MarkDH

I think this is a premature diagnosis! Seems more likely the problem was something other than overheating.

I saw a report of an Nvidia RTX5070Ti overheating after SU5 was released for MSFS 2024. Normally this would not have caught my eye, but I have that same card and just yesterday I experienced the same thing myself. I ran MSFS 2024 for the first time in a few weeks, and of course it installed SU5. It worked fine and I ran for a good few hours, but then I started to get rendering glitches that persisted after restarting MSFS and even rebooting the PC. I suspected temperatures and after finding a way to monitor GPU stats I could see MSFS was basically working the GPU hard all the time and temps were in the up to mid-60s sort of range. The card was very hot to the touch, but all fans were working. If I let it cool down things went back to normal. This was running VR headset (Bigscreen Beyond 2) but the high temps persisted when not in VR mode. Is it plausible that this is a software matter and not a physical cooling deficiency? I have seen other reports, but it's the internet and I worry that I am seeing one person's experience reported in different places. I guess if it's software it will get fixed in due course. If I see this consistently than maybe I can try falling back to MSFS 2020 for comparison.

I don't have the SU5 beta but ny advice would be to use Opentrack, which is a lot clearer and has access to all the available functions. It looks like Asobo is converging on all the same settings available in the tobii/Opentrack setup, but they are still falling way short. The curves, for example, appear to be missing, and the ability to set up asymmetric axes. The tobii web site gives very clear instructions on the one-time Opentrack setup. Once you have it installed, you get a much clearer options page in the tobii GameHub. Asobo should also rename the 'eye/head tracking ratio' option to acknowledge that eye tracking and head tracking are distinct functions with no real connection. In the Opentrack/Gamehub setup you can turn either on or off independently. Likewise 'camera boost ratio', which is even more cryptic than just calling it 'camera boost'. Assuming 'Camera Boost' still works the way tobii describes it, If you have this slider set to anything other than 0 the 'eye tracking vs head tracking ratio' will be ignored and eye tracking will not move the view at all. 'Camera boost' basically uses eye tracking as a YES/NO flag to modify the 'sensitivity' (i.e. amount of motion scaling) of an axis depending on whether your eyes are looking straight ahead or not. (Most of the Camera Boost settings are also missing from the Asobo setup page.) If you want to see some of this in the sim, my recent video demonstrates and compares the setup features available for tobii and TrackIR.

I wonder if you have the 'head auto center' option turned on? That would make it appear to drift if you were looking at your EFB for more than a few seconds. Otherwise, I haven't experienced this but I also haven't flown a whole lot with the Tobii beyond testing it out. Maybe I will experiment and report back.

I have done a moderately detailed comparison of the Tobii and TrackIR in the following video. Many thanks to @andy1252, who generously donated his Tobii so I could make this!

I had the original TQ6 and if I remember correctly the tension adjustment needed you to exercise the levers throughout their entire range ten times or so to redistribute the tension evenly. That said, I found the TQ6 levers a little bit stop/start in operation for small changes, so I'm not sure I would buy another one. The Neo looks eye-wateringly expensive but I see it also appears to have lift gates for the beta and reverse zones, which is new.

This is the purpose of the 'True View' flag in the TrackIR software. If you have True View turned on, the translation frame of reference is clamped to the rotation frame. If you don't have True View set, the two reference frames remain independent. It sounds like the Tobii software doesn't have an equivalent to this setting (or maybe you don't have it turned on!)

It's under 'experimental' at the bottom of the settings menu. It's very easy to do. The following video shows it for MSFS 2020 but I think it's just the same in 2024.

If it behaved as you describe, this would mean that zooming in and out was just the same as moving the camera forwards or backwards, which it manifestly is not. You are also using 'FOV' very loosely (FOV is an angular value measured from a given position, so it doesn't change when you move the camera forwards of backwards.) Here's a challenge: (1) Take a screenshot (of any scene). (2) Zoom in substantially (enough to make an obvious difference). (3) Move the camera - in any way you want - and try to get the view you captured in step (1). You will not succeed. For small zoom changes you may get something that looks close to the original screenshot, but they will not overlay exactly. For larger zoom changes, they will be completely different, and in particular you will see less distortion further away from the centre of the image. BTW, I can see only two material differences between the XPlane dialog above and MSFS 2024: (1) you can't set the horizontal and vertical FOVs independently (does this have any practical use?), and (2) you cannot set the FOV manually in degrees (so you have to set it by adjusting the zoom).

I would be surprised if you're seeing increased distortion when you zoom IN.

I don't think you will get anywhere on this assumption unless you are prepared to be convinced otherwise. In FSX, P3D v4 and X-Plane 11 the Zoom numbers are proportional to (because they are derived from) the FoV. Yes, the algorithm by which the default zoom numbers are derived varies, and in the Microsoft sims is tweakable by changing the WideViewAspect flag, but this does not affect the spatial qualities of the scenes you will see if you zoom progressively in and out. (The only real effect of setting WideViewAspect=1 is it lets you zoom out further. I expect this is the same in MSFS 2020 and 2024, although I have not investigated.) P3D (at least to version 4.5-ish, where I stopped) has an extra variable if you use ViewGroups, which allows you to select a projection algorithm from flat, cylindrical or spherical, but I think this only compensates for the angling-in of multi-screen display setups. In FSX (and I think X-Plane, at least to v11) the projection algorithm is always 'flat', which I think is why things distort as you move further away from the centre (vertically as well as horizontally). Yes, I think you can change the horizontal and vertical FoV independently but I don't know of circumstances where it would make sense to do this. I looked briefly at your two screenshots but they seem to be taken from two different places, so I doubt you can draw any useful conclusions from comparing them.

No. It's the difference between a single sample and an average. A point sample may be very atypical, but the key feature of this product is averaging of many historical samples.

I haven't seen much coverage of these panels, which are all-metal construction and very good quality. They come with default profiles for the popular sims and are completely user-programmable. They are Arduino-based and need MobiFlight (free) to operate and to customise. The panels have been around for a year or more, I think originally as 'CircuitXL', now 'Circuit Avionics'. You can see my reviews below, and there are some more videos on my YT channel showing them in use.

The Honeycomb Alpha is a good step up from Saitek. At $300 for both the Alpha and (I assume) the Bravo throttle it sounds like great value. (Also worth noting the XPC is the 'deluxe' version, with contactless sensors and XBox compatibility.) I would not recommend the new Alpha LITE yoke, which feels very notchy in the pitch. The original Alpha was more uniform.

Brief update, including how to access the onboard calibration!