Detecting short circuits
What is a short circuit?Edit
A short circuit is an electrical circuit that has an improper path to ground with no/low resistance. This usually happens when a component such as a capacitors or MOSFET dies, creating a path to ground. The failure can be caused by corrosion, blown capacitors, and so on.
If you are finding that a voltage rail in your device does not come on (stays at 0 V) or does not reach its specified voltage, the problem can be a short on that rail, although you should check if the rail is being enabled first. The first thing to do when you suspect a rail to be shorted is always to visually inspect the board and look for things like corrosion, burnt or damaged components, discoloration, etc. To check if the rail is shorted, you can use a multimeter in either diode or resistance mode. You should always put the negative (black color, reference point) probe on the lower side of the rail (GND most of the time), if you turn the probes around you can measure low resistance because of internal protection diodes in ICs even if there is no short. If there is a 0 Ω (or 0 V) reading from the rail to GND there is definitely a short. If the reading shows non-zero values, it depends on the voltage and current rating of the rail, if it is a short. The following table shows some general values for good (not shorted) rails in MacBooks:
|Device||Rail Name||Voltage/Current Rating||Ohm Reading||Diode Reading|
|A1466 2013-2014 13” MacBook Air 820-3437||PP3V42_G3H||3.425 V / 300 mA max||195 kΩ||0.325 V|
Please note that these values are rough guidelines, your readings may vary a bit even when the rail is good. Also, these values do not represent the actual resistance of the rail at operating voltage.
There can be shorts that are not detectable with a multimeter. This can be when the short is only there when there is a significant voltage on the rail (simply speaking).
Locating a shortEdit
If you have a shorted rail in your device, the next step is to locate the short. Again, look for anything burnt or discolored, solder bridges (if the device has been worked on before), maybe a chip with a little hole in it (things like this take some practice to find). If you find anything like that, remove it and check if the short persists. If a visual inspection does not reveal anything, you can try powering up the device (with caution) and checking if there are any components on that rail that get significantly hot. Remove any shorted components found and check if the short persists. Be careful, as not every component that gets warm is necessarily shorted. As a rule of thumb, anything that gets too hot to touch is probably dead. Usually, the shorted component is the one that gets a lot hotter a lot quicker than the rest.
If you can't find any (more) shorted components with the above-mentioned methods (because the rail does not supply enough current to heat up the short), but the short still exists, you might need to Inject voltage into the motherboard. The idea behind this is that in the area of the short circuit there is more current flowing, resulting in more heat being generated. Be very careful when injecting voltage, you can easily damage your device when you use too much voltage or current. Do not inject many amps of current into a low power rail that usually only supplies a few milliamps. NEVER inject more voltage into a rail than its specified voltage.
Another method you can use is while having power injected into the shorted line you can scan the board with a thermal imaging camera and locate the hottest component in order to locate your short. Then remove the component to check. A similar method is also by using freeze spray or alcohol and noticing which component heats up soonest.
When you still are not able to locate the short, it means that the short is either very low resistance or non-existent. Retrace your steps, remove any wires and bodges you added, check if the short is still there. Visually inspect the area around the rail. Maybe another rail, that is being created from your "shorted" rail, is shorted and causes your problems? "Dead" shorts (very low impedance shorts) are usually caused either by solder bridges or by power components (MOSFETs, diodes, large capacitors). MOSFETs and diodes are usually heatsinked very well, therefore it is harder to see when they dissipate power from a short. Components like the CPU or GPU always measure like a short circuit, so shorts on power rails for them (e.g. CPU_VCORE) are very hard to detect (as a reference, a CPU that uses 25 W at 0.8 V core voltage has a "resistance" of just 25.6 mΩ and takes over 30 A, you can easily imagine how much current it would take to heat anything up in a rail like that).
Short Circuit ProtectionEdit
Virtually all power supplies (voltage rail generators) use some form of short circuit protection in the way of an overcurrent protection. The following lines should give a short overview of the most commonly used protection circuits:
The simplest way for overcurrent protection is to use a resistor after the power supply. This is most often used in low power circuitry, where the current draw is very low (e.g. logic signals, low power battery gear like a TV remote). It is extremely reliable and cheap, but terribly inefficient and because of the voltage drop across the resistor unsuitable for rails that require precise voltages. Another problem with resistive overcurrent protection is that even in the event of a fault, the power is not turned off, it is just limited to a maximum value determined by the resistor (that limiting value should be at least four times (!) the nominal power draw for efficiency reasons).
The next method is fusing. Here, a device physically cuts the power if the current draw exceeds a certain value. The main advantage to using fuses is that the fuse physically cuts the power and therefore prevents any further damage to the shorted device, which is a main concern in high power circuitry (like house wiring or the electricity grid). It is used in high power circuits (more than 100 W), where resistive protection is not economically viable or possible. The main disadvantage is that (most) fuses are single use, meaning after an incident the fuse has to be replaced. There are also resettable fuses (circuit breaker) or self-resetting ones (Poly-Fuse, only for low power!).
Another method of short circuit protection, which is commonly used in switching supplies, is active current limiting. Here, a controller measures the current draw on a supply (rail). If a set current is exceeded, the controller either reduces the voltage (lab power supply) or shuts the supply off (most switching supplies other than lab power supplies). If the supply is shut off, the controller decides if and when the supply should turn on again. Most supplies are configured in a way, where the supply shuts off when an overcurrent (short circuit) situation is detected, and turned on again after some time. If the short has not been removed, the supply turns off again. This can be heard as a "clicking" sound in higher power supplies (also called "power cycling"), and see on voltmeter as a non-stable voltage output.