F\A-18 Cluster Bombs - Page 2 - ED Forums
 


Notices

Reply
 
Thread Tools Display Modes
Old 09-17-2020, 11:30 PM   #11
LastRifleRound
Member
 
Join Date: Apr 2008
Posts: 839
Default

Quote:
Originally Posted by Frederf View Post
Historically killing armor with unitary bombs was poor. Accuracy of unguided bomb delivery and needed to kill a tank were incompatible. It was like shooting skeet pigeons with a rifle.

CBUs are shotguns. They bring the required accuracy to the level of performed accuracy. The bomblet still has to hit the tank to damage it. Take the footprint of a tank and divide it by the circular area of the impact zone and multiply it by the number of submunitions. If that number is 0.5 then you have a 50% chance of a submunition hitting the tank. Very quickly you realize that it becomes a question of how many CBUs does it take to kill one tank and not how many tanks can be killed by one CBU.

Unfortunately our control over the dispersion parameters is limited especially in the Hornet. We can't change HOF and instead of the 247 submunitions in the real weapon how many are in DCS? 21. So take your probability of hit and multiply it by 21/247. If you drop 8 CBU-99 at once you get 168 submunition total or less than 247 of one weapon.

I did a drop where at least 5 Mk 118s hit a T-72 (dropped 8 CBU) and it was killed about four times over. If the Mk 118 hits a tank that tank is really hurt. The significant part of that sentence is "if."
8 bombs to 5 hits is very low. Were these delivered in a dive or level? the pattern seems to close up when delivered in a dive as per the tutorial with the recommended 1200ft HOB
LastRifleRound is offline   Reply With Quote
Old 09-18-2020, 12:24 AM   #12
Frederf
Veteran
 
Join Date: Dec 2008
Location: Campbell, CA
Posts: 4,576
Default

In a 30-45 degree dive CCIP. The 8 CBUs only produced 168 submunitions total and as the HOF is fixed to 1500' minus however more it falls in the next 1.2 after that. The dispersion is pretty large and the tank is pretty small. Maybe there were 10 hits, I didn't count exactly, at least five for sure. Even at minimum time scale it's hard to be sure since they pass through the tank object.

That's like a 5% hit rate which is as good as I can get it practically with reasonable delivery and no control over the dispersion characteristics. And if those 8 CBUs were delivering their realistic 1,976 submunitions then that 5% hit rate would be 99 hits which is more than more than enough.
Frederf is offline   Reply With Quote
Old 09-18-2020, 02:06 AM   #13
Flagrum
Veteran
 
Join Date: Feb 2013
Posts: 6,241
Default

Quote:
Originally Posted by Frederf View Post
We can't change HOF and instead of the 247 submunitions in the real weapon how many are in DCS? 21. So take your probability of hit and multiply it by 21/247. If you drop 8 CBU-99 at once you get 168 submunition total or less than 247 of one weapon.
Afaik the visual effects are reduced to 20, but simulated are still all 247 bomblets. That is at least how I interprete the bombs_table.lua:
Code:
           cluster = {
                count        = 247,
                effect_count = 20,
            
                wind_sigma  = 50,
                impulse_sigma = 2,
                moment_sigma = 0.0001,
             }
__________________
Did you know? True Harrier fans contribute to the AV-8B Community Bugtracker!
Flagrum is offline   Reply With Quote
Old 09-18-2020, 06:17 AM   #14
Bunny Clark
Member
 
Bunny Clark's Avatar
 
Join Date: May 2019
Location: Minneapolis
Posts: 464
Default

So, I'm still frustrated by this.

I made 10 drops in CCIP with pairs of CBU-99s (so, 20 bombs total). All were visually on target, which is to say the BMP-1 I was attacking was within the pattern of visual explosions. I got 5 kills. So half the time two CBU-99s with a centered blast pattern over the target failed to destroy a BMP. That's not too impressive.

Maybe I'm still doing something wrong? It seemed I had better luck with steeper dive angles, like around 40 degrees, but it was hard to tell.

I also tried in a few drops in Auto with a FLIR designated target point. Every single release fell short by about 100 feet.
Attached Thumbnails
Click image for larger version

Name:	Screen_200918_003627.png
Views:	39
Size:	393.5 KB
ID:	247687  
Bunny Clark is offline   Reply With Quote
Old 09-18-2020, 06:31 AM   #15
Frederf
Veteran
 
Join Date: Dec 2008
Location: Campbell, CA
Posts: 4,576
Default

Modifying it changes the MK20 but not the CBU-99. So I set effect to 1 and got 1 visual object but effectiveness equal to before as far as I can tell. I'm still a little confused how the non-effect is calculated. Is it a random Monte Carlo type trajectory calc or even distribution or purely probabilistic in a volume? If I set effect = count can I inspect hit/no-hit visually? I don't know.

The fuze is still wrong because it's 1500' but it's waiting for 1.2s to detonate instead of 0s and I couldn't get it to zero even editing the file. The dispersion is very high for a non-spinning canister. There doesn't seem to be a spin value anywhere unless one of those L I Ma Mw impulse values control that.

MK118 aren't the super deadliest AT weapons with 150mm or more RHA equiv penetration which I expect is significant but not overmatch for T-72 roof armor. BLU-97 on the other hand should mulch T-72s at ~150% explosive weight and about double the penetration.

Really I'd want to change spin and HOF to really narrow down the footprint and get the density up.
Frederf is offline   Reply With Quote
Old 09-18-2020, 07:10 AM   #16
LastRifleRound
Member
 
Join Date: Apr 2008
Posts: 839
Default

Quote:
Originally Posted by Bunny Clark View Post
So, I'm still frustrated by this.

I made 10 drops in CCIP with pairs of CBU-99s (so, 20 bombs total). All were visually on target, which is to say the BMP-1 I was attacking was within the pattern of visual explosions. I got 5 kills. So half the time two CBU-99s with a centered blast pattern over the target failed to destroy a BMP. That's not too impressive.

Maybe I'm still doing something wrong? It seemed I had better luck with steeper dive angles, like around 40 degrees, but it was hard to tell.

I also tried in a few drops in Auto with a FLIR designated target point. Every single release fell short by about 100 feet.
There is a current bug another user and I reported (it was acknowledged by the team) for all bombing with AUTO mode. If you are not at a specific airspeed for the altitude you were at for release, the release point is wrong and you will miss. Falling short indicates you were faster than the algorithm calculated for. Use CCIP only for now if you're going to attack in a dive.

EDIT:
Have you tried this same thing with Mk20's? That's what I usually use, maybe they are different in the sim for some reason.

Last edited by LastRifleRound; 09-18-2020 at 08:36 AM.
LastRifleRound is offline   Reply With Quote
Old 09-18-2020, 12:32 PM   #17
Flagrum
Veteran
 
Join Date: Feb 2013
Posts: 6,241
Default

Quote:
Originally Posted by Frederf View Post
Modifying it changes the MK20 but not the CBU-99. So I set effect to 1 and got 1 visual object but effectiveness equal to before as far as I can tell. I'm still a little confused how the non-effect is calculated. Is it a random Monte Carlo type trajectory calc or even distribution or purely probabilistic in a volume? If I set effect = count can I inspect hit/no-hit visually? I don't know.
I would assume, that the bomblets are simulated all the same, no matter if visible or not.

To observe visually the outcome if you set effect = count might be difficult. Even if your computer can handle it (slideshow during the explosions was the reason ED reduced the visual effects back then), the visual effects are quite large and will probably obscure the actual hit point quite a bit.

But maybe it's worth trying to investigate it further with TacView? Does TacView show individual bomblets? If so, only the visible ones or all?

Maybe try this (or I will this weekend, hrmmm): build a target area with targets tightly packed so that every bomblet should hit something. We should get a hit count per target/vehicle and could calculate the actual bomblet density if we know the (top)surface area of each target/vehicle.
__________________
Did you know? True Harrier fans contribute to the AV-8B Community Bugtracker!
Flagrum is offline   Reply With Quote
Old 09-18-2020, 08:23 PM   #18
Frederf
Veteran
 
Join Date: Dec 2008
Location: Campbell, CA
Posts: 4,576
Default

That's a reasonable assumption. I might edit the Mk20 down to count 1, effect 1 and delivering against a parking lot of T-72s to see if the damage matches the effect. So I did that



This looks like ~25% damage for a single impact. I managed to hit a tank a second time in a row (not bad for one submunition at a time) and it was slightly less than 50% health remaining. In my several attempts and watching the impact at 1/64x time I noticed that damage only occurs with hits and hits don't happen unless you're inside a certain radius. So it's possible to intersect the 3D model but not hit. The bubble is about the size of the turret, maybe a bit bigger.

Fire it up again with 247, squeeze together the tanks touching track to track, drop two REs for quantity 494 total. After a good pass where footprint overlaps parking lot I get 6 kills, 11 damaged (9 green, 1 yellow, 1 red).

Tanks killed
#2-10 hitx7
#2-6 hitx7
#2-9 hitx7
#3-6 hitx7
#3-7 hitx7
#3-8 hitx7

Damaged (health remaining)
#1-7 hit x1 75%
#11-3 hitx2 75%
#12-4 hitx2 75%
#14-4 hitx2 75%
#2-1 hitx2 75%
#2-3 hitx2 75%
#2-4 hitx6 15%
#2-5 hitx4 45%
#2-7 hitx2 75%
#3-4 hitx2 75%
#3-5 hitx2 75%
#4-10 hitx2 75%

Undamaged
#8-9 hitx1

I don't know if Mk 118 (7) means that seven Mk 118 hit that tank or if it's counting the HEAT warhead as a hit and maybe a second hit due to being in the fragmentation envelope. The sum of all the scoreboard hits is 74 out of a total submunitions of 494 or about 8% (really 4%) which might be reasonable or not. I dunno I had a grid of 5,500m2 parked 190 T-72 which have a nominal area of 23.6m2 each for 4500 m2 total. The CBU footprint had good but not perfect alignment, maybe 60-70% had this tank park in it so probably about 600 Mk 118s fell where the tanks were which covered 80% of their parking lot.

Since 74 hits (divided by two since indirect hits count as hits) occurred when one might naturally expect about 600 hits it's fair to conclude that the hit box for the tank is significantly smaller than the hull dimensions although density certainly isn't uniform so not as much as the raw numbers might suggest. All of the kills happened at the center of the footprint after all.

Considering that it seems to be possible to kill a T-72 in about 3.5 direct hits from a Mk 118 and yet some tanks are surviving 6 hits that suggests that those are 3 direct hits and 3 splash damage. That means the T-72 hit box is tiny compared to the hull dimension rectangle. I understand hitting the corner mud guard isn't a critical thing but you pretty much have to hit the turret to register a hit.

Killing a T-72 with a Rockeye II has the following issues:
  • F-18 aiming and/or canister fuzing makes accuracy difficult (bugs)
  • T-72 are either surviving up to 3 direct hits (pretty sure "6" means 3 direct plus 3 indirect)
  • T-72 hit box is a lot smaller than the hull dimensions depending on the above math.
  • Sub munition spread is non-configurable and quite wide (and uniform) for anti-armor work ~75m radius making density quite low compared to required for a kill.
Based on the available delivery accuracy and the required density to kill even when delivering in pairs I would crank down the spin if possible (I think Rockeye are fixed at 1500 RPM while CEM can do 0,500,1000,1500,2000,2500 RPM configurable) and lower the HOF. One source says a 500' HOF gives about a 40m radius which is kinda what we're seeing here. I wonder if the distribution is more Gaussian than uniform though. Uniform that's 1 bomblet for every 19m2. Even against T-55 it takes 2.5 direct hits to kill.
Frederf is offline   Reply With Quote
Old 09-18-2020, 10:26 PM   #19
Stubbies2003
Member
 
Join Date: Aug 2012
Posts: 345
Default

Quote:
Originally Posted by Frederf View Post
Unfortunately our control over the dispersion parameters is limited especially in the Hornet. We can't change HOF...

Completely true that wen have no direct control over the height that the weapon opens we do have direct control over speed and angle on weapons release which indeed has a marked effect in DCS on the footprint. Steep and fast drops result in very concentrated footprints that drive up your need to be accurate to score a hit. This makes sense with the 1500 feet + 1.2 seconds. The weapon has traveled farther prior to opening on fast drops thus smaller footprints. Steep dive angles also add to this effect by not having much movement over the target while dispersing compared to shallow dive angles. Slow and shallow drops can produce a very big footprint indeed.
Stubbies2003 is offline   Reply With Quote
Old 09-19-2020, 07:45 PM   #20
Bunny Clark
Member
 
Bunny Clark's Avatar
 
Join Date: May 2019
Location: Minneapolis
Posts: 464
Default

I did some testing, and the Mk.20 does seem to be more effective than the CBU-99, which doesn't make sense.

I'm still not finding it very effective. A single Mk.20 will kill a BMP about 75% of the time, and a pair of Mk.20s will kill a BMP nearly every time so long as my aim is decent. But a pair of Mk.20s seems to only be about 50% effective against a T-55.

At that point, against single tanks, Mavericks are more efficient. And GBU-12s are far more efficient, but also a more modern solution.
Bunny Clark is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

All times are GMT. The time now is 01:02 AM. vBulletin Skin by ForumMonkeys. Powered by vBulletin®.
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.