# A Modern Submarine Descends Below Crush Depth-What Happens? **URL:** \ **Category:** Factual Questions\ **Created:** [September 5, 2015, 12:28pm UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002 "2015-09-05T12:28:00Z")\ **Posts on this page:** 7\ **Page:** 3
**Author:** ![Ex\_Bubblehead](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/ex_bubblehead/32/3397_2.png) [@Ex\_Bubblehead](https://boards.straightdope.com/u/Ex_Bubblehead)\ **Post date:** [September 8, 2015, 1:31pm UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002/41 "2015-09-08T13:31:16Z")
> [@Xema](#): > > This seems a bit strange: If a human can consistently do better than an autopilot, the obvious step is to carefully record and examine exactly what the human is doing, then design a control system that closely matches this. If it were that easy it would have been done already. It isn’t. ---
**Author:** ![Habeed](https://avatars.discourse-cdn.com/v4/letter/h/a587f6/32.png) [@Habeed](https://boards.straightdope.com/u/Habeed)\ **Post date:** [September 8, 2015, 1:32pm UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002/42 "2015-09-08T13:32:12Z")
> [@Xema](#): > > This seems a bit strange: If a human can consistently do better than an autopilot, the obvious step is to carefully record and examine exactly what the human is doing, then design a control system that closely matches this. Exactly, that’s what I thought. Keep in mind, Bubblehead was on a series of submarines where the newest was commissioned 1975 for the last sub in the series, per wikipedia. The control solutions you had available would have been something like a slow digital machine with limited memory, or some analog controllers that would have been limited in how complex you could make them because more analog stages mean more internal circuitry noise. Also, analog PID has the problem that the coefficients are hard to change on the fly. With that said, various solutions do come to mind that would have worked with equipment of that vintage : another issue is that defense contractors aren’t very smart or creative for institutional reasons. (the engineers working for them are but they would have been hamstrung by red tape) Another issue is that the equipment cabinet all this stuff has to fit in is huge with that vintage of controls, and a bit of leaked seawater or condensation might cause it to fail. Humans don’t fail if a trickle of seawater or a few drops of condensation get on them. We can do far more with a chip the size of your fingernail today, and it emits so little waste heat you can seal it all hermetically so not a drop gets on it even if the sub itself floods. ---
**Author:** ![Elendil\_s\_Heir](https://avatars.discourse-cdn.com/v4/letter/e/7cd45c/32.png) [@Elendil\_s\_Heir](https://boards.straightdope.com/u/Elendil_s_Heir)\ **Post date:** [September 8, 2015, 2:03pm UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002/43 "2015-09-08T14:03:52Z")
> [@Fear\_Itself](#): > > There were some pretty good answers [the last time you asked this question](http://boards.straightdope.com/sdmb/showthread.php?t=553491). :dubious: Heh. I think you scared him away. ---
**Author:** ![robby](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/robby/32/11048_2.png) [@robby](https://boards.straightdope.com/u/robby)\ **Post date:** [September 8, 2015, 4:47pm UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002/44 "2015-09-08T16:47:45Z")
> [@Fear\_Itself](#): > > There were some pretty good answers [the last time you asked this question](http://boards.straightdope.com/sdmb/showthread.php?t=553491). :dubious: Exactly. There was a very similar discussion in that thread. > [@Habeed](#): > > …Yes, but _why_. The “ex bubblehead” I personally talked to thought it was because the shielding on the reactor itself was inadequate. That the meters of clear space around the reactor itself cuts the dose down by a factor of 8 or so due to 1/r^2 law, then the bulkheads and doors cut the dose down another 2 or 4 or so (depends on their thickness and composition), so it’s only about 1/32 the dose standing in the areas the crew are allowed to stand, even if they are taking a smoke break leaning against the reactor compartment bulkhead. All of this may be true, but it does not mean that the reactor shielding is “inadequate.” The reactor compartment is a relatively small compartment, and is not designed to be accessible while the reactor is operational. It would add a great deal of additional weight and complexity to make the reactor compartment safely accessible during reactor operation, and it’s simply not necessary. The “why” question I would ask you is why you think the reactor compartment needs to be accessible during reactor operation? For example, part of the secondary shielding consists of the main emergency diesel generator fuel storage tank, which is placed just forward of the reactor compartment bulkhead. It’s designed to attenuate neutrons, which are primarily produced when the reactor is operational. When the reactor is shut down, the neutron flux is greatly reduced, and neutron attenuation is less of a concern. Also, while the reactor core itself (where the fission process takes place) is heavily shielded, the primary coolant loops which exit the reactor core are not. These loops contain short-lived radionuclides which contribute to making the reactor compartment a high-radiation environment while the reactor is operational, but safe to enter when the reactor is shut down. > [@Habeed](#): > > Exactly, that’s what I thought. Keep in mind, Bubblehead was on a series of submarines where the newest was commissioned 1975 for the last sub in the series, per wikipedia. > > The control solutions you had available would have been something like a slow digital machine with limited memory, or some analog controllers that would have been limited in how complex you could make them because more analog stages mean more internal circuitry noise. Also, analog PID has the problem that the coefficients are hard to change on the fly. > > With that said, various solutions do come to mind that would have worked with equipment of that vintage : another issue is that defense contractors aren’t very smart or creative for institutional reasons. (the engineers working for them are but they would have been hamstrung by red tape) > > Another issue is that the equipment cabinet all this stuff has to fit in is huge with that vintage of controls, and a bit of leaked seawater or condensation might cause it to fail. Humans don’t fail if a trickle of seawater or a few drops of condensation get on them. > > We can do far more with a chip the size of your fingernail today, and it emits so little waste heat you can seal it all hermetically so not a drop gets on it even if the sub itself floods. Part of the “red tape” you disparage is to ensure that the electronics are reliable in all foreseeable conditions, including battle situations. That being said, there have been some trends toward more automation in recent decades. However, the submarine force still tends toward far more more manual control than you might expect. Manual operation means there is one less thing to break in a casualty or combat situation, and the numerous personnel on board provide lots of people for damage control, as well as preventative and corrective maintenance while underway. ---
**Author:** ![Dr.Strangelove](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dr.strangelove/32/6613_2.png) [@Dr.Strangelove](https://boards.straightdope.com/u/Dr.Strangelove)\ **Post date:** [September 8, 2015, 9:03pm UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002/45 "2015-09-08T21:03:12Z")
> [@Ex\_Bubblehead](#): > > It was that bad. No amount of programming can match the light touch of a trained and experienced sailor on the controls. Perhaps that’s been overcome with the newest boats. I wouldn’t know, I served on the Skipjack, Permit, and Sturgeon classes during the Cold War. I definitely believe it; I just mean that it sounds more like a design problem than anything. The famous PID controller was largely developed for an automatic steering system for the US Navy. I’ll quote the Wiki page: > [@](#): > > One of the earliest examples of a PID-type controller was developed by Elmer Sperry in 1911,[5] while the first published theoretical analysis of a PID controller was by Russian American engineer Nicolas Minorsky, (Minorsky 1922). Minorsky was designing automatic steering systems for the US Navy, and based his analysis on observations of a helmsman, noting the helmsman controlled the ship based not only on the current error, but also on past error as well as the current rate of change;[6] this was then made mathematical by Minorsky. > [@](#): > > Trials were carried out on the USS New Mexico, with the controller controlling the angular velocity (not angle) of the rudder. PI control yielded sustained yaw (angular error) of ±2°. Adding the D element yielded a yaw error of ±1/6°, better than most helmsmen could achieve.[8] > > The Navy ultimately did not adopt the system, due to resistance by personnel. Similar work was carried out and published by several others in the 1930s. So it’s definitely possible, even way back the 20’s when the “computer” was entirely mechanical (or pneumatic, hydraulic, etc.). Which isn’t to say that it makes sense on a sub or that the Skipjack system was anywhere close to as good as it should have been. ---
**Author:** ![Habeed](https://avatars.discourse-cdn.com/v4/letter/h/a587f6/32.png) [@Habeed](https://boards.straightdope.com/u/Habeed)\ **Post date:** [September 9, 2015, 12:40am UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002/46 "2015-09-09T00:40:49Z")
> [@Dr.Strangelove](#): > > I definitely believe it; I just mean that it sounds more like a design problem than anything. > > The famous PID controller was largely developed for an automatic steering system for the US Navy. I’ll quote the Wiki page: > So it’s definitely possible, even way back the 20’s when the “computer” was entirely mechanical (or pneumatic, hydraulic, etc.). Which isn’t to say that it makes sense on a sub or that the Skipjack system was anywhere close to as good as it should have been. Yeah. One thing I thought of when it came up was a PID controller using analog circuits (vacuum tubes or transistors) and a slow digital computer that would analyze the data and auto-adjust the PID coefficients.(it would either switch in and out banks of resistors or it might turn a pot with a stepper motor) That way you get near instantaneous response (analog circuits have very little propagation delay) and the benefits of digital analysis. That might have worked, dunno if the computers you could cram into a sub in 1970 could hold enough memory to do this. ---
**Author:** ![Xema](https://avatars.discourse-cdn.com/v4/letter/x/9de053/32.png) [@Xema](https://boards.straightdope.com/u/Xema)\ **Post date:** [September 9, 2015, 2:30am UTC](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002/47 "2015-09-09T02:30:36Z")
FWIW, Googling “submarine autopilot” produces more than enough links (e.g. [this one](http://www.raytheon-anschuetz.com/fileadmin/content/Downloads/Brochures/submarine-3d-steering-control.pdf)) to suggest that this is now mainstream technology. [Previous page](https://boards.straightdope.com/t/a-modern-submarine-descends-below-crush-depth-what-happens/730002.md?page=2)