Troubleshooting Basics (generally small electronic components)[edit | edit source]
Introduction[edit | edit source]
This article is a starting point that can and should be expanded an elaborated on. Please be aware there might be omissions or oversights, and not all instructions may apply to your specific issue or device. Before all, apply a generous amount of judgement.
General Advice[edit | edit source]
- Try the easiest thing first. Check power source. Reboot.
- Don't trust the other guy (the one who worked on it before you).
- 90% of problems are human error or oversight.
- Build a conceptual map of how the device should work, and how the parts that do work connect to the ones that don't.
- Verify the problems that it got returned for. Run self tests if possible. Verify for yourself that the user properly identified what the issue is—they likely have worse troubleshooting skills than you and may make mistakes.
- Get enough sleep. If something doesn't make sense, move on and come back after a rest with a clear mind. 9 times out of 10 you will find what you were missing before.
Formalized Troubleshooting Framework[edit | edit source]
These "best practice" steps have proven themselves in many real world situations. The first steps are analytical and preventive actions that can (should...) always be taken, regardless of what the issue on hand is.
- Make a conceptual map/block diagram of the intended functionality and how different parts of the system interact.
- See if you can obtain any documentation for the device, from user manual to service manual. Having anything at all can help you understand how the device is intended to work, regardless of your expectations.
- Often times, block diagrams are included in operators manuals in the troubleshooting section, and almost always in the service manual. They may also be included on schematics. These should be your bread and butter for approaching unique problems.
- Document the case as presented by the client.
- Verify that the user has done the troubleshooting properly—they may misunderstand how something is supposed to work, and lead you down the wrong path.
- See for yourself that the issue the user is experiencing is what they think it is. This could change how much they are willing to pay, especially if they underestimate what they think is wrong.
- Note the failure and repair history of the equipment.
- Analyze and document the current state of functionality as positives or negatives.
- Use the block diagram you constructed earlier to verify that each stage of the device is working as it should. When you find where something is wrong, you know that all of the previous stages work correctly and you won't have to waste time backtracking.
- Make notes of all the areas that are misbehaving—they may be correlated. Take a step back and see if the problems could be related—do they share any functionality?
- Approach this as methodically as possible.
Now you have a dossier for the equipment that you can refer to for your technical analysis.
Before you begin working on anything, do the following:
- Disconnect the equipment from power.
- Remove all batteries and anything extraneous (hard drives, optical drives, etc)
This will minimize the risk of cumulative and/or collateral damage, as well as ensure that no other parts of the system are interfering with what you're working on. If you find that it works but stops working when you add subcomponents back in—there is your clue. The key in this whole process is to approach everything as methodically as possible. It will save you a lot of time and headache when you will be able to easily follow your previous steps and verify to yourself/others what the problem is. If you can't understand or explain what you did, the problem likely isn't fixed (at least permanently).
Initial Overview and Analysis[edit | edit source]
First things first—try to boot it up. Does it turn on? If no, and this matches the problem that the user brought it in for, then you're good. You can probably just open it up and start taking a look. If that isn't what they brought it in for, make sure the user knows that there may be some additional issues. Primary troubleshooting in this area is changing out the power supply with a known working one, and checking external fuses if applicable. If that doesn't fix it, you'll need to start opening things up.
Start with a complete visual inspection. Look for corrosion, rust, burn marks, or smoky smells. You don't have to do this under a microscope but be sure to be thorough—small part failures can be just as consequential as large part failures. There may not be a singular fault: some defects are compound, involve multiple components, can be on both sides of a board, etc.
The most primitive block diagram you can create for the device consists of the power supply and whatever functionality the device has. Obviously, if the power supply is bad, then you will not be able to do anything. Therefore, in many cases, this is the best place to start. With a power supply that shows voltage and current, check what the machine draws. If it's drawing close to 0 mA, it's likely that a fuse or some other blatant component failure has occurred (like a trace being blown open) because no power is getting into the device. If you find that the device is drawing excessive current, it's likely you've got some shorted component. It's possible for this to be caused in the power supply, but it's also possible this is happening in another place on the board. If possible, disconnect the power supply and see if the power drawn is still excessive. If you find that the power supply is only dragged when other parts of the system are connected, then your problem is elsewhere. if the problem persists, then you know you have to start poking around the board. Approach this methodically—check each power rail to see if the test points measure appropriately, and if they don't, check to see if the rail is shorted to ground. At this point, this is all work that you can do without a schematic or drawing of the device—you should be able to identify basic blocks of the system just by visually looking it over (lots of caps/transformers is the power supply, etc). After this brief look, you should have a good idea of where the problem likely lies and can be addressed from a more informed position.
Document your findings.[edit | edit source]
A simple sketch/diagram of where the visually-damaged components are is enough, and is useful further on in the process and to communicate with the user. Always make a note of any parts you removed or swapped out with the designation so that you can remember to replace them later when you are done fixing the issue.
If visible defects are found, verify individual components and adjacent/connected components with electronic measurements to determine which ones are faulty and to what extent. A component may look bad, but be working fine, such as a connector that is slightly bent but is not otherwise compromised. If no visible defects are found, proceed to doing measurements on the electronics.
Making the most of your electronic measurements[edit | edit source]
An important part of taking measurements is to understand both how the measurement is being made as well as how the circuit you are probing looks to the meter or scope.
Basic Electronic Components[edit | edit source]
Modern electronics are composed of a few types of devices, arranged in a schematic to create functional systems (blah blah, I know). However, a fundamental understanding of how these electronic components work can help you more efficiently find and resolve issues, as well as throw out red herrings if you're wiling to give practical theory a bit of thought when you're working on something.
Passive components[edit | edit source]
Resistors, capacitors, and inductors are all common passive components. This means that they are not powered, and rather are strategically placed around the schematic to create the desired effect. These components can all be diagnosed with the power off and your meter in resistance mode (yes, even for capacitors). You cannot probe with a meter in resistance mode while the circuit is powered—your measurements will be meaningless.
Capacitors and inductors are reactive components, as their design exploits electromagnetic properties allowing energy storage as electricity (electric fields: capacitance; electrons/charge) and magnetism (magnetic fields: inductance; magnetic flux/"magic"). Resistors are designed to avoid these reactive properties, instead to impede the flow of charge in a static, non-reactive, consistent manner, limiting flow proportional to the voltage. To do so, resistors dissipate as heat the energy absorbed from the electrons flowing through, again proportional to the (the square of the) voltage.
Passive components are generally constructed of "normal" materials, classified as conductive (conductors) and non-conductive (insulators).
- Resistors (component designation R) can be used to create voltage drops and condition signals. Typically, when they fail, they will burn on the inside and either become an open circuit ("OL" on your multimeter) or short circuit (0 Ω), although partial failures are possible. it is not always easy to measure the resistance in circuit because of other nearby components (like capacitors) which charge up and will distort the measurement. However, measuring them with the resistance mode should give you an idea of whether the part is in good health or not most of the time.
- Capacitors (component designation C) act as small "tanks" of energy. They can be used for a variety of things, including signal filtering, power supply decoupling (reducing ripple), and AC coupling signals. You can't measure capacitance in circuit because of the parasitic capacitance in many of the traces as well as the behavior of other components. However, this doesn't really matter because their failure mode is much more straightforward—they just get shorted. Often times, this causes a problem because they are used as power supply decoupling, meaning they are connected from the supply rail to ground—meaning a short basically dumps the power supply entirely. They do not always appear burned or broken when failed. A trickier issue with these is that all decoupling capacitors are connected between the supply and ground, effectively putting them in parallel—meaning if there is one failure, it will appear as all of them failing (until you remove the specifically broken part). This can be tedious to track down.
- Inductors (component designation L) are small turns of wire. They work based on the electromagnetic field produced by current flowing through the wire. Inductors are generally used power supply filtering to remove high frequency content (like in switching power supplies) and oscillator circuits. Since they literally are just turns of wire, they will look like a short on your meter. However, this also means that they generally do not change value and when they fail, they will just fail open (OL). These generally will cause you the least amount of problems.
Active Components[edit | edit source]
Components designed to exploit the physical properties of both conductive materials and semi-conductive materials, such as carefully crafted and organized silicon and other semiconducting elements, are active components. These components use sub-atomic physical properties to control the flow of electrons in highly orchestrated (e.g. processor/digital logic) and controlled (e.g. FET/power/signal control). Such components and circuits may become so sophisticated, "integrated", and functional, that as part of their design a "power rail" or "power supply" (or multiple) are required inputs.
Active components are a bit more advanced and can be very difficult to make accurate measurements on while in circuit. Often times, the best way to measure these components is to turn the device on and measure their behavior (with a DMM in DCV mode or an oscilloscope) while the system is trying to work.
- Diodes (component designation D) are composed of a P–N junction. This doesn't mean anything to you and doesn't have to. The way they work is they will pass current in one direction, and not in the other, and they create a voltage drop across them in the process of conducting. The side of the diode that is marked with a stripe is the output side of the diode. You actually can test these with the device turned off if your meter has a diode mode. The meter will display a voltage—this is the voltage drop across the diode. The voltage drop varies depending on the type of diode, normal diodes have 0.6–0.7 V, schottky diodes 0.3–0.5 V and LEDs anywhere from 1–4 V, depending on their color. If you measure a diode in the reverse direction (with the probes facing against the conduction direction), then your multimeter should show a high voltage drop (>1 V) or "open" (depending on your model). This is because no current can flow and the shown voltage is the applied voltage from the meter. If you measure 0.3 V or less, then your diode has likely shorted and should be replaced. When measuring a diode, it is always recommended to measure in both directions across it—this will guarantee that you haven't made a mistake in your measurement and will also make confirm that there is nothing else in the circuit (like another diode, facing the reverse way!) messing up your measurement. Keep in mind that LEDs (light emitting diodes) and laser diodes are both types of diodes and share the same characteristics—they can fail just like any normal diode can.
- Transistors (component designation Q, usually) are composed of two diode junctions and therefore come in two flavors: NPN and PNP. The difference between these is that one of them requires sourcing current to the base in order to conduct and the other requires sinking current from the base in order to conduct. These are generally used in digital signal switching and amplification circuits. They are a bit tricky to measure in circuit unless you have a good handle on your theory. The easiest way to diagnose if these are healthy or not is to turn the device off, and measure the resistance between each set of leads (B–E, B–C, C–E). If you measure anything below a few kΩ, it could possibly be bad. However, there are transistor circuits that involve feedback resistors with low values, so you should try to verify if that type of topology is present before declaring the part is bad. You could also just remove the part and measure the pads on the board to see if the resistance is the same with the part removed—this would indicate that the part itself is fine. Note that their behavior is driven by current and not voltage, so you will have a very difficult time measuring these in circuit unless you lift up traces, It's generally not worth the hassle—just verify they look OK with the resistance measurement.
- Integrated Circuits (also known as ICs, with component designation U) are collections of all of the above parts into a subsystem packaged into a single chip. If you suspect that you have an issue with an IC, you should find the datasheet and look at the pinout. The datasheet will describe what the chip does, as well as how to integrate it into a circuit. From this, you can also find what voltages and signals should be present on all the pins. Sometimes you will be able to read the part number off the chip, and in other situations, you may have to refer to the schematics in order to see what part is used. You cannot generalize that ICs do a certain thing, as their functionality varies widely from power supplies to microprocessors to amplifiers.
Miscellaneous Components[edit | edit source]
- Relays (component designation K, sometimes) are isolator switches. These are generally usually used to protect a subsystem from some signals (power rails or high voltage) until some criteria has been met (like transient voltage spikes during startup). These have a normally open (NO) and normally closed (NC) state. Verifying these work is a matter of checking the input signal and seeing if the output is switching between the open and closed states.
- Fuses (component designation F) are used to protect the circuit in case of failure from overcurrent. They are made from a small film or wire that has a current rating matching the rating of the fuse, and the part will fail open circuit (OL) if the current through the part is exceeded. These are generally located around power supplies, laser diodes, and backlight drivers. The only measurement you need to take to verify that the part is working as normal is to make sure it reads under a few ohms.
- Connectors (component designation J) are physical connectors. when these fail it usually is from a pin getting bent, or something like that. It's different from connector to connector. It is also possible in the case of chemical spills for the plastic in the connector to become slightly conductive and leak signals from one pin to another, although this is a bit more uncommon. Usually a visual inspection is enough to tell if the part is good or not.
Basic Electronic Measurements[edit | edit source]
Most measurements you will need to make can be done with a DMM (digital multimeter). A DMM allows you to measure a variety of signals, which vary from meter to meter. However, a basic understanding of how the meter is making its measurements is necessary in order to properly use it and get useful readings. Another generally useful tool is an oscilloscope, which you can use to look at dynamic signals beyond the bandwidth of the multimeter.
Multimeter Probes[edit | edit source]
You need 2 probes to do any measurements, in any mode, period. One is for the positive side of the signal, and the other is for the COM ("common") side of the signal. The reason that it is COM and not ground is that the DMM is a floating reference—voltage measurements always need to be taken with respect to some point. The two probes set the point that you are trying to measure against.
This is a really important thing to understand. Suppose that you want to measure a 6 V signal. If you probe with the red probe on a 6 V test point and the black one on ground, you will find a reading of 6 V. However, if you were to measure with the red probe on a 18 V signal and the black on a 12 V signal, you would find that this also reads 6 V (18 V − 12 V = 6 V). So, the meter is not always referenced to ground, and this is why you need to pay attention to where you are probing. The probes will also have different functions based on what type of thing you are trying to measure. For voltage measurements, the probes are in parallel with the circuit under test. For current measurements, the probes are in series. So, if you don't pay much attention when setting up your measurements, it's very easy to end up with useless information about what you are looking at, and even damage your circuit beyond the state it came to you in.
AC and DC Volts Mode (ACV/DCV)[edit | edit source]
This mode is how you would measure voltage with your multimeter. It is an easy mode to understand and use. In this mode, the circuit basically appears in parallel with your meter. You can just probe the area of interest with the red probe (inserted in the V socket) and set the black probe on ground or another point of reference. Generally, the multimeter will have an impedance of >10 MΩ, meaning that for measurements smaller than this (<~2 MΩ) your measurement should not effect the circuit or your reading. If you are having a hard time finding a ground point, trying grounding the black probe on the chassis or a screw on the device—usually this is a grounded point and can be used just as well as a test point.
Resistance Mode (Ohms)[edit | edit source]
This mode injects a current through the two probes (small, typically 1–10 mA) and measures the voltage developed across the load. for resistive loads (like resistors!) this follows Ohm's law: V = IR. Because the DMM is good at measuring voltage, and we know the output current (or at least the meter does) this gives us the resistance of the component under testing. Of course, there are some traps here that you should know about before you are complaining that your measurements don't work correctly. First, because it is outputting a current, this mode will tend to charge up capacitors. This is why sometimes when you are trying to do a reading, the value will drift up as the capacitor soaks up the current and develops a larger voltage across it. Another side product that can happen from this mode is that the current being injected into the circuit can leak into diode junctions and (very slightly) turn on diodes/transistors, causing the voltage to discharge away. So, if you want to take good resistance measurements, just keep these things in mind and try to probe mindfully to make the most of your measurements. The most accurate way to make resistance measurements is to remove the component and test it out of circuit.
Diode Mode[edit | edit source]
Diode mode is basically resistance mode, but sources a voltage on the positive probe and measures the forward voltage drop on the negative probe. For a healthy diode, this is somewhere in the ballpark of 0.6 V–0.7 V but depends on the diode (see the specific datasheet for more information). By measuring the voltage drop, you can infer if the diode is healthy or not. once you do this, you should also swap the probes and do the measurement backwards to make sure there is not leaking in the reverse bias direction—this is a likely fault for diodes.
Capacitance Mode[edit | edit source]
This mode outputs a frequency on the probes and tries to measure the voltage change relative to frequency, and uses a bit of math to figure out the capacitance. This is generally a pretty useless mode—it's unusable in circuit, so it only works on parts that you've removed and generally capacitors don't fail by drifting in value—you only need to know if it's shorted or not, so you're better off doing resistance measurements in circuit.
Current Mode[edit | edit source]
In this mode, you need to move the probes over to the other socket. The reason for this is that while in voltage and resistance measurements, the meter is in parallel with the circuit—in current measurements, the meter becomes part of the circuit—it exists in series. When the meter is in current mode, the two leads are basically shorted together. This means that you have the very real possibility to damage your circuit if you try to probe something like you would a voltage signal, and instead end up dumping the power supply onto a sensitive chip or straight into ground.
The way the measurement works when in current mode is similar to the resistance measurement. There is an internal resistor with a small known value (to not limit the normal flow of the circuit). As current flows across the resistor, a voltage is developed across the two sides, and this is measured by the voltage part of the meter. Since we know the voltage and the resistance, this gives us the current through the device (again, refer to Ohm's law). This is likely the least useful mode and most likely to damage your device, so make sure you know what you're doing when you are setting up experiments to measure current.
Once a fault has been located, verified and isolated, the next step is the actual repair, which of course is a procedure in itself.