TACX Fortius and GC - initial setup

[Anders Tarlebø Løvik]

Hi - need some guidance on how to get this set up correctly. 

Have followed the advice found here and elsewhere on the net, current status is at the moment:

Have installed GC

Installed Fortius drivers and firmware (green light on the 1932)

In the add device wizard (after it asks for firmware) I cannot find any line for Fortius under the virtual section - is it supposed to be blank/none??

what about the power /cadence/ speed options at the bottom

Appreciate any tips

@

[Eric Christoffersen]

Hi!

I just checked my own private build to see what you're talking about and found that the fortius choices weren't there for me either!

Very strange as my defines and everything suggested it should be (it needs GC_HAVE_LIBUSB, which I have.)

I touched the file (devicetypes.cpp) and rebuilt and the fortius choices magically appeared.

There is maybe a problem with the build dependence?

Where did you get your bits?

Eric

[Ale Martinez (Please don't email or cc me)]

El viernes, 3 de enero de 2020, 14:42:05 (UTC-3), Anders Tarlebø Løvik escribió:

Hi - need some guidance on how to get this set up correctly. 

Have followed the advice found here and elsewhere on the net, current status is at the moment:

Have installed GC

Installed Fortius drivers and firmware (green light on the 1932)

In the add device wizard (after it asks for firmware) I cannot find any line for Fortius under the virtual section - is it supposed to be blank/none??

Yes, virtual power is only for dumb trainers, to estimate power from speed based on known power/speed curves.

[Mick Drake]

Hi Anders

did you get this working OK?  I have my fortius installed and working but the power fluctuates a lot ... I have trued averaging over 5 seconds but still quite erratic?

TIA

[nicholas colloff]

I have also been trying to link my TACX Fortius trainer to GC via T1932, but software doesn't seem to be able to find a port. I have followed all the advice I've read abot, but some of the files mentioned, I can find e.g. GC_HAVE..... Where can I find this?

[Ale Martinez (Please don't email or cc me)]

As far as I can recollect the last change to Fortius support is from january 2016 and it includes this driver you should install: https://github.com/david-waterworth/GoldenCheetah/blob/7f8305c2b34904960e7cd976f62ea578691ec727/contrib/GC_Fortius-VR_libusb-0.1_drivers.zip and this thread has some new information from that point: https://groups.google.com/d/msg/golden-cheetah-users/8EDSFJiRLGU/boqmuINdAgAJ, but it seems no developer has access to a Fortius anymore to be able to give some kind of support, may be other user can help.

[Tim Sinclair]

I used to successfully use my Tacx Flow/iFlow via USB in Sept 2018 with Golden Cheetah by using the exact same Fortius drivers that Ale has linked to. That was on a Windows 10 x64 system.

I hope you manage to get yours to work :-)

[Mick Drake]

I have a spare one in my loft LOL...

[nicholas colloff]

Thanks Ale, for your rapid response. I have downloaded the ZIP file ad extracted it, followed the instructions and "a problem was encountered while attemptig to add the driver to the store". Looking at the text in the setup file, it seems to be referencing Windows NT. I note that Tim Sinclair has used, in 2018, the same files on Win10 x64 (same as my system) successfully, so not sure what's going wrong. Any ideas?

[Ale Martinez (Please don't email or cc me)]

In the linked thread there is information about how to add the certificate so that driver signature can be verified.

[Mick Drake]

Good morning

I have re read the linked thread (many of the posts are from me) and I dont think there is a mention of the video chart / overlay working.  My install does seem to work in ERG mode which is possibly all that ever worked, maybe the video chart option was not coded?  Only guessing but could be the reason why I am having no success?

On Friday, 3 January 2020 17:42:05 UTC, Anders Tarlebø Løvik wrote:

[Ale Martinez (Please don't email or cc me)]

El lunes, 29 de junio de 2020, 13:30:58 (UTC-3), Ale Martinez (Please don't email or cc me) escribió:

In the linked thread there is information about how to add the certificate so that driver signature can be verified.

Specifically slightly adapted, from this message:

I use zadig to build the installer. It generates a digital certificate for the cat file and self signs it
So this certificate needs to be loaded into the windows certificate store in order to install the driver. You can extract it yourself by:
1) Right click on VR-Interface.cat and select properties
2) On the Digital Signatures tab select the certificate (USB\VID_ etc.) and click Details
3) View Certificate
4) Select Details tab then Copy to File
To install you need to run the following from an elevated command prompt
> certutil -addstore "TrustedPublisher" VR-Interface.cer
> certutil -addstore "Root" VR-Interface.cer

After that you should be able to install the driver.

[Mick Drake]

Hi I disabled driver signing on the PC.. installed the driver and the VR-Interface is shown in Device manager when I plug in the fortius

[Ale Martinez (Please don't email or cc me)]

Added a summary to FAQs for future reference: https://github.com/GoldenCheetah/GoldenCheetah/wiki/FAQ-TRAIN#how-can-i-use-my-fortius-trainer-on-windows-10

[nicholas colloff]

Hi Ale. I have followed the instructions to the letter. I'm not an expert in Operating Systems, or programming in general, but I can follow instructions. Managed to install the driver, (I think), after getting the system into Test Mode (after switching off security in the BIOS). A message appeared stating that the driver was of unknown origin. I selected "continue anyway", and I had no error message, so assume the driver was installed. Hooked it up to GC but it still cannot find the port for the Fortius interface. Tried every USB socket,

but it was still a no-go. I'm just assuming now that I don't have the necessary skills to be able to get this working, I must be doing something wrong, or missunderstanding something.

A shame because I would really have liked to try out GC, but not much good If I can't link to it.

[Mick Drake]

I finally used the "driver signing" instructions" which worked fine.

If you plug in the fortius then go to device manager you should see VR-Interface.. you can then look at the properties and see if the driver is there.  It probably is not, even if it is there the interface is flaky so give it a go and let us know if it works for you.  Mine drops out at random with a USB 115 error.  Others report it has run for an hour one day then only a few minutes the next...

As Ale stated previously non of the developers have a fortius so unless someone with some programming skills picks this up I think we are stuck :(

[nicholas colloff]

Yes. If I go to Device Manager and view Device by Container, I can see the VR Trainer, but see no mention of anything else, details of driver etc. I've always been able to see the VR Trainer, but have no idea how to see whether the required driver for GC has been installed. It's very frustrating.

[Mick Drake]

"right click" on VR interface

then properties

will show this

[Mick Drake]

Right Click on VR interface

Properties

and it should show this if its installed.. if not it will say no driver

On Sunday, 12 July 2020 12:05:17 UTC+1, nicholas colloff wrote:

[nicholas colloff]

Thanks Mick, that's REALLY useful. I currently see the following, which means The VR Interface is still on the TACX driver. Is it possible to save this driver so I can revert back to it for use on TTS4 if the GC proves to be problematic. Still not sure why the GC driver is not linking. Would updating the river give me this option?:-

[Mick Drake]

A "long time ago" I was told its not a good idea to have TTS and GC USB drivers on the same PC... Yes you still have the TACX driver installed

If you cannot test on another PC I think I would stick with TTS.. There has been no recent comments on this thread suggesting that GC traon modes and the fortius is stable.  I have recently been in contact with a guy testing this on windows 10 and then 7 and he still has no reliable working system directly in GC.  I have reverted to TTS and just upload my data to GC

[Michel Dagenais]

I have run the Fortius with GC on Linux and the installation went fine. However,  I had serious problems with the slope control and reported power. Since I did a fairly detailed analysis of the Fortius protocol back in 2010 to insure that it could eventually run on Linux, I could get back to my notes and try to dig into the problem. 

Here is what I found:

- The unit does not show much wheel resistance.

- The code uses byte 46 as the pedal sensor signal, used for feedback to the trainer and insure that the trainer software is still working correctly, something called a watchdog in computing jargon. The traces that I took show that byte 42 is the pedal sensor, while byte 46 is the brake status. Once you change from byte 46 to 42, the trainer starts to better respond to changes in slope.

With the pedal feedback corrected, other problems become visible:

- The slope felt is about double the slope requested. In the code, the slope is converted with (1300 * gradient + 507) while in my traces the conversion factor seemed to be (gradient * 650).

- When a gradient of 5% is requested, a value of 4% is sent to the device. This is because the code converts to int before multiplying by 650, thus something like 4.99% is truncated to 4%.

- The reported power is all over the place. I came to the conclusion in 2010 that the value obtained must be the force on the brake unit. You can then multiply by the speed to get the power. With this, the values reported seem much better. However, we would need someone with a powermeter to validate the metrics reported and possibly fine tune the equations.

- The Tacx software on Windows polled the trainer every 100ms. If you go faster, you do not get new values from the trainer which replies with 24 instead of 48 bytes, and you simply waste CPU time. The current code has a 10ms cycle, it could be increased to 50 or 100ms.

I will prepare a pull request for these changes. However, I would like to hear feedback from developers who have played with the Fortius driver who may have more insight. More importantly, if the Fortius currently works like a charm for you, let me know before I submit changes that may break your setup. Alternatively, if all you get is the type of problems that I am reporting here, let me know as well. Thanks!

On Saturday, April 25, 2020 at 2:37:15 AM UTC-4 micky...@gmail.com wrote:

Hi Anders

did you get this working OK?  I have my fortius installed and working but the power fluctuates a lot ... I have trued averaging over 5 seconds but still quite erratic?

TIA

[Dave Waterworth]

Hi Michael

I did some work some years ago on the Fortius Controller. My specific contribution was to fix up the code to read both the short message (from the controller I believe) and the long message (from the brake - this wasn't being read properly) as well as some modifications so that the code didn't miss button pushes from the controller (originally unless you pushed the button during the read it would miss it). I also spent a lot of time on the driver for windows - changes to windows driver signing made it really difficult to get to work consistently and I eventually moved on to other things (a Tacx Neo in particular)

I always thought there were issues with the actual reverse engineering of the brake message. I certainly suspected that the brake reports force/torque and not power - I did a bunch of experiments and modified the code but then my laptop died. If I recall my electrical engineering days correctly torque is proportional to current in an electric motor so it would make sense if it were engineered that way. It may appear correct in power mode if you peddle at a constant cadence. I did some tests where I set the power to 100W and started at 50rpm, increasing cadence at a regular interval - the output did seem to change with cadence and it should be more or less flat.

I never got slope mode working, I used a USB sniffer and did think there was something wrong with the mapping of the bytes to different variables but could never put my finger on what was wrong. The existing factors are probably wrong because of the torque vs power issue.

I think at the time the polling time worked ok on my laptop but it makes sense to slow it down if the brake responds better. 

I think it's worth submitting the pull request