Garmin Rally Power Meter In-Depth Review (SPD/SPD-SL/LOOK KEO)

Garmin has launched three new power meters today, all under the new ‘Garmin Rally’ brand, which supersedes the Garmin Vector power meter lineup. The Rally series includes three pedal bodies at this point:  Shimano SPD-SL (road), Shimano SPD (typically off-road/MTB), and Look KEO (which they previously had).

However, if there are two takeaways to know from this review, it’s this: The Garmin Rally series power meter is but a spindle that fits into multiple pedal platform types that you can choose from – at least, as long as those choices are Look KEO/SPD/SPD-SL. Which of course, is the vast majority of the endurance sports pedal market, save Speedplay (and Wahoo just solved that last week).

The second thing to know is that if you already have Garmin Vector 3 pedals, then you’re also able to take advantage of this pedal portability, and switch between all of the three pedal types above – complete with an even newer battery cap design.

Beyond that, this review will of course cover the ins and outs of the Garmin Rally series of power meters (which, is effectively Vector 4 by another name). I’ve been using all three pedal body types over the last while, getting in boatloads of rides indoors, outside on the road, and then outside off-road. I’ve also converted a pair of Vector 3 units to Rally SPD bodies too – to test how that works. You can watch my full quick guide to converting your pedals here, and soon also a dedicated post on it.

Or, you can get the entire Garmin Rally In-Depth Review in video form with lots of sweet cycling shots b hitting the play button below:

In any event, Garmin shot over these media loaner Rally pedals/bodies to test, which will go back as usual. Though, the Vector 3 sets I used for conversion tests are my own that I’ve had for about three years now. Once I’m done with these loaner Rally pedals here shortly, I’ll box them up and send them back to Garmin. Just the way I roll. If you found this post useful, consider becoming a DCR Supporter which makes the site ad-free, while also getting access to a mostly weekly video series behind the scenes of the DCR Cave. And of course, it makes you awesome.

What’s new:

Temporarily setting aside the added pedal bodies, the Rally sets aren’t appreciably that different from Vector 3 – however, they do have a few minor tweaks to them:

– Switched to pedal threads for the pedal body: This basically increases durability, but also helps protect a bit against overtightening
– Adopted latest gen battery caps: In the never-ending Garmin Vector battery cap saga, the Rally pedals use the most recent Vector 3 caps, which are about 6-9 months old now. Existing Vector 3 owners can always call Garmin support to get these shipped out free if they don’t have them already.
– Switched to CR1/3N batteries: While Vector 3 has supported this battery type for a while, this is now the official recommendation from Garmin, as opposed to dual LR44 batteries.
– Minor changes internal to spindle: Slightly newer hardware components that get better standby/sleep time than previously, which should give you longer overall battery life.

However, the big-ticket item here is that you can swap between the three pedal types easily, using a so-called ‘Conversion kit’. In fact, there are basically 6 core products that Garmin is launching, and then the single-sided variants of them too. Here, let me explain. First, you’ve got the three pedal types:

Rally RS: Shimano SPD-SL Road variant (get it? RS = Rally Shimano)
Rally RK: Look KEO variant (RK = Rally KEO)
Rally XC: Shimano SPD MTB Variant (XC = Cross Country)

Then, you can buy *any* of the three sets as:

A) Dual-sided full power meter ($ 1,099-$ 1,199) – Rally RS200/XC200/RK200
B) Single-sided power meter ($ 649-$ 699) – Rally RS100/XC100/RK100
C) Pedal body conversion kit ($ 199-$ 249)

Make sense?

Plus, atop all that, Garmin Vector 3 is compatible with any of the conversion kits. So if you’ve got a Garmin Vector 3 pair today (which is Look KEO), but you want it to be SPD for your gravel bike or mountain bike – then you’ll just need to buy the conversion kit.  As with before, they also still have upgrade kits (from single-sided to dual-sided – $ 549-$ 599), replacement bodies, and all the usual parts and such in case you need them. Here’s a simple chart that Garmin has showing the core models:

Finally, for lack of anywhere else to note it – Garmin seems to know they’ve got a Vector reputational issue to deal with. In fact, in their press materials, they lead with just how much testing they’ve put into this series, and how broad the testing has been. They have 125 riders around the world that have been using Rally, completing 6,000 rides for 110,000 hours of usage and 125,000 miles (200,000KM). They said this includes rides in rain/snow/sun, and in temperature ranges from 0-105°F (-18-40°C), and notably including someone who rode 430 miles over 47 hours. I can only assume that person couldn’t get his route to load on his GPS, and just kept on riding or something.

Of course, the proof is in the pudding. And certainly, I can show you plenty of data from this review (both good and bad data), and others can do the same. But only time will really tell. There are people, especially between year 1 and year 2, who rightfully gave up on Garmin power meters. Inversely, there’s also plenty of people who never had any issues at all. Of course, it helps that this is basically just Vector 3+. And by now, Vector 3 is pretty well stabilized, even if it did take nearly 4 years and a lot of pain and anguish for a lot of people.

Unboxing:

Each of the Rally pedal kits is essentially the same, so I’ll unbox one, and then show galleries for the others. First up, said box:

Inside, you’ll find the pedals atop some protective foam stuffs:

And then below deck, you’ve got the cleats, washers (more on them later), the manual, as well as some spare o-rings.

And here’s a closer look at the Rally KEO pedal bodies:

Notable is that if you can’t figure out whether a pedal is a Look KEO or Shimano SPD-SL (as they look very similar), the RK or RS name is written in super-tiny letters on the back:

Alternatively, here’s the two from a top-down standpoint.

Note the little peaks/tips of the silver strike plate in the SPD-SL variant as the main giveaway.

Anyway, here’s the quick-start guide. While you can activate the pedals with the Garmin Connect app (so that they get firmware updates), there’s no absolute need to, unlike some power meters. Of course, you’ll want to update the firmware from time to time, so that you can get any bug fixes.

In the box, there are the washers, which can be placed in between the crank arm and pedal, in the event you have clearance issues (either chain/derailleur clearance, or crankarm inset clearance).

And then the two o-rings are for the battery cap, just in case you need them. The batteries come pre-installed in the pedals. Also, there’s mounting hardware for your shoes, and then a set of cleats:

Again, the Shimano SPD-SL variant is precisely the same in terms of box components.

Meanwhile, here’s what comes in a pedal body conversion kit – in this case, for the Rally XC (SPD). Essentially it’s the pedal bodies, new battery caps (in case you’re coming from Vector 3), and washers/o-rings:

However, you’ll notice that the pedal bodies are lacking the pedal spindles. Again, you can’t just buy a $ 200-$ 250 conversion kit and have a power meter. You need the spindles from a Vector 3 or other Rally set.

In any event, I dive through the full conversion process in my post/video here.

Size & Weight Comparisons:

(Left to right above: Favero Assioma LOOK-KEO, Garmin Rally LOOK KEO, SRM X-POWER, PowerTap P2 LOOK KEO)

Next, let’s do a quick size and weight comparison. Here’s the two road Garmin Rally pedal bodies, plus Vector 3:

  

Then for comparison, here’s the Favero Assioma (LOOK KEO-ish), PowerTap P2 (Look KEO-ish)

 

Then we’ve got the SPD variant of Rally, plus SRM X-Power’s SPD option. Note that slight dust from actual use on both is included free of charge:

 

Oh, and some random weight tidbits:

Garmin Look KEO cleats/mounting weight: 37g per side
Garmin Shimano SPD-SL cleats/mounting weight: 40g per side
Garmin Shimano SPD cleats/mounting weight: 35g per side
Planned Wahoo Speedplay POWRLINK ZERO reference: 216g = 138g per pedal + ~78g mounting (cleats/baseplate/screws)

This seems as good a time as any to talk about stack heights and q-factors. Thus, here’s your q-factors:

Garmin Rally (all units): 53mm (55mm with spacer)
Garmin Vector 3: 53mm (55mm with spacer)
PowerTap P2: 54mm
Favero Assioma: 55mm
SRM X-POWER: 54mm

And then here are your stack heights:

Garmin Rally RS (Shimano SPD-SL): 12.2mm
Garmin Rally XC (Shimano SPD): 13.5mm
Garmin Rally RK (LOOK KEO): 12.2mm
Garmin Vector 3: 12.5mm
PowerTap P2: 14mm
Favero Assioma: 10.5mm
SRM X-POWER: 10.5mm

And then some actual pedal heights for the MTB side with some comparison heights:

Garmin Rally XC: 40mm (power pedal)
SRM X-Power: 33.7mm (power pedal)
Shimano PD-M505: 37.5mm
Shimano XT PD-M8100: 32mm
Shimano XT Trail PD-M8020: 32mm

And for lack of anywhere else to put it, max rider weight:

Garmin Rally Series: 105 kg (231 lbs)
Garmin Vector 3: 105 kg (231 lbs)
PowerTap P2: No practical limit according to PowerTap
Favero Assioma: 120 kg (265 lbs)
SRM X-POWER: No practical limit according to SRM

Then we’ve got side profile shots:

 

And a few more:

 

Ok, I’ll circle back here in a bit with a more detailed comparison post – like the power meter pedal shootout post I did a few years ago.

General Use Overview:

Getting the Rally power meter installed is the same for all three versions, and requires just a pedal wrench. If you’re coming to these from the days of Vector 1/2 where you needed a torque wrench, that’s not really necessary here. But yes, you’ll need a pedal wrench, because like all other pedal-based power meters, the spindle houses the electronics (so you can’t stick an Allen wrench in there). But I suspect most people considering a pedal-based power meter will have a pedal wrench. If not, they cost $ 15-$ 25.

For the washers, you don’t need them by default. However, if the spindles are too close to the chain, less than 2mm, then Garmin recommends using a washer on each (to balance out the q-factor). My chain clearance is plenty, so I didn’t need it on any of my bikes. Also, yes, I cleaned my bike after this shot. Just not before. I ain’t got time for that.

You can simply thumb tighten the pedals on initially, and then once snug just grab the pedal wrench to go all the way. Technically speaking, you’ll install the pedals to 34Nm (25 ft-lbs) of torque, which is likely a bit tighter than you’re used to installing pedals. So don’t go crazy, but if you lack a torque wrench, just give it a bit of extra snugness. Pedals will tighten over time by design, so at worst it just means it might not be as accurate until you get some nice hard sprints to tighten things up.

 

Once installed, you can use the Garmin Connect app if you’d like to, to pair it up and check for firmware updates. It’ll also set the crank length – in case you’re using a head unit that doesn’t support that. The crank length is *incredibly important* on a pedal-based power meter, and especially if you’re not running 172.5mm cranks. It uses this length to calculate power. All Garmin bike computers/watches support doing the crank length on the GPS unit, but not all apps/devices do. So by using the Garmin Connect app, you basically hard-set it in the Rally pedal so other devices don’t have to worry about it.

  

Note that on a dual-sided Vector system, the left side is considered the master/primary pedal, and all communications funnel from the right side to the left side. Further, the ANT+ ID and BLE name shown on a bike computer/watch/app will match the ID shown on the left spindle, just next to the side of the crank arm.

For calibration, you can do it from the app or your bike computer. It’s a good idea to do it now, and then Garmin recommends doing a calibration before every ride, though after the temperature stabilizes (about 10-15 mins in most cases). This is one area I dive into in more depth later in the accuracy section. I saw a couple of quirks here that I haven’t seen on most other power meters in a while.

Here’s doing the calibration on the app:

  

On your bike computer or watch, you’ll go to the normal sensors menu to pair up Vector 3. As long as it supports ANT+ or Bluetooth Smart power meters, you should be good. If on a Garmin device, it’ll see that it’s Vector and ask if you want to set it up:

As part of that, it’ll ask for your crank length. Most other bike computers will do the same. This will overwrite what you put in via the app, but hopefully those two match.

Once done, you’ll do a calibration using a Garmin Edge device. Most other GPS bike computers work the same way. As always, ensure no load is on the pedals and that you’re unclipped.

 

With all that done, you’re done and ready to ride. However, for most power meters it’s good to do a few hard sprints to settle things – ideally 2-4 sprints and then ideally a few minutes of riding in there. Then do another calibration. With the Rally pedals it’ll automatically determine what’s called the ‘install angle’, which basically means it figures out the exact position how the pedal is screwed into the crank arm. Ultimately, there’s nothing you need to do here manually – it just happens in the background, and then may send a brief notification to your bike computer that an ‘Install angle’ has been computed/determined.

If you were to move the pedals to another bike, you should ‘reset’ the install angle on your Garmin or app, so it knows to recalculate it. You can see the ‘Reset Install Angles’ option here, and I just hit it after moving these pedals to this bike.

The side of the pedal is where the battery compartment/pod/door is. This is the latest gen battery door, and matches what is I think the 4th generation doors for Garmin Vector 3, from 6-8 months ago. For those not following that saga, the one-paragraph version is that Garmin had significant troubles when switching to the LR44 coin cell batteries, for which you’d use two of them. These batteries differed globally far more than Garmin expected, and also expanded/contracted, causing battery contact issues inside the battery pod, which in turn caused dropouts/suckage. That in turn led some Vector 3 owners wanting to throw their units into an industrial blender.  There were other pedal thread material issues too. Garmin has gone through numerous battery pod designs, and the most recent one aims to handle all these better. At this point, issues with new Vector 3 pedals are basically non-existent.

The new pods/doors on Rally match the last set of Vector 3 pods, and as noted earlier in the post – if you’re on Vector 3, you can get these new pod doors too. Anyway, as part of that, they’ve switched to using CR3/1N batteries across the board, which is basically a double-stacked LR44 battery. You can see it below (left is Vector 3, right is Rally):

This reduces contact-type issues with two batteries. While I’d probably have suggested they go rechargeable, they say they’re confident in this battery cap design. The benefit for consumers is I suppose a much longer battery life, in this case a claimed 120-150 hours per charge. Further, the only difference Garmin noted in the newer Rally spindles is updated components which should get most people much closer to that 120-150 hour range, which was apparently heavily driven by how much your pedals slept (non-usage), more than straight usage. In this case, the newer components in the new spindles sleep better, thus saving more battery life. They must not have any small children at home.

Also, I noticed the baseplate of the battery contact area is slightly different too. It’s gone from an H-shaped contact pattern to a four-pronged uppy-thing. Again, left is Vector 3, right is Rally:

Also, as noted, the pedal bodies now have metal threads vs the plastic threads, which should significantly reduce people threading their pedals from overtightening, an issue with earlier Vector 3 (they had iterated the thread materials once or twice already in Vector 3 in the first year).

Ultimately, as with Vector 3 – we won’t know for a while on Rally. Remember that with Vector 3 it took nearly 6 months for issues to start cropping up, because that’s about when most people started swapping out the Garmin-provided batteries for random ones from your corner store. Now of course, since these are the same pods that Garmin has been using for 6-8 months already (after 3 additional years of learnings on Vector 3) – it hopefully reduces the change of issues here. All I can say is that I’ve had zero dropouts/wonk on any of my rides here on Rally – and then we can check back in later this fall or so and see if that holds true.

In any case, by now you’ve probably stopped reading and have already paired up your pedals – but just in case, if not, the pedals pair on both ANT & Bluetooth Smart. In the case of ANT+, you’ll have unlimited connections.

In general though, you’ll virtually always want to pair on ANT+ over Bluetooth Smart, because the spec for power meters is more advanced/standardized there, and you’ll get more information – notably Cycling Dynamics. From an ANT+ standpoint, the unit broadcasts power balance, ANT+ power pedal balance, ANT+ pedal smoothness, and ANT+ torque effectiveness – those are all baseline specs you’ll get on pretty much every ANT+ bike computer. However, if you’ve got a bike computer that supports ANT+ Cycling Dynamics, you’ll get added information like platform center offset/seated & standing time, and power phase. Here, let me show you.

First up, is all the charts you get on a normal ride – you’ll see the cadence/power/balance, and then the seated/standing position bits of Cycling Dynamics are mixed in there:

As you scroll down, you transition entirely into the Cycling Dynamics related bits (Platform Center Offset, Pedal Phase):

Then, down below in Garmin Connect you’ll see much of that same info summarized:

But there’s also a secondary tab there, to see all the Cycling Dynamics information:

And then this same info is seen on your head-unit mid-ride as well, using the Cycling Dynamics page on a Garmin device:

Now – whether or not you find any value in any of this, that’s pretty darn debatable. I think there’s some benefit in bike fitting, and certainly some benefit in injury recovery. However, after years of having Cycling Dynamics in my ride, I virtually never look at it, nor really know what to do with it. Garmin has a few minor things they note, especially around PCO – but ultimately, there’s yet – 7 years later – any studies I’m aware of that actually demonstrate how, from a science standpoint, any of this is going to make us faster in training or racing (or more efficient). Lots of ‘might & maybe’, or ‘coulda/woulda/shoulda’, but very little concrete evidence.

Looking at other pairing things, if you use Bluetooth, you’re going to get total power, cadence, and power balance – assuming the head unit supports power balance (most do). You’ll likely use Bluetooth to pair to something like Zwift or TrainerRoad. Here it is enumerated and paired in Zwift for both power and cadence as a single source (versus dual-sources for most other pedals unless put into a specific mode):

And here it is enumerated and paired in TrainerRoad as a single source, for both power and cadence

 

Both correctly pull in the power from both sides of the pedals, so you don’t have to worry about toggling any settings in the app like some pedals.  I’ve also paired it up to a Polar Vantage M2 GPS watch, and it paired up to it just fine, though, I did get data dropouts from it exactly every two minutes – which to me implies a Bluetooth power spec negotiation issue. Given Polar has a long history of not playing well with power meter pedals specifically (see: Any PowerTap P1 or P2 pedals, the SRM pedals, etc…) – I’m going to guess this is more of the same.

Next, rounding home here, I haven’t had any durability issues at this point across any of the pedal bodies, including the SPD ones while mountain biking. While I don’t have the super-rocky terrain of high mountain territory, I do have enough things to whack the pedals on during my recent rides. I’ve clipped more than enough tree roots on them, and a few rock clips too. I also managed to crash with them and had the bike (+ me) land directly on the right side pedal. No issues.

Further, in Garmin’s media presentation, they noted they use a cinder-block test where they have a machine that fires the pedal at the cinder-block representing a pedal strike on a rock:

I don’t have such a cool machine – so I’ll have to defer for now to that and check back in a year from now or so to see if my pedals are still alive on my MTB (after I go out and buy my own).

Converting Pedal Bodies:

Now, I’ve got an entire post – and video – showing how to do this. So I’m not gonna repeat that here, in an effort to save digital pixels or something. But you can hit the play button above, or here to see the full post on it (post coming shortly!).

However, the main takeaway you want to know is that you can convert any Rally pedal to any other pedal type, simply by buying the conversion kit (which is $ 200-$ 250). While that might seem a bit pricey for just a pedal body, if you consider the cost of a typical premium pedal, it’s not that bad. Plus, it’s a heck of a lot better than buying multiple power meter sets.

The process takes me on average, about 3-5 minutes. However, Garmin doesn’t really see it as a ‘do this every week’ kinda thing. Rather, they see it as a seasonal thing – and in doing it, I agree. While it’s easy to do, it’s an operation that if you tried to do it every week, I think you’d eventually wear down some of the pieces from ‘over-use’, such as the internal nut/bolt that connects the battery compartment to the pedal body, or the tiny jewelry screws that connect the battery compartment to the internal nut.

Anyway, you can see all that in the post/video in plenty of detail. And again, this also applies to Garmin Vector 3 – which you can convert to any of the Rally pedal body types.

Power Meter Accuracy:

I’ve long said that if your power meter isn’t accurate, then there’s no point in spending money on one.  Strava can give you estimated power that’s ‘close enough’ for free, so if you’re gonna spend money on something it shouldn’t be a random number generator.  Yet there are certain scenarios/products where a power meter may be less accurate than others, or perhaps it’s got known edge cases that don’t work. Neither product type is bad – but you just need to know what those use/edge cases are and whether it fits your budget or requirements.

As always, I set out to find that out.  In power meters today, one of the biggest challenges is outdoor conditions.  Generally speaking, indoor conditions are pretty easy to handle, but I still start there nonetheless.  It allows me to dig into areas like low and high cadence, as well as just how clean numbers are at steady-state power outputs.  Whereas outdoors allows me to look into water ingest concerns, temperature and humidity variations, and the all-important road surface aspects (e.g. vibrations).  For reference, the Garmin Rally series has a claimed accuracy rate of +/- 1.0%.

In my testing, I generally use between 2-4 other power meters on the bike at once.  I find this is the best way to validate power meters in real-world conditions.  In the case of most of these tests with the Rally series pedals I was using these other power meters or trainers concurrently in three basic configurations:

Road Bike #1 (Canyon Ultimate CF SL) – Mostly road + indoors testing

– Rally RS200/RK200/XC200 (used all variants here)
– Quarq DZero power meter
– PowerTap G3 hub power meter
– with Wahoo KICKR V5/2020 smart trainer (when indoors)

Road Bike #2 (Giant DEFY) – Baseline indoors testing

– Rally RS200/XC200 power meter (SPD-SL & SPD dual-sided)
– Stages LR dual-sided power meter
– with Tacx NEO 2 smart trainer (when indoors)

Mountain Bike (Canyon EXCEED CF SL) – Mostly mountain bike testing

– Rally XC200 (SPD dual-sided)
– Quarq DZero DUB power meter
– 4iiii Precision Pro dual-sided power meter (on XX1 crankset)
– with Wahoo KICKR CORE smart trainer (when indoors)

So…yeah, I’ve got a lot of data across all the pedal types as well as in my conversion post I do a comparison on Garmin Vector 3 converted to Rally XC too. I’m going to try and distill all these data sets down though since I don’t really think you want to look at that much data. But fear not, if you do, I’ve provided links to it.

I’m going to basically chunk this section into:

A) Indoor data
B) Road data
C) MTB/off-road data

I’ll mix the pedal types between them. I have two sets of pedal spindles, and three pedal body types (all of them). So I’ve been moving spindles/bodies around as I see fit to make it all work.

First off, indoors. Starting off with something straightforward, a Zwift ride on the RK200’s (Look Keo variant), compared to a Quarq DZero & Wahoo KICKR V5/2020. This was the Titans Grove loop, with rolling terrain climbing up over the ridgeline and back down again, with a few sprints tossed in. Here’s that data set:

As you can see, the three units are all very very close, with the KICKR V5/2020 being slower/lower than the rest of them, as expected. If we look at one of the sprints, smoothed at 5-seconds, you’ll see very close alignment. There’s a momentary connectivity dropout from the Quarq to the head unit recording it, but setting that aside, these all look virtually identical.

Cadence is very similar as well – minus a few connectivity-looking dropouts that we see on the Quarq. Not sure what was up that day.

Next, let’s switch over to TrainerRoad and a pair of Rally RS200’s (Shimano SPD-SL road variant). In this case, this was some 3-minute intervals on a Tacx NEO 2 and Stages LR crankset. Here’s that data set:

As you can see, the NEO 2 was surprisingly a bit higher than I’d have expected, by about 5-7w (on ~320w). One could also argue the Stages LR (dual-sided) and Rally RS200’s were lower than expected. Generally speaking, the NEO 2 is a pretty stable beast since it has no calibration option and is well known for being pretty much spot-on. Granted, this is out in the shed with wonky temperature shifts, on this first day of spring.

In any event, cadence is virtually identical on this. The NEO2 of course has estimated cadence here, and you can see when that falters a bit:

Looking at the ride power curves, the Stages LR & Rally RS200’s maintain near identical values, even if the NEO2 seems offset:

Of course, whether or not the NEO was annoyed at me that afternoon seems to depend on what I was asking it to do. Just prior to the above set, I had the XC200’s on the same bike/trainer, doing some 30×30’s. That was before I then moved the spindles to the SPD-SL pedal bodies. So again, NEO2 vs Stages LR vs XC200’s (SPD pedal bodies), here’s that data set:

And frankly, that’s looking pretty darn happy if you ask me, as is the cadence below it. And generally speaking, 30×30’s can be fraught with ways things can go wrong, though this seems super clean.

But now we’ll take a left turn, into what is arguably the singular set that gives one pause. This was a TrainerRoad workout on a pair of RK200’s (KEO dual-sided). What’s notable about this is that in virtually all my power meter and trainer testing, I like to do a test where I specifically don’t calibrate at the beginning of a ride, to see how it handles. In 2021 (or frankly, for the past 2-4 years), virtually every power meter on the market has some form of active or passive temperature compensation, to correct for shifts in temperature (such as the day warming, climbing up a mountain as it cools, or going from outside to inside or vice versa).

And that’s exactly what I did here. The *day prior* I did an outside ride on my bike in the afternoon. Then, the bike sat downstairs in what is effectively a heated garage at the studio, but slightly cooler than the upstairs where my trainers are. So the next day I took the bike upstairs, put it on the trainer, spun the pedals a few times to wake up all the power meters to ensure head units were ready, and then left it alone for 15-30 minutes before starting my ride without re-calibrating. And unfortunately, this was the result – with Rally in purple offset above everything.

It didn’t compensate for the temp shift, clearly. And in theory, it should have. In talking to Garmin, Rally will both actively compensate for temperature shift mid-ride, but also will do its automatic zero as it goes to sleep (so the night before, but also slightly before my ride too). When it does its zero, it has logic to check if it perhaps has shoes on it, or is leaned against something – things that would apply for weight impacting the zeroing.

Either way, it didn’t compensate. Thinking it might fix itself by the next day, I left it alone again. And nope, still off on a quick test ride. You can see this easy trainer test below. The first half of the ride I just got on and started riding. Then I paused and calibrated (I also waited a couple of mins, as apparently Rally writes extra diag data to the file when paused, and Garmin wanted to see some of that data). But after the calibration (zero), you can see all is well again.

So, in my mind – that’s not ideal. You can see the Quarq DZero handled this just fine. And frankly, so does every other power meter I’ve tested in the past few years. And in this case, Rally had numerous instances to sort itself out. Thus, I’m less confident in how well it can actually temp compensate over time.

I asked Garmin about this, and this is what they had to say on how to handle calibrations when you take a bike inside/outside (Note: I’ve re-ordered these four paragraphs to make more sense, as they were part of two different e-mail responses but the content is otherwise untouched):

“For the best results you will still want the bike to have fully adjusted to the temperature it is in and then do the zero offset calibration. We expect that to take about 10mins per 10°C. If you were to calibrate immediately, you would be setting the zero offset based on a temperature that’s likely much closer to what it was inside.

There is a difference between our zero offset calibration and temperature compensation. The zero offset is influenced by temperature at the time it is done which is the meaningful difference we expect between your rides after looking through the data. Our temperature compensation is then used to ensure we maintain that offset throughout a ride if the temperature changes to protect against drift.

Additionally, our Auto Cal feature does have some limits on how much it is allowed to change the zero offset which we believe you would have been outside of from your outdoor to indoor ride. The reason for that is we do not want to accidentally get a bad zero offset in the middle of a ride if the customer stops and happens to say, have one of their pedals leaned up against something.

Like I said, this is an area I’d like to improve going forward and we have some ideas around how to do so. I don’t want our customer to need to know all of this and even more importantly, we want to do everything we can to prevent a ride with inaccurate power data that can then influence the rider’s training guidance and our quantification of their fitness.”

The TLDR version is to ensure you use the zero offset option, ideally about 10-15 minutes after a temperature shift – kinda like the ‘olden days’. And in doing that for all my other rides, the data is spot-on perfect.

Ok, so, let’s head out on the road for some data there. First up is a roughly two-hour road ride on a pair of RK200’s, albeit with a short mid-ride break for some filming (also a great unintended way to test that the auto zero/temp comp stuff isn’t doing weird things mid-ride). This is compared to a PowerTap G3 hub wheelset, and a Quarq DZero power meter. Here’s the data set:

As you can see, at a high level, it looks pretty darn close (spoiler, it is). Here’s a few sections close up. First, this nice steady-state chunk:

While it is very close, there are a few minor imperfections there – and in some cases the Rally seems a few watts higher than the others – but they all float very close to each other. And the same is said for the second half too:

Cadence is virtually identical across them too, save the PowerTap’s estimated hub cadences when you reduce torque/power for a split-second. In general, we actually see Rally pick up these faster than Quarq and PowerTap, which is notable (see how the yellow dips further than the other two, when they dip – note these are smoothed, hence why the dips don’t hit the floor).

Next, we’ve got another outside ride on the RK200’s. Same grouping of Quarq DZero & PowerTap G3. Here’s that data set:

Boring alert: Virtually identical again, even across 8 sets of sprints. Here are the first four sprints:

As always, getting power meters to agree to an exact value on a short 800w+ sprint is challenging, due to the precise sub-second time alignment issues with different power meters and head units and how they all transmit/receive. But nonetheless, it’s really damn close.

Same goes for the 2nd set of sprints. There does seem to be a handful of dropouts on the PowerTap G3 recording here, I know I changed the battery around this ride, but can’t remember if it was before or after.

As expected, the power curves here look awesome:

Next, a road ride back over on the Shimano SPD-SL pedal variant, on my primary road bike compared to the Quarq DZero & PowerTap G3 hub as usual. Here’s that data set:

It looks really darn good. Super crispy. Very happy with this. Especially since I had just moved this set to a different bike right before this ride. I merely did 3 short sprints, then calibrated (zero offset) again, and called it done.

The power curve below is pretty much spot on as well:

So with all the roadie adventures covered for now, let’s head off-road with my mountain bike. In this case that’s configured with a dual-sided 4iiii Precision Pro power meter on an XX1 crankset, with a Quarq DZero DUB spider-based power meter on there as well. But I’ve had some connectivity issues between the two sides of the 4iiii unit lately, so while I’m troubleshooting that, I’m just running the 4iiii unit as a single-sided (left-only) power meter. No biggie, but just wanted to be clear on that.

Note, analyzing power meter data from mountain biking is frankly…a hot mess. Mostly because the types of trails I have here in the Netherlands means very few climbs – thus, the data s constant surge and coast (or brake). Whereas something like high-mountain terrain you’re going to get longer ascents and more ‘clean’ data to look at. Ain’t nothing clean here.

Welcome to the party, here’s your first data set:

Oh, and I’m being kind to you – that’s smoothed at 20-seconds to make it barely readable. So let’s just zoom in on a few chunks, since that’s the only viable way to tackle this. Here’s a random chunk that’s somewhat steady-state:

As you can see, the Rally XC200 and 4iiii left-sided unit seem to roughly agree the most. Whereas the Quarq seems a bit higher than the others. For all these sets, I did calibration in two groups. First, after arriving at the MTB location, I took the bikes out and let them chill outside for 10-15 mins, then I did a zero offset. Then, 10-15 minutes into the ride, I did another zero offset.

Now, if you remember back to a week or so ago, I had the SRM SPD pedal power meter, and you’ll remember the Quarq also *appeared* to read higher there too than the others. However, what’s when I had smoothed turned way up (20-second smoothing). Take a look at it when I yank all smoothing off:

They start to look pretty darn similar, huh?

Here, let’s zoom in more – to just a 45-second section from above. What you see starts to explain it – just like with the SRM SPD pedals. At the second to second level, they’re all actually very similar. However, the moment I make a surge – the Quarq responds faster – and usually higher. What I suspect is happening is the exact same as with SRM: Everyone else is being slightly more cautious in their rough-terrain algorithms than Quarq.

So when the ground noise increases, it appears that Quarq/Garmin/4iiii are basically being more hesitant on the surges, and recording a lower power value, whereas Quarq is letting it fly- and doing so faster. As a result, once you apply a longer average to it, those higher surges pull the whole line up above others.

Now – does that mean Quarq is more accurate? I’ve got no idea.

And frankly, I have no way to test that. That’d require incredibly complicated test rigs that can somehow determine force while being vibrated senseless like outside on rough terrain. Yet, it explains why on the same bike on smooth terrain, things all clean back up again nicely.

Anyway, here’s another example of MTB data analysis hell:

And then cleaned up with a 20-second smoothing average:

And it’s just like above. Any time my power significantly increases, the Quarq goes higher than the others, while at steady-state, it’s relatively the same. And because of the smoothing, even reduced to 10-seconds below, you’ll almost always see the Quarq as higher for any shifts of power. Now again, that could be totally 100% correct. It is what it is:

Note, before I forget, I did see a single clip-in spike oddity here – an incorrect 1,100w clip-in spike on the Rally XC200’s with an Edge 830.

And you can clearly see it’s still reporting 0-value for the cadence at that second, thus, an improper spike:

However, while Garmin might be more conservative than Quarq on power spikes, Garmin is clearly faster on cadence pauses. You can see in the cadence valleys below that Garmin goes lower, faster, to each time I briefly stop pedaling. It’s not really a meaningful thing in most cases, but just an interesting data thing I noticed both indoors and out.

And finally, for those that want yet another data set, you can check out this MTB data set as well, for more messy but normal mountain bike data to analyze.

So – with all that, where do we stand on overall power accuracy? Well, in short for both road riding and indoor riding – it’s great and spot-on with what I’d expect. I mean, assuming you calibrate (zero offset) if you move between substantially different temp ranges at the start of a ride. Everything I see in this realm looks great, and honestly basically what I’d expect from Vector 3 previously. I didn’t see any dropouts or such that would indicate any sort of battery cap/pod type issues – or any dropouts indicating connectivity issues at all on ANT+. I only paired via Bluetooth Smart once, to a Polar Vantage M2 watch for a ride, and that did see drop-outs – though Polar and pedal-based power meters via Bluetooth have a long and complicated relationship – and given both products released within one hour of each other, I couldn’t exactly tell either company about the issues to have them troubleshoot. I’m not worried about that though on the Rally front, I’d strongly suspect that’s a Polar BLE-spec negotiation thing (again, knowing Polar’s pedal pairing limitations historically).

For mountain biking, it’s a bit more complicated. Like with SRM a few weeks ago (or even the 4iiii on this set), it’s clear that Quarq is far less conservative with its power values during surges on rough terrain. On smoother terrain, all units are equal. But as the terrain variability increases, Garmin seems to become more conservative with power values, and won’t show as high of power values (or, Quarq’s are incorrectly high). Again, I have no way of knowing which are truly correct here. And given the variability of dirt/roots/rocks/mud/sand/etc on any section of trail, there’s no viable way I’m aware of to mathematically reverse engineer these power values like you can out on the road using something like Virtual Elevation from Robert Chung.

Which isn’t to say Garmin is inaccurate – not at all. It’s simply to note that it agrees with 4iiii (and by extension SRM’s X-Power SPD pedals), and that Quarq is higher. I don’t believe I have any viable tools in my arsenal to deduce exactly which one is 100% correct in off-road conditions. But for everything else that I do have plenty of tools for – it’s spot-on.

(Note: All of the charts in these accuracy sections were created using the DCR Analyzer tool.  It allows you to compare power meters/trainers, heart rate, cadence, speed/pace, GPS tracks and plenty more. You can use it as well for your own gadget comparisons, more details here.)

Pedal Power Comparison:

I’ve added the Garmin Rally series into the product comparison database, to help you compare it against other power meters. You can always make your own comparison charts in the product comparison database, but for simplicity I’ve included a couple of main comparisons below. Notably, I’ve shown it compared to the Favero Assioma (LOOK-KEO) and SRM X-POWER units, as the PowerTap P2 series has been discontinued.

Note that aside from pedal interface type (SPD/SPD-SL/LOOK KEO), all remaining specs on Rally are the same across all units. And again, you can mix and match your own comparison charts in the product comparison database.

Summary:

There’s no question that Garmin sees Rally as the one-size-fits-all answer to pedal-based power meters. And the mix and match pedal body concept makes a ton of sense, the idea that you can buy one Rally power meter set, and move them between bikes with different pedal types seasonally is something that will undoubtedly appeal to a lot of people. And basically, unless you’ve got Speedplay or Time, this pretty much covers most of the most types people use.

And while Garmin’s Rally XC200’s with SPD compatibility makes it an option for mountain bikes, I suspect the biggest focus area there is really gravel bike conditions more than nasty rock-laden terrain of some mountain bike trails. Certainly, I’ve been using it mountain biking just fine, but for most people it’ll be cheaper to go with a crankset-based power meter on their mountain bike – unless they are actively moving the pedals between bikes with different bottom bracket situations. But hey, it’s an option.

Obviously, the elephant in the room continues to be Garmin’s history with Vector 3. And no amount of shorter-term testing from me on Rally can really solve for that in comparison to 6-12 months of longer-term pounding/testing. Sure, design-wise, it’s effectively the same spindle inside the same pedal body (now with metal threads) with the same most recent battery pods that have been out for 6-8 months already on Vector 3. So that risk is reduced substantially, but I certainly wouldn’t blame anyone for saying they’ll check back in 2022 to see how they are going.

But for those looking for an SPD-SL option, there’s…well…no other options in the pedal-based realm. And for those wanting to shift pedals between pedal types, this seems like your best and only option as well. If you’re on LOOK KEO and trying to decide between Favero Assioma or Rally, I’d say it’d honestly depend on whether or not you plan to shift to other pedal types down the road. If not, I’d probably save the cash and go Favero Assioma (since you can literally buy yourself a Garmin Edge 530 GPS with that savings). Whereas, if you plan to mix and match pedal bodies, then I’d probably go Rally.

With that – thanks for reading!

Product Reviews – DC Rainmaker