014 04:47, Łukasz Wasik <pankr...@gmail.com> wrote:
> Someone said that Airborn mode is better for fast planes. Pedestrian mode is
> better for copters. Is it true?

The defaults (used to be?) Plane used Airbourne 4G, 'Copter was
Airbourne 1 or 2G, Rover was Automotive. Tridge and I (mainly
Tridge!) did some testing 6 months ago and found that all three
probably should be using Airbourne 4G. 4G is the closest we've got to
telling the GPS not to filter at all. This is important for us
because 1) we have our own filters (INAV, recently EKF) that have more
information and can filter much more effectively than the GPS and 2)
because filtering inside the GPS introduces latency that degrades the
performance of our filters.

Things like "better accuracy for position hold" might be true if we
didn't have INAV (or if your frame vibrates heaps, in which case fix
that!), but in the general case it's best to just let INAV do its
thing.

Ben.

[Peter Plischka]

Ah, thank you Ben,

I had not considered that.

But we can still experiment with the Min SV Elevation Mask.
I would be glad if Andrew makes it a parameter.
That would help a lot.


regards Peter

[Ben Nizette]

Yep, for sure. It's only recently that the code's been changed to
allow parameters attached to GPS, it should be an easy-enough coding
task now.

Ben.

>>
>>
>> regards Peter

[Paul Riseborough]

Flew the M8 today on a 3DR quad with a combination of throwing it around in hybrid mode, auto and loiter. It got 14 to 15 satellites and during the 2x15min flights experienced no glitches. Unfortunately my Pixhawk did not log the 6H plugged into the serial 4/5 port, however I have never got more than 11 satellites using a 6H and I have never done two consecutive flights without a glitch of some sort so it appears to have outperfomed the 6H in that area.

Looking at the EKF data I think the latency on the M8 data (relative to the IMU)  is different to the 220msec value set by the EKF_POS_DELAY and EKF_VEL_DELAY parameters, which were optimised for the 6H. This is causing the GPS innovations to rise during manoeuvres. I will have a look at finding a better value for EKF_POS_DELAY and EKF_VEL_DELAY using the replay facility.

I was using a 10 degree mask angle, turned SBAS off (it can cause trouble in Australia) and was using the aircraft <4g motion filter setting.

[Craig Elder]

Thanks for the great news Paul

Did you have GPS_TYPE2 set to 1 or 2?

--

[Paul Riseborough]

I had rolled back to a previous set of params and forgot to check it. Looks like it was set to zero for those flights.