I design new equipment for manufacturers. For the past 40 years, I have designed products that consume significant current. These products are used on high voltage power mains and low voltage systems. I also designed (and was responsible for production of) automotive gauges and test equipment, including instrumentation (meters) in products marketed by national firms. You can skip all the technical stuff in the beginning, and just move on to battery wiring or to trunk mounted battery.
Conductors and Resistance (background)
All conductors have some resistance to electrical current flow. Resistance in conductors and connections is actually the ONLY thing that limits current flow through wires. Conductor resistance causes voltage drop along the wire's length, and the current through that resistance, and the resulting voltage drop across that resistance, causes heat. Resistance, voltage drop, heating, and current are all closely tied together with unbreakable electrical rules. (There is something else that limits current, reactance. Reactance is like resistance, but reactance limits current without heating. Reactance only applies to rapidly changing currents. Steady currents ware not affected by reactances.)
To keep things simple, let's consider steady slowly-changing loads like sensors, starters, lights, and other slowly-changing or steady dc loads. In the context of steady or slowly-changing direct current systems, we can ignore reactances. (We cannot always ignore reactances when dealing with sharp pulses or rapidly changing voltages, such as pulsed fuel injectors or ignition system pulses.)
This means in DC systems such as automotive wiring, or in normal home powerline frequency AC circuits, we have two things to worry about:
DC resistance is directly proportional to conductor length for a given conductor cross sectional area (conductor size) and material, and inversely proportional to conductor cross sectional area for a given length and material. A ten-foot long #14 gauge copper wire, with a certain current, will drop ten times the voltage as an identical one-foot long wire. Voltage drop is always the load current times the path resistance. The longer a given size and material path, the more voltage loss.
People like to tell us electricity is like water in a pipe, but that really is not a good picture to have. Here is why...
People sometimes say wires are like pipes, and the inside dimension controls the amount of current flow, just like a pipe.
If we have 40 pounds of water, a small "jet" or nozzle, or a small section of pipe of almost any length, will limit the maximum possible liquid volume we can ever flow with that pressure. This is NOT true with electricity! With electricity, we can flow any amount of current through a wire unless the voltage drop is so high the voltage to run something becomes too low. We can run any amount of current, up to the point where the wire overheats and melts. Once again, the two things that limit the amount of current a given wire can carry are:
Current in a short length wire or conductor, if system voltage is high enough, is limited by the safe wire temperature rise. A small size wire will not restrict current to certain amount, like a pipe does with water. Here is an example:
A number 10 ga copper wire has about .001 ohms per foot of length. If we run 100 amperes through one foot of #10 ga, voltage drop would be (100 amps*.001 ohms =) 0.1 volts. The heat, however, would be 0.1 volts times 100 amps, or 10 watts. This would make the 1 foot of wire pretty hot after a short amount of time, even though voltage drop of 1/10th volt is barely noticeable. This is actually how fuse links work. The short length of small gauge wire reaches a fusing current, and the conductor melts.
Typical Fusing or Melting Currents of AWG copper wire, room temperature, no airflow
Let's say the load on that wire is a giant spotlight with 0.12 ohms resistance. With 12 volts at the light, the spotlight current would be 12/.12 = 100 amperes. The spotlight power would be 1200 watts. Our #10 wire would add 0.001 ohms, reducing voltage by 0.1 volts. This would make load current 12/.1201 ohms = 99.17 amperes. Load power would be 1180 watts, instead of 1200. This is hardly any change in the spotlight performance, even though the current passes through a very small wire.
That certainly would not be true with water. If we passed water from a large hose through a section of very tiny pipe, we would have almost no water flow at all!
We can see wire size does not limit current or power unless the wire is so long that the wire's resistance is a significant percentage of the load resistance value, or the wire melts. A hose or pipe, on the other hand, limits liquid flow based on size and pressure.
Shorter Wires are normally Limited by Heat
This is because the total wire resistance and current determine voltage drop. Your house wiring, for example, is normally sized by the heat that can be safely tolerated. Electricians and building codes use a certain wire size for a certain current with little regard to voltage drop. This is because they assume the wire is a reasonable length, and the real worry becomes preventing house fires. At 240 volts, there would be little worry about loss of five volts through a long wire's resistance. The length would spread heat around, and voltage drop is not a significant percentage (about 2%) of total voltage for most household loads.
In some cases, especially where voltage is lower or wire length is longer, voltage drop can be a significant problem. Vehicle voltage, for example, is only 12.6 volts when running modest loads from a fully charged lead-acid battery. While a good alternator is charging, 12V vehicle voltage can be slightly over 14 volts. If we allowed the same voltage drop as house wiring by setting wire size with wire tables used for house wiring (by safe enclosed wall heat), we can easily run into situations where voltage drop is excessive. If the resistance of a wire or conductor dropped five volts in a 14-volt system (a 36% drop), the end result would almost always be devastating.
Because automotive voltage is low, we must pay attention to both heat and voltage drop in high current wiring.
Some articles on Internet claim a vehicle ground should involve a large copper buss connection. While well-intentioned and laced with true statements, the overall conclusion and some of the practical advice is not true. The length of a ground stud or bolt is so short that resistance and voltage drop does not matter very much. The biggest issue is having a good, reliable, corrosion-free connection to the body shell and ground lug. If the connection stud were three feet long, it would be another story. If you doubt this, look at the bolt size in modern car batteries. Current comes out through a fairly small terminal and bolt, neither of which is generally copper. As a matter of fact, ALL of the connections in the battery are lead, which is a terrible conductor! Why would any logical person go through the bother of attaching a large surface area, short length, copper ground plate to the body shell when the battery and all the connections inside the battery are lead?
It is easy to make insignificant things a worry, the old "making a mountain from a mole hill" adage. Ground stud or ground conductor length is usually so short and so wide or thick that the material resistance in the ground stud is one of the least significant factors. Corrosion and clamping pressure is much more of a factor than use of low resistance materials like copper. While copper terminals can be good in some applications, like insulated power entrance bolts to very high current devices like starting relays, starters, or alternators; copper studs or copper buss bars are normally a very poor general application choice for body-to-lug or frame-to-lug bolted connections.
To avoid galvanic corrosion, we do not want two metals with significantly different electro-chemical rankings in contact in the presence of salt and moisture. Materials choice varies with the environment and application, and is far more than material resistivity alone! Bare copper is generally a poor choice for a bolted ground connection point in a car. Copper in direct pressure contact with steel will erode the steel, and form insulating oxide layers. We are much better off using proper stainless or galvanized fasteners, even though material resistivity is higher.
For a good ground connection, use a fastener that:
Avoid soft fasteners, fasteners that corrode easily, or fasteners that cause the vehicle metal to corrode. This rules out copper or aluminum in direct contact with a steel or galvanized steel body.
Heat also causes resistance to increase, but this is normally a small or insignificant change in proper sized and length wires. The change, near room temperature, is about 1/2% per degree C. Hotter wires have higher resistance, which is another reason to oversize long wires. Heat related changes are negligible on short length wires, because they do not have that much resistance to start with. Increasing almost nothing by a few percent is still nothing.
Long Starter Wires Are Limited by Voltage Drop
Alternators and starters are usually the two things with the highest current. An alternator might be required to output 100-150 amperes (which is always much more than 2-3 horsepower load on the drive belt), and a starter might require as much as 200-300 amperes. High current means we have to be very careful about wire length and quality of connections.
A vehicle normally gets 100% of its electrical energy from the alternator. The battery supplies cranking power, and also supplies electrical power when alternator output falls below 13 volts or so. Too much voltage drop in alternator leads makes a vehicle battery dependent. At the same time excessive alternator lead resistance makes the vehicle battery dependent, excessive alternator lead resistance slows battery charging. Because of this, we ideally want the alternator to feed the power distribution point of the electrical system. This is why factory alternator leads almost always run to the common point where electrical feeds branch, and NOT to the battery post.
Long battery leads can cause slow cranking when hot. Long battery leads can also ruin dynamic regulation of alternator voltage, and slow battery charging. When battery leads are long, they should not be sized for current. Long leads should be sized for voltage drop. Long battery leads almost always have to be significantly oversized. Number 4 might be fine for a six-foot starter lead, but it is almost always too small for a trunk mounted battery.
Running a ground lead from the engine back to a trunk mounted battery is a common battery wiring mistakes. This adds unnecessary resistance. The same is true for running the alternator output wire back to a trunk mounted battery.
Voltage drop for each 10 feet of wire (ONE direction) at normal starter currents:
*also typical maximum resistance from Ford Mustang trunk pan to inner fender, with clean tightly bolted connections. Most of the chassis resistance loss is in the sheet metal near the bolted connections, so front to back connection distance makes only very small changes in resistance
** for negligible hot or extreme cold cranking speed change with good battery and starter, assuming good connections and ground return
Connection and ground return resistances add to this. If you have #2 fifteen feet long, the wire loss would typically be about 5%. The return loss would be about 2%. You would lose about 7-8% total of battery voltage, plus connection losses. This will noticeably slow a hot starter, but should not cause start problems with a good battery and starter.
Frame or Body Connection?
In a vehicle, the steel shell is almost always better than a sub-frame or frame connection point. A typical body shell has large stamped areas. These areas are normally welded at multiple points to surrounding panels. Body panels and the body shell have a very large surface area. The large surface area provides solid electrical connections. Because area is so wide, even a thin panel has a very large cross-sectional area.
Sometimes frames are isolated from the body by rubber mounts. My F-250 truck has virtually no cab-to-frame or fender-to-frame electrical bonding, so the wide body shell does not contribute to electrical ground conduction cross section.
In the case of a unibody, the frame is an isolated stamped structure. It is spot welded to the body shell at a few dozen points. The (sub)frame is probably OK, but it is certainly no better than any other panel. There is no reason to use the (sub)frame on a unibody.
If the body shell is a non-conductive material like plastic, carbon fiber, or fiberglass, use the metal frame or a very large auxiliary ground buss conductor.
Avoid ground connections through bolt-on or hinged sheet metal, such as fenders, hoods, trunks, and doors.
Unless you have a fiberglass or non-conductive body shell, or have something isolated on rubber mounts like a truck, don't waste time and material running extra ground lead length to a frame or cage. Such connections, because of the extra path length, usually have more resistance than short direct connections to the shell. Try to use radiator supports, inner fenders, firewalls, floors, and other large permanently integrated parts of the vehicle that are close to the thing you want to ground.
With almost any automotive electrical system, a very good, reliable, battery-to-chassis-ground connection is critical. Automotive manufacturers almost always have the very heavy battery negative lead directly to the engine block or some very solidly attached major component. This is because alternator charging currents and starter currents, the two most significant current paths, must be kept out of small wiring. High currents must not depend on or pass through unreliable connections. Unreliable connections not only include physical connections we visually recognize as "terrible connections", but also include engine components that are not securely and solidly grounded to the engine block. One example of a poorly grounded part might be a manifold, which floats from the heads and block by gaskets. The only ground path to the block is through the manifold bolts, which can become corroded or insulated with thread sealers.
Loctite and other thread locking compounds make pretty good insulators. I've seen thread locker compound completely isolate bolts or screws from the components they thread into. Even when physically tight, most thread locking and some thread sealing compounds can electrically insulate bolts and screws. The general rule is if a compound penetrates and hardens, it can electrically insulate an otherwise tight connection. Do not rely on connections or rely on ground current paths that might run through penetrating thread locking compounds, like Loctite.
On the other hand, silicone grease and anti-seize compounds will actually preserve electrical connections. They keep air and moisture away from the bare metal areas that physically touch.
Abnormally heavy loads, such as heavy-duty cooling fans, exceptionally large audio amplifiers, or two-way radio systems, must also have reliable ground paths to or from the battery negative. Unless a heavy external load is ground isolated, the external device should not attach directly to the battery heavy cable. The battery should be grounded to the vehicle chassis, and external devices with negative leads connected to chassis paths should have negative grounds connected solidly to the vehicle chassis.
Battery currents and other high currents to alternators or high current loads (like starters or cooling fans) must never be allowed to flow across high resistance ground paths. Allowing high currents to flow across poor connections causes service related issues, like slow engine cranking and unreliable accessory operation. Poor ground connections can permanently damage expensive parts, such as wiring harnesses and electronic engine control systems.
Front Battery System
Battery negative grounded to vehicle chassis through smaller, but secure, secondary lead for accessory and light currents. Never omit this battery-to-chassis wire!
Battery cable negative heavy lead runs directly to block or to a large metal-to-metal contact major engine component (not insulated by gaskets). Never ground the very heavy high-current lead to vehicle chassis with front mount batteries. Always ground the heavy lead to the block, bell housing bolt, or some other solidly-connected (no insulating gaskets) engine component.
The rear-of-engine to firewall ground braid, like the hood ground people often remove, is primarily for electrical noise reduction. This ground lead can be a smaller braid to an unused head accessory mounting hole, or to a bell housing bolt. Never put ground lugs under a tension-critical bolt, like a head bolt. Lugs are too soft and will yield, causing unreliable bolt torque.
The alternator grounds to the block or head through alternator brackets. The alternator is somewhat less critical than the starter current grounding point, because peak alternator current is less than peak battery current.
Always remember, battery currents can be as large as several hundred amperes. We do not want even a fraction of that getting into bearings or into electronic engine controls. Front-mount batteries should always use a very heavy negative cable to the engine block, or to some other engine part like a bell housing that is reliably and securely bolted to the block. The large negative cable is the primary path for starter currents and alternator battery-charging currents. While generally not the first choice, unused head holes are usually OK for these connections.
A critical path that is often ignored, or accidentally removed, is from the battery negative or engine block to the vehicle chassis. There must be a reliable battery negative-post to chassis connection. Without a reliable connection, the vehicle chassis can electrically connect to the block through engine and transmission bearings. This can not only damage bearings through arc pitting of bearing surfaces, it can also cause intermittent or fluctuating voltages to appear in critical or sensitive vehicle electronic systems.
Electrical potential difference between the block and vehicle chassis will often show as voltage impressed upon ground leads of electronic controls and sensors. Even a small voltage difference between the engine block and vehicle chassis can change sensor readings at the EEC.
The battery-negative to chassis connection for a front mounted battery only needs to be 50-100 ampere size. This is because, with a heavy battery ground cable to the block, starter current and charging current does not pass through the vehicle chassis. The small battery-to-ground connection runs accessories that return through the body shell to the battery and alternator. These accessories include lights, electric motors, audio equipment, gauges, and electric pumps. These shell currents normally amount to less than 40 amperes.
Some time before I purchased my used 1989 Mustang, someone replaced the battery cables. They neglected to use a two-wire negative cable terminal with the smaller chassis ground wire. To repair the missing ground in the front mounted battery situation, I soldered flexible braid to deep lugs (McMaster-Carr is a good source).
Heatshrink sleeves reduce flexing at the connection points, preventing fatigue failures. The connections are cleaned, and then coated with pure clear silicon grease (tune up grease or dielectric compound) in contact areas. The grease keeps water and air away from pressure connections, keeping them clean.
The very thick, black, negative battery post lead goes directly to the engine block. This is the lead that carries alternator and starter currents. These currents can be in the hundreds of amperes.
All vehicle accessory power, including lights, radio, ignition, fan motors, wipers, and computer typically run negative from the vehicle chassis. The negative battery post small lead should be securely grounded to a solid piece of sheet metal that is welded to the main vehicle chassis structure. In this example the radiator support works. Whoever replaced battery cables in my 1989 Mustang never grounded the battery negative to the chassis. Don't make that mistake! Working does not mean it is right!
Without this critical ground from the battery to the chassis, voltages to the computer and other important electrical devices become unreliable. To be sure the engine block stays at chassis, I also have a wide ground braid on the rear of the head to the firewall. This makes sure engine sensors are at the same ground potential as the computer chassis grounds.
Welded large panels anywhere in the chassis of a unibody vehicle have very low resistance. Despite what some people might claim, the unibody chassis makes a great low resistance ground path all through the vehicle. You will probably never do better than the body shell for connection path resistance.
Trunk mounted batteries in metal chassis vehicles are a very special wiring situation. With a rear-mounted battery, we have to get the high starting and charging currents to the engine block with minimum resistance and high current capacity. Do not run a ground wire to the battery unless the vehicle chassis is an insulating material, like fiberglass. The vehicle chassis should be the ground lead. The battery should be high-current grounded to the vehicle chassis with short direct leads. The engine block should be high-current grounded to the vehicle chassis.
This is how I handled my car, and it starts in any weather at any temperature just as a front battery would.
The engine block is grounded to the chassis through a heavy "starter size" cable.
The battery should be grounded through a fuse link or fusible braid that fails at less current than the main battery feed cable. This way, if the positive cable shorts to the chassis, the negative lead disconnects the battery. Nothing else connects to the negative post.
Alternator output, 12V fuse block feed, and battery share a secure common connection point.
To increase traction, and to have more room and less weight in the front, my battery was relocated to the trunk. The extra mounting space allowed use of a large, deep-cycle, marine battery. I selected a deep-cycle battery because normal sealed lead-acid batteries permanently lose charge capacity with any excessive discharge. Even just one single deep discharge can ruin a sealed lead-acid battery's ability to store charge. Deep cycle or deep discharge batteries are much more immune to unintentional deep discharge.
Prior to clamping ALL bolted or clamped connections (except solder points) are cleaned and given a light coating of clear silicon dielectric grease. This grease is also called "electrical tune-up grease". Anti-seize compounds or Noalox also work to prevent corrosion, and preserve electrical connections. Be sure the grease you use is compatible with the metals.
Battery hot wire is 2/0 (two-ought) welding cable, about 15-feet long. Welding cable has a tough heat-resistant jacket, and is extremely flexible for its size. This cable runs directly to the starter relay battery post on the starter. The lug at the starter, since it is exposed to water and dirt, is soldered on. The lug is close fitted to the welding cable. I flow rosin core solder into the lug, totally flooding the connection with solder. To keep the cable's insulation from melting during soldering, I wrap the cable insulation with a wet rag.
The 89 LX had a fairly heavy wire from the start relay to the big terminal on the starter solenoid, perhaps number 1 or 2 gauge. The old heavy starter feed wire from the starter relay is now shared with the large lead from the rear battery at the starter solenoid post, and the original starter feed wire now back-feeds the electrical system. The smaller wire with push on lug, as was used originally, runs the starter solenoid coil. It is switched 12 volts from the start relay.
Battery ground straps are 3/4 wide by 1/16th inch thick tinned strap with carry current rating of 150 amperes. Two equal lengths are run in parallel. These straps have low enough resistance to not drop excessive battery voltage. They do not heat under normal conditions, but will act as a fuse if the welding cable should ever short to the vehicle chassis (such as in an accident). The bottom lug is crimped and soldered. The battery post connection only crushed by the battery terminal cable clamp. Some sort of reliable fusing is mandatory for fire protection. The fuse links must melt before the 2/0 welding cable melts, and must melt before the battery significantly heats or melts.
The battery box is homemade. I cut, bent, and welded it here after an unsuccessful job (it leaked) done by a local welder. This box is water tight aluminum, and it is held down by six 1/4-20 stainless screws. Stainless screws with stainless star washers (to bite into the metal) sandwich the vehicle chassis. They are nutted to the trunk floor metal, so they act as studs. They are also coated with silicon grease. These "studs" hold the box and also provide the ground connection. This creates six parallel ground connections to the vehicle chassis. A sealed plastic box fits over the battery and vents outside the car to a low-pressure area, preventing hydrogen gas accumulation in the trunk.
The battery is a large heavy Marine deep-cycle battery. The battery is mounted on the right tire side, mounted as high as possible, and positioned back as far as possible to place maximum weight on the right rear tire. This caused a significant increase in traction with the street radial tires. This lead-acid deep cycle battery was selected because it has a very low internal resistance. Standard batteries lose ability to store charge if allowed to discharge below 12.6 volts. One deep discharge cycle can ruin some regular batteries. A deep cycle will not lose charge storage capacity when allowed to drop below 12.6 volts.
This is how I just did a 1994 Mustang the same way as my 1989. This is the trunk floor hole pattern. Notice we wire brushed the area around the holes until the galvanizing shows.
Close up of a hole in the 1994:
The starter and power wires prepared with protective sleeving and lugs for the 1994 ready to fit the starter terminal. I soldered the lugs on:
I cut the old battery negative lead and added a good high-current wire terminal clamp for the high current ground connection. I used a handy copper compression lug for metal compatibility from the copper battery lead to the lug, although a soldered lug like on the starter would have probably been better. I coated all connection points with silicon dielectric grease. The actual body connection is through a 5/16 stainless bolt with a stainless steel tooth star washer directly against the sheet metal on the dry side, a stainless spring washer on the bolt head, all buffered by stainless flat washers. There is grease between all layers to seal out moisture.
This stack forms a tightly-connected stud for connections to sheet metal. A second nut with spring or split-ring tensioning lock washer and stainless flat washers holds and buffers the copper clamp or lug, preventing direct contact between copper and the galvanized steel car body. This area was also cleaned and coated with dielectric grease before bolting. This is the proper way to do a ground stud on sheet metal.
Use dielectric grease on all electrical joints, and between sheet metal and any washers. Dielectric grease will keep moisture out and preserve the connections.