A way to use new smart trainers?

Post Reply
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

A way to use new smart trainers?

Post by Russ »

Feature request for the new smart trainers with ANT+-FEC protocol.

In thinking about it, these new trainers such as my new Tacx Bushido Smart trainer transmit power data in ANT+ format.

Seems it would be easy to add a trainer selection based on current outside ride capabilities, to collect rear wheel speed and cadence from on bike sensor and power data from the trainer, thus not needing a -FEC version of the protocol (short of that).

The elevation, slope, wind would all be bogus as long as the power data were sourced from the trainer (ala DPFM) then the TSS, NP KJ data should all be correct along with speed, power and cadence - no?

I cannot yet test out if it will just work because I am waiting for repair for my new trainer so I am back on my dumb Kurt Kenitic.

Comments, thoughts?

Thanks,
Russ
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: A way to use new smart trainers?

Post by Velocomp »

Could you please explain again? I've read your post and cannot figure out what you are saying. Sorry!
John Hamann
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Ok, my brain damage does cause problems :-)

I will try another rewording approach...
I have purchases a Tacx Bushido Smart trainer which uses ANT+-FEC protocol via ANT+ and BLE.
The ANT+, as I understand it, sends power data out in standard power ANT+ format, like a DPFM.

I like collecting my trainer data with the iBike, and have been doing so for my ole Kurt Kenetic Road Machine for years
with the trainer mode. Works great even when I am running some virtual training app on the ipad, the iBike data is better
and the CSV is great for GoldenCheetah, etc. I would like to continue collecting the power/speed/cadence, NP, TSS, KJ
data with the iBike. I do not desire the iBike to 'control' the smart trainer, only suck off the power data.

I would imagine that a simple trainer profile is impossible due to the changing load that the virtual app commands the
trainer to apply. But I would think it would be fairly easy (no?) to use a normal iBike bike profile to capture the speed,
cadence and power from the bike and record it and override lack of accelerometer derived power. I can ignore the
slope, elevation and wind in this sudo trainer mode. Perhaps it would just work, I have no experience with a DPFM
to finish figuring this out experimentally and I am waiting on parts for the Tacx. Perhaps it would need a firmware
tweek for this to work.

Just asking and hoping.
Thanks John and Happy New Year,
Russ
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Happy New year!

I actually have a Bushido smart, only 4 weeks, and a DFPM (Power2Max). I would help you out no problem, but I don't understand what you are saying.

I can tell you a few pseudo-random things and hope it answers your question.

1) F-EC is part of the ANT+ FIT profile. It is only useful for control. As the Newton has an interval training optional it would be really nice if it did do it (preferably a virtual ERG mode using power in), but I imagine that might be too hard to do...
2) F-EC and Power are separate ANT+ profiles
3) the trainer can be controlled and send out power/S&C both in ANT+ and BlueTooth
4) you will need a phone or tablet that can do BTLE to calibrate it
5) when comparing the DFPM and the power from the trainer there is an error (even if it is just calibrated), although it seems quite consistent. I don't honestly know which is best for power: bushido or Newton. I haven't had a look, but I wouldn't be surprised if there isn't a virtual power profile for the Bushido Smart yet, in which case you can send power ANT+ from the trainer to the Newton. I wouldn't trust the calibration of the trainer too much though - make sure tyre pressure etc. are identical anyway. The Bushido has 'progressive resistance', which I think means resistance is non-linearly related to speed. I'm sure it can be graphed, but I wouldn't bodge on another trainer profile.
6) I haven't tested it but I suspect you could teach the Newton the details of the trainer's S&C profile (and Power) and it would hopefully autoswitch; you would then have power, Speed, Cadence and all those other metrics you want. I think Velocomp should be able to confirm. If you leave the S&C sensor on the bike I would move the magnets to avoid waking up that sensor... It is possible, if you really stomp, to get the back wheel to slip, but overall I think the speed and Cadence, from the trainer, are probably quite accurate.
7) on an entirely different matter; the Bushido has a virtual flywheel and seems, to me, to work much better when the wheel is moving at high velocity, so stay in a high gear for even low cadence work. It also smooths out changes in Braking when at low cadence, although I think this is again related to the virtual flywheel.
8) the Tacx App itself has limitations. For me I find it most annoying that it will not take any data that it can provide itself e.g. Power. It ignores my DFPM. I can see no reason, however, that you could not cycle up a virtual mountain in the App whilst logging everything on the Newton. The best app I have found is however TrainerRoad, but it is for training and the video implementation is poor.
9) the cycleops virtual training app is the second best and there is a 12 day free trial period, so you can have a mess around.
10) the Zwift program needs a computer. Golden cheetah also needs a computer, but is free and has F-EC. In ANT+ 'pairing' is not real pairing; e.g.. You can send ANT+ power both to Golden Cheetah and to the Newton. You could control the training session using Golden Cheetah, but log everything on the Newton for analysis in Isaac...

That is everything I can think of. Quite happy to test something else if you would like, but please explain exactly what and I will do it (rest day on Monday, so should be doable then). As I have a DFPM I think I can probably create a power-speed curve; Will need to look at the maths used in Isaac, as I'm not sure if it assumes linearity or can do curves...
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: A way to use new smart trainers?

Post by Velocomp »

This discussion is getting beyond my pay grade... :-)

You can do curve fitting with Isaac!
Attachments
Screen Shot 2016-01-02 at 6.11.29 PM.png
Screen Shot 2016-01-02 at 6.11.29 PM.png (78.73 KiB) Viewed 50662 times
John Hamann
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Hello EHB!

Thanks for all the info, you do answer some of my questions which I pose in relative ignorance have a broken Bushido :? I knew the Bushido should send out power as ant+ and possibly cadence but I had not heard mention of speed, so sounds that the iBike could simply pair to the Bushido and collect all the data. I am not sure why the bike s&c would not be fine for the speed and cadence part.

My best understanding is that the Bushido, in stand alone offline mode, does have some kind of power curve that seems not to have been published. If you were interested in taking the necessary measurements to develop a profile for that, it would be of some use for offline trainer use such as when the Internet is down of the trainer is on the road someplace, but I would not expect to use that very frequently.

I am paid up for awhile with Kinomap Trainer and hope to use that with the Bushido when I get it working. I can use it now with my old Kurt Kenetic Road Machine.

So the challenge is can I get the iBike to collect the data when the Bushido is power controlled by the app? The part I do not quiet understand is with no physical accelerations that the iBike can sense with it's accelerometer, can the power in data via ant+ overide the power data that the iBike would calculate from speed, CRR and CdA alone?

I hope this is clear enough, I am still working through some of the understandings so I am groping at the expression (obviously).

Thanks so much for you input!

Russ
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Hi Russ,

Sorry to hear your Bushido is broken. One of the ends of mine was pushed in and I had to hook it out. In not sure the padlocks good enough.

if you turn on the trainer mode then the Newton will not calculate power using CdA etc. It uses when velocity in a similar way that TrainerRoad and Golden Cheetah calculate 'virtual power'. This can be done in both Isaac and on the Newton itself.

Thus, just turn on trainer mode on the Newton.

As regards the Speed and Cadence sensor on the Bushido, I was merely suggesting using to take advantage of the Newton's Automatic profile selection....e.g. I think if you turned the magnets around so the bike's sensor stayed asleep and cycled on the Bushido the Newton would see the ANT S&C from the Bushido, switch profiles automatically and look for the power meter using without you having to do anything. I haven't tested this, but Velocomp can correct me if i'm wrong.

I don't think you will have much challenge getting the Newton to record the power whilst the App controls F-EC. Firstly F-EC and Power are not the same thing. Secondly ANT+ does not have proper pairing. It is called pairing, but isn't. If the Bushido broadcasts ANT+ then loads of devices can pick it up at once. I have often recorded rides on a computer, bike computer, iPad and multi-sport watch.

Velocomp very helpfully attached a screen shot of the formula used for trainer mode. I am not due to be cycling on Friday, so will probably be able to do the test then. Whether it will fit with the formula in Isaac will depend on how the Tacx Engineers have programmed the progressive function. There are four functions so hopefully it should be ok, unless Tacx have done something really weird.

BW, E
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Hello again EHB,

Sounds great, John did not say in this thread but before, he has said that if you develop the profile for a trainer with your dpfm, send to Velocomp and they will usually add it to their distributed list!

In any case, I would add it to mine if you so kindly provide the data! If so, please describe your tire pressure and approximate screw adjustment pressure (turns) so I can try to duplicate.

Thanks in advance!
Russ
Pete
Posts: 110
Joined: Wed Mar 05, 2014 8:09 am
Location: Melburn, AUS

Re: A way to use new smart trainers?

Post by Pete »

Why can't you just pair your smart trainer to the iBike as if it were a DFPM and not be in trainer mode?
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Sure Russ, I was planning on noting tyre pressure (prob 105PSI) and getting the calibration as near to the middle as I can. As you'll note, as the Bushido has no flywheel to push, it is hard to manually test for slip, which is why I think middle on calibration is best. If I can I will attach a copy of the screenshot.

I've also thought that the progressive nature of the Bushido is such that it can be changed by firmware updates....so I'll document the firmware version.

Pete, the plan is to use the broadcast ANT+ Power. This is a backup in case the broadcast part of the Bushido stops working. Also, I would be interested to see the what this 'progressive' coding of the firmware will look like. One of the advantages of the Bushido is that it is free of wires, so it could be taken out and about, in which case you are more likely to use it in the free mode; I'd like to know how it is going to react.
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Pete,
"Why can't you just pair your smart trainer to the iBike as if it were a DFPM and not be in trainer mode?"
This is one of my questions, I see Newton+ will, with "ON/OFF DFP(M) DISP" setting display DPFM power in.
I also see from those Newton+ instructions that Newt will record both it's view of power and the DPFM (power in from trainer).

Since I do not have a dpfm and my smart trainer is currently awaiting parts, I cannot 'see' a sample DPFM 'signal', as Isaac calls it, to evaluate ability to show and export into CSV file the DPFM data over the iBikes' own measurement (which would be largely bogus in this case).

So I look forward to EHB's measurements for offline use! And will experiment when I get my trainer working to see if there is a useful option there.
Specifically, again, my challenge, to have Goldencheetah use the DPFM trainer power instead of the iBikes own... to be determined.

Regards,
Russ
Pete
Posts: 110
Joined: Wed Mar 05, 2014 8:09 am
Location: Melburn, AUS

Re: A way to use new smart trainers?

Post by Pete »

OK, well if/when you can get it recording, your CSV file will have Newton power in the 'Power' column and trainer power in the 'DFPM" column.
For export to GC you will have a manual step to copy/paste the DFPM column data to the Power column because GC won't load the DFPM column
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Thanks Pete,

That makes sense. I am not a spread sheet wonk, only basic skills, but it would be great if someone with one of these trainers wrote a conversion sheet that would get the time sequenced speed, cadence, power and build one suitable. I am thinking in the process set the wind to zero (assumed for virtual rides) and then back calculate and insert slope and elevation data from the power, CdA, CRR, weight and starting elevation values! Just dreaming at this point :-)

Russ
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

I've noticed a variability between my DFPM and the Bushido. The DFPM automatically recalibrated each time I stop (for 3sec), which I've tried and there is no step change in the difference.

I thought for a while it was room temperature, but that doesn't seem consistent. I was worried for a while that it just wasn't consistent enough, but have changed my mind now.

My current thinking is that it is related to the Bushido warming up. I've done a few sessions in ERG mode with no correction now. The general thing is for there to be a 20-30W difference at the start. A surprising amount of time later the two measurements come together, but it is like 13 minutes later (and not just messing around - even solidly at 60% of FTP). After that point the error is about 12W.

It also seems to happen again quite quickly if you stop.

I am assuming the same will be true when relating wheel speed to power (the software doesn't graph speed).

So I would suggest a 15 minute warm up and not chilling out too much between intervals.

I attach a screen shot of today's session. The spikey bits are because I was doing drills or removing clothing. The blue filled bit is what the Bushido thinks is targeting as power. The yellow line is a smoothed DFPM trace.

Edit: it won't let me upload a picture, as the screen shot is too many pixels. I will need to see if I can simplify the image.
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

iBike_01_13_2016_1905_18_km_Tacx.csv
First of two, 40 minute on Tacx Bushido
(261.4 KiB) Downloaded 455 times
iBike_01_13_2016_1948_9_km_Tacx.ibr
Second Bushido with tweeks to CRR and CdA
(57.81 KiB) Downloaded 483 times
EHB,

I appreciate the feedback/experience data! And I finally got my part and my Bushido is now functional. My two back to back test rides, after installation the part which came with 3.1.13 firmware (latest) was interesting. The iBike recorded it's view of power based on more or less flat road and the profile values while the DPFM power data from the Bushido recorded as you explained, in its own column in the exported CSV file.

Also, I guess since the test ride was reasonable flat, I was able to 'tweek' CRR and CdA values to nearly match the two power records as displayed by the power meter comparison for what ever that is worth (or not worth). I will try to post the screen shot and will post the with tweeks ibr file for those interested in looking.

My perceived exertion was much like you described ie, exertion higher than the Tacx reported. I did a few 300watt intervals near the end and they actually felt about right. Now I understand that better, thanks to you detailed study of your data, that the unit was probably warmed up enough to be more correct. I road two different 'virtual' rides in Kinomap (after calibrating and checking with Tacx utility and app). The first, when viewed in compare to power meter data, came more together after 18 minutes, in my case. Possibly because I road at lower average power than you for an extended warm up for the unit. There was, in the second short ride which I am posting, about a 10 meter descent for first 8 minutes then flat with a small 10 meter climb and descent at about 18 minutes so basically, a flat ride.

I was able to move the power data from the DPFM column to the main power column as you suggested for Goldencheetah!

Also I see a forum topic on Tacx site you may want to weigh in with your observations!
http://forum.tacx.com/viewtopic.php?f=6 ... 2&start=20

Oh well, my screen shot is too big and I haven't figured how to crop in Libra Draw so no screen shot but ibr files, I will post both
so the warmup data coming together can be seen in the first 40 minute one followed by the 20 minute on I tweeked.

Regards and Thanks!
Russ
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Hi Russ,

Glad to hear you are up and running!

Incidentally, you know Golden Cheetah can do F-EC control of the trainer? There is also a soon to be improved workout generator. Admittedly, rather more boring than a real simulation ride. The current workout generator is a bit clunky and is hidden in the tools option (Mac).

Although, there is a way to create a control file from a GPS route! I haven't done it yet, but will be for upcoming races/sportives. I believe you just need a free account with FastestSplit.

As it happens I am doubly glad you are up and running. I've just got my tyre at 105 psi and the calibration at exactly middle, but I can't get it to let me cycle with less than about 230W. Even in lowest gear (39x28) and with cadence of 50.

As the creation of a trainer profile needs really low speeds, like 3mph, I'm not sure it is possible. I will have a go again tomorrow to see if it needs to forget the last setting or something, although I tried having a break. Currently thinking that for warming up purposes it may be necessary to have a phone, as 230W minimum and a cadence of 50 is not a warm up.
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Incidentally, in the TACX forum you linked to I suspect the big differences between the trainer and DFPM are right at the start when it is cool. I get similar errors right at the start.

I've done lots of cycles and at the end of the cycle the error is consistently 12-13W (Tacx low). You just need to ignore the first 15 min or so and not stop much between intervals.

I've been told by an independent company not to use their calibration and to use the Tacx iPhone/iPad App itself. Not sure why. The TTS software they are on about is notoriously rubbish; I suspect it is behaving badly too.

I guess I haven't tried the new Open BT control that came with the last firmware update, but the BT card in my Mac is too old to do BT4.
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

EHB,

I rode today, first a 'warmup ride' two times over total 21+ minutes, totally flat course. At about seven minutes in, the two power graphs in compare power meter function came together with my high CRR and CdA tweeks. My second ride immediately after was Montigo Bay in Jamaca, it had a few small hills two of any significance. The first of the two put out 12% for only a few seconds and my cadence dropped to 45. This bike is not geared as low as my outside bike but I was surprised as I ride a few 12% short rollers around my area and one 16% occasionally. Later the last 'steepish' hill was 6.6% for a 1' 30", cadence dropped to 64 and it credited me with about 190-210 watts during the climb. I should have been able to easily hold 250 or so for that (219 FTP - iBike profile needs update). Outside if I hold up to 275 or so I can recover well, much above 300 watts and I have recovery issues (I am an old man :-). I figure my CP (Goldencheetah, Dr Phil method) to be around 234 watts.

So what I am saying is it seemed to still be under reporting power to some degree for me after the warmup. I will tinker with the calibration some and see what happens on that.

I use a mini ipad retina and it is so convenient I have held off setting up a laptop for Goldencheetah to link. I did buy an ANT+ usb dongle in case I need to do that.

Oh, the power compare looks right weird when hills are involved as iBike interprets the speed drop as a power drop :-)

I look forward to any more conclusions you draw from your testing and use!

Russ
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Hi Russ,

I've noted it doesn't seem to like slow cadence as much, but I haven't really tested it much at low cadences. I tend to keep at least at 80 and mostly spend time between 100 and 110. I can imagine it doesn't do low cadences as accurately; how do you feel the electronic flywheel works at those cadences? I've only tried a couple of hill sims, when I was testing apps. I thought it felt jerkier than in real life. I am hoping that gets improved with a firmware update.

I think there is a markedly different physiological cost between
1) X Watts and jerky
2) X Watts and smooth

The ANT+ protocol only requires power once each second. That is not actually very often and can miss a lot. I think ANT+ can go up to 4/sec, but DFPM I've seen just broadcast the same number and only change each second.

It is also true that there is a different physiological cost between
1) Y Watts and no drill
2) Y Watts and drill (most obviously single leg drills)

and
1) Z Watts and seated
2) Z Watts and standing

I would not be surprised if the power numbers it is giving you were okish, but it was doing something to make it harder for you. Can you do anything to adjust the gearing and see if it feels better?

NB: there are supposed to be 1-2% drivetrain loses in power between a crank PM and a Powertap Hub, so Newton to trainer should be even more. Also FTP inside is often lower than outside; I've read a lot of hypothesis why (e.g. temperature, rotational momentum etc.). The gap still seems a bit big though.

I've tried the isolated trainer again and am pleased to say it was much better today. It doesn't like calibration being done and then the device being removed. I did try waiting 10s or so yesterday, which didn't work, but today was a completely different thing. So that is one important lesson.

It is very non linear! Essentially at low speed the brake is off. Once you go faster it racks them up. I'm having a rest for an FTP test at the weekend, so I didn't go that high, but it certainly shoots up.

so:
105psi Schwalbe Trainer Tyre 23mm
calibration exactly in middle of the two lines yesterday (nothing changed since)
15 minutes warm up (approx 150W) and no break!
latest Firmware (20 jan 2016)

I chose to use the trainer's PM and Speed. I thought using the trainer's PM would add consistency (as the use for this is if the trainer's PM can't broadcast) and I've never done a roll out with this wheel/tyre so if I used the bikes speed it could be wrong.

I've sent the file to Velocomp but will attach one to this post.

All in all it is good. It does mean you can take the Bushido with you in the car and use it for warmups without your phone (as long as you keep the speed below 17mph or so). Just don't re-calibrate it before the warm up.
Attachments
Bushido Smart power and speed.ibr
(227.71 KiB) Downloaded 415 times
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Hello EHB,

Regarding your flywheel question, from the forum it was not clear what units had and did not have the flywheel. I had gained the impression that it was in old non-smart units but then removed but perhaps they added it back in.

Well, after I inquired about my replacement part suitability (it was an older serial number and all), they asked if I had a flywheel or not and asked I send a picture of internals. They then assured me I have the correct part and I saw, and they said, I do have a flywheel. I must say it is not a large flywheel though. So I do not know what the true shipping history with or without said flywheel truly is, I only know that I do have one!

Thanks for your further analysis and observations, comments. And yes I am aware of varied physiological costs and trainer differences in general. I started with a Fluid2 and moved to the Road Machine for the 'known' power curve which turned out to be any of several :-) I also experienced the ~15% perceived difference and adapted to it. I found with my Road Machine that I could run about 150psi and a light contact and nearly erase the ~15% perceived difference. I just had to not jump on it too hard. Not jumping on it too hard had a consolation in that iBike trainer function being pure speed based could not credit me with massive accelerations to get up to higher speed quickly. So I opted for smooth accelerations of a steady power application until I achieved the power level of target. I wonder and have not yet tried to see if this approach may help with the Tacx.

I also am not sure that I have had an 'off' day or two on the Tacx. All this may drive me to go ahead and purchase a DPFM!
Anyhow, probably snowed in (Mid Atlantic U.S.A.) this weekend, so if snow shovel does not take too much out of me perhaps I will get a new outlook on my Tacx :-)

I will update if I find anything worth posting.

Thanks again and especially for the data!
Russ
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Hi Russ,

I used to have a Tacx TT2200 FLOW. It gives a power output to its head unit. I realised it could be consistently good, but, even with roll downs, it was very sensitive to how it was setup. I bought the DFPM whilst I had it.

They will have improved things for the Bushido Smart, but it'd expect it would be still quite sensitive.

I have the impression there is a small fly wheel too. It feels too heavy to not have one, but it is lighter than my old trainer and the maximum possible wheel diameter is less. If you look at the angular momentum on the Tacx website they must be doing something extra. I think they originally possibly tried to get rid of it completely.

The combination of a DFPM with the Newton is cool. Live CdA and the time gained is a really good prompt to stay aero. It is also reassuring, as you can compare Newton to DFPM and tweak it so they match almost perfectly. Then you can move the DFPM to another bike. As well as the ability to tweak, it also gives confidence in the Newton's output, which I think some people are going to lack. In-fact I am a little surprised ibike don't offer an option to borrow a pair of really easy to use Powertap P1 pedals, or similar, for a couple of weeks at purchase. I suppose they could lend a really expensive SRM, but then there might be a pain with bottom bracket adapters, crank lengths, BCDs etc.
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Hello EHB,

I used your measurement ride to setup my Newton trainer profile and rode yesterday. In the compare power function, the Newton tracked just below the Tacx reported power. For some reason Libra Calc refused to start (a reboot later fixed that) so I imported the CSV unmodified into Goldencheetah. I was pleasantly surprised to find that GC used the power meter data column and not the Newton column. The percentage difference was about 4% on average power so that must be (hard to see values) the difference in compare power.

I suppose that 4% is a difference due to calibration. Since this is a dedicated bike on my trainer, I get away with using an old tire which is a Continental Gaterskin 700x25 I think (hard to read now) at 105psi. The difference may also be because of a difference in rolling resistance causing a variation from your measurements.

I can send (or post) the jpg I took of the flywheel, should you want to remove the end and investigate, let me know.
It is obvious, though, when you remove the end.

So the net of the matter with this new protocol is YES the iBike collects the data with or without using the trainer mode
and GC will import and use the ANT+ power in without bothering to edit the CSV file!

Caveat is with power controlled from an app, especially with hills, the Newtons calculated power will move inverse to actual power required for the hills, looks weird on compare power. Still GC will get the right stuff from the ANT+ power records.

Thanks again, great work!
Russ

Edit: To be clear, this ride was no controlling app, so the Tacx Bushido power curve (as measured by EHB) was in use
and it was simply like a perfectly flat ride.
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

I had a thought. The size of the real-flywheel is pretty obvious. I don't know if your previous trainers had a roll down calibration, but I had a Tacx Flow T2200, which had a display of power. It would use roll downs to calibrate. The Bushido smart however rolls down much quicker, suggesting that the real flywheel is smaller, although it still goes on a bit, so there must be some sort of real flywheel.
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Hi EHB,

Yes the flywheel is quite small, I will try to attach my jpg so you can see it.

Also, I had run 105psi with my Conti Gator's when I did the cal and test ride with no app control and your trainer tested coefficients (after loading file into Isaac of course) and my cal result was about 25% from left into the green band. Last ride I dropped pressure to 100psi and no change in the screw position, followed by a cal with the Tacx cycling app and it moved to about 50% (middle) of the green band. I was however using the Kinomap Trainer app with a totally flat warm up 10 minute ride.

Interestingly, the difference between avg power from iBike trainer mode and that collected from the Bushido was about 24% or up 20% from the 105psi ride. Both cases the Bushido power was higher. I do not know if the app controls flat at a different power for the speed, but I suspect the that is the case. However I now have an experiment in mind, better with a DPFM, if you are interested. I may try with the black-box approach if you are not interested.

My question is does the Tacx calibration only help gain the mechanically desired CRR or does the Tacx actually change values internally as well?

I think you could measure power after cals at several cal points in the green band and possibly variations by changing tire pressure up or down an increment or two without recal and post offsets of power measurements. If you do this, your results would probably be of particular interest over in that Tacx forum.

Ahh the flywheel picture is too large to post, ping me with your email if you want it and I will send it that way.

Thanks again!
Russ
EHB
Posts: 129
Joined: Sun Oct 26, 2014 5:52 pm

Re: A way to use new smart trainers?

Post by EHB »

Hi Russ,

I've been thinking about your idea. I think I understand (mostly).

So the thing we would be varying would be the position of the green line. To do this we would adjust rolling resistance by altering tyre pressure. Is that correct? How many? In thinking about 4. Possibly it would be easier to do quickly by turning the dial.

And then how many points on the power curve do you think we need? Doing similar to the Newton curve would take ages. I reckon cut out the readily low powers and change from speed to power. I guess maybe 100W 200W 300W for two minutes each. I would do 400 but in envisioning this for a rest day, so it needs to be easy. I'll average the last minute. I will try to stay in the same gear. I can simultaneously measure power (ANT+) from the DFPM and the trainer. Pretty sure I can do that in GC.

That would total 12 measurements. Each of the 4 runs would need 15 minutes warm up. With some brief messing about between runs makes 80 minutes, which sounds reasonable for my next great day (sunday next week). I can use the week to try and find a single gear I can do 100-300W in.

Is that roughly what you had envisaged?

(BTW my email is eerke@icloud.com).
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Hello EHB,

I just forwarded the flywheel picture via email.
I also will now send another email with more on my ideas
since they may simply weigh this thread down for now.
Post what ever my help, of course!

Thanks,
Russ
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: A way to use new smart trainers?

Post by Russ »

Now that I have had a few rides with this Tacx Bushido using the Newton6 in trainer mode with a profile developed from EHB's posted test data, I have new observation to post!

It seems that the NP, IF, TSS that the Newton displays at the end of the ride is based on the ANT+ power in from the Bushido. This is not the case for the uploaded data which has those metrics from the trainer profile power function. So you must not reset the Newton until you have recorded those metrics, if you want them.

They are likely different if you are using a virtual trainer application to control the power load of the smart trainer as opposed to a 'free ride' on the Bushido with no application having control so that the power curve matches (more or less) the reported power from the trainer.

It would be nice if Isaac could preserver these values and have them be a choice or second display from the upload.

Hope this helps!
Russ
Post Reply