GPS module showdown - BN180, BN220, BN880, 7M, M8N, M8P

Main results at 21:58
*** Please see follow-up video with correct data: • GPS module showdown do...

[@WizardTim]

I’ve used the BN-880 quite a lot and really like it (especially for the price and availability), gets much better performance than the 7M in my experience especially with obstructions (although note that here in Brisbane, AU we get on average 10 more GNSS satellites visible with better DOPs than NZ). Although last year I bought a preproduction sample of the ublox 9 series which does 4 concurrent constellations which is crazy impressive, not sure if there’s any premade modules for it though and of course being export restricted buying them officially is a pain. Also, you said your axis was 1/100,000th of a degree for lat/lng, how many decimal places are you recording and at what float precision? When I have to deal with high precision lat/lng data I almost always use integers for the decimal places and often to 9 places (1/1,000,000,000th of a degree) otherwise I find too many errors are introduced from the limitations of floating point numbers. The gaps in your heatmaps and blocky line plots make me think you aren’t using the full resolution those modules can output. Also, in case you didn’t know the Beitian modules use receivers from ublox, the BN-880 uses the M8030 which is what the M8 modules use.

[@iforce2d]

I'm using the NAV_PVT message which provides a 32 bit integer representing lat/lon * 10^7 That number is then divided by 10^7 to get a 32 bit float representing degrees. There may be a small precision loss in the division, but I can see from looking at the original integers that there are still only the same limited number of unique raw values available. The 7 decimal places doesn't necessarily mean that the module actually gives 7 decimal places of real precision. I would use double precision if I needed higher resolution and the GPS module could actually provide it, eg. for something like this: kzbin.info/www/bejne/eZDMgnl8fMyLhqM

[@WizardTim]

@@iforce2d Interesting, I usually use the GGA NMEA message, most modules let you configure the precision from 1 to 8 decimal places (although few actually have 8 full decimal places of unique codes). Looking back at my BN-880 datasets they have a resolution of 1.6e-7 degrees which is about 1.5 cm. Considering you plotted the points on Google maps and measured a rough spread of 3.3 m and then showed the scatter plot at 8:37 with approximately 33 unique points horizontally it would be fair to estimate your resolution is closer to >1e-6 degrees (>10 cm). This doesn’t change the results of your showdown but might be causing some of the oddities with your data visualisations. Just did the math for it, the 32-bit float is your limiting factor for resolution and is why your longitude values have less resolution. Storing -40° your error will be 1e-6, storing 175° your error is 5e-6.

[@iforce2d]

I wrote a long post about how you were all wrong, but I became less and less sure of myself as I typed.... It might be worth doing a test where the sbusmixer writes the raw integer to the file instead of a float. This will be necessary anyway if end users want to use an sbusmixer with a GPS that does provide more accuracy. If I get a different outcome with that method, I'll have to do this whole comparison all over again.

[@iforce2d]

Yeah you're right. Gonna have to make a follow-up for this.

[@WizardTim]

@@iforce2d Sorry for not being wrong and ruining your fun XD. I’ll also mention that I have a couple geodetic markers near my house that I’ve tested GNSS modules with, not sure if you’re near any but it might be worth checking, it’s publicly available on: . I’ve found the Google maps imagery can be out by a fair margin even in greater metro regions let alone the sticks.

[@RobisonRacing68]

We had an eclipse the other day. That last animation was so cool! I mean, it was all cool but the animation really brought it home. This was cool Chris.

[@KenColangelo]

Thanks for the comprehensive video. I'd love to see what would happen if all of the antennas and logging equipment were on a turntable with the antennas raised up as much as possible away from all of the circuitry to ensure isolation, and then the turntable rotated in one direction or another. I've seen novel behavior from GPS antennas that are simply rotated without being moved.

[@Etyop]

I love your videos. They are so pleasant and relaxing to watch, all by being very comprehensive. Thank you

[@goobisoft4873]

Awesome video loved it i used m8n on my arducopter powered drone and i also tried i mav but unfortunately inav was giving so many trouble even with so much tunning it didn't hold well but the arducopter performance with m8n getting around 15 sat mostly was jaw dropping for me atleast for how accurate it did pos hold rtl and circle mode it was flawless loved m8n and arducopter

[@stevietee10]

I use the 880, flying in the Midlands, UK, getting 22-24 sats with no problem, love it :-)

[@dempa3]

Very useful comparison! Thank you! By the way, I have many times wondered why you haven't pursued a DIY HD FPV system solution, like OpenHD or EZ Wifi Broadcast, for some of your models. It seems like something you and your audience would find interesting.

[@TheWebstaff]

Thanks for the tip!

[@lasersbee]

Wow... Great showdown and tests... Always wanted to see the differences in those modules. Thanks.

[@1Esteband]

Thank you for such in depth analysis. I always wondered and you answered every question I had.

[@blacklion79]

uBlox rocks for many years now!

[@tinplategeek1058]

For once I actually remembered that in the UK we had a partial eclipse last Thursday morning and the clouds stayed away for once. Great to catch one.

[@mike_oe]

I've had several BM180s die on me. The still communicate via Tx/Rx, but are unable to get a lock. And they heat up. I've started to use 4.3V instead of 5V

[@MertUluc]

Really interesting work! I remember the car top gps timelapse you did a few years back (M8P?) and the accuracy was always around 1.2 - 1.3 meters. I thought the consistency was impressive. That's only slightly worse than an expensive handheld garmin gps unit when you leave it outside for 24 hours. Most of the time it pins the location to 2-3 feet accuracy very quickly and stays there but occasionally I can get it to report 1 foot or so overnight.

[@jeffmerlin2580]

You should have used ECEF (X, Y and Z) instead of Long/Lat. Probably not that easy to get from the GPS, but uBlox devices can output ECEF in binary mode. ECEF=earth-centered, earth-fixed, and gives a 3D coordinate where 0,0,0 is the center of the earth. "Earth-fixed" means that the coordinate system rotates along with the earth. ECEF is really all you need. The altitude can be derived if you know the shape of the earth. Not great news for the Flat-earthers by the way. ;-) For a stationary test, you just need to know the distance from sea level to the center of the earth and use simple math. Stationary ECEF coordinates gives a "ball" of points. Long/Lat have different scale and that combined with the resolution is why you see the "weird" measurements. Great video by the way!

[@flat-earther]

hi jeffmerlin, have you become a flat earther yet? If not I suggest watching _Level with Me (2023)_ by Hibbeler Productions

[@Terry_Fritz]

I usually know when an eclipse is happening and more than 50% of the time it's overcast. That night here in Tasmania it was perfect. Your luck is amazing. I have two Beitian BN-880Q and they give identical results with the units up on masts 100mm above the top plate of the quads, well away from interference. On a clear day 18 satellites is normal and picks them up from cold in under a minute.

As always, a great video with a lot of effort, thanks for that! I'm looking forward to the next one... Keep it up!

[@tenns]

nice little rocket lab easter egg

[@bbjunkie]

An interesting follow up to this would be to force the BNxx modules to use GPS only and see if the results changed at all.

[@trevorvanbremen4718]

I'm glad I read some OTHER comments first before asking WTF you were doing in Mahia... LOL

[@Gosuminer]

Thank you for this thorough experiment. The Beitian modules can also use the Galileo network when configured accordingly (Betaflight has the CLI command ublox_use_galileo) which would give you additional sattelites. I don't know if this would improve accuracy, though.

[@bradleyr4451]

Hi nice video, wonder how the accuracy changes for about a week of time with sat's moving around; But will this or any other small uC board of this class do the RTK correction in realtime using correction station info?? Have you made a station or have one near by? Did this years ago but it was really up to the antenna, use a cheap one and no Fix, good $$ one and pulls in much better. Wondering where that was such a nice area reminds me of home

[@landlocked4771]

That was really interesting thing to see, the animation at the end is really good. Thank you for sharing.

[@kermets]

Great video Chris.....I have notice my M8 doesnt get many Sats but still works Great

[@ryanthalman2904]

Also for geographic heat maps, consider using QGIS

[@changeagent228]

The end part with them moving is useful to help visualise the accuracy performance. I primarily run TBS M8 v1's now after having a ton of issues with M8N's losing battery charge and cold starting every time. Also even with a good battery they would sometimes take a couple of minutes to fix home. When I switched to the TBS M8 it finds a home fix in less than 30 seconds consistently. It's also glonass only which I wonder if that helps when I read about gps jamming lol. 🤔

[@윤상황-p9h]

Thank you for good information. It would be even better if you tested the same product two of each at same time.

[@timvukman4734]

Thank you for taking the time to test all of this. It is very helpful information

[@georgef7754]

Great video thank you. I am still playing with a 6M.

[@alexscarbro796]

The discolouration on the BN220 antenna looks like simple oxidation of silver plating. This will likely have no impact on the performance of the antenna.

[@sUASNews]

Shew you know your stuff

[@argentinomacrifuevidaltamb3772]

Excelente trabajo. Gracias por compartir. Saludos

[@AerialWaviator]

Thing you could avoid the whole sizing of latitude, longitude (pixel) issue at given position on the globe by using a distance, bearing from a reference location. ie: relative vs absolute reference. Knowing data is within X meters of an actual positions is more useful that knowing its +/- 0.xxxxº of latitude, longitude.

[@AerialWaviator]

Fascinating findings, and explorations. Great explanations. I find the number of satellites to be of less interest. More interesting would be which satellites each GPS module is seeing / using at a given period of time? (Or, which satellites is a GPS module rejecting from its calculations? This might be an indicator to quality and complexity of software on a given module) Generally the higher in the sky GPS satelites are to a given earthly location, the more precise the time signal, as it traveling through the least amount of atmosphere and that susceptible to less random distortion effects. [ Side note: GPS satelites far apart with low communication paths to each other are actually being used to study Earths atmosphere thickness and properties in real time. The time signals are compared with modeled data for atmospheric lens and ionic effects ]. So as you concluded, more GPS satellites does not mean better data. You might try using an opaque plastic storage container as your test housing. Using the lid as a base, and the opaque container as cover. (a fan in the bottom would help to reduce humidity from the warmer daytime air that is trapped. I'm guessing in your case, the wood held moisture that was being released as it cools. Using a non-water holding base should help solve the dew annoyance) For a more precise reference position, you might check to see if any land-survey markers at the corners of property lines in you area. Also if any historic points of interest nearby, they often have a documented precise position. (as do road intersections, but require more time to research). An airport, or airfield will also have a precise location and altitude marker. (only useful if wanting to compare data to a very precise reference of +/-1m vs +/-10m) If curious ... you might check if there is any value in using the average of pair of GPS modules (of similar type) placed a meter apart (like if mounted on each wing). Would an average position value be any more accurate vs. the individual data stream? (lat/long/altitude) Mentioning, as professional land-survey equipment tends to have dual GPS units.

[@ryanthalman2904]

You only need 4 satellites for a 3d lock. Beyond 4 more satelittes it calculates all of the possible combinations and averages those to get the "real" value. As another commenter mentioned, it is of more value to compare using the same set of satellite systems. Or run a test on each set separately. There will probably be disagreements between the glonass, gps and Galileo systems. I also wonder if you system interprets or has any of the WAAS (wide angle augmentation system) corrections built in or interpreted (these generally get broadcast along with the signals to perform corrections. The best way that i can put it would be more than 4 satellites means that if one satellite drops or you miss reception, you don't lose information and position (it's about redundancy).

[@GiBBO5700]

I see you work at rocketlab now! Nice Easter egg lol

[@tymoteuszkazubski2755]

I think better units using fewer satellites might have to do with better rejection algorithm. They can get a better and more consistent fix with fewer, but higher quality signals by consistently and reliably rejecting signals that would lead to fix degradation.

[@stephengloor8451]

Latitude is the direction that the distance between degrees is changes as the circles get smaller going up in latitude. Longitude stays the same.

[@iforce2d]

Latitude measures angle away from the equator, ie. horizontal lines when looking at a typical map. The distance between one degree of latitude and the next is always about 111km. Longitude is the vertical lines, which all come together at the poles.The distance between one degree of longitude and the next is about 111km at the equator, and zero at the poles. en.wikipedia.org/wiki/Longitude#Length_of_a_degree_of_longitude

[@linecraftman3907]

8:00 might want to hide those coordinates in the url bar!

[@avejst]

Great test Thanks for sharing your experience with all of us 👍😀

[@Qpwrtm]

Lower std for the longitude might be connected with the orbit inclinations relative to your location? Like maybe satellite is passing mainly across one of the axis and you get better precision with that

[@iantcroft]

Really good test thanks. Would you test other GPS modules to compare and show which is most accurate? The Quectel modules are low power, acquire fix quickly and low cost, especially the L70, the L80 is not bad either. Interesting how more satellites doesn’t necessarily mean higher accuracy. Thanks.

[@RandomTorok]

Hi there; I've been watching a lot of your GPS videos. I was given a Beitien BN 880 and I created a GPS out of it. It always seems to have the Time, Lat and Lng correct. However the Alt and Sats can be all over the map. I live about 200 feet above sea level but at times my GPS will be showing 8000 feet and then a few hours later it will be showing 97 feet. At times the number of Sattelites will be in the thousands. Any thoughts on what might be wrong. The whole rig is on a window sill that faces north and the gps module is pressed to the window with some blue tack. I've tried both an Arduino Uno and a Mega. I'm interested in any comments you or your viewers might have.

[@iforce2d]

Sounds like you are reading some values as the wrong data type, eg. signed instead of unsigned, or from the wrong location in the message structure.

[@TheWebstaff]

Haha yeah not all modules are equal. They come with no flash, some flash, loads of flash. There are loads of clones and loads of customs modules too. I remember the fun I had messing around with loads of these modules. I ended up with a ublox 6 module being one of the best. It was also the only one I pulled from an OEM device which could mean it was the only real one. Have you had a play with the ublox software it's got a hell of a lot of configuration options. Including enabling other sat configs. Oh and you can enable the marine radio mast stuff so you can use land beacons as well as sats

[@EinzigfreierName]

These test conditions are pretty ideal. So the total number of satellites is not that important. 7-8 nicely spaced all over the sky are already enough to give you a solid fix that will not really improve with more satellites received. Things look different in real world conditions where you have obstacles blocking parts of the sky, interferences and mulipath situations. If you only have 7 satellites and lose 2 of them, accuracy will go down a lot. That's basically the big advantage of these newer modules like the Neo 8M and 9M that can receive multiple satellite constellations simultaneously. There are so many satellites to choose from that it is much less likely to end up with only 4-5 of them being received and accuracy going down or even the fix being lost. Also, the antennas make a big differences.

[@MessingWithCode]

Thanks So much

[@hoggif]

I wonder how much different gps systems make a difference. At least on my BN-220 Beidou and Imes were enabled by default. Needed to disable both to get Galileo support as well. (You cannot enable galileo without disabling Beidou and Imes. Those are probably usefull mostly if you're in or near China and Japan).

[@evo-labs]

Whoever named Arduino code a 'Sketch' needs to be publicly shamed 😁 interesting results though, thanks!

[@tk-maker]

can you test on these GPS module ATGM336H GT-U8, GT-U13, VK2828U7G5LF (i use v.kel for autonomous drone, seem ok)

[@brkfpv60]

Nice comparation thank you

[@ПавеЛСемашкО-ы5ц]

Mega research! Cool bro!

[@skycrewfpv1896]

Haha, you got me :-) RocketLabs !!! well done, well well done :-D

[@CatzRuleZWorld]

If you do this again, it would be cool to include the Matek M8Q. I assume it should be similar to your M8N but who knows.

[@iforce2d]

According to the datasheets the M8Q module itself looks very similar to M8N. The main difference would probably come from the antenna on that Matek one, which seems a bit different to most of these ones I have - the ceramic looks kinda thicker?

[@tntv8834]

I have the an220 gps, what more I need to buy for connect it to my FPV airplane?

[@iforce2d]

You would need a flight controller, eg. that can run Arduplane or iNav.

[@bostonmacosx]

Hello there....I was wondering what is the overall accuracy of these devices. I need to get a ground RAV to a specific spot in my yard I was looking at the ublox m9n possibly... I need to be within 3-5 feet.. of where I want it to be and can't afford the $200+ for RTK stuff.... Can you advise?

[@iforce2d]

It's unlikely you could count on that reliably from any of these modules. Here are some results from further testing I did, after correcting precision issues. This is a 10 hour test, distances are in meters. Keep in mind this is with the best possible conditions in very rural location, and well away from buildings. www.iforce2d.net/tmp/gpsFaceoff2Distances.png The M8N might be worth a try I suppose.

[@bostonmacosx]

@@iforce2d People in general have a messed up idea about GPS accuracy...

[@iforce2d]

I posted a follow-up video that might give you a better picture of how the results stack up. kzbin.info/www/bejne/l5OwoIWAYpaspLs

[@DoctorMikeReddy]

Loved the video at the end. How did you plot it?

[@iforce2d]

I wrote a program that generated 8500 .png files and then used ffmpeg to make a video file.

[@jort93z]

10:00 pretty sure that is a heatmap and not a histogram. But interesting video!

[@pablosiga6082]

iforced2d , thanks for the video. Now i realize why ardupilot have issues with some GPS . So what GPS for ardup do you recomend at last?

[@iforce2d]

I use BN220 on almost everything these days. Well... 'almost everything' means planes for me recently. For multirotor or ground vehicle I think I would use BN880 or the M8N puck since they have magnetometer as well.

[@mistercohaagen]

In your crosshair comparisons, why are the pixels not square? Or are they?

[@iforce2d]

9:33

[@mistercohaagen]

@@iforce2d Interesting; en.wikipedia.org/wiki/Mercator_projection#Distortion Especially since it's apparent at such a small scale.

[@markgreco1962]

About how much time do you have in this gps video. GREAT WORK!

[@iforce2d]

Depends how you calculate the time I guess. Started making it right after the 'follow-me' video about 3 weeks ago, I think I ran 6 tests in total, summing to around 24 hours. 6-8 hours debugging and dealing with UART issues and BN180 quirks. Probably about 12 hours fiddling around with spreadsheets/charts and programming the bitmap generator, arranging screenshots and slides etc. A few hours for recording the outdoor video clips and post-narration, an hour or two for editing. Overall been working on the sbusmixer for about 15 months. Sometimes I think it would be nice to be a gaming KZbinr or streamer, where a 29 minute video takes about an hour to make :)

[@markgreco1962]

@@iforce2d impressive to say the least

[@loodwich]

So you record this video on the 26th of May, isn't it?

[@iforce2d]

The first part, yes.

[@loodwich]

@@iforce2d I like drones, but I also like astronomy, and that lunar eclipse was only for Oceania because you are from New Zealand was a good pinpoint. Like your video, thanks for that analysis and the information.

[@iforce2d]

Here's some astronomy related old videos of mine: kzbin.info/www/bejne/m2mzdJicm851mrc kzbin.info/www/bejne/pYbJlIKDfZpkd8k kzbin.info/www/bejne/nmPciq1_hMyHiZI kzbin.info/www/bejne/f3-8XoqQaJeMgtk kzbin.info/www/bejne/iZOvaWeNpcmioZI

[@loodwich]

@@iforce2d I didn't know that, I found your channel when you did the homeslice 1, just when I decided to take back the hobby RC planes, and I found that now is much more interesting... Telescopes, I don't have any now but I can use a 50cm internet controlled made by one of my friends that he made for this Ph.D. thesis... many of my friends are in astrophysics, astronomy, or robotic rovers to Mars.

[@gizfpv]

Haha, satellites are easier to see at night. Hard to argue with that logic

[@spleenware]

Can you paste a link to the logger devices you use?

[@iforce2d]

www.banggood.com/CJMCU-717-OpenLog-Data-Recorder-Flash-Recorder-Sensor-Module-Support-64GB-Micro-SD-Card-p-1461328.html Unfortunately they probably don't have the arduino bootloader so you might need to burn that yourself. Use a UNO bootloader, not pro-mini etc. See the review by 'ram777'. Pinout here: www.thebackshed.com/forum/ViewTopic_mobile.php?TID=11225&P=2

[@skelethornbro]

While all that.. I'm still flying with Neo6m originally for arduino.. :'(

[@bob2859]

Regarding BN180 forgetting settings: your module's battery might be dead.

[@iforce2d]

Although I can't find an official spec sheet, many sources say that it has no flash memory to store user configuration, which matches my experience. I think the battery must be for saving the satellite almanac, otherwise it would not be able to start up in only 30 seconds. There's a comparison table here, not sure where they got it from though: www.amazon.com/Geekstory-Navigation-Raspberry-Aircraft-Controller/dp/B078Y4XZN9 en.wikipedia.org/wiki/GPS_signals#Almanac

[@skyforce1983]

Hi I'm not being rude or disrespectful and not telling you what to do but I am really interested to see how the gps with r2k actually fairs up the reason is because they CLAIM CENTIMETER LEVEL POSITIONING and honestly that's a big super huge claim to make I do own the here 2gnss but I only fly fixed wing can you do a test right there where you just did your testing and actually see what it does cause honestly I'm calling 🐂 💩 for their claim? Thanks in advanced

[@iforce2d]

kzbin.info/www/bejne/eZDMgnl8fMyLhqM kzbin.info/www/bejne/pGmzhqyFqc56gdE kzbin.info/www/bejne/j6Col6SQqpd1d8k

[@skyforce1983]

Holy crap that's accurate thanks for the reply

[@RobB_VK6ES]

Make a surveyor your new best mate. Their GPS stations are the dogs danglies