Smart trainer?

steffenson

Hey Cyclists,

I am having some accuracy/consistency issues with my old Cycleops mag trainer (12+ years old) with Zwift. It sometimes jumps around between displaying 0w, 90w, 200w and 500w when I am spinning a steady cadence and holding what should be 200w (or so). I also have my Garmin cyclometer running as backup and sometimes that reads fine and other times it also skips around. I got a training tire to reduce slippage and have replaced the battery in my speed/cadence sensor multiple times but still have issues. Sometimes it also seems when I run my Bluetooth speaker it also causes more frequent issues, but sometimes not. I thought I had solved most of the issues last winter but after I went for a 30min spin this morning Zwift had my average speed at 9mph instead of around 16pmh because of all the time spent at 0w (did not seem to be any over wattage issues today, though last winter jumps to 400w for 15 seconds or so was the most common issue). Not sure if it is the trainer, the speed/cadence sensor, Bluetooth connection, the ANT+ receiver, etc but it is very frustrating!

So, what are people’s thoughts on smart trainers and running Zwift?
I was looking at Tacx Vortex T2180 ($323 usd from http://www.bike-discount.de/en/buy/tacx-t2180-vortex-virtual-smart-trainer-246764?currency=5&delivery_country=191 ) The variable resistance of any smart trainer + Zwift looks nice/fun plus I would hope this would work better than my current set up.

The Tacx NEO smart looks really nice (direct drive means no tire slippage which might solve 1 of my current problems???) but even at a discounted $1,125usd (same website as above) is a bit too costly - so the soon to be released Tacx Flux at around $900usd looks more appealing.

Thoughts? Experiences? Recommendations?

TheBigYin

well - for a starter, the man who's tried pretty much everything is worth a read if you've not already done so...

http://www.dcrainmaker.com/2016/10/annual-winter-2016-2017-bike-smart-trainer-recommendations.html

as to personal recommendations - I went for one that he "sort of" mentions - the Tacx Bushido... similar to the Vortex smart you mentioned, apart from it doesn't need a power supply and it does higher resistance for hill emulations - it's not necessarily about the maximum resistance a trainer can give, more about the maximum resistance it can give at lower "roller" speed - so, if you're (say) riding up a 10% gradient, with my weight and level of fitness, I know I'll be going maybe 8kph - 15kph tops if I'm "eyeballs out" - there's a calculator on the Tacx website that shows what resistance wattage the trainer will give at a given speed before it "tops out" - http://www.tacx.com/en/products/trainers/bushido-smart#tab_2

for the Bushido, at 15kph it'll give 650w or thereabouts, and at 10kph, 435w. Now, that's pretty close to what I'll do in the real world maximums, and indeed Tacx reckon it'd be good for around 11% maximum before "going virtual".

For the Vortex Smart, at the same speeds the max. resistances are something like 125 and 200w - that'd be more like a 3% gradient...

for the top of the shop Neo, the figures are even more extreme - they reckon that it'll go to 25% gradient for a 75kg rider (though there doesn't seem to be an interactive graph for it like the wheels on trainers get...)

("Going Virtual" - basically, your trainer can't emulate the slope on the training program or virtual world, so it "gears down" your screen speed to something it can cope with... which means, if you're in "ergo" mode, and riding in a group ride, the people on Neo's and Kickr's will just disappear into the virtual distance every time a hill appears - so you end up having to go "non-ergo", and basically you lose a great deal of the "realism" factor... )

Basically though, the wheels on Tacx Trainers share a frame from the Vortex up to the bushido and i-genius... it's pretty much boulder-solid - i'm upwards of 120kg, have been upwards of 170kg and haven't managed to do any physical or electrical damage to two of them - first a non-smart Bushido, then the Smart Version - actually in a couple of ways, the "Smart" was a small downgrade - the "flywheel" on the original was much more pronounced, which helped smooth out the pedalling feel and stop the whole "slammed the brakes on every time you stop pedalling for 2 seconds" experience - BUT the smart control aspect more than makes up for that.

If you're thinking of using either Zwift or TrainerRoad, then a "smart controllable" trainer that'll work in "erg" mode (i.e. the software controls the resistance to give you a set power requirement to maintain your current speed at current virtual slope... or, in plain words "hills get steep for you, and downhills get easy" - or "intervals at 250w for 6 minutes with 2 minute rests at 150w" are strictly enforced rather than you having to "hunt" the power out manually.) is pretty much a must IMO.

As to wheels-on or wheels-off - by preference, I think I'd prefer wheels-off - BUT, and it's a J-Lo sized but... There's pretty much only 2 wheels-off trainers i'd consider - and that's the Tacx Neo and the Wahoo Kickr. They've now both been around long enough to have the bugs shaken out (first 6 months or so with the Neo were a bit dodgy as Tacx's manufacturing got to grips with producing something so complex compared to it's wheels-on items) - indeed, at the moment, I'd still go for a mk1 version of the Kickr, as I'm pretty conservative when it comes to spending a grand on anything that's only been on the market a short time, and hasn't had time yet for the production gremlins to start to crawl out of the woodwork...

Big thing with any of the smart trainers though... get it from somewhere with a VERY good returns policy, there's quite a large "DOA" percentage for all the kit as far as I can make out, though some marques are worse than others...

So... my verdict on a Bushido Smart (T2780) - 7/10, Good - Solid, Wire-free, mains-free (works as a pre-TT wamup in a car-park if you need to), Works with most software I've tried. Bad - Wheels on so power "calibration" is down to how accurately you've set it up (as with all wheels on frames), Not exactly quiet, Tacx Apps and Software are "idiosyncratic" and the DVD/Download rides are fantastic, but bloody expensive) Would I buy again? I'd probably save and go for the Neo, to be honest, or, more accurately, whatever the Neo has morphed into in 3-4 years from now when I've worn the Bushido out...

steffenson

Thanks for all the info!
Buying from a 'local' shop makes a lot of sense (unfortunately as Tacx trainers seem to be a lot more here in the US). I had not seen dcrainmaker's new list of trainers so thanks for that link as well. Lots of down time research and reading to do now

steffenson

Do you have any connection/accuracy issues with Zwift and the Bushido as if you were getting major tire slippage (when you aren't)? Or does your wattage output stay very consistent when at a steady cadence?

ntnunk

As usual, I'll have to pretty much just point to @TheBigYin's post and just go with "what he said."

In addition to that, I personally have spent untold hours on TrainerRoad and Zwift with a simple Kurt Kinetic Road Machine, both with and without a power meter on my bike. Since then, I've put many hours in on both a Tacx Vortex Smart and now a Tacx Neo.

Definitely go with the local purchase option if possible. My Vortex was always rock-solid, and my wife still uses it to this day, but my buddy has had his Vortex replaced under warranty once, replaced the brake unit on the warranty replacement, and now the new brake unit has gone bad. That being said, he's eagerly awaiting the delivery of the brand new Genius Smart Tacx customer support has sent him to replace his problematic Vortex.

I've had my Neo almost a year now and I can say unequivocally that I prefer direct drive to wheel-on trainers. First, even with the Road Machine and the Vortex, I was always pulling my rear wheel off because I didn't want to trash my carbon wheels and nice Conti GP4000 tires on the trainer so I had an older aluminum rear wheel for trainer sessions. Second, the feel and power meter accuracy on the Neo is much, much, much improved. My Neo very consistently reads 3.5% higher than the Quarq Riken power meters on both road and TT bikes. Any software I used for erg-mode training (normally TrainerRoad, sometimes Zwift), I can just set my FTP in the software 3-4% higher than my "real" FTP from my powermeter and my intervals will be bang on accurate.

Regarding smart trainers and Zwift, it's a match made in heaven. Zwift is pretty fun anyway, but adding a smart trainer definitely takes it up a notch. I recommend it highly.

Last, with both the Neo and the Vortex, power numbers are relatively steady in Zwift. You mention your "Garmin Cyclometer" having problems as well. That being the case, I'm assuming you're connected via ANT+ rather than Bluetooth Smart. Be aware that ANT+ transmits at 2.457 GHz and WiFi is generally between 2.4 and 2.5 GHz. I used to use my laptop with TrainerRoad, connected to my home WiFi. I would always run an MP3 player on the laptop while I was riding, connected via the network to my music folder on a different computer. Everytime the MP3 software pulled the next song across the network it would effectively jam the various ANT+ transmitters (i.e. my sensors) and cause both TrainerRoad and my Garmin to lose signal. If you have this issue you can normally solve it by forcing your WiFi access point to move to a different channel.

TheBigYin

steffenson wrote: »

Do you have any connection/accuracy issues with Zwift and the Bushido as if you were getting major tire slippage (when you aren't)? Or does your wattage output stay very consistent when at a steady cadence?

Once the trainer roller tension has been "calibrated" with a "rolldown test" using either the Tacx App (free on iOS or Android) or the Tacx TTS software (not free, in fact it's bloody expensive, but it's fantastic for the DVD Real Life Videos) - then I don't get wheel slip even doing some of the TTS DVD's that include sections up to an emulated 18% slope... Wattage reading is within 3-4% (on the low side) of the figure from two "real" powermeters - Stages and a 4iiii - they're both single sided crank based ones, and considering that they both read on my left leg which I KNOW is weaker in output than the right (lots of wattbike analysis by my physio. puts it at 47%L/53%R all the way up to FTP+20%, after that, it sort of evens out...)

I generally use whichever software I'm training with to record my rides, and have the Garmin paired to the crank powermeter as a "safety net" in case of PC crashes, which allows me to compare the wattage from the two sources - Smart Trainer and Power Meter. Only time I pair the powermeter in Zwift is when I'm going an actual FTP test - because I want the FTP to be entirely representative of what I'll be doing out in the real world...



As to connection issues, I use either the Suunto Ant+ dongle on a long USB extension cable, to allow it to come up and sit half way between the cranks and the trainer (for FTP runs where I'm reading the Powermeter) or right next to the Trainer's resistance unit (for normal Zwifting etc.)... Last year I had pretty much no problem whatsoever with signal dropouts, but this season i've had a few 1-2 second "drops"... I must have a chat with my neighbour, see if they've swapped wireless routers recently, or bought a new DECT cordless phone (as both of these could mean I might need to swap my own WiFi channel to see if I can improve connection quality... there's an awful lot of different signals all competing for that single section of bandwidth that the trainer software relies on being perfect. Personally, I can't understand why the "mains powered" trainers don't just have a bloody CAT5 socket on the back and cut out the Wireless altogether. I mean, there's already wires going in there, what's the problem with one more...

cloggsy71

TheBigYin wrote: »

I can't understand why the "mains powered" trainers don't just have a bloody CAT5 socket on the back and cut out the Wireless altogether. I mean, there's already wires going in there, what's the problem with one more...

Agreed! No more to add, other than I have an updated Tacx Genius which now talks to Zwift and a Wahoo Kickr.

Personally, I prefer the Kickr (Mk. 1) as it has an actual Power Meter and not a Software one like the Mk. 2 and it does away with the rear wheel.

ntnunk

TheBigYin wrote: »

Personally, I can't understand why the "mains powered" trainers don't just have a bloody CAT5 socket on the back and cut out the Wireless altogether. I mean, there's already wires going in there, what's the problem with one more...

I don't know of any trainers that use WiFi, so a CAT5 jack on the trainer wouldn't help. The wireless signaling from the trainer is all sensor-related, i.e. Bluetooth Smart or ANT+. The computer that's running whatever software you might be using is generally the source of the interference issues, since frequently it's a laptop connected to your home network via WiFi. This is why my dedicated "training" computer is a desktop model and is connected to the network via Cat5.

Regarding connection issues, I'm currently using a Tacx ANT+ dongle, one that came with the Upgrade Smart package, along with a USB extension cable. I had a really bad time with connection issues in Zwift last winter but finally traced it down to the Garmin USB dongle I was using. For whatever reason, Zwift seemed far more susceptible to the dropouts the dongle was causing than TrainerRoad was. I could do an hour on TrainerRoad just fine, but an hour on Zwift would typically require at least one, and often two or three, restarts.

cloggsy71

ntnunk wrote: »

... but an hour on Zwift would typically require at least one, and often two or three, restarts.

Oh, I couldn't be doing with that...

ntnunk

cloggsy71 wrote: »

Oh, I couldn't be doing with that...

It was annoying as Hell. I swapped emails with Zwift support for weeks over that issue. I was convinced it was a Zwift problem because TrainerRoad worked fine. I ultimately started seeing a few similar, though much less disruptive or serious, issues with TrainerRoad. Then I got the Tacx dongle and the problem went away completely.

It was annoying enough that I mostly stopped using Zwift during that period and I really never have gone back to using it with the frequency that I did prior to that, though in truth there are other issues at play there as well. Turns out that I really only like Zwift for certain types of rides. If I'm doing serious intervals I find it to just be a distraction. I'd much rather use TrainerRoad.

TheBigYin

ntnunk wrote: »

I don't know of any trainers that use WiFi, so a CAT5 jack on the trainer wouldn't help.

yeah, what I was saying was more "stop dicking about with ANT+ or Bluetooth and just make the damned thing a wired network device" - i've got the house wired with Cat5 (had it run in when I had the house rewired maybe 20 years or more ago...) so I don't really have to "do" wireless networking apart for the Laptop, iPad and Mobile - all the other IP devices (Sky+ boxes, smart TV's, Media Streaming box etc, are on cabled connections

35dollars

This is probably a daft question, but if I'm using a smart trainer (a Tacx Vortex, say), does that report the speed from the rear wheel, or does that just provide resistance so that I also need a speed sensor?

ntnunk

35dollars wrote: »

This is probably a daft question, but if I'm using a smart trainer (a Tacx Vortex, say), does that report the speed from the rear wheel, or does that just provide resistance so that I also need a speed sensor?

I think all of them will report at least speed. Many of them report cadence as well, though it's kind of computed or estimated. It's usually pretty close though, in my experience.

TheBigYin

35dollars wrote: »

This is probably a daft question, but if I'm using a smart trainer (a Tacx Vortex, say), does that report the speed from the rear wheel, or does that just provide resistance so that I also need a speed sensor?

well - if I pair the Tacx Bushido smart with my Garmin 520, it'll feed back speed, power and cadence figures should I desire to use them. By default preference the 520 will take the power (and maybe, haven't checked) cadence from my power meter however, so I would have to de-select the power meter as a sensor if I wanted to use the Tacx's figures...

so, no, there's no real need for the GSC-10 on the chainstay (or the newer ones on the hub and cranks...) unless you feel you must - though if you do a lot of "urban canyon" type rides in the centre of town, it's often a good idea to have the speed sensor on there in case you lose gps signal...

oh, and yes - the cadence is "estimated" by measuring the "pulses" in your pedalling, i.e. the less smooth your pedalling style is, the more accurate the cadence figure comes out - the better your souplesse, the less accurate it gets.

ntnunk

TheBigYin wrote: »

well - if I pair the Tacx Bushido smart with my Garmin 520, it'll feed back speed, power and cadence figures should I desire to use them. By default preference the 520 will take the power (and maybe, haven't checked) cadence from my power meter however, so I would have to de-select the power meter as a sensor if I wanted to use the Tacx's figures...

I believe the 520 will prefer power meter cadence if available.

TheBigYin wrote: »

so, no, there's no real need for the GSC-10 on the chainstay (or the newer ones on the hub and cranks...) unless you feel you must - though if you do a lot of "urban canyon" type rides in the centre of town, it's often a good idea to have the speed sensor on there in case you lose gps signal...

FWIW, I have GSC-10s on CX and MTB due to the havoc slow, tight 180 degree turns tend to play on GPS speed and distance. For the road bike I use strictly GPS speed. I have one of the new hub-based speed sensors on my TT bike for the sole, solitary purpose of collecting speed and distance data on my rollers. It'll probably get removed for outdoor rides.

TheBigYin wrote: »

oh, and yes - the cadence is "estimated" by measuring the "pulses" in your pedalling, i.e. the less smooth your pedalling style is, the more accurate the cadence figure comes out - the better your souplesse, the less accurate it gets.

That's been my experience. The smoother your pedaling, the less accurate the cadence meter. Even at my smoothest, it's generally been usable. In that regard it's kind of bad because it could be seen to be "encouraging" bad pedaling technique.

35dollars

I strongly suspect any cadence measure of my cycling would be very accurate