Naval Fire Control Questions

[DesertFox]

Thought I would bring this up here.

I have a person on another board claiming that you perform terminal guidance for the SM-2 and ESSM using only the SPY-1 and not the actual designators. This goes against the understanding I have of much of what I have read.

Anybody read anything on this subject

[Praet]

I have a person on another board claiming that you perform terminal guidance for the SM-2 and ESSM using only the SPY-1 and not the actual designators.

SPY-1 can provide midcourse guidance (either through target position updates the missile interprets with its own inertial nagigation system or full guidance commands), but for terminal guidance of SM-2 and ESSM illumination by a seperate fire-control radar is required.

[Josh]

That is my information as well, from every open source I have ever seen. The only exception would be the new active version in development I think to be named SM-6 no yet in service.

 

I have been told that when used against surface targets there is a way to create a ghost target at zero altitude on top of a ship and that SM-2 can be guided this way but I've never confirmed that with any other source so AFAIK for surface use the directors still need to be radiated on the targets.

[DesertFox]

I am kind of wondering if upgrades may allow it for much finer targeting with later marks of the SM-2 missile.

The people who i am discussing on another board tell me that they got the information from Stuart Slate, the same person in the Naval Technical Board.

 

Edit: Wanted to add that an FC on another board seems to argue against the idea as well with

 

Most likely it's a misconception, ESSM and SM-2 receive mid-course guidance from the SPY-1 and only switch to X-band illumination in the terminal phase. S-band from SPY-1 isn't high resolution enough for terminal phase, the average miss distance would be greater than the range of the missile's fuze.

Edited May 19, 2010 by DesertFox

[FITZ]

I am kind of wondering if upgrades may allow it for much finer targeting with later marks of the SM-2 missile.

The people who i am discussing on another board tell me that they got the information from Stuart Slate, the same person in the Naval Technical Board.

 

Edit: Wanted to add that an FC on another board seems to argue against the idea as well with

 

If they got that information from Slade then they didn't understand what he was talking about. Stuart knows full well that SM-2 and ESSM employ semi-active radar homing for the terminal phase of the engagement. Since SPY-1 CAN NOT PROVIDE continuous wave illumination of the target the missiles seeker requires that is what all those big round AN/SPG-62 dishes are for on the ship. The much more recent Dutch APAR unlike SPY-1 can provide terminal phase target illumination.

[DesertFox]

This is suppose to be a quote directly from Slade:

 

True, but one of the neat tricks with AEGIS is that one illuminator can control several missiles. Also, SM-2 doesn't need an illuminator for final stage homing (exactly why not depends on the version of SM-2 in use but SM-2 has been independent of that since the late 1980s). Now, using an active target illumination does greatly increase accuracy from a PK of around .6 to .9 against cruise missiles so its a prefered option. Standard is expensive after all.

 

There were some other quotes but this is the only one which seems to be a concern

[Praet]

Well, you can of course try to get the missile close to the target via midcourse updates/command guidance so that the proximity fuze triggers. That is consistent with what Stuart Slade wrote (although I'm suprised that the kill probability is supposed to be 0.6, as DesertFox I'd assume it should be lower).

But that does not equal target illumination for semi-active terminal homing. SPY-1 simply can't do that.

[DesertFox]

Thought I would add this here instead of starting a new thread

Why do you think that Australia chose Aegis/SPY-1D over APAR?

Is there an advantage for Aegis/SPY-1D over APAR

First, I am thinking as an older system, it may be more mature technology

Second, I am thinking it might be cheaper

Third, I was thinking that they want to use US hardware to keep a better relationship but that does not seem to always be the case. Much of the other hardware on the vessel is not US hardware such as sonar.

[Guest aevans]

Real naval fire control...

[ScottyB]

Well, you can of course try to get the missile close to the target via midcourse updates/command guidance so that the proximity fuze triggers. That is consistent with what Stuart Slade wrote (although I'm suprised that the kill probability is supposed to be 0.6, as DesertFox I'd assume it should be lower).

But that does not equal target illumination for semi-active terminal homing. SPY-1 simply can't do that.

 

When I was on Scott during the NTU tests, we used mid-course guidance to hit a drone using only the track info from the SPS-48.

 

Scott

[Guest aevans]

When I was on Scott during the NTU tests, we used mid-course guidance to hit a drone using only the track info from the SPS-48.

 

Scott

 

Command guidance is hardly new or unreliable. It goes all the way back to the Nike.

[sunday]

Real naval fire control...

 

Excellent videos!

[Steven P Allen]

Excellent videos!

 

Absolutely wonderful stuff!

[Guest aevans]

Absolutely wonderful stuff!

 

Thank my dad. He sent me the link. He's heavily into all that Second Industrial Age tech. In his mind, anybody can write an algorithm, but it takes a real man to synthesize an abstract mathematical concept into concrete machinery. I have to admit that the engineering solutions were pretty slick.

 

As a computer programmer, the constant maintenance of state is pretty impressive to me. In modern "real time" systems, you spin around in an endless loop, constantly resampling your data feeds and recalculating the problem state. In an electro mechanical fire control system, the state was just the combined position of all of the components at any given time. (It took up to 30 seconds to settle state down and start "tracking" a new target, but once you were there, the machine could keep up with changing inputs.) And it wasn't just the computer that was in a continuously updated state. The rotors in the control synchros were physically following the rest of the machine through feed shafts. When the computer's state changed enough to move a rotor to the next position (or the previous one) the new state of the system was transmitted to the corresponding synchro on a mount. If all of the mounts were in "Auto", then the mounts just became extensions of the fire control computer's state.

Edited May 21, 2010 by aevans

[APF]

The old systems surely looked impressive, but don't forget what a single mistake in the underlying algorithm meat: if you got lucky you *only* had to refabricate the part of the machine which was done wrong (in case of a slightly wrong exponent, for examle), if the shit hit the fan you'd have a lot of really expensive yunk in your ship/tank.

 

And while it looks as if they would calculate in real-time in reality they couldn't: like everything else a mechanical input needs a driving force to change its state, and together with the elasticity of every system it will lag behind until enough torque is build up to accelerate/move the components.

 

Much the same holds true for the supposedly infinite resolution: the resolution of the system is limited by noise/hysteresis and friction between the parts.

 

While this might be negligible, or rather should be designed to be lower than the inaccuracy of the sensors/ actuators, it's still there, in much the same way as you'd design a electronic FCS loop time to be short in comparison to the changing inputs.

 

Not to mention that I wouldn't want to have to add for example firing solutions for a 16' base bleed projectile into the MKxx calculator of an Iowa.

 

Greetings

[Guest aevans]

The old systems surely looked impressive, but don't forget what a single mistake in the underlying algorithm meat: if you got lucky you *only* had to refabricate the part of the machine which was done wrong (in case of a slightly wrong exponent, for examle), if the shit hit the fan you'd have a lot of really expensive yunk in your ship/tank.

 

They couldn't even design the machine, much less implement it, until the algorithm was verified by real world experience.

 

And while it looks as if they would calculate in real-time in reality they couldn't: like everything else a mechanical input needs a driving force to change its state, and together with the elasticity of every system it will lag behind until enough torque is build up to accelerate/move the components.

 

There was enough power in the system that once it was tracking, it was considered capable of keeping up with inputs. They certainly didn't miss what they were shooting at, and only got more accurate with the introduction of better and more continuous data (i.e. radar).

 

Much the same holds true for the supposedly infinite resolution: the resolution of the system is limited by noise/hysteresis and friction between the parts.

 

While this might be negligible, or rather should be designed to be lower than the inaccuracy of the sensors/ actuators, it's still there, in much the same way as you'd design a electronic FCS loop time to be short in comparison to the changing inputs.

 

Nobody said it had "infinite" resolution. It had good enough resolution that resolution wasn't an issue, if properly calibarated. It wasn't an infinite state machine, but it was a very fine state machine.

 

Not to mention that I wouldn't want to have to add for example firing solutions for a 16' base bleed projectile into the MKxx calculator of an Iowa.

 

Not difficult at all -- just stick in an appropriately cut ballistics cam. They had several in each main battery computer, to account for different projectiles and powder charges. The secondary battery computer (Mark 1) was designed for a standard projectile and powder charge, but included a star shell computer. If the specifications of the standard 5" projectile or powder charge had ever changed, they would have stuck in a differnt ballistics cam as a field mod.

Edited May 22, 2010 by aevans

[ScottyB]

Command guidance is hardly new or unreliable. It goes all the way back to the Nike.

 

Never said it was new or unreliable. I thought it was pretty interesting that the missile hit the target without a director during the terminal phase.

 

Scott

[Guest aevans]

Never said it was new or unreliable. I thought it was pretty interesting that the missile hit the target without a director during the terminal phase.

 

Scott

 

But that's the point I was making by bringing up Nike -- it was strictly command guidance, but it could definitely hit a target. The roughly contemporary Soviet SA-2 was also command guided, and also hit its targets, including high performance aircraft. To me that's considerably more interesting, given the state of technology when Nike and SA-2 were developed.

 

If the AEGIS system equiped for command guidance couldn't get a command guidance hit, at least on an aircraft that wasn't maneuvering too much, now that would be interesting.

Edited May 22, 2010 by aevans

[APF]

They couldn't even design the machine, much less implement it, until the algorithm was verified by real world experience.

 

And thats the difference to todays digital computers: Now we only need to know the input and output parameters to design the machine, the algorithm is implemented (and changed) in seconds by means of a memory stick (or something like that).

 

Nobody said it had "infinite" resolution. It had good enough resolution that resolution wasn't an issue, if properly calibarated. It wasn't an infinite state machine, but it was a very fine state machine.

 

40kYards in steps of 25 yards (Dreyer table) is some 10 1/2 bits. An el cheapo DVM has 3 1/2 digits, thats some 11 Bits resolution. Maybe I misinterpreted your original post, but to me it looked as if you wanted to say that the old analog computer were better than todays digital ones. While they were masterpieces of engeneering they're still outdated by todays standards and the reason is *not* that we're to lazy to mount some balls into cams.

 

Not difficult at all -- just stick in an appropriately cut ballistics cam. They had several in each main battery computer, to account for different projectiles and powder charges. [...]

 

You're right, that was a bad example. A better one for the difficulty to change the inner workings of an analog computer would be an either forgotten or (at the time of design) unknown dependency of a variable, something along y = a + bx+cx^2 instead of simply y = a + bx, or y = a + bx + cz. Things which might well creep up once the firing range increases.

 

Greetings

[ScottyB]

But that's the point I was making by bringing up Nike -- it was strictly command guidance, but it could definitely hit a target. The roughly contemporary Soviet SA-2 was also command guided, and also hit its targets, including high performance aircraft. To me that's considerably more interesting, given the state of technology when Nike and SA-2 were developed.

 

I agree, the SA-2 and Nike were primitive compared with an SM-2 but did pretty well. I guess, to me, perhaps the performance of the SPS-48E was very impressive.

 

Scott

[DesertFox]

I though that with Nike, there was a large reliance on nuclear warheads and you don't need to come very close with those?

[Guest Charles]

I though that with Nike, there was a large reliance on nuclear warheads and you don't need to come very close with those?

 

I suppose it would depend on the target(s). A large concentration of Soviet bombers, one would launch a Nuke tipped Nike (assuming the balloon has gone up here). A single target and I would hope that a standard Nike is launched.

I think that we need to look at time frames here, when discussing the older gen SAM's and their respective FCR.

 

Charles

[JOE BRENNAN]

Nike Hercules had a nuclear warhead option, but the earlier and (more) widely deployed Nike Ajax was conventional only. It had three warheads along its length, detonated together by ground command. And as already mentioned, besides the Nike missiles, the Soviet V-75/Sa-2, by far the most fired heavy SAM in combat, was also command guided. There's no issue of whether or not you can command guide a SAM to hit an airplane, you obviously can, it just has the basic drawback that expected absolute miss distance increases as range increases, whereas a semi-active seeker gets more accurate as the missile approachs the target regardless of range.

 

Joe

Edited May 23, 2010 by JOE BRENNAN

[Guest aevans]

And thats the difference to todays digital computers: Now we only need to know the input and output parameters to design the machine, the algorithm is implemented (and changed) in seconds by means of a memory stick (or something like that).

 

Digital computers are a different technology. And the ease of loading programs has nothing to do with the relative ease of analysing and programming updates. They don't make changes in modern fire control code like Web developers alter sites on the fly. They undergo so much formal analysis and review that the time it would take to load a solution, mechanical or digital, is a negligible part of the process.

 

40kYards in steps of 25 yards (Dreyer table) is some 10 1/2 bits. An el cheapo DVM has 3 1/2 digits, thats some 11 Bits resolution. Maybe I misinterpreted your original post, but to me it looked as if you wanted to say that the old analog computer were better than todays digital ones. While they were masterpieces of engeneering they're still outdated by todays standards and the reason is *not* that we're to lazy to mount some balls into cams.

 

First of all, I'm talking about USN mechanical fire control computers, not a Dreyer system.

 

Second, there's no such thing as half a bit or half a digit. Maybe you only need to use a fraction of the values available to you with a given number of bits or digits, but you have to allocate all of the space.

 

Finally, I never said mechanical systems were "better". I was commenting on their ingenuity as an implementation. Even though they were called "computers", in a sense they were more like an abacus -- concrete models of the problems input into them. Better than an abacus, they were automatically updating models. When the gun mounts were under automatic control, they became very real extensions of the model. I find the ingenuity behind that type of solution both technically impressive and theoretically intriguing. That doesn't mean I think they were the last word in fire control.

 

You're right, that was a bad example. A better one for the difficulty to change the inner workings of an analog computer would be an either forgotten or (at the time of design) unknown dependency of a variable, something along y = a + bx+cx^2 instead of simply y = a + bx, or y = a + bx + cz. Things which might well creep up once the firing range increases.

 

If an analog computer is what you have, then you have to suck it up and make the modification.

[APF]

Digital computers are a different technology. And the ease of loading programs has nothing to do with the relative ease of analysing and programming updates. They don't make changes in modern fire control code like Web developers alter sites on the fly. They undergo so much formal analysis and review that the time it would take to load a solution, mechanical or digital, is a negligible part of the process.

 

Today. But at a time where an additional square root in an equation meant redesigning and replacing a ton of high-precision mechnical parts in *all* instances of the computer it meant a lot of hours of skilled labour in addition to the analysis and review. Even today theres a huge difference between changing a piece os SW and adding 'just' that single connection which went missing on the PCB, oh, and once you're on it add that capacitor over there as well for better noise resistance. On some 20 PCBs. BTDT.

 

First of all, I'm talking about USN mechanical fire control computers, not a Dreyer system.

 

Second, there's no such thing as half a bit or half a digit. Maybe you only need to use a fraction of the values available to you with a given number of bits or digits, but you have to allocate all of the space.

 

I used the Dreyer table because I found a value for it. If you've got an accuracy for the american FCS feel free to provide it. Though I don't think it'd have a higher resolution because range adjustments (those I've herad of) were usually in hundreds of yards so there's really no need to provide data in the tenth of yards frame.

 

The term "3 1/2 digits" is a technical term; those DVMs have a display which shows +/- 1999 values, hence 4000 possible values or 3 1/2 digits in adspeech.

 

Finally, I never said mechanical systems were "better". I was commenting on their ingenuity as an implementation. Even though they were called "computers", in a sense they were more like an abacus -- concrete models of the problems input into them. Better than an abacus, they were automatically updating models. When the gun mounts were under automatic control, they became very real extensions of the model. I find the ingenuity behind that type of solution both technically impressive and theoretically intriguing. That doesn't mean I think they were the last word in fire control. [...] If an analog computer is what you have, then you have to suck it up and make the modification.

 

Yes, the underlying work to be able to implement them was ingenious. I wonder wether they were build as modules or wether they were build to fit into the smallest space possible.

 

As to the differentiation between abacus and computer:I disagree. The difference between the Mk?? and todays computers is that we eventually figured out how to 'reprogram' them in a faster and more economic way than a complete rebuild, much like the difference between TTL logic and an FPGA. AFAIK the Mk?? was a fully automized system which was able to compute a firing solution from external sensors without additional operator input 1), much like todays systems, and I wouldn't anchor a term like 'abakus' on the implementation.

 

Greetings

 

1) Within the limits of its time, i.e. the magazine temperature had to be set by hand, ditto air pressure and -temperature etc. But that's just superficial and a designers choice of that time: with 1000+ men running around someone was bound to have the time to read a thermometer and turn a dial.