What is dual frequency GNSS aka multi-band GPS and why is it so accurate

Garmin Epix 2 & Fenix 7 - Is the new 'Multi-Band' the best-ever accuracy from Garmin? or from anyone, for that matter? Must Read: Garmin Fenix 7 Review Must Read: Garmin Epix 2 Review - Best SportsWatch Ever This article looks in detail at 9 performance tests that strongly indicate the true accuracy of Garmin's absolute-best mode of GNSS/GPS accuracy which they call All + Multi-Band. That mode is only found on the Sapphire models of Fenix 7 and Garmin Epix 2 In a nutshell: Garmin has just released the most accurate 'gps' results ever for me on a sports watch. There's still a sting in the tail, so if you want to know what that sting is or if you want to know exactly how good your Garmin can now be then read on. For once I'm going to be very nice to Garmin because they deserve it. This post is an expansion on a similar one from Feb'22 where I looked at the All-Satellites single-band mode for the Fenix 7. Single-band mode is found on all other Garmin models but All-Satellites is new to the Fenix 7 and Epix 2. Background In late 2021, the Coros Vertix 2 debuted the Airoha/MediaTek AG3335M GPS chip and, more recently, the Garmin Fenix 7 & Epix Gen 2 saw Garmin ditch the Sony chip in favour of, presumably, for the exact same Airoha chip that Coros uses. Every new sports watch and GPS chip seems to claim to be more accurate than the previous one but the truth is far from that. Indeed, over the last 5 years or so the focus has been on sports watches using reduced power components that extend battery life. As a consequence, accuracy suffered, my Fenix 6 Pro (six) being a prime example of accuracy meh-ness. There are many factors that determine GPS accuracy, Garmin's latest advances simply reflect a brute force approach to attach to even more satellites and, where possible, to make two connections to each individual satellite that allows them. All Systems (Constellations) & Multi-Band GNSS (Dual-Frequency)? Airoha brings us two things of note that should each increase accuracy on your Garmin - multi-constellation GNSS and dual-band reception. Garmin calls it All Systems + Multi-Band, which is confusing as multiple constellations are used and only two frequencies (bands) are used. Terminology: Constellation = System Frequency = Band GNSS and GPS were previously used synonymously, the former is the correct word. Assume that 'All' Systems means GPS (USA), Russia (GLONASS), Galileo (EU) and Beidou (China - now a global constellation). The source of accuracy from dual frequency Here we look at the impact of multiple satellite systems and multiple broadcast frequencies (bands)...ie the everything-turned-on mode but I won't look at the extra battery juice needed to do that, that's a post for another day. Accuracy is increased by the Airoha chip in two, new ways a) a greater chance of receiving more satellite signals and b) the opportunity to determine which signals are most likely to be accurate. Important: If you want the most accurate Garmin Fenix 7 or Garmin Epix 2 then you must get the Sapphire model. Multi-band/dual-frequency is only found on the Sapphire models even though the chip is probably identical across all Fenix 7 derivatives. All-system GNSS previously meant two systems ie GPS+GLONASS or GPS+GALILEO. Both combinations offer more satellites for your watch to find...the more the merrier when it comes to delivering accuracy in dicey reception conditions. Garmin and Coros's new Airoha chip goes one step further than the previous generation of chips and allows you to connect to ALL Systems. We can take the word 'All' to mean what it says but in the context that there are different constellations available in different parts of the world - some constellations have a global reach and others have a regional reach. Some of the satellites in these constellations have the ability to emit two signals - L1 and L5 - ie two frequencies or two bands, whichever, you want to call them. And some of those satellites have that feature turned on! (Some don't). In one sense, each satellite is now like two satellites because of the dual bands but, more importantly, the different frequencies will be refracted and reflected differently to each other and travel subtly different distances between an identical source (satellite) and destination (your watch). So if your watch finds a mismatch from any one satellite in the distances reported by L1 and L5 then that information could be disregarded its data as inaccurate. Garmin Epix 2 Sapphire Accuracy (Fenix 7 Sapphire also) OK, I've bored you enough with satellite stuff. I've performed some running workouts over the past few weeks where I followed the same routes as for my All-Satellites mode test a while back on the Fenix 7. The earlier test compared the GPS-only mode to the all satellites mode and today I'm just writing up the dual-band tests for Coros Vertix 2 and the Epix 2 Sapphire, I'll assume that the Fenix 7 Sapphire performs the same as the Epix 2 Sapphire. I ran within a metre of the previous tests and in any case, the satellites will be in completely different positions - so bang goes the scientific element of this test but I'm happy enough with the results. Test 1 - Suburban Grid - Dual Band (Vertix vs. Epix 2) vs Single Band, All System (Fenix 7) So here we have the Garmin Epix 2 and Coros Vertix 2 cranked up to the max on the same day. In some of these images, I've added in tracks made a few weeks ago with the Fenix 7 on a single band, multi-system with supposedly slightly lower accuracy. At a zoomed-out level, these tracks are all good. In other tests on this route, I frequently see one or two of the watches with either jagged lines or lineds cutting the tight corners. Not so here. It's seemingly perfect (enough ;-) ) [gallery size="medium" columns="6" ids="71420,71419,71418,71423,71421,71422"] However, when zooming in, the details show that the Epix 2 is definitely better than the Vertix 2. When comparing the Epix 2 Dual to the Fenix 7 Single, it is very hard to notice if the Epix is better. On the same run, I've also compared Epix 2 to the Apple Watch 6. Again, their accuracies are hard to split at the high level but the Epix definitely produces a slightly superior track at the detailed level. Remember the Apple Watch 6 (six, not seven) has one of the best GPS receptions on the market. These comparisons are comparing the crème de la crème. I'm splitting hairs. It's hard. BUT: Just ask yourself, "Will this make any difference to me in real-world scenarios?" Take Out: This is easy GPS territory. Dual-Band tech shouldn't make an iota of difference here. But it seems to make accuracy a little better with straighter and more precise tracks. Even with the same chipset, it seems that Garmin has eked outa little something extra over Coros. This time. Test 2, 3, 4 - Cycling There's nothing to see here folks! These 3 rides are all perfectly acceptable with the new Stages Dash M200 being the least accurate - although even that is still accurate enough for my cycling. Any extra accuracy the Epix 2 might have is irrelevant for road cycling, IMHO. Ride 1 [gallery size="medium" ids="71427,71428,71429"] Ride 2 - Richmond Park and back home The only small point of interest here is the third image where the Epix 2 is clearly the best at going each way through a 60m long tunnel. Does it really matter? A: Probably not. [gallery size="medium" ids="71430,71431,71432"] Ride 3 - Box Hill, Surrey The Richmon Park ride and this ride up Box Hill in Surrey are both on some of THE most cycled roads in the UK - - I try hard to get representative samples!. The vast majority of cyclists in the UK rarely if ever complete TdF stage rides - even for fun. Tens and tens of thousands of you use these exact same roads. The Epix 2 probably is better here in rural, hilly England at the detailed level but ...#Shrug. [gallery columns="4" size="medium" ids="71436,71435,71434,71433"] Test 5 - Tall Buildings - All Systems + Multi-Band on the Coros and Epix 2. Running close by tall buildings and encountering 6/7/8-storey high buildings in nearby Kingston is the closest I'm going to get to downtown Manhattan. When I previously did this exact same route for the Fenix 7 and Coros Vertix 2 (both on single band), it was hard to pick a winner and harder still harder to see if the Fenix 7 and Vertix 2 were better than the Apple Watch 7 (AW7, seven). This time around there clearly are differences at the higher level on the first image. You can see the Apple Watch is way off on the top middle section and also in the bottom right section yet, in the latter case, AW7 is correct and the others are wrong. You can check out the various twists and turns of the detail of the route where I have made various annotations. No doubt my annotations are hard to understand but the bottom line is that the Garmin Epix 2 Sapphire has the best track BUT it is still fallible and clearly makes mistakes when sometimes the others do not. [gallery columns="6" size="medium" ids="71442,71441,71440,71439,71438,71437"] If I re-use the same Epix 2 Multi-band track and compare it to the Fenix 7 (single band) from a few weeks back we get the image below. Again these tracks are noticeably different. But as I analysed each bend and each straight I heard myself almost literally alternating in my head...red, better - blue, better - red, better - blue, better - red, better - blue, better. If I had to pick a winner it would be the Epix 2 but mostly because of the way it has smoothed the track not because it was inherently more accurate. [caption id="attachment_71444" align="aligncenter" width="921"] Epix 2 Dual Red, Fenix 7 Single Blue[/caption] One of the issues we look at in built-up areas is the effect of reflected signals from tall buildings. Just like your image appears to be behind a mirror so the GPS track is similarly affected by tall buildings although not quite in the same way. You will often find that if you run on a path past a tall building on one side and a road on the other side, your GPS track seems to throw your route out into the middle of the road. This is the effect of reflected signals travelling subtly further and the same principle is applied when in mountainous areas. So. That was my Manhattan test. Or as close as I get to it. Somewhat annoyingly DCR did an actual Manhattan test and published his results a couple of days ago. Grrr ;-). His report is definitely worth a read if you ever regularly run in such massively built-up areas. Much of his run test is in Central Park which has benign conditions for GPS so pay close attention to the other bits where he runs between skyscrapers and through shops! those are the unique and interesting bits, although perhaps only experienced by a few of us regularly. Take Out from my tests: Epix 2 Sapphire with Multi-Band is better but not much better than anything else on test. The biggest surprise for me is that the Epix 2 and Vertix 2 both have multiple instances of their tracks being thrown away from tall buildings. This is precisely the sort of behaviour that multi-band should eliminate. Clearly, it doesn't. At least not yet and I have found more of the same behaviour, multiple times in other running tests. Test 9 All-Satellites 10-mile test Finally, this is my formal test over a 10-mile route that has some deliberately-tricky GNSS reception issues in places. The Epix 2 (multi-band) has done this test twice and the Fenix 7 (single band) once. More precisely I've done it 3 times! That's 30-odd miles just on this one test. Hopefully, that's enough for all readers! Full Methodology, results and data files here. The primary test for the Epix 2 (dual) when worn on the most favourable wrist received the best-ever "score" of any sports watch, ever. The Fenix 7 (all systems, single) turned in the third-best performance from a Garmin watch and the Epix 2 (Dual) when worn non-optimally on the right wrist produced a fairly good result. On this one occasion, Garmin's new Multi-Band performance was better than the Ambit 3/Polar V800 from pre-Industrial Revolution days but even those watches had off-days too. Garmin and Coros with the new Airoha (MediaTek) chip have clearly closed the gap. With a bit more accuracy-tinkering applied to their algorithms, Garmin and Coros might, just might, take running GPS accuracy to a slightly higher level. [caption id="attachment_71448" align="aligncenter" width="1024"] Clicks to full detail and source files on DCRAnalyzer[/caption] Other Tests I have several other tests I could show but I've reached my 2500 word limit and this takes a long time to do for a relatively small number of interested people. Please help support the work here when you can. Thank you. I have run in some densely wooded areas and I think that the Epix 2 Multi-Band mode was best. However, every watch's track differed and even the maps do not have correct paths so I can't say if Epix 2 is better for sure. I think it is. So What? Battery-saving technology seems to have finally caught up with historical accuracy. The 'so what?' is a tricky one to quantify. Whilst I don't especially care about a pretty line on my Strava tracks...others do. Whilst I use a footpod to get accurate instant pace...others still want to get as much accuracy as possible just from their sportswatches. Whilst I don't need precision for routing and navigation...others do. And whilst I don't really care about logging my total workout distances, some of you treat that with religious fervour. The point is that we all want accuracy for different reasons so be kind if someone wants something different to you. Maybe you're both right? Almost certainly, Airoha has a chip that can match or beat the performances of yesteryear, although it's not quite there yet in GPS-only or All Satellite+Single Band modes. Will you ever get an accurate instance running pace? Hmmm. Probably not always, but you will start to get it more often. As always, if you want consistently accurate instant pace and distance then you need to use Stryd, if you want an acceptably pretty Strava track then do yourself a favour and save some battery juice and use the GPS-only modes. Those recommendations are unchanged. Garmin product execs have recently been talking publically about their devices now having spare power to throw at the GPS chip. This is great news, with the spare power that is now available, Garmin and Coros can start to interrogate the chip data more intensively and use cleverer, more mathematically-intensive methods to weed out bad data points. The result could be more accurate instant pace and more accurate overall distance measurement. The generation of GNSS chips might start to give us decent accuracy in forested areas but accuracy in urban canyons is probably still a tantalising step away from where we are today. Most disappointingly for me, there are definitely shifted GPS tracks caused by nearby tall buildings. That's one of the exact problems this tech was meant to solve in urban canyons - it patently doesn't solve it. Q: Is 'All Systems + Multi-Band Mode' more accurate than other modes for the Epix 2 (Fenix 7)? A: Yes, just, but it matters to a small number of people. Q: Do you think the Epix 2 in All-System, Multi-Band mode is the best ever? A: Yes. Just. It will have off-days and can certainly be bettered. Q: Will you use Multi-Band mode? A: No! I'll save some battery for a later day and use Stryd for instant pace (find out more here). Q: Will MultiBand mode be on the upcoming Forerunner 955 and Edge 1040? A: I'm reasonably sure they will have Multiband capability with the Airoha chips. It's debatable whether these new models will have Sapphire screens so does that mean Multi-Band will be disabled? ie will this be a marketing/premium decision rather than a technical one. This post will very soon become subscriber-only. Fill your boots now! Note: I have tested the Fenix 7 but not the Fenix 7 Sapphire, I assume it has the same GPS performance as the Epix 2 Sapphire which I have tested extensively.