DSC_2647_thumb

Last week Shane Miller posted his review of the TechnoGym MyCycling Smart Trainer.  For those not familiar (which is almost all of you), TechnoGym makes a direct drive trainer, sorta like the Wahoo KICKR, except missing all the stuff that makes the Wahoo KICKR appealing. It doesn’t have ANT+ FE-C or Bluetooth Smart FTMS, doesn’t work properly with Zwift (or any other app) on Windows. In order to make up for your sad panda feelings, they charge you significantly more instead – twice as much as the Wahoo KICKR, coming in at $ 2,159USD. It’s like icing on the cake and all! (If you want all the fireworks, you can start Shane’s video at 9 minutes and 26 seconds)

While TechnoGym has success in other product segments and markets, they of course completely missed the mark here.  Not because they made inferior hardware (by all accounts, that part’s just fine), but because they made a trainer that attempted to ignore the standards used within the trainer market and pretended like things such as ERG mode aren’t important.  The last time someone tried to ignore standards, it didn’t end terribly well.  Nor did it end well for the last company that tried to do that.  And those products were a fraction of the price.

Shane did a good job of pointing out that most of this is easily fixable via software updates, if the company wanted to.  Of course, the trainer has been on the market in some capacity since last fall. One of the core reasons I didn’t bother to review it was because when I looked at the spec sheets last fall and then looked the price, I just laughed.  I could have ultimately written a review in under a paragraph, whereas Shane has far more patience than I.

But that got me thinking – what should be in a trainer in 2018? What’s considered baseline in any smart trainer? Be it hardware or software, this post is what happens when I get on a bit of a rant and have 2 hours on an airplane without internet access. Let’s get rollin’!

Required Smart Trainer Features:

DSC_1918

These days, I roughly bucketize trainers into price bands.  These bands shift downwards each year, on average by about $ 100 per year once above about $ 700-$ 800.  Said differently: What was once a $ 1,200 trainer 3-4 years ago is now an $ 800-$ 900 trainer. There are nuances to that of course (primarily around flywheel weight), but the general gist of that statement is true.

Note that the definition of a smart trainer is generally agreed upon to be a trainer that can have its resistance controlled via a 3rd party app or device.  Meaning that it can specify a wattage (i.e., 250w) or gradient (i.e., 5%) to simulate varying conditions and workouts.  You can, however, have trainers that may not allow control, but do broadcast data, which I’ll address later.  I’d argue those are not exactly smart trainers. We could call them wannabe-smart trainers (examples are the Kinetic inRide-equipped trainers, the Lemond Revolution with PowerPilot, and some Elite units with Misura accessories).

But let’s not get distracted.  Here’s roughly how I bucketize pricing (oh, and these buckets are only valid for 2018, they’ll shift again in 2019 and so on):

Spinny things: Sub-$ 300 trainers with virtually no electronics
Budget Smart Trainers: $ 300-$ 500
Middle of the Road Smart Trainer: $ 500-$ 999
High-End Smart Trainers: $ 1,000+

Again, the price line gets a bit messy as you approach $ 1,000, and will continue to get messier this year as well.  But that won’t matter in this piece.

See, as far as I’m concerned, once you cross the $ 300 threshold into a ‘Smart Trainer’ category, your trainer, without question, must have the following:

Control: ANT+ FE-C (Fitness Equipment Control)
Control: Bluetooth FTMS (Fitness Machine Service)
Broadcast: ANT+ Power (with speed baked in)
Broadcast: Bluetooth Smart Power (with speed also baked in)

In the event you have a non-controllable trainer, then the trainer should at least do the broadcast portions above.

However, some smart trainers go beyond that with a few optional tidbits:

Broadcast: ANT+ Cadence (either within the power channel or separately)
Broadcast: Bluetooth Smart Cadence (also either within the power channel or separately)

In general, when cadence is broadcasted, it’s an estimate. It’s usually good enough for most applications, but it’s super accurate in certain spots – such as if you sprint (with cadence increasing to say 120RPM) and then quickly back-off to light pedaling at say 60RPM.  Most trainers will incorrectly handle this from a cadence estimation standpoint.

Next, you can even double-down one step further, as Tacx does.  All of their trainers also transmit:

Broadcast: ANT+ Speed & Cadence sensor
Broadcast: Bluetooth Smart Speed & Cadence sensor

Why do this? Well, their goal was to integrate with head units and apps that may not support power meters.  When they first started doing this years ago, there were more head units that didn’t support power (such as something like a Garmin Edge 25).  Whereas these days, it’s fairly common for lower-end units to support power, such as the Lezyne’s units, Polar M460 and even Garmin’s new Edge 130.  Still, I appreciate it. It doesn’t ‘cost’ them anything to add it, and it gives consumers flexibility in certain situations.

2018-03-25 20.01.29 2018-03-23 10.34.42

So why am I bringing this up at all?

Because the reality is that there are trainers that aren’t following that logic above.  In general, they do one of the following misdeeds:

1) They don’t broadcast power as a power meter: This is problematic for the countless devices that don’t support ANT+ FE-C or Bluetooth FTMS.  For example all Polar devices, all SRM head units, all Suunto devices, all Lezyne units, all Garmin wearables, boatloads of apps, and countless other things.  It also means you can’t do a workout in Zwift using FE-C and also concurrently record it to your Garmin Edge device so you get ‘credit’.  That’s because even if it supports FE-C, you can’t have two devices controlling FE-C at once. It becomes a tug-o-war.  Power should always be broadcast as a baseline. Period.

2) They don’t properly implement ANT+ FE-C: This happens less and less these days, but it’s still an issue. Especially with the calibration routines.  Look, if you’re a trainer company and can’t figure this out there are plenty of avenues to get assistance (many of them free).  You can go to ANT+ themselves and they’ll help you out.  You can go to the leading app companies out there like TrainerRoad and Zwift, and they’ll often help you out (I know TrainerRoad helps out numerous trainer companies in sorting through protocol things).  Or you can go via companies that specialize in consulting in those areas, such as North Pole Engineering, which behind the scenes works with many companies implementing both ANT+ & BLE in their products [Note: This isn’t sponsored in any way, I just tired of companies shipping broke-ass FE-C implementations].

3) They don’t properly implement Bluetooth FTMS: This is far more frequent, mostly because FTMS is far newer. It’s barely a year old in terms of companies leveraging it.  Elite was the first last year, alongside TrainerRoad.  Nowadays it’s fairly common for newer trainers, but even Wahoo and CycleOps don’t implement it.  Their logic is that everyone is already using their (Wahoo/CycleOp’s) SDK’s in their apps, so there’s no reason to re-invent the wheel.  Which is true.  But it’s also true that it makes app developers lives more difficult, since it’s yet another company to support via Bluetooth Smart.  Versus just using one standard. I’d also remind Wahoo that their entire basis as a company was trying to make things open and supporting standards.  That’s why the Wahoo Fitness Key started, and why the KICKR was an open platform.  Still, I’m more worried today about non-Wahoo/CycleOps companies that screw-up on their implementations of FTMS, because it means you can’t use TrainerRoad or Zwift on your iPad or phone.  Like ANT+ FE-C, you can pretty much go to all the same peoples, except talking to the ANT+ folks won’t help you much here.  But otherwise, all those same companies are experts in sorting this out if your product is having issues.

TacxNeo-1

Which brings me to the final point, and one that I find I’m having to remind trainer companies of more and more: Just make sure the darn thing works with at least the two leading apps out there.  To make this really simple for trainer companies, I’m going to give you a secret. It’s pretty dirty secret, so brace yourselves.  Here’s exactly how I test trainers:

A) I use Zwift on iOS/Apple TV via Bluetooth Smart (free riding)
B) I use Zwift via PC over ANT+ FE-C (ERG workout, specifically usually ‘Jon’s Mix’)
C) I use TrainerRoad on iOS via Bluetooth Smart (ERG workout, specifically the DCR 30×30 Trainer Test)
D) I use TrainerRoad on PC over ANT+ FE-C (ERG workout, specifically the DCR 30×30 Trainer Test)
E) I validate that the trainer shows up on a Garmin Edge device via ANT+ FE-C (sometimes I’ll use a Wahoo ELEMNT too)

There’s more nuance than that in terms of the types of workouts I do, I’m looking at response times of course and such. But I’m making sure those five core scenarios above work.  I also often validate it shows up in Kinomap and sometimes Sufferfest.  Of course, if you’ve read any of my trainer reviews, you’d know my relatively predictable test pattern. By and large, it hasn’t changed.  Do note that I said ‘Bluetooth Smart’, and technically not ‘Bluetooth Smart FTMS’. That’s because while I’d like it to be via FTMS, if it shows up in TrainerRoad and Zwift and works via Bluetooth Smart (such as Wahoo or CycleOps), I’m not going to get too fussed up as to how it did that at the moment.

It doesn’t matter if the trainer cost $ 500 or $ 5,000 – if it can’t do those five things above, it’s a non-starter. End of story.

Yes, really, end of story. I won’t recommend it in 2018 if it can’t do those. Be warned companies.

Optional Hardware Smart Trainer Features:

TrainerKICKRWahoo3

Of course, software is only one part of the situation.  You can have great software but crap hardware.  It’s actually more rare these days, but it does happen on occasion.

From a hardware baseline standpoint, I actually don’t have any hard and fast requirements.  That’s because most hardware specifications are based on hitting a certain price point.  For example, flywheel weight (which largely dictates how much road-like feel there is) is usually heavier in more expensive units.  It costs more to make a heavier flywheel, and also costs more to ship one too.  Especially since flywheel weights are many kilos, not just a few grams.

And even within that, there is a nuance to hitting price points.  For example, companies that make you install the legs on the trainer are often doing so to reduce shipping sizes, and thus shipping costs. Or a variant of that whereby they’re using cheaper materials which can’t support the weight in one configuration, but can in another – again, requiring different setups.

Same goes for aspects like direct drive trainers.  A few years ago those floated in the range of $ 1,200 (Wahoo KICKR, initially).  These days they’re down to $ 899 (Elite Direto and Tacx Flux), and I suspect we’ll continue to see downwards shifting here in pricing and options.

DSC_2148

Still, there are some core things I like to see in any trainer above $ 300 from a hardware standpoint:

Important/Useful Baseline Features:
A) Support most common bike types: Be it a road bike or a mountain bike, it should work with it
B) Properly support common thru-axle bikes: Again, similar to A above, and not really an issue in most units I’ve seen from 2017 and beyond.
C) If a wheel-on trainer: Must have a calibration/roll-down procedure
D) If a wheel-on trainer: Ideally has some automatic press-on knob, like what CycleOps Magnus has
E) If a wheel-on trainer: Ideally has quick-lever for connecting bike to frame, like what KICKR SNAP has
F) If requires a power cord: Ideally has a trip-safe connector, like what Wahoo trainers have in case you trip over the cable it doesn’t rip it out of the trainer itself.
G) It shouldn’t tip over: If I, as a 6’2” tall dude, go to mount my equally high-saddle bike, I shouldn’t tip the trainer over. Go find a tall dude and their tall bike and test it.
H) LED status lights: This may sound obvious, but having some form of lights to assist troubleshooting is useful
I) Trainer legs should lock in place, or otherwise be stiff enough to not snap your fingers off if you pick up the trainer by the legs
J) If a heavy trainer (i.e., 20KG+), it should have a handle to move it around)

Next, a brief word on claimed accuracy (and tested) accuracy.  When I look at accuracy claims, it’s roughly tied to price point (though, I think even those days are going away). I don’t care whether or not you claim to have a power meter inside your unit. I just care that it’s accurate across a broad range of circumstances. After all, there’s certainly examples of trainers that don’t have power meters that are more accurate than those with them. So it’s a meaningless thing to me. Here’s what I’m expecting in trainers today:

Accuracy Baselines:
Under $ 600USD: +/- 5%
Between $ 600-$ 999USD: +/- 3%
Above $ 1,000USD: +/- 2%

That probably gets into response time items as well, but I don’t want to specify a hard limit there. I’d say that my previously mentioned 30×30 Trainer Test tends to weed those out pretty quickly.  But if you can’t do a 30×30 ERG trainer workout and have the wattage respond and stabilize within 2-5 seconds, then it’s too slow (unless you’re talking 100w to 1,200w or something).

After this, we get into the wide world of premium features that exist in some way in the market already.  Though, as we know, nobody has implemented everything yet in a single unit.

Cheery On Top Features:
A) Silence: Pretty self-explanatory – as quiet as a Tacx Neo, or even as totally silent as a STAC Zero
B) Simulate road terrain: Like the Tacx Neo does (as explained here)
C) Rock back and forth: As Kinetic does with their Rock & Roll base, which works with a few of their trainers
D) Simulate climbing/descending: As the Wahoo CLIMB accessory does with Wahoo’s trainers
E) Requires no power plug: Self-generating power, though, I prefer both a plug if you want it as well as self-generating when you need it. A bit easier for app pairing.
F) USB/Wired connectivity (for control/data): While I personally find this mostly useless, some companies have it on their trainers, and approximately almost no companies support it in software. I suppose it’s the thought that counts. Either way, I see the time has mostly passed for this.

There are of course probably other features that others do – like the Tacx Isokinetic modes for example, but I think the above lists are really the bigger differentiators right now.

Of course, each feature in the above list has its pros and cons.  Some people like some things, others hate them (or find them meaningless). That’s fine. Ideally, those features can be turned on or off as required, allowing the user to decide what matters to them.

The Future:

DSC_2144

Now the good news is that most trainers I’ve tested over the past year conform to these standards.  There are exceptions, however. For example, the JetBlack unit I tried a month or two ago had standards issues, as did the far more expensive TechnoGym unit that Shane tested that I mentioned at the beginning of this post.  In the case of both of these units, I (and to the same extent Shane) are even harder on them because they cost more.  They weren’t $ 300 budget units, but rather they were $ 1,200-$ 2,150 high-end units, designed to be the most premium offerings you can buy.

Still, I’m relatively confident that as we go into the trainer announcement season of 2018 that we’ll see most companies do this right.  With Eurobike less than two months away, those trainers are starting to come together as companies get down to crunch time.  Even more so since Eurobike is a month and a half earlier this year than in years past (now early July vs late August).  While I expect most (if not all) trainer companies will announce products at Eurobike, I also foresee some stragglers that may slide to later in the summer or Interbike (mid-late September).

I’d argue though that given what I see upcoming for 2018, companies that don’t at least announce their wares by early August will be at a significant disadvantage going into the Fall 2018 timeframe, as most consumers will have placed orders by July putting themselves in a ‘safe’ position on the backorder/pre-order queues (or at least, that’s what I’d recommend).

And, as always if you’re trying to decide whether to pick up a trainer now or wait to see what comes up, my guidance stays the same each year: If you want to use a trainer before October, then buy now. If you don’t mind not having a trainer until October (or even November), then you can always see what comes down the pipe.

And as always, expect my full 2018 trainer buyers guide in the fall, unless I’m highly confident everything has been announced by August, in which case you might see it then.  Till then, here’s last years trainer guide.

With that – thanks for reading!

Product Reviews – DC Rainmaker