Pointblank
Senior Member
War isn't fair. Your exercises should reflect this. You take any advantage you got over your opponent, even if it is a small one.
New J-10 thread II
- Thread starter wlchang
- Start date
- Status
War isn't fair. Your exercises should reflect this. You take any advantage you got over your opponent, even if it is a small one.
The article never mentioned the J-10s breaking the rules, so all is fair and kudos for the J-10s training to fight below the belt like they are at war. it should be mentioned that the J-10 pilots used their knowledge of the terrain to fly low and hide beneath the radars so they can jump on the Su-27s. While the exercise does not prove the J-10 being superior to the Su-27, it does show that one set of pilots have aggressive initiative that makes them better than the other set. Seems to me the decision to gift the J-10s to the 44th Division was well deserved, while the fact that it was an instructors regiment that got their asses handed must have been disheartening for the PLAAF.
One must remember that as a note, J-10 groups have been playing "Blue Flag" aggressors to the PLAAF. Which means they are simulating Western type pilots and aircraft. The practice started early in the J-10's deployment with the J-10s that IOC'ed in the FTTC.
Now for other things. Lets take a look at the list.
"- Advanced Radar with Mecanical Sweeping (or PESA/AESA), with Standard Pulse Doppler radar capabilities plus, NCTR, TWS, LPI, Terrain Following, "Fox-3" Capability with ability to guide several missiles on several Targets, High speed Microprocessor"
Not all the airplanes you list have all the capabilities here. Anyway, the J-10's radar appears to meet most of these, based on what we have seen with other Chinese radars. The LPI is probably pulse compression, which has been featured with other Chinese radars before. Terrain following has been seen on the JH-7A's radar. Multiple engagement is probably around two to four, with TWS at least 10 or 16 targets. The radar is a mechanical slotted array planar using pulse doppler and monopulse. Chinese radars since the late 1990s have been benefiting from COTS technologies.
"- Full Digital FCS, with analogic backup."
Full digital quadruple redundant FCS, but no analog backup.
"- Digital Stores, Mission, engines and instruments control systems."
Check.
"- Advanced Electronic War system, with chaff/flare programs, automated responses to threats, 360 ° Coverage, use of digital jamming techniques, algorithms used to treat electromagnetic waves."
Based on what we have seen on the JF-17, the J-10 should have no less. In fact, added to that, the J-10 also has an optical MAWS, that along with the RWR, provides an automated response to missile threats. The EW located on the tail of the J-10, and probably the same one on the JF-17, appears to phase shift a jamming beam towards the offender, cued by the RWR.
"- Ability to do both Air-to-Air & Air-to-Ground missions.
- Data Link system."
Check and check here.
"- Sensors data fusion"
This is a bit vague as the list includes planes that don't seem to fit here well either.
"- Use of Smart Munitions (GPS or Laser bombs/missiles)"
Check, though you have to change GPS with Beidou.
"- Advanced Navigation system with GPS."
Check, change GPS to Beidou. It may use GPS during peacetime though.
One must remember that as a note, J-10 groups have been playing "Blue Flag" aggressors to the PLAAF. Which means they are simulating Western type pilots and aircraft. The practice started early in the J-10's deployment with the J-10s that IOC'ed in the FTTC.
Now for other things. Lets take a look at the list.
"- Advanced Radar with Mecanical Sweeping (or PESA/AESA), with Standard Pulse Doppler radar capabilities plus, NCTR, TWS, LPI, Terrain Following, "Fox-3" Capability with ability to guide several missiles on several Targets, High speed Microprocessor"
Not all the airplanes you list have all the capabilities here. Anyway, the J-10's radar appears to meet most of these, based on what we have seen with other Chinese radars. The LPI is probably pulse compression, which has been featured with other Chinese radars before. Terrain following has been seen on the JH-7A's radar. Multiple engagement is probably around two to four, with TWS at least 10 or 16 targets. The radar is a mechanical slotted array planar using pulse doppler and monopulse. Chinese radars since the late 1990s have been benefiting from COTS technologies.
"- Full Digital FCS, with analogic backup."
Full digital quadruple redundant FCS, but no analog backup.
"- Digital Stores, Mission, engines and instruments control systems."
Check.
"- Advanced Electronic War system, with chaff/flare programs, automated responses to threats, 360 ° Coverage, use of digital jamming techniques, algorithms used to treat electromagnetic waves."
Based on what we have seen on the JF-17, the J-10 should have no less. In fact, added to that, the J-10 also has an optical MAWS, that along with the RWR, provides an automated response to missile threats. The EW located on the tail of the J-10, and probably the same one on the JF-17, appears to phase shift a jamming beam towards the offender, cued by the RWR.
"- Ability to do both Air-to-Air & Air-to-Ground missions.
- Data Link system."
Check and check here.
"- Sensors data fusion"
This is a bit vague as the list includes planes that don't seem to fit here well either.
"- Use of Smart Munitions (GPS or Laser bombs/missiles)"
Check, though you have to change GPS with Beidou.
"- Advanced Navigation system with GPS."
Check, change GPS to Beidou. It may use GPS during peacetime though.
Last edited:
RedMercury
Junior Member
Concept of data fusion is trivial. I can write a kalman filter in 6 lines of matlab which does effectively fuse data from multiple inputs. You'll have to be more specific about what data fusion you mean.
Do PLAAF adapted russian's standard measurement of RCS( 3m RCS) or the US (1m rcs)?
alright, listening to the guy talking about J-10's avionics now from
- he said that avionics are very import, but are also very expensive
- J-10 cost 200 million Yuan each, export to Pakistan would probably be 300 million each
- out of that, 100 million is spent on the avionics, takes 10 years to develop
- one is DCFS (Digital Flight Control Systems), which allows pilot to control the plane with displayed information
- talks about the FBW computer, which does 500 K calculation per second (that's not even 1 MHz!, what?)
- other is mission computer, you just plug in the mission and the fighter will tell you what to do
- then talks about INS (inertial navigational system), you can go from Beijing to Shenzhen with this and only be off by 500 m
- and the other is CADC (central air data computer), it is used to detect the temperature, dew point, wind direction/speed and stuff like that.
- other is SMS (store management system), this is the system controlling the launching of missiles and bombs (it has to calculate how high you are launching certain weapons and your speed to achieve optimize results)
- radar - can detect up to 150 km (here is interesting part, he mentions that F-22 can only do 170 km, but that is vs 1 m^2 target, I doubt J-10 is facing that size to get to 150 km, maybe 3 m^2 or 5 M^2?)
- CNI - communication/navigation/identification, plane has wide band, narrow band (different types of waves that it's sending and receiving), didn't really talk about datalink or IFF, but I'm guessing that's part of this
- EMS - (electronic counter measure system?) - not much said
talks about a fighter jet is like a networked computer system with many sub parts, each has a computer that can do half to 1 million calculations per second. Also, they take up very little space.
That's about it on J-10's avionics.
- he said that avionics are very import, but are also very expensive
- J-10 cost 200 million Yuan each, export to Pakistan would probably be 300 million each
- out of that, 100 million is spent on the avionics, takes 10 years to develop
- one is DCFS (Digital Flight Control Systems), which allows pilot to control the plane with displayed information
- talks about the FBW computer, which does 500 K calculation per second (that's not even 1 MHz!, what?)
- other is mission computer, you just plug in the mission and the fighter will tell you what to do
- then talks about INS (inertial navigational system), you can go from Beijing to Shenzhen with this and only be off by 500 m
- and the other is CADC (central air data computer), it is used to detect the temperature, dew point, wind direction/speed and stuff like that.
- other is SMS (store management system), this is the system controlling the launching of missiles and bombs (it has to calculate how high you are launching certain weapons and your speed to achieve optimize results)
- radar - can detect up to 150 km (here is interesting part, he mentions that F-22 can only do 170 km, but that is vs 1 m^2 target, I doubt J-10 is facing that size to get to 150 km, maybe 3 m^2 or 5 M^2?)
- CNI - communication/navigation/identification, plane has wide band, narrow band (different types of waves that it's sending and receiving), didn't really talk about datalink or IFF, but I'm guessing that's part of this
- EMS - (electronic counter measure system?) - not much said
talks about a fighter jet is like a networked computer system with many sub parts, each has a computer that can do half to 1 million calculations per second. Also, they take up very little space.
That's about it on J-10's avionics.
Pointblank
Senior Member
The original computer in the F/A-18A Hornet was as powerful as a Commodore 64. The latest computers in the F/A-18C Hornet is about as powerful as a Pentium I or a 486. There’s a very good reason for this apparent lag in technology. What the new fighter aircraft computers lack in computer power, they more than make up for in ruggedness. Your laptop or desktop computer might be powerful but, how long would it keep working at 35,000 feet, in -50°C tempuratures, and under a force of five or six times normal gravity?
Yup , if I recall well, the F-22 & Rafale CPU's have approx. 200 MHz of frequency.
tphuang said:
It is due to the fact that the computer is only computing the parameters for flight controls, it's using pure processor language, so don't need too much iterations, etc...
For example, on the F-16C for the rudder control, there is nearly 60 fonctionnal blocks without counting filters. The FCS is just "reajusting" the inputs done by the action of the pilot on the stick, nothing else.
Crobato said:
I'm not pretty sure of that, because there is often 3 Digital wires and 1 analog, and it's very important to have an analog way.
Crobato said:
I could be wrong with the Terrain Following function (not present on F-16C/D and F-15S).
To explain the sensor data fusion, I'll take an example:
The sensors data fusion means that each contact or signal identified as a potential target or threat by every enough accurate sensor is projected on a tactical situation display, you can have data from the Electronic War System, Radar, Data Link or IRST, not neccesary all "sensors" but 2 or 3 "sensors" in the majority of cases. This data can be exploited to engage those targets and potential threats, or give backup to the Electronic War Sys.
It's enough ?
RedMercury
Junior Member
You're not even talking in terms of probability. I'm on a whole other level. 
Edit: To elaborate: the fusing of data from multiple sources can be done in an ad-hoc manner or done using proper mathematics. The former any programmer can hack together. The latter has come about with the recognition that any real world problem requires consideration of probabilities and the subsequent advancement of Bayesian techniques in AI. This means using mathematical models to interpret data. The theory is pretty well understood. The challenge is approximating the computationally intractable calculations. This is where most machine learning research is these days: better and faster approximations. This is where faster and more clever algorithms meet faster hardware to tackle larger inference problems. Once you do this, the rest, displaying it is just user interface design.
An example: why did the tail slide maneuver break radar locks back when it was first revealed? Radars operate in search and or track modes. In search, finding the target is actually pretty low probability, because there is so much volume to scan. Once you find the target, you want to lock on to it, so you don't have to find it again. This is called tracking. The most basic tracking algorithm you can imagine is one that keeps a guess of where the target is. The algorithm has a model of how the target moves, i.e. for an aircraft, it moves generally in straight lines, but has some chance of turning. This is quantified with a probabilistic model of movement. Maybe, say, 90% of the time the aircraft is moving straight, and 10% of the it is turning. Of course, a real model would be much more complicated, take into account the typical piloting maneuvers, etc. Using this prior knowledge of the movement, called a "movement model" or "system evolution model", the radar has a good idea where to look for the target, so it has a much higher chance of getting another return, instead of just scanning the sky during search mode.
Now, the tail slide maneuver probably was not anticipated in the movement models of the time, so when a Mig or Su performed it, the radar expected the plane to keep moving forward, so it scanned in front of the plane. However, not getting any returns, it loses lock and has to go back to searching again. How was this fixed? You could add a probability of a backward movement happening, esp. if the target is flying upwards (or if the radar believes the plane is such and such model). And you could go to more complicated tracking methods where more than one guess of the target location is kept in memory at a time (at the cost of more computations to track the target), with the ambiguity eventually resolved with the arrival of more information.
Edit: To elaborate: the fusing of data from multiple sources can be done in an ad-hoc manner or done using proper mathematics. The former any programmer can hack together. The latter has come about with the recognition that any real world problem requires consideration of probabilities and the subsequent advancement of Bayesian techniques in AI. This means using mathematical models to interpret data. The theory is pretty well understood. The challenge is approximating the computationally intractable calculations. This is where most machine learning research is these days: better and faster approximations. This is where faster and more clever algorithms meet faster hardware to tackle larger inference problems. Once you do this, the rest, displaying it is just user interface design.
An example: why did the tail slide maneuver break radar locks back when it was first revealed? Radars operate in search and or track modes. In search, finding the target is actually pretty low probability, because there is so much volume to scan. Once you find the target, you want to lock on to it, so you don't have to find it again. This is called tracking. The most basic tracking algorithm you can imagine is one that keeps a guess of where the target is. The algorithm has a model of how the target moves, i.e. for an aircraft, it moves generally in straight lines, but has some chance of turning. This is quantified with a probabilistic model of movement. Maybe, say, 90% of the time the aircraft is moving straight, and 10% of the it is turning. Of course, a real model would be much more complicated, take into account the typical piloting maneuvers, etc. Using this prior knowledge of the movement, called a "movement model" or "system evolution model", the radar has a good idea where to look for the target, so it has a much higher chance of getting another return, instead of just scanning the sky during search mode.
Now, the tail slide maneuver probably was not anticipated in the movement models of the time, so when a Mig or Su performed it, the radar expected the plane to keep moving forward, so it scanned in front of the plane. However, not getting any returns, it loses lock and has to go back to searching again. How was this fixed? You could add a probability of a backward movement happening, esp. if the target is flying upwards (or if the radar believes the plane is such and such model). And you could go to more complicated tracking methods where more than one guess of the target location is kept in memory at a time (at the cost of more computations to track the target), with the ambiguity eventually resolved with the arrival of more information.
Last edited:
speculator
New Member
G's wont affect computer components. even my ipod can sustain 50g and still work. (says on the manual) as for -50, most computer chips work better the colder they are. it is not unheard of for people to cool CPU's by using liquid nitrogen or dry ice. altitudes do not affect components, computers don't need oxygen or pressurization.
thus i have no idea why 50 million dollar jets lag behind so much on electronics.
RedMercury
Junior Member
Redundancy and resistance to EMP. Redundancy requires, at least in the traditional paradigm, 3 times as much circuitry to do the same thing. The 3 independent systems vote on the solution, so even if one fails the correct answer is still given. To resist EMP, either systems must be heavily shielded in faraday boxes or the line widths must be wide enough to conduct high currents without melting. These factors are contrary to the design considerations which drive commercial technology.
- Status