Jump to content

Recommended Posts

Posted

For JDAM / JSOW weapons, I cannot make use of more than 1 program for a single target.

- Using targeting pod (Lighting) , identify ground target and designate using TDC depress. This sets TOO1 for the selected weapon correctly. Switch to TOO2 and the co-ordinates from the target are listed already. Slewing the target pod does not update TOO2 target coordinates. If I select designate on a new target, it will update TOO2 to the new target however it also replaces TOO1's target coordinates. If I select undesignate, the TOO1/TOO2 both lose their target. This renders having two programs for TOO (TOO1 and TOO2) fairly useless.

Various videos have demonstrated how to use TOO1 and TOO2 to target multiple impact points. Similarily, multiple impact points just based upon station number was possible and a total of 8 could be individually targeted using a comination of station and TOO1/TOO2). None of the methods in these videos work any more.

Wags had mentioned in his video posted June 2, 2020 that we would be able to target multiple impact points for our JDAMS in TOO (Video link included at correct time stamp). This functionality which Wags mentioned did get put in ( as evidenced by instructional videos from other users) but is now broken / removed?

Are there any plans to fix / restore this? Thanks!

 

  • Like 1
  • Thanks 1
Posted

It doesn't work like that for months. ED changed that and consider it "correct as is" although it's BS. Now you can attack only one target in TOO

 


 

Posted

Yes it has changed I remember doing it this 'old' way. I recommend using markpoints to do this now. Bit of an extra step but fairly intuitive. Markpoints had not been modelled when that video was done. 

  • Like 1
Posted (edited)
8 minutes ago, Devastator said:

if that is correct as it is, then why there is TOO1 and TOO2 in first place? What is the usage of them, can someone from ED also shed a light on this one then?

Good point. Unsure. I seem to remember someone saying that this TOO process was debated. Although it can be done IRL it was unlikely and unusual so no-one was really sure. But I would agree that the way it used to be made sense and seems to fit with the way everything works. Kind of similar to target points in the harrier. 

Edited by Hoirtel
Posted
27 minutes ago, Devastator said:

if that is correct as it is, then why there is TOO1 and TOO2 in first place? What is the usage of them, can someone from ED also shed a light on this one then?

Don't ask me, ask BigNewy, but ha can't or won't answer that question.
IMO it's illogical and non sense, but ED has its own way....

  • Like 1
Posted (edited)

There are light clear, highly detailed and very well written documents that describe exactly how the JDAM works in the Hornet. It's also described that the reason why TOO isn't used doctrinally has nothing to do with the avionics or implementation but rather the fact that on board sensors are very limited when it comes to generate coordinates without too high of a TLE. An ATFLIR autotrack with lazing at low slant range and high grazing angles is the only authorized way to do this accurately. But yes, the initial implementation was correct, both TOO targets can be saved to the desired weapon and undesignate doesn't remove the coordinates.

 

Using JDAMs in TOO mode is less of a thing in real life because the vast majority of targets that an aircraft would employ weapons against are fixed buildings and the US has a good enough intelligence apparatus to obtain mensurated coordinates that allow a low enough TLE to achieve the desired weapon effects. What I suspect is that this was changed on purpose to be inaccurate so that ED can avoid the accusation of using illegally obtained material to simulate the Hornet. Same thing was said about MSI, same thing was said about EOM mode,  the RWR functionality, same deal with the flight model and so on. The Hornet is a sensitive, classified aircraft where many systems are changed in game to avoid getting into confidentiality issues.

Edited by Fromthedeep
  • Like 3
Posted
12 hours ago, Fromthedeep said:

There are light clear, highly detailed and very well written documents that describe exactly how the JDAM works in the Hornet. It's also described that the reason why TOO isn't used doctrinally has nothing to do with the avionics or implementation but rather the fact that on board sensors are very limited when it comes to generate coordinates without too high of a TLE. An ATFLIR autotrack with lazing at low slant range and high grazing angles is the only authorized way to do this accurately. But yes, the initial implementation was correct, both TOO targets can be saved to the desired weapon and undesignate doesn't remove the coordinates.

 

Using JDAMs in TOO mode is less of a thing in real life because the vast majority of targets that an aircraft would employ weapons against are fixed buildings and the US has a good enough intelligence apparatus to obtain mensurated coordinates that allow a low enough TLE to achieve the desired weapon effects. What I suspect is that this was changed on purpose to be inaccurate so that ED can avoid the accusation of using illegally obtained material to simulate the Hornet. Same thing was said about MSI, same thing was said about EOM mode,  the RWR functionality, same deal with the flight model and so on. The Hornet is a sensitive, classified aircraft where many systems are changed in game to avoid getting into confidentiality issues.

 

Screaming at the wind but why have a TOO mode in the first place then if targets are always pre planned. And since sensitive info is classified, what does it matter how DCS chooses to model the Hornet systems, who is to know if it's wrong or if it's right besides those sworn to secrecy. If DCS took stabs in the dark at it, or best assumptions, and got it close to right, that means the government comes along and says, no that's too accurate even though you don't actually know whether it is or isn't so change it. That just tells us it was probably right in the first place. Not sure why something like TOO 2 would be classified but whatever. I mean, program target, drop bomb, bomb goes to target, that's just common sense.

Posted
6 hours ago, rfxcasey said:

have a TOO mode in the first place then if targets are always pre planned

The avionics are not necessarily going to 100% reflect how fleet jets operate or what their TTPs are. There can be a lot of superfluous features or things that may sound nice to an engineer but it doesn't actually work out in practice. With that being said, TOO can be used but realistically speaking it has very limited use. If you're going up against any kind of moving targets, conventional JDAMs are not going to be useful. If you're going up against any kind of static target, a factory, a radar site, a fuel depot, a bunker or whatever else, you'll know where it is in advance. You can get much better coordinates through the intel apparatus than you'd get with ownship sensors. 

The only situation in which you'd probably use this feature is CAS. The issue however is when using onboard sensors to designate for a JDAM, the coordinates will only be accurate in certain conditions. No matter the jet, you can't willy nilly designate something and expect to get useful coordinates. If you're using the TGP, you'll have to adhere to specific guidelines and you'd only designate at a certain range, altitude and with a specified angle looking at the target. If you're outside of these parameters and you designate an apartment complex at 40 nautical miles at 30k feet with a few degrees of look down angle, I can guarantee you that you won't get accurate coordinates for the object you want to hit. In a CAS wheel, close in with an autotrack and laser ranging, I can see it working out. But if the coordinates given by the JTAC are more accurate, even in this situation it'd be better to punch them in and drop it in PP.

 

6 hours ago, rfxcasey said:

And since sensitive info is classified, what does it matter how DCS chooses to model the Hornet systems, who is to know if it's wrong or if it's right besides those sworn to secrecy

Even guessing at classified systems can be very problematic because the people who would investigate the issue do have the ability to check out how the real system is supposed to work. And then it's on you to prove that you didn't use illegally obtained material when modelling the system. How do you prove a negative? That's the issue and that's exactly why ED refuses to even guess at sensitive systems, like the IFF or electronic warfare capabilities. Things change really quickly if you're just Joe Schmoe talking about things on Reddit and if you're an international military contractor with hundreds of employees selling actual products. The other big issue is that info on the JDAM in the Hornet is in fact available and if you model it in an accurate way, it's very simple to assume that illegal material was used to model the system.

  • Like 1
Posted
19 hours ago, Fromthedeep said:

...What I suspect is that this was changed on purpose to be inaccurate so that ED can avoid the accusation of using illegally obtained material to simulate the Hornet. Same thing was said about MSI, same thing was said about EOM mode,  the RWR functionality, same deal with the flight model and so on. The Hornet is a sensitive, classified aircraft where many systems are changed in game to avoid getting into confidentiality issues.

 

Then why always keep it on the dumb downed side...  why not create a super functional system, unmatched by any other DCS platform :biggrin:. Who knows, some contractors might try to copy and then ED will get them in trouble. Oh wait... things are changing a bit in the world :argue:

Posted
4 minutes ago, Gripes323 said:

Then why always keep it on the dumb downed side...  why not create a super functional system, unmatched by any other DCS platform

You win some and you lose some. Anything with contrast lock, TGPs, coordinate generations, lasers and ground radars are much better than in real life.

Posted
Even guessing at classified systems can be very problematic because the people who would investigate the issue do have the ability to check out how the real system is supposed to work. And then it's on you to prove that you didn't use illegally obtained material when modelling the system. How do you prove a negative? That's the issue and that's exactly why ED refuses to even guess at sensitive systems, like the IFF or electronic warfare capabilities. Things change really quickly if you're just Joe Schmoe talking about things on Reddit and if you're an international military contractor with hundreds of employees selling actual products. The other big issue is that info on the JDAM in the Hornet is in fact available and if you model it in an accurate way, it's very simple to assume that illegal material was used to model the system.
I don't see the issue with IFF whatsoever. They could at least get the CP side functional. For AI it can be real simple, you set only returns from the set code be to displayed on the AZ/EL. Nothing wrong with that. Even just general mode 4 can be done just fine. We know there's codes, we know how they are displayed in the aircraft, what an interrogation looks like, we know that the codes can automatically be switched at set times or geo locations. Non of the IRL backend system stuff matters for this.

Say we know that RWS sub mode 2 allows for better detection of helicopters, increasing detection range by 15nm. That can be stimulated right? Nothing about how the radar gets that result matters.

It's still insane to me that we know much much more about the capabilites of the F-35s sensor fusion than a 2005 Hornet

Sent from my SM-G975U using Tapatalk

Posted
3 hours ago, Fromthedeep said:

because the people who would investigate the issue do have the ability to check out how the real system is supposed to work. 

You are making a very big assumption there that I don't think that you can back up and thus would be misleading.

 

And as Hulkbust44 says, there is enough open source knowledge on many systems to implement many things that are not implemented in DCS.

Unfortunately, as I don't think anybody can prove with OS info whether TOO is correctly or wrongly implemented right now, I think this is not one of those things.

Guest
This topic is now closed to further replies.
  • Recently Browsing   0 members

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