Jump to content

Su-27 Bug: No radar when inverted <1500m


BlackPixxel

Recommended Posts

In DCS, Su-27, Su-33 and MiG-29 all lose their complete radar functionality when they fly below 1500m with >120° roll angle. You can observe this behaviour in the attached trackfile.

 

This behaviour is wrong and is based on incorrect assumptions.

 

Incorrect assumption #1:

The roll gimbal limit is 120°, so the radar does not work beyond 120°.

 

Wrong! Roll gimbal limit of 120° means that the radar will no longer be aligned with the horizon when exceeding 120°, but it does not mean that it incapable of operation because of that.

To proof that the radar is able to operate when not aligned with the horizon you have to know that the radar is using no roll stabilisation at all in all other modes than BVR, so its antenna is never aligned with the horizon in those modes and still works.

 

Incorrect assumption #2:

Because the guard antenna (used for sidelobe compensation) is fixed to the roll gimbal with a very slight downward angle, and does not follow the radar antenna in elevation and azimuth, it will be rotated away from the ground when exceeding the roll stabilisation limit of 120°. The assumption is that it then can no loner cover the sidelobe returns from the ground.

 

This is also wrong!

The compensation antenna is very small, much smaller than the one used on the F-15C radar. So it has a very wide beamwidth. This wide beamwidth is required because it has to cover the whole azimuth and elevation range of the main beam.

The F-15C guard antenna already has a beamwidth of about 50°. The Flanker/Fulcrum has a much smaller guard antenna and as a result an even higher beamwidth. So it can definetly cover a wide range below the nose, left, right and above the nose of the aircraft.

 

The same priciple with a non-moving compensation antenna is also used on the IFF of Sukhois and MiGs. The main IFF beam is moving with the scanning reflector, while the sidelobe compensation is done with a fixed secondary antenna. This IFF is also able to cover the whole azimuth and elevation range (by adjusting gain of the compensation channel depending on main beam deflection), and so will the compensation channel of the main radar.

 

In addition to that, the behaviour is not even in line with the only "official" statement by ED about this, which is in the F-15C manual.

There it says that the loss of tarets due to rotating away the guard antenna from the ground only affects rear hemisphere tarets. In DCS it affects all target aspects. And what is stated in this manual is wrong because even when rolled the antenna covers the whole area in front of the aircraft due to its high beamwidth.

 

There is a high chance that the incorrect radar behaviour was implemented into DCS because of this thread:

https://www.secretprojects.co.uk/threads/mcdonnell-douglas-f-15-avionics.53/

Here someone working on implementing radar and ECM into computergames wrote a few posts, possibly an ED employee?

 

 

This unrealistic limitation needs to be removed from the russian aircraft, unless ED comes up with correct explanation for it. So far it is just based on incorrect guesses.

radarlock.trk

Link to comment
Share on other sites

  • ED Team

We have no data to say it works any different from our ingame implementation, which we are currently happy with.

 

Any data relating to this will be classified.

 

Thanks

smallCATPILOT.PNG.04bbece1b27ff1b2c193b174ec410fc0.PNG

Forum rules - DCS Crashing? Try this first - Cleanup and Repair - Discord BIGNEWY#8703 - Youtube - Patch Status

Windows 11, NVIDIA MSI RTX 3090, Intel® i9-10900K 3.70GHz, 5.30GHz Turbo, Corsair Hydro Series H150i Pro, 64GB DDR @3200, ASUS ROG Strix Z490-F Gaming, HP Reverb G2

Link to comment
Share on other sites

There is no data to proof that it should work like the ingame implementation in the first place.

 

You have to understand that this instant radar loss when rolling is a massive issue during air combat. Having such a drastic and unrealistic disadvantage just because of some incorrect assumptions is inacceptable.

This inverted radar loss affects almost EVERY fight with the DCS Su-27, because of the weak R-27 behaviour almost every fight ends closer to the ground.

 

If it was realistic, then why does none of the available real manuals warn the pilot of this? Do game developers know more about the radar than the engineers that made them?

 

Why is it not even working as it is explained in the F-15C manual?

 

There is plenty of data showing that the ideas behind the ingame implementation are wrong, such as the high beamwidth of the compensation antenna, which makes the tailchase effects described in the F-15C manual nonexistent.

The concept of a narrow beamwidth guard antenna that only covers a small spot on the ground already fails when the aicraft would pitch down its nose by a few degrees. Or when the radar is actually scanning and changing its main beam direction. Then the guard antenna wold no longer cover the same sidelobes as before.

 

And even if the compensation antenna would be a low beamwidth antenna (which defeats the concept of the compensation antenna itself), then it should only affect a tailchase scenario, as MiG-29 documents clearly say that the spectral component of the target return in front hemisphere is outside of the spectral component of the ground return. So ground clutter from the sidelobes can not affect the radar in such a way as it is currently implemented.

cWKcm05.png

 

This radar issue must get reviewed. DCS claims to be a simulator, so the players are asking for realistic behaviour.

Link to comment
Share on other sites

We have no data to say it works any different from our ingame implementation, which we are currently happy with.

 

Any data relating to this will be classified.

 

Chizh once posted an image like this to prove the loss of radar when inverted:

LIddZei.jpg

 

It clearly says that the antenna is stabilized in roll during SCANNING operation. And just in BVR scanning, not in modes like vertical scan. In vertical scan the roll gimbal of the antenna remains locked.

 

You can also see the gimbal limits of single target tracking. Note how no roll gimbal is mentioned for STT. If it was necessary to keep the radar aligned with the horizon during single target tracking, then it would have been mentioned here.

 

But let's go back to BVR scan mode. What would happen if you bank beyond 120°? Why not take a look at a full fidelity module with a very similar radar design (in terms of being a twist cassegrain with a roll gimbal), the MiG-21:

 

Note how the return from the ground during the 140° bank is no longer on the bottom of the radar, but it represents something looking like an exponential curve.

ZF44orF.png

 

Why does this happen? The radar in the MiG has a roll gimbal limit of somewhere around 70°. Up to roll angles of 70°, the radar will be aligned with the horizon. Beyond that, the radar horizon will be rotated relative to the real horizon. The elevation and azimuth gimals keep moving in their usual fasion. Elevation only goes up and down to switch bars, azimuth only goes left and right along the bar.

 

Here is a visualisation of how the scan pattern is aligned in this MiG-21 example. The bar scan pattern is rotated by 140° (bank angle of the aircraft) minus ~70° (roll gimbal limit of the aircraft). So the scan pattern is rotated by 70°. This causes the radar to pick up very close ground return on the left, and the more it goes to the right the more distant the ground reflections get, until the radar is looking above the horizon where it no longer receives main lobe ground reflections.

aS7B6oO.png

 

As a result of the rotated radar horizon in BVR mode, a different portion of the sky will be scanned. This means targets far on the left or the right of the intended scanzone may not be covered, but targets below or above scanzone could appear. Targets in the center of the scanzone will show up as usual. So a loss of radar scanning ability when inverted is unrealistic, it just changes the placement of the scanzone in the sky.

 

This rotated radar horizon would of course affect the radar in all altitudes. In the DCS Su-27, Su-33 and MiG-29 the odd radar behaviour that this thread is about only happens <1500m AGL, so the rotated radar horizon described in this post is not what the DCS Su-27 is trying to "simulate".

 

So the issue can only be related to the compensation antenna. I will show in the next post why this will also not affect the radar operation.

Link to comment
Share on other sites

The second "explanation" for the inverted radar loss of N001 and N019 described in the F-15C manual is the way the compensation antenna is implemented.

 

The compensation antenna is used for sidelobe supression. It has a wider beamwidth than the main radar antenna and receives the ground returns from a very wide area of the ground. The compensation antenna uses a seperate receiver channel.

The radar can now compare the returns from the main channel and from the compensation channel. If a return from the compensation channel is within a certain ratio to the corresponding return of the main channel, this specific return must have entered the main radar through one of the sidelobes and can be ignored. This way sidelobe returns are supressed.

 

 

The compensation antenna in the russian radars is a yellow horn antenna that sits in the top of the antenna housing. It is attached to the roll gimbal of the antenna, but not to the elevation/azimuth gimbal of the moving reflector. So it follows the radar in roll, but not in elevation and azimuth.

PstjKwo.png

 

The mistake that the DCS developer made was expecting te guard antenna has a very narrow beamwidth. Because the horn antenna is pointing down by a very few degrees (maybe 5°), he thought that the antenna would only gover a slight beam below the boresight, similar to the following illustration:

bertW1P.png

 

Now when the Flanker goes completely inverted, the radar antenna would be rotated by up to 60° compared to the true horizon. This would mean that the guard antenna no longer points on the same area as before, but left/right of it and a little higher. So the DCS developer thought that now the guard antenna could no longer receive the correct reflections and that now te sidelobe supression would fail.

 

This idea can already be disproven by a simple thought experiment:

What would happen if the Su-27 would pitch downwards or upwards by just a few degrees? The radar would still point at the very same target, but the compensation antenna would point at a completely different spot on the ground. So according to the DCS developer the sidelobe supression would fail:

BXRNiBU.png

 

In fact, also different target elevations would already make the sidelobe supression fail if the concept of a narrow beam guard antenna was the case. Because the main radar antenna would move up or down and illuminate different regions through the sidelobes as well as receiving from different regions through the sidelobes. But the compensation antenna would point at the very same spot in these cases.

 

 

Thankfully there are also photos of the compensation antenna, so we can estimate its dimensions and disprove the theory of that DCS developer using basic antenna design rules.

How many compensation antennas can be very roughly fit into the diameter of the front reflector? About 34:

U7nMXYt.png

 

The nose cone diameter at the poisition of the front reflector is about 0.65 m. There is some space between nose cone and the radar, but let's just asume a generous 0.6 m diameter. 0.6 m divided by 34 gives our compensation antenna a vertical aperture of about 2 cm. The width is a a little more than the height, so roughtly 2.4 cm.

The N001 and N019 are X-Band radars. The center frequency of X-Band is 10 GHz.

 

Horn antenna characteristics can be calculated using simple formulas, which is done on the following website:

https://www.rfwireless-world.com/calculators/Horn-Antenna-Calculator.html

 

We get the following results:

aa1voj5.png

 

The 3dB-beamwidths (which is the angle between the two points where the antenna has half of its maximum gain) in azimuth of more than 85° in elevation of more than 75°. But even above these angles the antenna can still be used.

Let's update the illustration with these values:

l0OqFIy.png

 

It should be very clear that the aircraft orientation does not matter. The beamwidth of the compensation antenna covers the sidelobes of the main antenna in any case.

 

 

Conclusion:

 

The current implemetation is wrong and based on an assumption that has been very clearly disproven. The compensation antenna covers the sidelobes during a wide range of aicraft orientations, and as result the N001 and N019 radars will not lose lock when the aircraft banks at low altitude, which currently happens in DCS.

Link to comment
Share on other sites

  • ED Team

We are happy with our implementation currently, but thank you for the feedback I will share it with the team.

 

Thanks

smallCATPILOT.PNG.04bbece1b27ff1b2c193b174ec410fc0.PNG

Forum rules - DCS Crashing? Try this first - Cleanup and Repair - Discord BIGNEWY#8703 - Youtube - Patch Status

Windows 11, NVIDIA MSI RTX 3090, Intel® i9-10900K 3.70GHz, 5.30GHz Turbo, Corsair Hydro Series H150i Pro, 64GB DDR @3200, ASUS ROG Strix Z490-F Gaming, HP Reverb G2

Link to comment
Share on other sites

Hi BIGNEWY,

 

When you say

We are happy with our implementation currently

 

to me it's a fairly ambiguous phrase without some rational for feeling happy with it.

 

It could mean that "We are happy with our implementation currently" because we believe it to be accurate...

 

...or it could mean that "We are happy with our implementation currently" because while we know it is inaccurate, we feel that the difference between this representation and reality is irrelevant to the SIM

 

...or it could mean that "We are happy with our implementation currently" because while we know it is inaccurate & has a significant impact on gameplay, we feel that fixing the representation would be too difficult to justify the time, effort and money needed to modify a change we made with all the goodwill in the world.

 

...or it could mean "We are happy with our implementation currently" because having it like this makes the DCS.Blue fighters much easier for casual users to fly successful sorties with and so enjoy their purchases & therefore more likely to make future purchases of DCS.NATO aircraft.

 

I have little doubt it's the first, & assume that pilots are explicitly told that the radar is only stabilised to +/- 120 degrees because after that point there is a degradation in performance (otherwise, you wouldn't want them worrying their crowded brains keeping tabs on roll angle during an engagement), but the binary implementation seems a bit unlikely.

 

To avoid confusion, could you be a bit more expansive about why E.D. are happy ?

  • Like 1

Cheers.

Link to comment
Share on other sites

  • ED Team
There is no data to proof that it should work like the ingame implementation in the first place.

There is data in the Su-27SK manual.

 

In order to avoid resetting the target from auto tracking with the leading channel of the RLPK (radar), fly the aircraft with a roll of no more than 90 degrees.

RLE_SU_27_kn1_152.thumb.jpg.b8d2eed3d073ce22a5c680bba37589e7.jpg


Edited by Chizh

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

  • ED Team

To avoid confusion, could you be a bit more expansive about why E.D. are happy ?

Because we have the most accurate simulation of Su-27 from all available commercial game-simulators on a PC.


Edited by Chizh

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

There is data in the Su-27SK manual.

 

In order to avoid resetting the target from auto tracking with the leading channel of the RLPK (radar), fly the aircraft with a roll of no more than 90 degrees.

 

Is that 90 degree roll to target, or to ground?

And if auto tracking resets, what happens?

i7-8700k, 32GB 2666Mhz DDR4, 2x 2080S SLI 8GB, Oculus Rift S.

i7-8700k, 16GB 2666Mhz DDR4, 1080Ti 11GB, 27" 4K, 65" HDR 4K.

Link to comment
Share on other sites

  • ED Team
Is that 90 degree roll to target, or to ground?

Ground.

 

And if auto tracking resets, what happens?

Radar goes to scan mode. Re-lock of the target is required.

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

There is data in the Su-27SK manual.

 

In order to avoid resetting the target from auto tracking with the leading channel of the RLPK (radar), fly the aircraft with a roll of no more than 90 degrees.

 

The MiG-29 manual says:

 

В связи с тем что информация по крену при радиокоррекции обновляется с частотой 18—20 Гц, сигналы радиокоррекции правильно выдаются только при угловой скорости вращения истребителя не более 60°/с (при применении ракет выпуска до июля 1986 г. — не более 30°/с). По этой причине не обеспечивается наведение ракеты Р-27ЭР на цель в инерциально-корректируемом режиме при выполнении истребителем противоракетного маневра типа «кадушка».

 

Due to the fact that the roll information during radio correction is updated with a frequency of 18–20 Hz, the radio correction signals are correctly issued only when the angular speed of rotation of the fighter is not more than 60 ° / s (when using missiles prior to July 1986, not more than 30 ° / with). For this reason, the guidance of the R-27ER missile on the target is not ensured in the inertial-corrected mode when the fighter performs the “coil” type missile maneuver.

 

When the MiG-29 does perform the described maneuver (high g barrel roll) it will enter inverted flight. Radar lock is kept, only radio corrections will not work.

So the DCS MiG-29 should not lose lock when inverted?

 

Both N019 and N001 are very similar, why should they behave different?

 

What do you think is the reason for the possible loss of lock beyond 90° in the Su-27SK?

 

Your quote from the manual comes directly after some paragraphs about engaging in ZPS with radar and/or OLS. Could the quote be about ZPS scenarios only?

 

 

In ZPS the lock is more affected by ground clutter, so an automatic transition to OLS lock might be required. With a radar only lock in ZPS the flight director ring brings the Su-27 in a position below the target for a radar lookup. Banking more than 90° could now result in the target being outside of the OLS angular limits (It can only look downwards by a few dregrees). The OLS would be unable to take over in case the radar lost lock (especially if the Su-27 would start to turn away after firing), and the target tracking would stop.

Maybe this is the reason for the recommedation not to exceed 90°.


Edited by BlackPixxel
Link to comment
Share on other sites

Because we have the most accurate simulation of Su-27 from all available commercial game-simulators on a PC.

 

That does not mean to much when DCS is basically the only serious one.

 

The DCS Su-27 is still lacking fundamental features (no MP group datalink, HUD symbology is missing alot of stuff etc. and let's not talk about R-27)

 

 

Anyway, about the inverted radar behaviour we will propably know more about it when Razbam releases their full fidelity MiG-23, which again has a very similar radar in terms of layout. I am sure they will do alot of research, so we can expect an accurate radar from them.


Edited by BlackPixxel
Link to comment
Share on other sites

  • ED Team
The MiG-29 manual says:

 

 

 

When the MiG-29 does perform the described maneuver (high g barrel roll) it will enter inverted flight. Radar lock is kept, only radio corrections will not work.

So the DCS MiG-29 should not lose lock when inverted?

Yes. At short time.

You can do it in the DCS and missile hit the target without problem.

 

The fact is that lock and target illumination are restored when the plane restored from an inverted flight. It is important that the inverted flight is not too long.

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

Yes. At short time.

You can do it in the DCS and missile hit the target without problem.

 

The fact is that lock and target illumination are restored when the plane restored from an inverted flight. It is important that the inverted flight is not too long.

 

I do agree with Chizh that it should lose the auto tracking after sometime as is now,

 

but I think that main issue here is this bug: https://forums.eagle.ru/showthread.php?t=253434 that even if you go over the limit for less the second the missile is trashed since it will just make a sudden and unexplained turn thus making the hit impossible.


Edited by FoxAlfa

-------

All the people keep asking for capabilities to be modelled.... I want the limitations to be modelled.... limitations make for realistic simulation.

Arguing with an engineer is like wrestling with a pig in the mud, after a bit you realize the pig likes it.

 

Long time ago in galaxy far far away:

https://www.deviantart.com/alfafox/gallery

Link to comment
Share on other sites

Yes. At short time.

You can do it in the DCS and missile hit the target without problem.

 

The fact is that lock and target illumination are restored when the plane restored from an inverted flight. It is important that the inverted flight is not too long.

 

Yes, there is like 4 seconds of radar memory, unfortunately only in BVR mode. In close combat modes lock is lost instantly (with no EO lock), except for vertical scan within 5 km. Vertical scan within 5 km can hold a lock inverted forever.

 

There is also another issue with this radar memory. When the target is within the EOS gimbal limits and hot enough to be picked up, then the lock will insantly transition into an EO only lock when going inverted. This means that the radar is no longer emitting for a moment, and due to the SARH missile bug the R-27ER will turn instantly into chaff while the radar is OFF. Forcing radar back on will not save them, they are still going for that chaff.

 

roll+R-27ER.trk

Link to comment
Share on other sites

This means that the radar is no longer emitting for a moment, and due to the SARH missile bug the R-27ER will turn instantly into chaff while the radar is OFF. Forcing radar back on will not save them, they are still going for that chaff.

 

That is interesting.

 

1) the radar and missiles does not have any memory of the target because missiles can't be recovered after inverted flight.

 

2) Missiles magically knows where chaff is and goes to it without guidance.

 

3) Missiles ignore the radio guidance (that is already well found pattern in the another thread) and "sucks" all chaff even when well outside the seeker/radar FOV and logic.

i7-8700k, 32GB 2666Mhz DDR4, 2x 2080S SLI 8GB, Oculus Rift S.

i7-8700k, 16GB 2666Mhz DDR4, 1080Ti 11GB, 27" 4K, 65" HDR 4K.

Link to comment
Share on other sites

  • ED Team
Yes, there is like 4 seconds of radar memory, unfortunately only in BVR mode. In close combat modes lock is lost instantly (with no EO lock), except for vertical scan within 5 km. Vertical scan within 5 km can hold a lock inverted forever.

You are incorrect. In the all close combat modes: 3 Vertical, 4 Bore, 5 Helmet the radar lock is not lost in inverted flight, at 10 km and closer.


Edited by Chizh

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

  • ED Team

There is also another issue with this radar memory. When the target is within the EOS gimbal limits and hot enough to be picked up, then the lock will insantly transition into an EO only lock when going inverted. This means that the radar is no longer emitting for a moment, and due to the SARH missile bug the R-27ER will turn instantly into chaff while the radar is OFF. Forcing radar back on will not save them, they are still going for that chaff.

 

This is a bug. Will be fixed.

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

You are incorrect. In the all close combat modes: 3 Vertical, 4 Bore, 5 Helmet the radar lock is not lost in inverted flight, at 10 km and closer.

 

The close combat modes will definetly drop the radar lock unless the target is within 5 km, not 10 km:

 

Link to comment
Share on other sites

  • ED Team
The close combat modes will definetly drop the radar lock unless the target is within 5 km, not 10 km:

I can not confirm.

 

In the attached track, I lock the target at a 10 km range. I turn inverted and launch the R-27ER missile. The target is not reset, the missile hits.

 

 

Track recorded into current OB build DCS 2.5.6.45915

Su-27 CAC Vertical inverted launch.trk


Edited by Chizh

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

  • ED Team
You are at 2000 m, at this altitude (>1500 m AGL) any radar mode will keep lock inverted.

I am at 1000 m AGL. There going inverted in close combat modes will cause the lock to drop unless the taret is within a range of 5 km.

OK

I tried below 1 km altitude and did not lose lock in the barrel roll after launch of 27ER at range more 5 km. Because radar is supported the OLS in hard cases.

 

Track recorded into current OB build DCS 2.5.6.45915

Su-27 CAC Vertical inverted launch below 1 km.trk


Edited by Chizh

Единственный урок, который можно извлечь из истории, состоит в том, что люди не извлекают из истории никаких уроков. (С) Джордж Бернард Шоу

Link to comment
Share on other sites

OK

I tried below 1 km altitude and did not lose lock in the barrel roll after launch of 27ER at range more 5 km.

 

Track recorded into current OB build DCS 2.5.6.45915

 

You lose radar lock while inverted, it switches into an EO lock because you are above 5 km range.

 

Point your nose down like I did, so that the EOS can not see the target when you do the roll. The lock will drop above 5 km if you do that.

 

I attached a track doing that.

Su-27invertedlock.trk

Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...