9
\$\begingroup\$

In most cars and trucks, a thick cable connects the battery directly to the starter (solenoid). Why don't these circuits have have any kind of overcurrent protection?

It is an accepted practice throughout the motor vehicle industry. Here are two separate standards that specifically exempt starter circuits from overcurrent protection. (These are both for watercraft where the consequences are even greater; you can't walk away from a fire at sea.)

Each ungrounded output conductor from a storage battery must have a manually reset, trip-free circuit breaker or fuse, unless the output conductor is in the main power feed circuit from the battery to an engine cranking motor.

These are just examples; I have no doubt that organizations such as SAE and ABYC have similar provisions in their standards. Millions of vehicles are wired this way.

All circuits, except the main supply from the battery to the starter motor and electrically driven steering motors, should be provided with electrical protection against overload and short circuit, (i.e. fuses or circuit breakers should be installed).

I am asking for the engineering rationale behind these exemptions. Even though the cable is much thicker than the starter windings, a mechanical failure or impact could still create a short to ground. The resulting current can easily exceed 500A and is enough to weld thick steel.

I understand that the starter demands far more current than any other circuit, but surely a cost-effective solution could be found -- such as a fusible link. Or am I mistaken?

Here are some possible reasons that don't make sense to me:

  • The cable is so thick (relative to the size of the battery) that it doesn't need protection. The starter will burn or the battery will explode before the cable melts. While this may certainly be true from a "protect the wire" standpoint, I believe this an even stronger reason for overcurrent protection on the starter circuit... to protect the entire vehicle.

  • The risk of a fault in this circuit is extremely unlikely. It is true that starters are sturdy devices and thick cables have more mechanical strength. However failures are still possible, and do occur from time to time in the real world. Furthermore the impact of a failure in this circuit can be catastrophic, leading to total loss of the vehicle or death. Therefore I'd expect the severity of the problem to overwhelm the (admittedly) low likelihood in a failure mode analysis.

Edit for future readers: Most of the answers focus on availability. That is very important, but a secondary reason coupled to the choice of a fuse as the protective device. A breaker would mitigate the risk of being stranded due to a nuisance fault. (Someone mentioned the possible loss of steering, but all production vehicles including the Infiniti Q50 still have mechanical backup.) Fortunately there is a concise answer that explains why even a breaker or fusible link wouldn't be suitable.

\$\endgroup\$
  • 4
    \$\begingroup\$ Your "seat of the pants" fault analysis leaves a lot to be desired. Through long experience, the sum of the cost * probability for all of the different failure scenarios (including the cost of false positives) has been determined to be such that additional protection is not necessary or even desirable. Sure, you can imagine a "worst case scenario", but you also need to understand just how unlikely that scenario is. \$\endgroup\$ – Dave Tweed Feb 13 '18 at 17:53
  • \$\begingroup\$ Your short circuit current of 500A is way out - a car starter motor can take 300 to 600A and Heavy Truck starters 800 to 1000A, so the short circuit current ... Well I have used a 12V battery and thick cables and an alternator brush (from a Lucas 16ACR alternator) to do spot welding of thin 1mm steel plate... Also, ALL the cables involved in the starter circuit have to be heavy, not just the one you mention : some solenoids are separate... \$\endgroup\$ – Solar Mike Feb 13 '18 at 19:04
  • \$\begingroup\$ I have seen thermal fuse like elements in a lot of starter relays \$\endgroup\$ – PlasmaHH Feb 13 '18 at 20:48
  • 4
    \$\begingroup\$ Would a fault even cause more current than the normal starting current? \$\endgroup\$ – τεκ Feb 13 '18 at 22:21
  • 1
    \$\begingroup\$ That is a slow reacting fuse, I think. And blows later than starter run time at starter current. Sounds like just the right solution. But it may not work in slightly different situations: different total cable length, different cable diameter, different starter, and more. \$\endgroup\$ – Volker Siegel Feb 14 '18 at 2:56
17
\$\begingroup\$

There's only any point in trying to protect a circuit with a fuse if there is enough headroom between the operating current, and the fault current, to ensure that the fuse won't blow in normal operation, and will blow under fault conditions.

Unfortunately, once you have included all the tolerances, there is no current level you can choose that is guaranteed to avoid nuisance tripping and still have a reasonable chance of working, should the starter motor be locked up for instance. A conventional starter motor is series wound, in order to pull a very high current at start-up.

\$\endgroup\$
16
\$\begingroup\$

Having a power fuse blow on the circuit that is going to the engine starter can result in a serious safety issue, especially in marine systems where the inability to start the engine can mean limited or no navigation capabilities. On cars, it can mean not being able to start in an emergency situation. Some cars used to have it. I once rear ended an old Toyota in the middle of a freeway curve that had one and the woman driver blew that fuse trying to re-start her car after it stalled, which killed her entire electrical system including her hazard lights, headlights etc. When my wife wanted to buy a Toyota years later, I asked if they still had that, they assured me that they did away with it because it was a safety liability for them.

\$\endgroup\$
  • \$\begingroup\$ I'm glad you mentioned liability; that's an important factor. However I'm not convinced that the need to start in any situation is truly the reason. (I agree that it's an important consideration in the context of nuisance faults.) After an ACTUAL fault, the starter might miraculously still work but I question whether testing it again would ever be in the safety interest of the operator. \$\endgroup\$ – csanders Feb 16 '18 at 6:40
6
\$\begingroup\$

I worked in automotive electronics for a number of years. From a safety point of view, you are not trying to "protect the vehicle", you are trying to "protect the person". Usually the person will be the driver or occupant, but sometimes the person may be a pedestrian (think collisions and crumple zones), a bystander (avoiding fuel tank explosions) or even emergency crews (electric vehicles need to not expose high voltages which could kill rescuers).

Most safety problems do not have a universally good solution though. What you get is a trade-off, and you choose the least worst option. There exist formal methods such as FMEA or FTA which enable you to quantify this, so that if things go wrong then you can prove you've followed best practise.

The most common problem you have is the trade-off of fault-tolerance versus availability. If your first reaction to any problem is to stop the car and force the driver to call out a recovery truck, this may initially seem like a good idea and the safest solution. I worked on a hybrid electric vehicle for Ford, where the software did indeed take this approach when it found itself in an uncertain situation, because the result could have been uncommanded vehicle movement. With the legal situation in the US, this seemed like the best approach.

During development, we talked to engineers from Volvo and found that they had a completely different approach though. The Volvo safety case is that unless you can prove that it is categorically unsafe for the vehicle to continue moving, you must not stop the vehicle. You may reduce the speed that it runs at, or reduce the power available, but the vehicle must not stop. Why? Because if your car cuts out in a Scandanavian winter, you've got about 2 hours before you freeze to death. Volvo's safety case said that it was preferable to have some risk of low-speed, low-power collisions if the vehicle develops a fault, and to accept that this will happen more frequently with mostly-non-fatal consequences, compared to the less-frequent risk of the vehicle cutting out in definitely-fatal situations. In addition, the driver can still take some action to mitigate the fault by turning off the ignition, even if the car is no longer responding properly to electronic control of accelerator, brakes and gear selection.

A fuse in the starter circuit follows the same rationale. What's the worst-case scenario without a fuse? Answer: the wiring or battery overheats and you get a fire in the engine bay which spreads to the rest of the car. Usually the battery voltage will drop before things get too bad. The driver can mitigate that scenario by not cranking continuously for extended periods. They can also mitigate that fault because there is a fire-resistant bulkhead between the engine and the passenger compartment, giving them ample time to escape the vehicle. Even in the case that central locking fails locked and powered windows fail closed, an emergency escape hammer makes it trivial to break windows and escape. (You do have one, right? If not, buy one - they're cheap.) The situation has defence in depth.

What's the worst-case scenario with a fuse though? Well, you can't start the engine when you need to. In a car, that could leave you stranded - and we're into the "freezing to death in 2 hours" scenario. On a boat, you're utterly stuck.

\$\endgroup\$
  • \$\begingroup\$ Thanks for this thoughtful response, it adds a lot of value. As I mentioned in my comment to J. Raefield though, availability doesn't seem like a strong enough argument to me. The relays that power the starter solenoid are protected by fuses themselves. If it's so imperative to start the engine I'd expect them to also be exempt, or redundant. My other problem with the availability argument is that it assumes the starter will still work after an actual fault. (That would have triggered a protective device.) Because normal starting currents are so high, any such fault is going to be a whopper. \$\endgroup\$ – csanders Feb 16 '18 at 7:21
3
\$\begingroup\$

Cars have warnings not to use starter more than x seconds and wait xx seconds to prevent battery over heating and gassing. As such the CCA rating must exceed the load so with adequate battery and cable design, a short is not possible unless human error ( drop wrench across battery terminals or ignoring safety warning on > 1 minute sustained starter useage under heavy load .) In this situation the starter motor temp will be high and then increase resitance and the battery will drain quickly but not cause a fire. If a battery fails with internal short, it will self discharge and risk of fire from sparks near H2 is possible but relatively low risk

Fuses are not suitable since they add ESR , reduce CCA capacity which is a loss of starting performance in cold weather. You would rather have your motor powered steering overheat than have the fuse blow when driving. So in short ( pun intended) fuses reduce reliability and safety.

E-cars are different and with denrites risking internal shorts so fusing is mandatory with temp sensing.

\$\endgroup\$
1
\$\begingroup\$

Cars and trucks have firewalls and the relatively safe place to stay in if the circuit protection prevents a start and not everyone lives in Scandinavia where freezing to death in an unmovable car is a concern.

There are circuit breaker switches can can easily be located in the space afforded by a car or truck that can be reset without calling for service. If the starter solenoid shorts closed and the starter remains engaged and running, unless the driver can hear (unlikely) the starter running (unlike a second engagement after the engine starts) the battery drains quickly and the vehicle stops running anyway. Therefore the argument that circuit protects prevents the vehicle from being started doesn't add up, once the battery discharges below a certain point the vehicle will stop running anyway.

Heavy gauge wire does not preclude an electrical fire, the wire doesn't burn but the accessories connected to them can and will. Take your typical accessory and look at the connector and then compare that to a 4 awg or 0 ga cable. It's not the cable that will catch fire.

Circuit protection can prevent the fire and also reduce damage to accessories. A starter stuck to run because a solenoid fused closed isn't going to last long trying to keep up with the engine running at 2500-3000 rpms or more. Far worse damage can happen besides a damaged starter. While.cirxuit protection is for safety, the additional benefit of accessory protection shouldn't be discounted in a narrow view of design.

\$\endgroup\$
0
\$\begingroup\$

It's an interesting thought. The cranking Amps specification of a battery is at 7.2V after 30 seconds, so the short circuit current can easily be over 1000A, but this current drops down rapidly as the seconds go by. I speculate that wires going to the starter may actually be too thick to be able to get hot enough to start a fire with a standard battery. Also, the wires are in the engine compartment and are not going by anything else that could melt. If the starter relay sticks engaged, by the time the starter motor over heats and shorts out, the battery's original short circuit current will be well below what could cause harm.

There have been other good answers such as cost to benefit, and the inability to have a fuse that would be small enough to protect the circuit and not burn out during longer cranking.

\$\endgroup\$
  • 1
    \$\begingroup\$ Googling for "car battery fire" yields plenty of results which disagree. \$\endgroup\$ – Dmitry Grigoryev Feb 14 '18 at 9:15
  • \$\begingroup\$ The results are filled with pictures of cars on fire and electric car fire incidents. Besides, it's not the battery that would catch on fire. \$\endgroup\$ – Alex Cannon Feb 15 '18 at 1:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.