The Tacx Flux Smart Trainer came on the market in December 2016 and started arriving to US customers on December 27th.  This trainer was a game changer this past fall, selling for roughly half the list price of similar direct drive trainers (discounted $709 price point  during the Black Friday/Cyber Monday sales).

A flag was raised on December 12th when Australian cyclist and blogger Shane Miller had an early unit (serial number ending #709) unexpectedly fail during a test ride (6:15 into video).

It became more concerning when Shane reported on December 17th that his #3396 replacement unit had failed (see 7min into video) with the same symptoms (slipping belt, burning rubber smell). He had inspected the internals of the unit and had not seen any apparent manufacturing differences from first unit.  This unit had be carefully prepared by the company prior to delivery.

I received my early production unit in the 1st batch of units reaching US customers on 27Dec.  It was a tough decision not to cancel the order given failure reports.

A Tacx Flux Owner Group created on Facebook discussed the issue. To their credit, Tacx  provided a transparent summary of their assessment of the situation to the group on January 11th:

Dear all,
Since the launch of the FLUX (end of November) there have been some FLUX trainers with quality issues (1,7% of all produced FLUX trainers). These issues found their way back to our company through social media, influencers like Shane Miller and our support department. Questions are raised about the quality of the FLUX and some of you have concerns about your own FLUX or future FLUX. Hereby we’d like to clarify the situation and restore the confidence in the FLUX of current and potential FLUX owners.
After analyzing the products with a defect that were send back to our HQ, we concluded that the problems can be brought back to two main issues. Below we explain our findings.
First issue
The first issue was caused by the tensioning wheel that might be assembled too closely to the steel chassis. During a workout and especially during sprints this wheel is pushed against the steel chassis, causing internal damages and the transmission belt running of. This problem will occur within the first hours of use, if it doesn’t; your FLUX is ok.
Please note this problem didn’t occur during our durability and stress tests, which was the issue with Shane’s FLUX during his sprint. After the issue was noted and resolved we made the necessary adjustments to prevent this from happening on all future FLUX products.
The number of products that got into the market and might have this issue is small, the current return rate is 1,3% for this specific issue.

Second issue
Analyzing the second issue was a bigger challenge as it was more difficult to reproduce this incident in order to find out what caused the belt to run of the pulley wheel. This second issue also occurred with a FLUX send to Shane Miller. When sending this second FLUX we double checked every assembly step to make sure this unit was perfect. Yet, this thoroughness has led to too much tension on the EDCO freewheel hub damaging the connector between the large metal pulley wheel and EDCO body. This weak spot cracked and caused a chain reaction displacing the large metal pulley wheel and letting the tensioning belt run of. This resulted in the grinding noise and smell noted in de video of Shane Miller. In order to prevent this issue from happening, we have tested the maximum torque on all the parts and adjusted this accordingly in the production process.
At the moment we only have a few cases with the same issue as Shane’s second FLUX. The vast majority of issues within this category is coming from the first issue.

Other issues
Besides the first and second issue we encountered some minor other incidents, but no structural problems.
Summary
The current total amount of damaged FLUX trainers known to us is 1.7%, with the majority coming from the first production batch showing the first issue. We are confident that most FLUX trainers with issues from the first production batch are currently known.
What can you expect from Tacx?
We are very committed in solving FLUX issues as fast as possible so you can start training again. Therefor we kindly ask all FLUX users experiencing issues to send an email directly to support@tacx.com. We will take care of your problem by replacing your FLUX unit for a new one. It will be very helpful for us if you return the damaged FLUX units for analysis, as we would like to be sure that the causality of defects stays within the first and second issue.
DO NOT OPEN A FLUX YOURSELF!
This is very a dangerous process as high voltage runs through the electronics! Also, it can cause other problems when not reassembling it correctly, like broken magnets and electronics. Please note we can no longer guarantee warranty when you disassemble a Tacx product. But above all, please avoid dangerous situations.
We would like to express our sincere apologies to all those who’ve encountered FLUX issues. We’ll do everything we can to replace your broken unit.
We’ll proactively share all product issues encountered at Tacx and are as transparent as possible in doing so.
Kind regards,
The Tacx Team

My unit (#230) failed after 47 miles.

Martin Smits, Business Development Manager from Tacx, subsequently shared on the DC Rainmaker blog comments that an “early unit” would be consider one with a serial number less than 1700.

Subsequently 70 failed units shared by customers in the Facebook Tacx Flux Owners Group and in comments on the DC Rainmaker blog Tacx Flux Trainer In Depth Review provided an emerging picture of the failure behavior of the unit. The initial reports were consistent with the summary from Tacx, but with subsequent failures occurring after 11Jan, the picture below has emerged:

There are many caveats to what can be inferred from isolated reports on the internet.

• How many high SN units are not in field?
• Are gaps in reported failure due to batches not in field or due to successful production?
• How many field units not ridden sufficiently?
• How many units not ridden at high wattage?
• How many failures not reported to FB group?
• Is reporting diminishing with late adopters?

Failure rates in various regions of production are noted on the plot.

Some observations:

• The reported failures run at 4.3% for 1st 500 units and then 2.6-6% for stretches under #2500.
• The failures appear to have persisted up to #2500 rather than the “early unit” 1700 threshold suggested by Tacx.
• Shane’s second failure came in a 7-unit, 2.2% failure rate stretch from #3300-3600.  No belt slipping/burning smell failures have been reported since this period.  Did Shane’s 2nd experience finally help resolve the issue for subsequent units?
• There is a 400 unit failure-free stretch prior to 3300.  Tacx informed a customer with a unit in that range that it included manufacturing improvements, so one might speculate that were pulled back from distribution.
• Although the data hasn’t been reported and tabulated, many of the reported failures took longer than the 1 hour suggested by Tacx to occur, particularly for riders putting out less wattage.
• At least ninecustomers have reported 2 failed units and one customer has had 3. That could point to an inherent failure rate of around 6% or that failures are more likely with certain types of customers — those that take their units out of their boxes, ride them, and ride them at a high enough wattage to trigger failure.
• Over half the units have failures within 8 units of themselves. That also points to a true inherent failure rate of about 7%.
• Customers in EU have reported great customer service directly from Tacx and rapid direct replacements. The ocean has gotten in the way of that for international customers.  Some early units remain on retailer shelves.

Bottom Line

I was driven to do the analysis above by need to decide whether to seek a replacement as well as intellectual curiosity about how much the cycling community could infer about manufacturing quality issues from crowd-sourced failure data.

Given that replacements have been slow to arrive in the US as well as the unclear picture whether certain riders might be doomed to multiple failures, I opted to go with and Elite Drivo on 28Jan.  I would consider a Flux for my spouse once the dust settles on this issue.

Seeing the emerging picture now that issues may have finally been resolved in production after the learning from Shane’s 2nd failure, I’m tempted to second guess that decision, but for now I’ll stick with the alternative but may consider a Flux as a 2nd unit for my spouse.

The picture is pretty clear for serial numbers under 3600, but not for those above.  At this point the highest reported serial number in the wild is 4404. Seven people have reported working units in that range and no belt failures have been reported above 3600.

Please comment below, on the FB owners group, or send an email to jackp@athletetechnology.com with your experience with those >3600 serial number units.

These phrases have never been spoken on the internet.

A coach had noted that data in earlier runs had showed deterioration of around 11% per doubling of miles. To quatify this I tried calculating a metric of “meters per incremental beat” (mpib) above an estimate of my standing heart rate (65). Plot below shows this run in those terms. I start at 3 meters/incremental beat and deteriorate to 2 after 9 miles. If I use my maximum heart rate of around 180bpm, that translates to 345 meters/minute or 4:40/mile, which is actually about the fastest pace I can hold for short spurts.

Learn to Run Barefoot with Lee Saxby and Terra Plana from GTB Goodtruebeautiful GmbH on Vimeo.

;

1. The par time is 2 hours plus your age in minutes, plus another 10% if you are female.  (2:48)
2. This will be achieved, after a year or so of effective training, if, in the last 13 weeks before the race – excluding the week of the race but counting the miles in all 13 calendar weeks prior to that – you’ve recorded an average of 55 miles per week.
3. More mileage than that contributes nothing. But for every mile per week less than 55, add one minute to your race time.  (3:28)
4. Provided, though, you’re not overweight. Par weight for men is 2 lbs per inch of height minus 5 lbs; for women 2 lbs per inch minus 15. Every pound over these weights adds 2 seconds per mile to your pace, say a minute to your marathon time.  (4:30)
5. Weather: add 2 seconds per mile for every degree (Celsius) by which shade temperatures exceed 18 deg C.
6. If it’s humid as well as hot, add 1 second per mile for every percentage point the relative humidity exceeds 65%.
7. On a windy day, add 1 second per mile for every mile per hour that the headwind exceeds 10mph : deduct 1 second per mile for every 2mph that the tailwind exceeds 20mph.
8. If the race director describes the course as “testing”, “undulating”, or more honestly “hilly”, add anything up to 5 minutes. e.g. the Sheffield course carries a handicap of 3 minutes for the 3 hour runner.
9. Now the positive bit. If you average, over the 13 weeks (see 2), one (short) hill times 10 repetition session or one (short) speed session per week, take off 4% from whatever 1 to 4 calculates to. For an average 1½ sessions per week, take off 8%. For 2 hill/speed sessions per week, take off 12%.
10. That’s it. But you can throw it all away by going faster at the start than the calculated pace and trying to put “minutes in the bank”. Remember minute banks charge at least 4 times as much for overdrafts as they allow on deposits, every second per mile you run too fast in the first 10 miles will cost you at least 4 seconds per mile in the last 10 miles.

Other Calculators

Jeff Galloway Magic Mile

MARCO method

Runner’s World

McMillian Running

Based on 2 distances

So far my Newton Running shoes came in at 985 while my New Balance MInimus Trail MT10’s seem to be 1005.

https://support.garmin.com/support/searchSupport/case.faces?caseId=%7B6e1aa610-6d25-11dc-6782-000000000000%7D