So I read an interesting article about GPS antennas called "Adding a GPS Chipset To Your Next Design Is Easy".
A few points to bring up that I have concerns with dealing with my M8N antenna.
1. Active vs Passive Antennas. Two paragraphs within the article describes the difference between Active and Passive antennas. According to CSG Shop's specification for the NEO-M8N it comes with a low-noise regulator and RF filter built-in. So I'm assuming that it is a active antenna.
2. Antenna's requiring adequate plane. If I read that document correctly, these GPS modules may require a GPS plane as they are installed on a PCB that does NOT have 40mm of side to them.
Quote: "Generally, patch antennas in the 15- to 25-mm size range with a least a 40-mm (on a side) ground plane will give the best performance in portable equipment, but this may be too large for your application. This could force you to look at smaller antenna topologies such as linear chip antennas."
3. The next concern is to mitigate the noise interference from FC, ESCs, and PDB. Since my Y6B is set up with a clam shell cover and my M8N is attached under and close to the all the electronics, I may need to develop a shield "ring" connected to the shield can and then connect that ring to RF ground through an inductor at a single point.
Quote: It's common in VHF and UHF RF shielding to connect all points of the shield can to the PCB's ground plane. This can be a mistake at GPS frequencies, since the open-air wavelength of a GPS signal is so much shorter than UHF. Depending on the size of the shield can, if there is current flow across the can, the shield can will be able to resonate near GPS frequencies resulting in interference or de-tuning of the GPS RF.
By developing a shield "ring" connected the shield can and the inductor, the inductor will filter any EMI-induced current flow. The ring connected to the shield can will prevent any current flows or resonation issues.
I'm not an electrical engineer and need guidance from those out there who are. Did I interrupted this correctly? and if so I could use some help with developing the "ring".
I'm beginning to understand more and more on these units. My first M8N was a 40mm antenna sized unit, but the latest ones I purchased are 25mm. So I now need to do a side by side comparison to see for myself that the 40mm antenna is better. However, the PCB board also needs to be larger to accommodate the ground plane. Without the larger ground plane, it very possible the back lope from signals or EMI interference can degrade front lope signal quality thus reducing overall performance.
Not to sound disgruntled, but I brought this matter up on January 31, even though it didn't start out about INAVErr. It garnered little interest for over a month. I wasn't looking for a fix, just wanted to know if it was a problem and why it's buried under PM and why, if 255 is bad, why it isn't flagged as a failure in the diagnostics drop down.
There are several discussions going on at RCG about this issue in various threads trying to find out what the number 255 represents and who was having problems using the M8N. Honestly, not one person has reported problems related to high INAVErr numbers. Some, or many by now, are concluding the INAVErr is as big a red herring as the loop errors because nobody has actually reported a crash or flyaway due to high INAVErr numbers. Of course up until recently very few knew it existed, so that may not mean anything.
Craig Elder said he's been trying to explain for months to people the M8N in it's bare form has problems, but I could not find any useful information concerning INAVErr before first posting two threads on the subject in late January. I literally had zero knowledge about INAVErr until later on, but still haven't seen the section of code that calculates the values. Fortunately a few others reported high INAVErr (and gaps in the time stamp/KMZ) and finally thanks to AKCopter's one simple question in an unrelated thread, Craig Elder spelled it all out......mostly.
At RCG several times as well, when it particularly came to a head in the RCTimer M8N thread here
Also at Arducopter Support forum- zero responses after two posts. Bionicbone, creator of APM Flight Log Analyzer also posted last year with no response.
What really bothers me is CSGShop brushed it under the rug when I emailed them weeks back and their knee jerk response was to blame Pixhawk, and didn't even attempt explaining why.
Before considering any mod, I want to know what could and "should" happen getting these high INAVErr numbers. Why is it actually a problem? Seriously. I still can't prove the occasional twitching is anything but a tuning problem. I've had the copter stay dead still for 2 minutes several times ~6 ft off the ground, all the while with INAVErr holding at 255. How do we know it's electrical noise from the other components and not just noise from the added satellites (TMI)?
Apologies for the rant. I understand and appreciate the hard work the devs put into Pixhawk, which I really like a lot. It's probably asking too much for them to get more involved, but IMO the LEA-6H has run its course and needs to be replaced by a better GPS system.
Read this post at RCG. He had two completely diverged INAVErr results with two different M8N's, but also no problems with either in the flights.
I recall you discussion on inav errors. At this point I feel we need to wait and see how the DEVS hammer out their red herring issue before pressing them further on excessive error reporting. For now, do you have any issues with Loiter, Auto, or Pos Hold?
My concern is with GPS interference at this time. Seems after reading u-blox's document on GPS designing considerations(Thanks Richard (DK) for providing that!), that developing and implementing a GPS system is not as easy as it looks. Many considerations of which two of them I identified should be considered to ensure the best quality reception of satellite signals. I think your issue is addressing a communication issue between the NEO-M8N and Pixhawk using the latest firmware.
Curious... What size antenna is your M8N? On my Y6B I have a 45mm antenna and on my Antenna Tracker I have the 25mm one. In this case size does matter in reception. My Y6B is also the dual band with no LNA where as the AT is using a triple band with LNA.
I'm planning to purchase another M8N triple band with 45mm ceramic antenna and strip off the one on my Y6B to perform side by side testing this next month. My plan is to run tests for 24hrs and perform analysis of both antennas for a baseline. Once this is completed, then I plan to conduct another set of tests using ground plane shield and maybe look into other options to improve reception and possibly mitigate EMI for onboard electronics such as FC, ESC, PDB, etc...
Your post is summurazing very well the situation :
today, after almost 1 month and a half you discovered this INAVERR problem (only reported in arducopter logs),
it is not possible to say is these INAVERR errors are pointing to a real functionnal problem that could cause crash or fly away during flights
With M8N GPS
Sometime INAVERR goes up to a max value of 255 (probably a 32 bits counter)
Sometimes goes up around 20
With 3DR GPS
This INAVERR has also been reported by several users:
I've started looking back at my logs, just looked at 20 of them for now, with flights > 15 minutes. Here's what I've found so far. Most have INAVerr > 0, with InavErr varying from 1 to 40, straight horizontal line suggesting the error happens before flying or right at the beginning. 19 logs out of 20.
I've got one scary one. Inaverr starts at 0, then goes to 40 about 30 seconds in the flight. Stays there for 5 minutes or so, then ramps up again to 100 right during the flight.
This was a while back and I don't remember the flight, except I know I had no issue. Go figure. This is quite disturbing.
I'd fall in that category. Great pos hold yet some errors.
Going to look into this more carefully from now on, + got lots more logs to look at.
I looked at many logs from the Y6B that had several >0 INAVErr but no indication of flight problems in missions.
Ha! My INAVErr is 255 start to finish with the current setup using the M8N, so I'd be in the category of "I'd be concerned". Of what I don't know, but those that say they've been using them without issues either aren't pegging at 255 or are playing Russian Roulette.
I looked at earlier logs with the 3DR GPS and it averages around 30 or so. Some have 0.
CSGShop advertised the M8N as the greatest thing since sliced bread and probably sold a large number of units. As they say Caveat Emptor.
What I'd like to see now is if and how people are making them work without high INAVErr. If not, then I'll just go back to the 3DR unit and live with not flying on some days.
I would suggest to open a problem in Github for arducopter code ( if not already done):
* Randy says it is a problem
* Last sentence below seems to point to an arducopter code problem
* As this INAVERR is discussed on several thread on DIYDrones and RCGroups, it will allow to gather all informations in one place and allow everyone to better foollow this INAVERR problem investigation
Reply by Randy on
an important reason to open a problem in Arducopter code is that INAVERR are also reported with 3DR GPS
I concur. However I believe your comments and concerns should be posted in DG's post here.
I set this discussion up to address EMI and RF reception issues.
OK Doug, will do that
I replied to DG as it was posting here about INAVERR errors from which all this started
Remenber that Craig Elder spoked about potential EMI / RF reception issues with M8N
after some discussion with UBLOX
If INAV Err is occuring with 3DR GPS, that does not necessarily mean there is a problem in Arducopter code. It's more likely to indicate the code is doing exactly what it is supposed to, which is filtering out implausible GPS data before it upsets the inertial navigation. Those errors are likely happening to EMI on the GPS.
He did and We have been talking offline about this. I've kind of decided to step up and help from a DIY perspective and not as a representative of 3DR. Those folks are eyeballs deep in many things and I feel as a hobbyist, that we can hammer out some of the issues together and provide value added feedback to 3DR and the forum.
Like I mentioned earlier. I motive at this time is to take a excellent product and hammer out the interference issues or at least identify solutions for all to consider when building and incorporating the M8N into their airframe.
And for the record, I'm also open to testing other GPS units but at this time, I've tested the LEA-6H, NEO-6N, NEO-7M, and NEO-M8N and I believe for my applications that the M8N is the better unit. But with all units, it really depends on what you're using it for, where in the world you live, and what you are willing to accept (quarks) with each unit and vehicle. I know there are others, but I have only so much capital to play with (mom gets the rest) and I tend to be somewhat selfish in picking and choosing what I want to purchase and test.
As I learn things, I won't be afraid to share my findings with all.
You bring up a valid point. It's quite possible that the EMI from the FC processor and possibly ESC frequencies could be interfering with the data communicated back to the FC thus it's possible that one is causing a cascading increase in iNav errors.
Could you relocate your GPS unit to a tall mast at least 50cm from FC and retest. It would be a temporary but could help us identify if the relocation far away from the airframe electronics helped to reduce the iNav errors? You may have to build a longer than usual cable.
Also may consider different lengths as well to show the different iNav errors.
The hope is to:
1. Identify if in fact an EMI issue is the culprit.
2. If so, at what distance away from airframe is optimal and practical.
3. And, if it's true this will help me with my project to develop a shield to ground out EMI issues in hopes that we all can keep the GPS close to the frame if possible.