Indicated Airspeed is no longer accurate SU5 all aircraft

[Dakfly0219]

Do you have any add-ons in your Community folder? If yes, please remove and retest before posting.
This problem exists for all aircraft in game, both default and community aircraft, but is only notable in aircraft that can approach or exceed the speed of sound at higher altitudes (above 30,000 feet). This will affect upcoming Asobo products like the F/A-18E model slated for release this fall and the hypersonic vehicle that was teased along with it.

Are you using Developer Mode or made changes in it?
To test this issue, yes.

Brief description of the issue:
Indicated airspeed is no longer accurate, in particular at near sonic and supersonic speeds. Indicated Airspeed appears to have been substituted for some sort of modified Calibrated Airspeed (TCAS), which becomes increasingly inaccurate as altitude and Mach number increase. Additionally, this value is now exactly the same as the CAS and IAS, which in the real world is not the case. More problematic, if any developer, including Asobo, were to develop hypersonic performance aircraft, the airspeed calculation is EXTREMELY divergent from expected correct values. At Mach 9 and 90,000 feet, the airspeed reading approaches 40,000 knots, which is far in excess of the True airspeed of 5300 knots and certainly the Indicated airspeed (which using real world calculations should be 780 knots). This speed is actually in excess of orbital velocity, so is clearly calculating incorrectly.

Provide Screenshot(s)/video(s) of the issue encountered:

Detail steps to reproduce the issue encountered:
Place any aircraft capable of supersonic or hypersonic speed at an altitude above 30,000 feet and increase speed to transonic, supersonic, or hypersonic airspeeds. Compared to real world charts and calculations, the returned airspeed on the gauge, which should be based on IAS, is not accurate and will make any attempt to pilot using IAS (for example, monitoring the Never Exceed Speed of a high performance aircraft) impossible. Aircraft don’t use CAS (and in particular whatever has been calculated to be this TCAS value) for the base line speed reference. These use IAS.

PC specs for those who want to assist (if not entered in your profile)

Build Version # when you first started experiencing this issue:
SU5

Are you on the Steam or Microsoft Store version?
Microsoft Store.

Did you submit this to Zendesk? If so, what is your ticket #?
No.

[OlieTsubasa443]

Hello Dakfly0219,

Thank you for this feedback. We investigated on this. TCAS and IAS relationship has not changed with Sim update 5 and is still calculated with a scale and offset indicated next to the TCAS in the debug window.

You are correct that the supersonic speed calculation has been improved for better accuracy below Mach 2 and up 50,000 feet. We don’t simulate accurately Mach speed as high as Mach 9 altitude.

However we will check this further.

[Dakfly0219]

Thank you for reviewing and considering my post. However, please forgive my bluntness, but it’s clear you didn’t read my post very carefully. Please re-read it. We strenuously disagree with the assertion that something didn’t change with SU5 as regards airspeed calculations. My dev team and I have been modeling supersonic fighter aircraft for months now and there was most definitely a change in how IAS is modeled in the game, and it is now modeled incorrectly, much to our disappointment. This change has made it impossible to accurately report overspeed conditions for aircraft traveling in the transonic to supersonic range (Mach 1+ to Mach 2). I used the prior Mach 9 hypersonic example to show how wildly inaccurate the game’s calculations become as Mach numbers increase, but this doesn’t mean that the problem doesn’t exist at more typical speeds for supersonic aircraft.

Also, I never said that the supersonic speed calculation has been improved for accuracy below Mach 2 and up to 50,000 feet. I said “Indicated airspeed is no longer accurate, in particular at near sonic and supersonic speeds.” In other words, anything above Mach 1 is inaccurate now when at altitudes above 20 to 30 thousand feet. An aircraft traveling at Mach 2 and 40,000 feet now over-reports indicated airspeed by at least a couple of hundred knots.

Below is a reference calculator and some formulas for calculation of indicated airspeed - KIAS, which is used by all aircraft (including modern supersonic aircraft) in setting in cockpit reference speeds such as Vne - which is a critical speed value for all aircraft to avoid airframe damage from overspeed conditions. It seems the game developers have substituted IAS in all airspeed gauges in all aircraft for a form of hybridized calibrated airspeed calculation that is calculating incorrectly above the speed of sound at altitude. While this may(?) have possibly resolved problems with airspeed calculations for low and slow aircraft (like the Cessna 172), it has created a problem for any aircraft that are traveling in the flight levels and near or above the speed of sound.

Airspeed Calculator (IAS/CAS/EAS/TAS/Mach) | AeroToolbox

Of particular interest in this calculator is the Equivalent Airspeed, which is essentially the calculation of IAS that aircraft display as the primary airspeed reading.

My dev team and I can further provide all sorts of documentation and evidence of this problem and what the expected correct values should be as well as how to calculate them in game.

[AwarePlot117729]

The airspeed calculations did indeed change in SU5. The incorrect compressibility correction that existed since at least FSX (and no doubt before) was corrected. I don’t know about supersonic speeds, but the subsonic speeds are now much more accurate than pre-SU5. The only remaining inaccuracy is the slightly off standard pressure vs altitude relationship that they’ve said will be fixed in WU6.

Airspeed indicators provide indicated airspeed. The indicated airspeed is not the same as equivalent airspeed. It differs from calibrated airspeed only by the position error, which is the error in measuring the free stream static pressure. In airplanes with air data computers (for example, all modern jet transports), the air data computer applies a position error correction before the airspeed is displayed. So, for all modern jet transports, the indicated airspeed is the calibrated airspeed. MSFS has the capability of simulating position error, so I’m assuming that works properly for those airplanes without an air data computer.

@Dakfly0219 , I’m confident you and your dev team can, for example, take the FBW A32NX up to 35,000 feet (where the MSFS standard pressure is correct) at M0.78 and verify that the indicated airspeed (which is KCAS) is 264 knots, which agrees with the airspeed conversion calculator that you linked. At ISA temp, this is 450 KTAS, and at ISA+10C it is 460 KTAS.

If there is a problem with supersonic speeds, please concentrate on getting that fixed. Do not have them make changes to the subsonic speed conversions, which are now finally working properly.

[Dakfly0219]

Thanks for your thoughtful response and for confirming the change in airspeed calculations. We referenced the EAS calculation from the aforementioned calculator as it’s a reasonable analog for broadly assessing the problem without getting into some pretty complex concepts, such as you’ve raised. We’ve only identified the problem as truly significant when at higher altitudes and in the transonic / supersonic regime. As a real world example, the F-16C operating manual references a Vne of 800 KIAS and / or Mach 2.05, whichever is reached first. A clean F-16C should theoretically reach that maximum Mach number at about 35,000 feet (or even a little lower) without coming close to exceeding its Vne of 800 knots. At 35,000 feet in ISA conditions the airspeed indicator should report approximately 650 knots. Whether the flight computer is adjusting airspeed for position error is less important in this discussion, when an F-16C, simulated in-game, will now report well above its Vne (approx. 880 knots) when it reaches its Mach limit of 2.05 at 35,000 feet, which is waaay off.

A thorough examination of speed calculations below Mach 1 and at low altitudes hasn’t been explored by us, because, like you, we don’t see enough deviation to warrant a trouble report. And at low altitudes near sea level, the speed calculations, even in the transonic regime do seem to be considerably more accurate than before and appear to have resolved some of those issues that also existed in FSX and Prepar3d. Our focus is on the problem of higher altitudes and Mach numbers near and above Mach 1, where this problem raises its head.

[Dakfly0219]

Here’s a speed debug screenshot showing this example in-game:

This is in ISA conditions. Note the game reported EAS value is approx. 660 knots, which is very very close to real world expected and is the corrected airspeed value also accounting for compressibility, instrument errors, etc… The indicated airspeed IAS is reporting identical to this TCAS (true calibrated airspeed, whatever that is) and is much too high for this altitude and Mach number - this is what is showing in the in game air speed indicator gauge in this situation. This would break our theoretical F-16C when in the real world, this should not be the case. If nothing else, the ASI should be pulling from the EAS value in this case to return an accurate pressure reading on the aircraft. As I’ve mentioned previously, the problem really starts around Mach 1 at altitudes where the low air pressure (high altitude thin air) difference is significant and the reported values in the ASI increase in divergence from expected values as Mach number increases.

You can see in my original Mach 9 example above, that the IAS and TCAS are identical and report a speed that would only be possible in outer space.

[trex5365]

This could also be the temperature problem causing divergence. We do know that changed a bit.

[Dakfly0219]

The temperature reported in the game appears to be accurate in this particular example. You can see in my second screenshot it’s reported at -54C which is consistent with ISA conditions. We don’t think this is the underlying problem, we think it’s a formulaic error introduced in an attempt to provide correct airspeed values in the subsonic realm.

[AwarePlot117729]

Well, now that is truly odd. If I use the same KTAS/KCAS conversion equation that I gave them to correct the airspeed conversion error pre-SU5, my result (of course) agrees with the airspeed conversion calculator on the website you linked in your previous message. For a true airspeed of 1184 knots for this condition, the calibrated airspeed should be about 860 knots. Under the pre-SU5 incorrect conversion, for the same conditions, the calibrated airspeed would have been around 1043 knots.

Unless you have an input for airspeed position error, the calibrated and indicated airspeed should be the same. The ASI should always be pulling from the calibrated airspeed and, if necessary, be corrected for position error. Equivalent airspeed is not known to the airspeed system and is not displayed.

[trex5365]

I was referring to temperature across shock wave, sorry for not being more clear.

It’s clear that some temp effects are wrong. Was wondering if a temp problem was creeping into the indication.

[Dakfly0219]

Ah, ok, thanks. Yeah, that could be, and temperatures and pressures are definitely part of proper supersonic airspeed calculations, but without knowing what the coding they implemented looks like that would be speculative at best.

[evanburnsdev]

I think the dynamic pressure calculations are to blame for what’s throwing off the high speed/high alt calculations.

The sim seems to be calculating dynamic pressure linearly based off of speed at a given static pressure. Which is clearly inaccurate, given:

(Source)
Showing dynamic pressure will increase linearly with static pressure, but exponentially with speed.

I’m not sure the formula the sim uses for CAS, but if it’s using dynamic pressure for the calculations, the formulas should look similar to these:

For subsonic speeds:

For supersonic speeds:

(Source for both formulas)

I’ve implemented these formulas into a wasm file(with a custom pow function due to webassembly precision), and comparing dynamic pressures at 60kft at M1.99, the difference in dynamic pressure calculations is significant.

In the screenshot below:
Imp_Pres = Impact(Dynamic) Pressure in hpa
Imp_Press = Impact Pressure converted to psf for comparison to the sim calculated pressure.
CAS SubS = Subsonic Speed calculation
CAS SupS = Supersonic Speed calculation

And with the sim’s calculation of dynamic pressure scaling linearly with speed, the divergence from this point increases exponentially.

[evanburnsdev]

Looking more into it, it appears the sim uses the incompressible fluid dynamics formula for speed calculation, where they should be using the compressible fluid dynamics.

[FlyingBear01]

That would seem to be a problem, since compressible fluid dynamics applies to transonic speed and that is where the jet airliners fly.

One would think it is straightforward to add the appropriate formulas to the implementation? One line code change (yeah I know, adding a semicolon somewhere is 1 week program management plus 10 seconds coding plus 2 minutes unit test plus 1 week QA plus another week to move the semicolon to where it was supposed to go in the first place…)

[trex5365]

My old Aerodynamics text is buried in a box somewhere. So I found these resources as well.

Might be too much information but the ratios ahead of and behind the shockwaves will be important to understand what is the probe actually sensing.

https://www.grc.nasa.gov/www/k-12/airplane/normal.html

ConfigAeroSupersonicNotes.pdf

[AwarePlot117729]

No, that’s not quite it. The sim does use compressible fluid dynamics equations for speed calculation just like the airspeed conversion website that the OP posted a link to. The issue with supersonic speed measurement, as you yourself posted, is that dynamic pressure cannot be sensed directly by the pitot tube because a shock wave forms in front of the probe. What the probe senses is the pressure behind the shock wave, so the effects of the shock wave must be taken into account when using the pitot tube measured pressure to determine the airspeed. The Rayleigh Pitot tube formula does this, but it is not a trivial matter. Solving it requires an iterative process.

[evanburnsdev]

Correct, the iterative process with the Rayleigh formula was the last formula I posted. You input the computed subsonic CAS into the formula to determine what your supersonic CAS should be.

Edit: But you did answer my concern as to whether or not compressible fluid dynamics were used for the CAS calculations. Thanks for that. I’m surprised they didn’t add impact pressure to the speed debug menu to avoid this.

[AwarePlot117729]

You have to iterate on CAS until the equation closes (CAS on the 2 sides of the equation being the same). Should close with only a few iterations.

I guess I don’t understand your adding impact pressure (dynamic pressure) to the speed debug window. It is there.

[evanburnsdev]

I was just a bit confused by the dynamic pressure label. I read dynamic pressure refers to incompressible fluids and impact pressure refers to compressible fluids. Either way the formula I listed above shows a different number than the dynamic pressure displayed on the speeds menu.

I did find a few errors in my code and verified that CAS lines up perfectly up to mach 1, but starts to diverge from there, most likely due to the sim not using the Rayleigh formula. I did find another version of that formula that doesn’t require iteration on an MIT paper Here.

I’ve plugged all of the formulas into Desmos. You can use the top two sliders(Mach and Static Pressure) to view how subsonic CAS and supersonic CAS are calculated. You can view it Here.

I want to reassert that these formulas perfectly line up with the sim’s CAS calculation up to Mach 1. It’d be awesome if the game devs could use the C_assup formula listed in the link I provided within an if statement when mach > 1. At mach == 1 both C_as and C_assup are identical, and diverge in line with the table shown on the last page of the MIT paper.

Thanks @AwarePlot117729 for pointing me in the right direction. If you disagree with what I’m saying, or see something wrong in my math, please let me know.

[AwarePlot117729]

Hi, yes, sorry I was too loose with the terminology (just plain wrong) for impact and dynamic pressure. They are only the same when dealing with incompressible fluids (speeds below around 200-250 knots). Impact pressure is the difference between total pressure and static pressure, whereas dynamic pressure refers to the kinetic energy of the airflow, and is given by the formula, 1/2 x density x true velocity**2.

The MSFS debug speed window provides the dynamic pressure. You’re right, I don’t see any readout of either impact pressure or total pressure as either a simulation variable or in any of the aircraft editor debug windows.