US F/A-XX and F-X & NGAD - 6th Gen Aircraft News Thread

[Gloire_bb]

Yes but we must remember, just because the US decided to forgo RCS for better maneuverability doesn't mean that Shenyang couldn't achieve both better RCS and maneuverability at the same time. That depends on its controls(all moving wingtips) and engines etc.

This raises some questions, even though I don't want to underestimate the US.

These 6th gen platforms are built mainly with the Westpac in mind. Its an extremely large area, and is static with lots of radar infrastructure and would have great visibility.

If I had to guess the PLA's doctrine, it would be:

1. Crippling first strikes, and continuous strikes on static enemy infrastructure and logistics.
2. Area denial through BVR combat.

Considering the above and China's advantage in BVR missiles, IMO, the chance for close range, dogfighting WVR engagements would be very low. Any that could happen would be around the Korean peninsula which would be highly exposed to an initial attack.

I also think the emergence of the J-XDS and especially the J-36 'Quasimodo' has put a tactical wrench in their plans. The J-36 will cover the BVR detection and engagement, while the J-XS will cover any WVR engagements + BVR too. Now, since the F-47 'Convict' (lol) seems similar to the J-XDS, any upcoming F/A-XX is definitely not going to cover for the J-36. I am not sure about the B-21 because it being subsonic is a big flaw, and US doesn't have a decisive BVR missile advantage (more at a disadvantage).

When I think about it, I am not exactly sure why the US would choose maneuverability here. A focus on RCS/operation range to limit the BVR and logistics disadvantage is the right way to go.

Again, its early, but perhaps the US got tactically outmaneuvered by the PLA (again)?

I just wrote a similar reply in ngad thread, don't want to repeat directly.

Disclaimer: it's speculation everywhere.
My current understanding is that LM proposal (more along the lines of expectations), aimed at mixing NGAD, F-35 and later added CCAs. It lost to Boeing deeply integrated NGAD+CCA, intending to replace them altogether as air fighting system. As such, whole system can be built on a higher level. But new main fighter is responsible for everything, including competitive BVR and WVR. No way out, it should maneuver.

Chinese 6ths can afford to follow original 6th gen idea, because J-20/J-35 pair is perfectly viable - adding Hi to them makes sense.
F-35, on the other hand, isn't up to the task, even in its future blk.5 form. Or maybe, customer doesn't believe LM will suddenly make this program work.

I.e. it's not unlikely Boeing proposal is significantly cheaper(per plane), but is radically more ambitious in terms of entire force air fighting capability. It's also more risky, in that nothing yet exists(compared to a system built on F-35)

F/A-XX is in its own realm, as it is designed as heavy "quarterback"(long range drone command post/stand off effects launcher), which will also eventually displace both Bug and F-35C.

i.e. my current theory is PLAAF and PLANAF building more or less symmetric 3-tier solutions.
USN and USAF build two very different asymmetrical ones. Lockheed approach (original NGAD) - lost.

Of course, usual disclaimer - everything above is but my speculation from available data points.

 

[AndrewS]

That inevitably begs the question: why would Boeing and the USAF prioritize maneuverability over achievable RCS by pursuing a design with canards when they can almost certainly engineer a design without canards?

The general consensus seems to be that sensor technologies for detecting and tracking VLO aircraft are evolving faster than VLO technologies that make aircraft stealthy or at least stealthier. As such, is the USAF going for a less stealthy design because they feel the "additional stealth" that can be achieved sans canards will largely be moot or otherwise make for poor ROI due to emergent sensor technologies?

I think there are two main reasons why absolute stealth won't be as important:

1. There are going to be a lot more stealthy aircraft in the air. But you can't really tell the difference between a F-47 compared to larger numbers of cheaper CCAs and Kratos Valkyries

2. With the proliferation of AESA radars and also platforms, there is going to be so much ECM and EW spoofing/warfare.
So lower levels of stealth won't make much difference.

---
So it isn't really the sensor technology getting radically better.
It's that there are going to be many, many more targets.

 

[Gloire_bb]

So lower levels of stealth won't make much difference.

Still, all new planes are stealthier than their predecessors(let's see for USN)
F-47, canards or not, is likely wastly ahead of 5th gen(though point of intermixing is indeed well taken - even if lower bands can detect, they won't identify).
Both Chinese ones are clearly aimed at broadband VLO.

 

[00CuriousObserver]

I haven't seen anyone seriously claiming that the F-47's canards are a major issue. After all, the J-20 has canards too.

And yes, it's technically true that not all canards are the same, and that they can be done right. But like... Let's be honest hereーmost of the people who joked about J-20's canards weren't exactly concerned with technicality.

Shifting the goalpost is the method here. "Canards aren't stealthy" --> "Well not all canards are the same, some can be stealthy"

 

[constalation]

I think there are two main reasons why absolute stealth won't be as important:

1. There are going to be a lot more stealthy aircraft in the air. But you can't really tell the difference between a F-47 compared to larger numbers of cheaper CCAs and Kratos Valkyries

2. With the proliferation of AESA radars and also platforms, there is going to be so much ECM and EW spoofing/warfare.
So lower levels of stealth won't make much difference.

---
So it isn't really the sensor technology getting radically better.
It's that there are going to be many, many more targets.

But you are forgetting AI. Machine learning is exceptionally capable at pattern recognition, so much so that we can access almost 'all-knowing humans' at a tip of a hat (LLMs like chatgpt etc.)

Pass radar/sensor data through a ML algorithm, and stealth or not, even the smallest differences are going to stick up like a sore thumb. Such algorithms can be very easily distilled and fine tuned for deployment.

 

[AndrewS]

Still, all new planes are stealthier than their predecessors(let's see for USN)
F-47, canards or not, is likely wastly ahead of 5th gen(though point of intermixing is indeed well taken - even if lower bands can detect, they won't identify).
Both Chinese ones are clearly aimed at broadband VLO.

That's a good point.

If you have clouds of manned 5th gen stealth fighters, those can still be detected with long-wave radars.
So even with canards, a 6th gen aircraft (such as F-47 or J-50) should still be stealthier.

So there isn't a requirement for the lowest levels of stealth.

---

But in comparison, if a 6th gen aircraft is operating by itself, additional stealth would be useful as there aren't many/any other aircraft to hide amongst. Nor as much ECM/EW support.

So I think this helps inform the design choices for the B-21 and J-36.

 

[AndrewS]

But you are forgetting AI. Machine learning is exceptionally capable at pattern recognition, so much so that we can access almost 'all-knowing humans' at a tip of a hat (LLMs like chatgpt etc.)

Pass radar/sensor data through a ML algorithm, and stealth or not, even the smallest differences are going to stick up like a sore thumb. Such algorithms can be very easily distilled and fine tuned for deployment.

All you get is a track. However, you don't know what it is.

 

[siegecrossbow]

I haven't seen anyone seriously claiming that the F-47's canards are a major issue. After all, the J-20 has canards too.

And yes, it's technically true that not all canards are the same, and that they can be done right. But like... Let's be honest hereーmost of the people who joked about J-20's canards weren't exactly concerned with technicality.

Shifting the goalpost is the method here. "Canards aren't stealthy" --> "Well not all canards are the same, some can be stealthy"

I think the concern is that having canards might not be as good for broad band stealth. Fifth gen are usually optimized for high frequency radar but Sixth gen should aim for greater radar band/all aspect.

 

[constalation]

All you get is a track. However, you don't know what it is.

Well, ML algorithms specialize in classification. Given a data input, after processing, they will assign a probability and degree of confidence to each of the categories, the highest of which will become the answer. I would therefore say, ML algorithms will be better at knowing what it is rather than tracking it.

If we take your example, we can have one algo to processes and separate input noise to check whether there is a valid radar signature. This signature can be upscaled, then be used to predict the target based on.

Predicting future paths (depending on region) will be the hardest imo.

Overall, I think the biggest obstacle will be in gathering training data. They will need to find a way to make a model with an approximately accurate signature, then catch radar signatures at different approaches, distances/altitudes , even weather if that applies. Depending on the radar types/sensors there might be data format differences etc. So will need transforming and cleaning. I think these will be the hardest actually.

 

[AndrewS]

Well, ML algorithms specialize in classification. Given a data input, after processing, they will assign a probability and degree of confidence to each of the categories, the highest of which will become the answer. I would therefore say, ML algorithms will be better at knowing what it is rather than tracking it.

If we take your example, we can have one algo to processes and separate input noise to check whether there is a valid radar signature. This signature can be upscaled, then be used to predict the target based on.

Predicting future paths (depending on region) will be the hardest imo.

Overall, I think the biggest obstacle will be in gathering training data. They will need to find a way to make a model with an approximately accurate signature, then catch radar signatures at different approaches, distances/altitudes , even weather if that applies. Depending on the radar types/sensors there might be data format differences etc. So will need transforming and cleaning. I think these will be the hardest actually.

The signal strengths are just too low.

I think you can use submarine warfare as an analogy.
You have sound waves of varying frequencies and a lot of background noise.
It's still really difficult to figure out if there is a real target, what it actually is and the location.

In the airborne realm, it's a lot more difficult because:
1. radars only work with a narrow range of frequencies
2. there will be way more targets and EW/ECM actively emitting