Monday, April 29, 2019

Torpedo Threat - Is It Really?

Today’s naval analysts and observers tend to think that torpedoes are wonder weapons that can’t miss, can’t be avoided, and sink any ship with a single hit.  Well, we’ve already disproven the one-hit, one-kill idea and the back breaking myth along with it.  Now, what about the can’t miss, can’t be evaded belief?  Is that really true?  Well, unfortunately, we have no data to work with but when has that ever stopped us?  Let’s see what we can logically surmise.

Why are torpedoes considered can’t-miss weapons?  It’s because of two attributes of modern torpedoes:

  • Speed – Modern torpedoes have sprint speeds greater than ships have.  Thus, it’s not possible to outrun one.  It is possible to outlast one if the detection occurs early and the geometry-range is favorable but that’s unlikely unless the sub launches from the edge of the envelope.

  • Guidance – Modern torpedoes have self-contained sonar sensors and wake homing guidance.  Thus, unlike during WWII where simply turning parallel to the torpedo was generally sufficient to produce a miss, maneuvering to avoid a torpedo will be far less likely to succeed.

Or so the story goes …

The speed/range/geometry issue is straightforward but what about the guidance?  Does it really work?  As I said, there is no data to guide us (sorry, that was unintentional).  Commentators attribute near magical, perfect performance to torpedo guidance systems but are they really that good?  Consider …

Air-to-air missile guidance systems certainly aren’t perfect and, historically, have achieved something in the vicinity of 20% success rate in combat.  Why would we think torpedo guidance systems are so much better?

Laser guided bombs are around 80% effective under perfect conditions and as low as 50% in scenarios with adverse weather or less than perfect release geometry.  Why would we think torpedo guidance systems are so much better?

Surface-to-air AAW guided missiles have a historic success rate of 5%-20%.  Why would we think torpedo guidance systems are so much better?

While none of those systems use the same sensors and guidance systems as a torpedo, making direct comparisons invalid, we can note that every guidance system tried has proven to be far less effective than advertised.  There is no reason to believe that sonar and wake homing sensor/guidance systems have some kind of magic performance that no other guidance system has.  It is far more likely that sonars and wake homing suffer from the same poor performance that every other system does.

While we have no direct body of data to work with, we do have a few related bits of evidence that we can draw inferences from.

The US Navy anti-torpedo torpedo weapon system was a failure in actual use conditions.  The culprit was the sonar systems which produced so many false alarms as to render the system useless.  That being the case, why would we think that a torpedo with it’s small on-board sonar won’t be subject to the same kinds of false signals?

During the Falklands conflict, the Royal Navy’s submarine Conqueror fired three 21 inch Mk 8 mod 4 torpedoes (conventional, non-guided).  The sub carried modern Mk 24 Tigerfish homing torpedoes but doubted their reliability.  NavWeaps website offers some insight:

… in a test performed in 1982 immediately after the Falklands War, two out of five Mod 1 [ed. Mk 24 Tigerfish] torpedoes fired at a target hulk failed to function because of bad batteries and none of the others even hit the target. This unreliability was well known in the Fleet, which is why ancient Mark 8 torpedoes were used to sink the Argentine cruiser General Belgrano. (1)

This finding introduces yet another source of torpedo failure: mechanical/electrical.  Torpedoes are just like any other piece of machinery.  They have a mechanical/electrical failure rate.  We’ve seen Tomahawk missiles fail to launch or fail immediately upon launch.  We’ve seen Standard missiles explode during launch (rocket motor failure).  We’ve seen missiles drop off aircraft rails and never ignite.  And so on.  Why would we think torpedoes are mechanically/electrically any better?

Certain Death?

All we have is circumstantial evidence but it’s pretty convincing.  Torpedoes are nowhere near the inexorable killing machines that commentators make them out to be.  The reality is that torpedoes will simply fail to acquire targets, miss targets, fail mechanically/electrically, suffer from false signals, and generally fail to hit their targets to a large degree.  Throw in torpedo defenses such as acoustic decoys, Nixie-like tails, aggressive maneuvering by the target, etc. and the success rate of torpedoes will be even lower.

All of this analysis is not to say that torpedoes aren’t a serious threat.  They are.  A torpedo, if it can hit its target, is a powerful weapon. 

The conclusion we should be taking from this is that while the torpedo threat is serious, it does not preclude surface ships from operating and surviving during war, despite the many claims to the contrary. 

This also suggests that the Navy should be conducting extensive torpedo performance tests.  Have a sub fire live torpedoes – with the warheads removed, of course – at ships and see what actually happens.  Yes, we may get some dented hulls but the chance to gather actual performance data and develop real defensive tactics is priceless.

It’s peacetime.  Now is the time to find out what works and what doesn’t. 


(1)NavWeaps website, “Torpedoes of the United Kingdom/Britain”, Home / Weapons / Torpedoes / United Kingdom/Britain / Post-World War II,


  1. While testing torpedos, it would be interesting to test RBU like devices, see how they work as a torpedo defence.

    Mark 14

    1. Absolutely! We need to test the full range of offensive and defensive torpedo performance. I'd love to see the Navy adopt an RBU-ish weapon. It seems to have multiple uses.

  2. Threat or not, Anti torpedo systems are evolving with the German Seaspider. USN ATT and the Russian Paket. Hopefully they will be far more fruitful.

    1. As I've noted before, the US anti-torpedo torpedo (ATT) system also performed well in initial testing but failed badly in real world application. We'll have to see whether the other systems succeed in real world scenarios.

  3. That's a great post. Often probability of kill is talked about in terms of air to air missiles but I don't know that I've ever heard it talked about with torpedoes.

    With all the wonder weapons that we have, and with all the hype that surrounds them, I sometimes wonder what the hype was about the Mk. 14 prior to WWII.

    1. I don't know about the hype but Wiki states that not a single live fire test was conducted of the Mk14 prior to the outbreak of war. That's simply shocking. On the other hand, we do almost no live fire testing today, either!

    2. As Anonymous states below, they do live fire testing against hulks several times a year. Usually at the major, multi-nation exercises. The MK-48 is a killer.

      Another reason the Brits didn't want to use the Tigerfish was that the warhead was rather small. It was designed to go after subs and on surface target, destroy the props/rudder. Its replace, the Spearfish, is fast (designed to go after Alfa subs) and with a big warhead.

  4. The Navy has used the Mk 48 in numerous sink exercises, which seems to happens once a year, maybe more. While likely not a test under combat conditions, it is a live fire test of the complete system.

    And, the Navy and Air Force do periodically test the rocket motors for some of their missiles as a quality control measure. Does the Navy do the same with their torpedoes?

  5. The problem I see watching the videos of MK48 shoots, the target isn't doing anything. USS Racine for example isn't moving, no speed, not sure if there's a decoy or not, it's just a sitting duck AND IT STILL TOOK 1 HOUR TO SINK after being hit with no crew onboard! Are we that sure Mk48 and other torpedoes are really that good? Those test shots dont look that realistic to me.

    1. What about shots from an angle other than 90 degrees, the perfect shot? What about the confusing effect of a loud wake sound from a vessel sailing at maximum speed? What about a maneuvering ship so that the torpedo has to change directions while it travels? What about torpedo countermeasures. And on and on. You're quite right, the SinkEx'es are about as unrealistic as you can get.

  6. "...while the torpedo threat is serious, it does not preclude surface ships from operating and surviving during war, despite the many claims to the contrary."

    Which claims would those be? I get the impression that Navy pays little mind to the torpedo threat.

    1. Across the Internet, commentators routinely make statements that surface ships can't survive because of the submarine threat. Closer to home, those same opinions are fairly prevalent on this blog. This is why I've done a couple of posts, now, on torpedoes - to try and more realistically assess the threat posed by torpedoes.

  7. Submarines carry acoustic decoys.. are surface ships too loud for acoustic decoys to work?

    I think I'd have more faith in a decoy torpedo making a big loud wake and pumping out noise, than a little interceptor torpedo trying to hit an inbound torpedo.

    1. That's a very good question about why surface ships don't use torpedo decoys. Of course, they do use towed acoustic decoys (Nixie, for example) but not the expendable submarine type decoys. I don't know why they don't.

  8. As pointed out, at regular sink-exercises torpedoes clearly have the power to sink ships if they hit them a few times - especially the heavy weight torpedoes of large ships and submarines. However, even in ideal circumstances this takes *time* using conventional warheads, and passive or active damage control can make a ship survivable against several hits from conventional torpedoes. Add to that existing decoys and sensors - plus a competent helmsman (sorry USN) - and large torpedo attacks may be necessary to sink even a destroyer.

    Of course, we should consider the threat of nuclear torpedoes as well because the Soviets have an overt interest in them and you never know what Xi has planned. The energy of a 500 kg payload varies from 500 kg of TNT to nearly 1 megaton of TNT, while the shockwave from just 9 kT at a 500m depth largely disables ships 2000 feet away laterally. You can't settle for decoys against that, and that's on the light side.

    If Xi sets his mind to it and we don't develop a robust underwater kinetic intercept capability quickly enough, then a couple packs of PLAN submarines will one day be able to decisively engage multiple CVBGs. On the other hand, if we develop this capability and our enemies don't step up their submarine warfare, then we've almost completely countered their existing submarine capability. An underwater AEGIS could also protect amphibious forces from submarines and naval mines, a badly needed capability.

    Unfortunately, I think the physics of water, sonar, and relatively stealthy torpedoes presents a massive technical challenge compared to aerospace intercepts of hot, loud air frames, which CNOps rightly points out are difficult in real circumstances. In particular, intercepts at sufficient ranges to protect from nuclear torpedoes may never become economical compared to the long range heavy torpedoes they must counter. A hedgehog to blow up underwater grid squares *near your ship* is relatively easy, adding to CNOps' point about conventional torpedoes.

    1. "at regular sink-exercises torpedoes clearly have the power to sink ships"

      You're aware, I take it, that in typical SinkEx'es the target ship is left with all watertight doors and hatches open to facilitate a clean and predictable sinking of the ship. No one wants a stubborn target that refuses to sink drifting into commercial traffic lanes! Also, by the time a torpedo hits a target in a SinkEx, the target has usually been hit by several to dozens of other, lighter weapons. Typically, the lightest weapons are fired first and then launches proceed sequentially with each weapon more powerful than the previous. It makes sense - you don't want the first weapon fired to immediately sink the target and deny all the other assets their shots.

      So, between no damage control measures, open compartments, and numerous previous weapon impacts, trying to assess the lethality of a torpedo in a SinkEx is a pretty much pointless effort. It's as unrealistic as you can get.

    2. "A hedgehog to blow up underwater grid squares *near your ship*"

      I love the idea of the Soviet RBU. It would seem to have utility as an area anti-torpedo weapon and, in fact, is claimed as such by the Russians. Whether it's actually effective or not is an open question. For a variety of reasons, I'd love to see the Navy develop an RBU type weapon.

    3. "You're aware, I take it, that in typical SinkEx [is]... as unrealistic as you can get"

      Yes, I wasn't explicit, but that's part of what I was referring to here:

      "even in ideal circumstances this takes *time* using conventional warheads, and passive or active damage control can make a ship survivable against several hits from conventional torpedoes."

      My point was never to disagree with the post's premise - in fact, I wholly endorse it. My point was that the scope of the torpedo threat is a lot broader (nuclear torpedoes), and the definition of success is as well (losing billion dollar ships is unacceptable even if it takes a couple dozen torpedoes).

      "I love the idea of the Soviet RBU."

      I can tell, you've said so in response to every mention of it here so far! I like it too, and I absolutely agree that it's a necessary capability, but I see it as the CIWS of the underwater AEGIS; the last hard kill countermeasure before damage control becomes the name of the game.

      An RBU doesn't have the range to intercept a nuclear torpedo at a safe distance, or to follow-up failed engagements with additional shots (counting a full volley as a shot), the latter of which is a serious concern even with conventional torpedoes.

      RBU-1000 has (reported) max range of 1 km, reaction time of 2 minutes from initial detection, 1 minute if some targeting data is already available, and reload of "less than three minutes". Mark 48 travels (reportedly) at up to 102 km/h, and we expect it (and other well designed torpedoes) to sprint the last 1000m of an engagement. After the explosions stop and your sonar operator gets some idea of what's going on underwater again, you've got 35 seconds left - assuming no actual downtime in detection capability caused by blowing stuff up *near your ship* underwater.

      What about using multiple RBUs? Sure, now you can take 2 shots against 1 torpedo (or *dense* swarm) with several large deck-mounted systems. I'll admit, it definitely beats taking a hit if you're a thin-skinned ship like Burke, but a longer range option is mandatory... if it's feasible.

      If it's not, then I refer back to CNOps' adage about ASW; don't play tag with submarines, attack their bases.

    4. "RBU … I see it as the CIWS of the underwater AEGIS"

      Exactly right. I would hope that if we were to develop an RBU-ish weapon, we'd give it a few improvements over the original Soviet design: proximity fuzing, very short range simplistic guidance??? (may be a reach too far considering this should be a basic, cheap weapon), faster reload, integrated fire control, larger pattern, etc.

    5. "proximity fusing, [terminal homing], faster reload, integrated fire control, [larger lethal area, and cheap]"

      I like it, but we both know complex fuses make the shots cost more, guidance systems (likely) unacceptably so, and the other features require (hopefully slightly) more complex, capable, and expensive launchers. If we accept the costs, it may still be viable or even necessary.

      The mortar array is a great launch system for firing a wide spread in a mostly set pattern. Some upgrades to its rate of fire are possible, as are the other proposed launcher improvements, but I'm not against bristling our hulls with dozens of simpler mortar arrays if that's more effective. Deck space is at a premium, but these could be mounted with little to no footprint (except RCS!!) at the ship's extremities and/or integrated in the hull (low RCS, slightly more $$).

      I don't bash torpedo guidance in general, ships various emissions are *usually* a lot louder than background, making the problem technically easy. All of that changes when the depth charges go off...

      After thinking about the problems of torpedo guidance, I've concluded this effect - if sustained - is probably more valuable than a depth charge's pK to hard kill, and the majority of our decoys and other soft kill countermeasures combined.

      Unlike missiles, heavy torpedoes actually have options in the terminal phase. Once (the guidance system assumes) the torpedo is directly under a ship, it can safely loiter for an extended period and use its sensors to discriminate real ships from decoys before detonating. If it misses a ship or decoy in a terminal maneuver, it has the guidance, hydrodynamics, and propulsion to maneuver to the target's predicted position before that prediction is useless. Both capabilities rely on endurance, which heavy torpedoes have in spades unless they're fired at the end of their envelope. Without something (exploding) to obscure their sonar and wake-homing, they get multiple attempts at an easy problem.

  9. If two ships travel in opposite directions in a circle what does that do to wake homing torps?

    1. I, too, have wondered about simple countermeasures like circling back across one's own wake. That would seem like it would completely confuse the torpedo. I doubt that wake homing is fine enough to distinguish multiple, crossing wakes. Yet more reason to conduct realistic testing which the Navy refuses to do.

    2. How to program a control system to defeat this tactic:

      1. take some aerial photos of ships in a variety of single-ship and multi-ship wake-crossing maneuvers.

      2. generate wake-homing sensor data that mimics these maneuvers (or substitute actual data collection for step 1, both are easy enough)

      3. Have a human annotate this data with the actual position and heading of the ships in each data set.

      4. Provide the data to a machine learning algorithm, or just manually analyze the patterns to develop "dumb" algorithms to deconvolute the wakes.

      5. Analyze the algorithms, try to break them (design countermeasures), make a few changes to counter the obvious countermeasures, then stick it in a torpedo.

      Of course, it's not exactly *easy* to actually do, but the problem is obvious enough and the solution simple enough that I strongly suspect that it has been done.

  10. A 21-in torpedo going 40 knots is a pretty small target to hit with a depth charge, which is basically what the Russian RBU delivers. The Russian RBU-6000 can fire a salvo of 12 rounds. To be effective against a torpedo, I would imagine having to fire many dozens to provide a good chance of hitting a torpedo.

    It would be interesting if swarming technology could be applied to anti-torpedo weapons. The idea would be to drop groups of RBU-like charges along the path of an incoming torpedo with each group positioning itself to provide the best chance of hitting the torpedo.

    One possible countermeasure to this would be for the incoming torpedo to approach the target ship from a much lower depth and strike the ship from below or rise a short distance from the ship and strike it amidship.

    1. You may be thinking of this a bit wrong. Think of RBU charges less like a depth charge and more like a carefully designed pattern of proximity fuzed charges. The ship's sonar (or multiple sonars in an ASW networked defense!) will provide the firing solution and the charges are dropped in a predictive pattern, boxing in the torpedo along its predicted path. Proximity fuzes eliminate the need for direct hits and the pattern ensures a good chance of a significant effect.

      Even if the torpedo isn't out and out destroyed, it may be sufficient to simply deflect the torpedo from its course/depth enough to break its target lock. A torpedo sensor has a relatively small field of view. Once deflected, it may well be unable to re-establish course/depth and reacquire its target.

      Viewed this way, the odds of success seem reasonable. Of course, that's just speculation on my part. Realistic testing would have to be conducted but at least it seems plausible.

    2. The problem with the anti torpedo system was the false alarm rate making it unusable. Using underwater explosive charges makes sense on the disabling side but does nothing to solve the problem which is detection and identification.
      I think that problem needs to be solved before any reliable torpedo defense is viable. There has been some scientific work published on samarium doped ferroelectric crystals which have a log magnitude improvement in signal noise ratios over current sonar piezoelectric crystals. You will laugh at this but it is work being carried out in the USA and funded by both the Department of defense as well as the republic of China! M
      Just unbeliavable.

      The problem is that they are large, expensive heavy and non disposable. I envisage relocatable sonar buoys being placed around carrier battle groups and being moved by helicopters to new locations as the carrier group moves. In otherwords create a stationary sonar net array around the force well away from the noise of the ships. Network the sonar arrays so an accurate sonar map is produced so that subs and torpedos can be detected at much larger distances so that the ships simply turn away in another direction and decoys are deployed either by the buoys themselves or by the helos. It takes a torpedo about 8 minutes to travel 10km.- plenty of time yo deal with it.

    3. "I envisage relocatable sonar buoys being placed around carrier battle groups and being moved by helicopters to new locations as the carrier group moves."

      We had this. It was called the S-3 Viking. It was used to lay chevrons of sonobuoys out ahead and around the carrier group.

    4. For those that are interested the group was from penn state and published in science daily march 2018.

    5. I am talking about a future technology using much larger non disposable sonar buoys. The sensitive crystals that are up to 99% efficient weigh up to 10kg. The current devices used are no more than 62% efficient. The viking deployed disposable buoys and did not reteieve them.

    6. How many sensors and helos would you envision needing to screen a carrier group?

    7. Just wondering out loud, but could a "smart" torpedo avoid or reduce the probability of being affected by such charges by slightly zig-zagging as it approaches its target? Obviously it wouldn't maneuver to the point where it loses contact with its target.

    8. I have thought about this as well. A screen would consist of buoys up to 40 nautical miles in front and 10 nautical miles behind and 20 nautical miles to each side for a total of 16 bouys in a grid. Any torpedo or sub getting to within 30 nautical miles of the carrier would be alwaysc within 10 nautical miles of a buoy. I would use 1 asw helicopter to process the data transmitted and perform the kills. 4 auxillary helicopters would be used to continually move buoys from the rear echelon to the front echelon as the carrier group moves forward. The array and buoys could weigh up to 2 tonnes each with sensor arrays up to 300 metres below the surface. The floating buoy would have a hook for deployment Communications technology as well as passive sensors to detect anti ship cruise missiles entering the networked area. The technology exists that the buoys would be both active and pasive sonar. They could also be configured to transmit targeting data to ssk an ssn using an acoustic link. Think of this as a very large and sensitive acoustic array that is well away from the ships surrounding the carrier. By being stationary, they are far more likely to detect the noise of a sub or torpedo. If a torpedo was detected it would give a 30 minute warning so the direction of the carrier could be changed to reduce the chance of intercept. I actually think a new cruiser needs to be developed and built that has enough deck and hangar space to operate the asw group efficiently at the speed of a cvn. I would give the same cruiser the role of ballistic missile defense and put a very good radar system on it. In other words it would be a ship that concentrates on only the defense of the carrier group.

      I think the carrier group without beefed up asw defense is an extremely vulnerable and costly asset that needs to be properly defended. I am a little shocked by the lack of asw development sibce the end of the cold war and see modern submarines having a feild day with the current cvn.

  11. The depth charges do not necessarily need to hit the torpedo. If set to explode between the torpedo and the vessel it would create a wall of disturbed water through which the torpedoes sensors could not penetrate for some time, giving the vessel a chance to manoeuvre or perhaps shut down its engines. To avoid German wake homing torpedoes in WW2 RN escorts would reduce speed to less the 7 knots. Below this speed the torpedo sensors could not detect the wake. Obviously technology moves forward but the principle remains.

    1. I wonder if an advanced version of the British Limbo depth charge system might work:

    2. Conceptually, that would work although you'd probably want a much bigger salvo capacity but, yes, a modernized version would be along the lines of what we're looking for.

  12. An aircraft is small, fast, and moves in three dimensions,
    If it dives, fires its after burners, and turns hard to the right, its in a very different place than if it climbs, air brakes, and turns left.

    A ship is big, slow and moves in two dimensions, flank speed hard right and all reverse hard left, take quite some time to move far apart.

    If you break an AAMs target lock 5 seconds before impact, it will most likely miss. If you blind a torpedo 5 second away, you arent going to get away, you are still where you were, unless it tries something clever.

    Belgrano was attacked by unguided torpedo, and destroyed, it was taken entirely unawares, and was unable to even signal it was under attack, or had been attacked, never mind activate active defences. One of its escorts was hit by a torpedo that failed to detonate, and it didn't realise.

    "This also suggests that the Navy should be conducting extensive torpedo performance tests. Have a sub fire live torpedoes – with the warheads removed, of course – at ships and see what actually happens. "

    That is perhaps the greatest idea I've ever heard, a blank torpedo would have no way of damaging a ship.
    Blank AAMS would still cause catastrophic damage to aircraft, blank AShMs would cause fires and superstructure damage, but a blank torp would pose no real threat.

  13. Much of the power of torpedos in many ways is psychological. A commercial ship like a tanker getting torpedoed even if doesn’t sink—in fact it’s doula single hit to a double hulled tanker—would cause insane amounts of over reaction.

    This is where I think a UUV could be useful. Imagine a large one being carried DSRV style on the back of a Virginia. A hundred miles off shore it launches the UUV. The UUV has surprisingly simple programming. It simply goes to a predetermined point and begins swimming a pattern in front of ( or maybe in) a large harbor, say Shang Hai. It has a simple passive sonar that listens only for really big ship sounds at relatively close range. When it gets a hit it launches its payload, two lightweight torpedoes which immediately go active and attacks the first thing it pings. The UUV turns off with a timer to turn back on and return to Pearl or another mothership.
    Basically it’s a mine that mimics a sub.
    Now you have temporarily bottled up a harbor as they sweep the hell out of it looking for a sub, and that Virginia is hundreds of miles away shadowing a boomer already.

    1. We already had that, essentially. It was the Mk67 Submarine Launched Mobile Mine (SLMM). It was, basically, a smart fuzed Mk 37 torpedo that would act as a mine and was launched from a sub, traveled to its destination, and then lay in wait for a target meeting the fuze specifications.

      A sub could, in theory, launch dozens of them and seal a harbor or navigational choke point while standing safely off.

      There are conflicting reports whether the SLMM remains in active service. There was, at one time, a program to convert the current Mk48 torpedo into SLMMs but I believe that was cancelled.

  14. Given the likely failure of a single torpedo to be fatal, I believe we need to deploy the same techniques we do with AShM, massed volleys. To do this I believe development of a VLS for torpedoes should be developed. It would basically be a variant of current SSBN or SSGN designs, but with torpedoes instead of missiles. It might even be possible to develop a system that can be loaded with either.

    Randall Rapp

    1. I'm not quite sure what problem you'd be solving? Subs have multiple torpedo tubes and can/do launch volleys of torpedoes.

    2. For the US Navy, I don't see a purpose.

      For a foreign Navy it might make sense to have a small sub that could broadside a dozen or more torpedos at a high-value target. Sealed shore-loaded torpedo canisters might offer simplified engineering and lower cost than conventional reloadable torpedo tubes.

      In short, a cheap 'pepper-box' sub that can put a hurt on large enemy ships or large groups of enemy ships.


Comments will be moderated for posts older than 30 days in order to reduce spam.