Tacx Trainer ANT+ FE-C Support?

[Leszek Prozorowicz]

New firmware for TACX Bushido/Vortex Smart add support for ANT+ FE-CE. It will be possible to use it in the GC?

[Jon Escombe]

Hi,

Not yet, but there is a feature request already raised for FE-C support.

Regards,
Jon

[_ JUM]

Hi
Any news about this issue?
Regards.

[stsc...@gmail.com]

If you build your own copy of GC, feel free to try the patch here:

https://github.com/GoldenCheetah/GoldenCheetah/pull/1448

[Mark Liversedge]

On Wednesday, 15 July 2015 06:57:17 UTC+1, stsc...@gmail.com wrote:

If you build your own copy of GC, feel free to try the patch here:

https://github.com/GoldenCheetah/GoldenCheetah/pull/1448

The only reason I haven't pushed this in straight away is that we are now feature frozen for v3.2 and so I will put this into v4.0.

I expect v3.2 to be released sometime in the next couple of weeks.

Mark 

[stsc...@gmail.com]

Sorry, I didn't want to come across as impatient, just wanted to update the mailing list side of things with whats going on over at GitHub.

[Mark Liversedge]

On Wednesday, 15 July 2015 19:50:54 UTC+1, stsc...@gmail.com wrote:

Sorry, I didn't want to come across as impatient, just wanted to update the mailing list side of things with whats going on over at GitHub.

No need to apologise ! :)

I was just getting my excuses in early !

Mark 

[Marco Silveira]

Hello 
Bkool just opened with ANT+ FE-CE.
ANT+ FE-CE suport will be great for a big number o bkool users like me.

Keep up the good work. 

Anything i can do to help? 

[Lee Springett]

 Hi,

Was there any update with this. I connected my BKool classic up last night and was detected under fe-c (not power as doesn't transmit a live figure) then just picked a random trainer from the list.

When I ran an erg file there was power data feeding into GC and the turbo resistance was changing but have no idea how accurate,

[Jon Escombe]

Hi Lee,

Yes - FE-C support was added with 3.3, good to know it's working with
the BKool.

For info, you don't need to pick a trainer from the list, as that's only
used for estimating virtual power for 'dumb' trainers. Tbh, not sure
what happens when you select one in this situation? (Worst case it will
make up some virtual power numbers which you won't want)..

Regards,
Jon

[Lee Springett]

Cheers Jon,

So just to confirm as its fe-c the power data going into GC and the resistance info being sent to the bkool with be accurate (as can be) without needed a specific power curve profile needed?

Thanks
Lee

Sent from my iPhone

[Jon Escombe]

Yes that's right, there's also some calibration support coming in the next release that will help with the accuracy..

Regards,
Jon.

[Todd Giorgio]

I'm using a Tacx Vortex Smart with GC 3.3.0 build 3933 and a Garmin USB ANT+ receiver. Everything works as expected...except...speed. After my GC session, I uploaded the TCX file to Garmin Connect and took the screenshot attached. It's not so much the absolute speed, but the speed fluctuations that are concerning. Note the lack of fluctuations in the power and cadence channels. Suggestions? Very reproducible and not a function of ANT+ receiver-to-Tacx distance - I've used a short USB extender and moved the computer so the receiver is within 10 cm of the trainer - no difference. Suggestions?

[Jon Escombe]

Are you paired with any other ANT sensors? If you can email me the
antlog.bin file for a sample session, I'll have a look at the data (this
file should be created in your working directory).

Regards,
Jon

[Todd Giorgio]

Jon...

Thanks for your quick reply.  While using GC and the VS, I had FEC and speed and cadence all 'on'. I also had a Garmin ANT+ heart rate strap connected. Should there be an antlog.bin for each ride, or one for the history of all rides, or one per TCX export. If the latter, is the antlog overwritten by subsequent TCX exports?  Sorry for all the Q, but I've carried out half a dozen tests and exports... And I want to send you a useful antlog... Todd

Sent using CloudMagic Email

[Jon Escombe]

Hi,

There should just be a single antlog.bin file, this gets (re)created at
the start of each training session - overwriting any previous content.

Cheers,
Jon

[Ryan Switala]

Are you sure the problem is GC and not, the tacx. I have the vortex and the speed numbers are garbage. Try connect the trainer to a different app (e.g phone, garmin etc) and see whether the same thing happens. If it does you know it has nothing to do with GC.

[Todd Giorgio]

Well, I guess that I'm trying to figure out if it is something wrong with *my* Tacx justifying a warranty claim or FEC teething pains between GC and Tacx.  So I don't know that GC is the problem.  I've done a bunch of tests and will post some of that here soon.  Any news from anyone else?  It would make me feel better if all of the Tacx Vortex Smart users observed similar speed weirdness on GC.  I think.  

[Todd Giorgio]

OK, so here's some data.  My ANT/FEC pairing screen from GC is attached.  I'm wearing a Garmin ANT+ HR chest strap and have a Tacx Vortex Smart powered up, green flashing lights (all ok) and have turned the cranks a bit to make sure that everything is awake.  You can see that I captured the screen while turning the cranks by hand - data appears in HR, power, and speed+cadence - and all of these are paired.  The FEC channel is also paired, but the value is 0. Note that the ANT+ ID for power, speed+cadence and FEC are all identical - is that a problem?

I did a short ride where the last 2+ min were constant velocity/cadence (using 100 W erg mode).  The TCX file is attached.  You can see the negative deviations in speed with no similar weirdness in cadence or power.

So...more info...Jon had suggested that I mark as 'unused' all of the paired channels except HR and FEC.  That did not work at all as I rcvd no data from that configuration, not even HR (!).  Interestingly, enabling HR alone works, power alone works, HR and power together works, but other combinations did not seem to work (or had these weird speed deviations).  So...is the problem that power, speed+cadence and FEC all have the same ANT+ ID?

Happy to do more tests...Todd

On Saturday, December 26, 2015 at 5:36:37 AM UTC-6, Ryan Switala wrote:

[Todd Giorgio]

Ryan...sorry for the barrage of messages - bad form to respond to myself, but here is some more data.  Same setup, but I've attached a FIT file from a Zwift ride.  Same computer, same ANT+ receiver, same FEC (and it does work).  Speed has no negative deviations.  Power, cadence work fine.  So here is an example of a receiver (software + hardware) that interprets whatever the Tacx is spitting out without those odd speed (negative) spikes).  Thoughts on this?  Todd

On Saturday, December 26, 2015 at 5:36:37 AM UTC-6, Ryan Switala wrote:

[Jon Escombe]

I would persevere with just pairing the HR and FEC channels. No good
reason for that not to work - is the setup I use with my Tacx Vortex..

Is not an issue that that FE-C value stays at zero (as the ANT id is
populated), is also fine that Power, S&C, and FE-C all have the same ANT
id - the trainer is just broadcasting on all three profiles.

Worth noting that Todd sent me his ant logs when paired with both the
FE-C and S&C channels, and the inbound data was all good - both channels
were in agreement for a 700c wheel size - with no obvious drop-outs.

Regards,
Jon

[Ryan Switala]

Hi Todd

Im certainly not knowledgeable enough about GC to help you technically troubleshoot it. Im also not near a pc over the next few days to look at the files. I have not yet tried my trainer with GC yet either.

All i can share is my experience with the same trainer. With my vortex i saw strange power and speed issues. The speed issues were particularly bad when the trainer was set to a certain power (i.e. powe/ergo mode, it seemed ok in say slope mode in the trainer). I contacted tacx support and they gave a file to help with firmware upgrade which helped with the power but not really speed. (I was less concerned about speed on the indoor trainer though so just left it at that)

It could be GC-tacx integration issue, but my bet is that it is the trainer itself.Given the thread in the tacx forum, where you are seeing the same thing recording via your garmin (and via tacx app) seems to bback up my hunch that it is the trainer itself not GC?

[William Black]

I've had a few trainees with Tacx trainers and all have given up on those due to erratic/erroneous readings.

Bill Black

> Date: Mon, 28 Dec 2015 12:32:25 -0800
> From: ryans...@gmail.com
> To: golden-che...@googlegroups.com
> Subject: Re: [Golden-Cheetah-Users] Re: Tacx Trainer ANT+ FE-C Support?

> --
> _______________________________________________
> Golden-Cheetah-Users mailing list
> golden-che...@googlegroups.com
> http://groups.google.com/group/golden-cheetah-users?hl=en
> ---
> You received this message because you are subscribed to the Google Groups "golden-cheetah-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-u...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Todd Giorgio]

I'll admit that some/most of my motivation here is compatibility with years worth of training data based mostly on mileage and speed, even indoors.  For a single trainer (all I've had until last week), the relative speed and distance provided a meaningful measure of workload.  Now, with a Vortex Smart - or any other adjustable resistance, FEC trainer - well, I *get* the approach to use power to estimate speed.  Still, how hard would it be for the Vortex Smart to send useful speed data on the ANT+ speed channel?  The 'problem', I think, is that within the constrains of an ANT+ speed sensor, both the transmitter and the receiver need to know the wheel/tire circumference for this to work.  That, if I understand correctly, might work with FEC, where the wheel circumference could be provided to the receiver and the receiver (which controls the trainer) sends that info to the trainer.  THEN, the trainer knows both the circumference of the spindle and the wheel/tire, does the math and sends a 'tick' for every wheel revolution.  Just like a magnetic speed sensor.  But what happens for a non-FEC receiver that is unable to send the wheel/tire circumference?  Then the trainer could default to 700/23C (or whatever the designers think most users will employ) and send the 'tick' per wheel revolution based on that,  In *any* case, it's difficult to understand the observed *variation* in speed measurement.  The pattern is not obvious, so it seems random.  But how can that be, knowing what little I know about ANT+?  Is this a Tacx hardware fault?  A 'programming' fault?  Todd

Bill Black

> To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-users+unsub...@googlegroups.com.

[Jon Escombe]

Hi Todd,

The good news is that there's nothing in the data you sent through that
suggests a hardware fault. The speed data coming from both channels is
correct, and doesn't contain the dips you're seeing once it's captured
in GC.

The speed data on the FE-C channel is in m/s, and so doesn't require any
special processing. However as you rightly suggest, the S&C channel has
to know something more - as this profile type only reports wheel ticks
and timestamps. For a real S&C sensor, only the head-unit needs the
wheel-size to determine speed. For the trainer based S&C sensor, the
trainer needs to know wheel-size as well (as it has to calculate the
virtual wheel ticks).

There is a mechanism to send the wheel-size from the head-unit (GC in
his case) to the FE-C device, but this code didn't make it into 3.3.
However, looking at your data, the trainer is defaulting to a 700c
wheel-size, and so everything lines up (see attached graph of the raw
ANT data).

So anyway, my best guess is still that GC is calculating conflicting
information for the two channels - i.e. a wheel-size configuration issue
on the S&C channel. Both channels will get fed into the telemetry, which
is then sampled at 4Hz, and averaged into the .json ride file. Is worth
noting that although both channels will nominally broadcast data at 4Hz,
there are housekeeping messages interleaved with this, so you will not
always be receiving new speed messages from each channel between each
sample.

Could you perhaps send me your configglobal-trainmode.ini &
athlete-preferences.ini files, and I'll see whether I can recreate here..

Regards,
Jon.

[Jon Escombe]

Hi Todd,

Just to confirm I can recreate the issue with your config files, I'll
take a closer look over the weekend and see what's up..

Regards,
Jon

[Jon Escombe]

I see what the issue is, and it's definitely with the S&C channel. Will
see about a fix, but in the meantime you could just not pair with that
channel as a workaround. For Tacx devices, power speed & cadence are all
provided over the FE-C channel anyway..

Regards,
Jon