Posted on 10 September 2014

What to Watch for when Turning on a SMPS for the First Time

Free Bodo's Power Magazines!




In the near future, SMPS will completely displace conventional supplies, hence also engineers who are no specialists will have to cope with them. Turning on a new or unknown SMPS, especially offline - SMPS, can have unexpected and disastrous effects. This article describes some dangers, precautions and procedures for setting up and testing SMPS.

By Dr.-Ing. Artur Seibt; Vienna

Basic Safety Considerations

This article is about how to start and test SMPS in laboratories. National safety rules for electronics laboratories have to be obeyed. Work on circuits above the low voltage limit is generally only allowed if:

  1. At least another person is present who knows what to do in case of an accident,
  2. The circuit under test is connected to the mains via a certified adjustable safety isolation transformer,
  3. All measuring instruments are connected to the safety earth.

SMPS for lower voltages may present no personal electrical hazard, but can well pose a fire hazard or release noxious gases. One should always wear at least spectacles. The isolation transformer cannot prevent harm by high voltages inside offline SMPS; the PFC output voltage of up to 420 V DC is mostly the highest one, but generally all voltages on the primary side are dangerous. Output voltages above the low voltage limit are seldom encountered. The iron rule: if the circuit is live, only one hand is used, for all operations which require both hands, the circuit has to be switched off. Beware of capacitors which may remain charged!

Electrolytic Capacitors

Electrolytic capacitors can explode and burn: in the best case, the safety vent will open and spill the hot aggressive electrolyte all over the circuit, it is advisable to scrap the electronics affected! In the worst case, the capacitor will fly apart and cause personal injury. Many small cylindrical types can expel the innards like a bullet. Pitfall: most large electrolytics, especially high voltage types, use the standard 10 mm pin spacing and can be inserted wrongly, this constitutes the worst case of abuse and results in failure within seconds, depending on the current delivery potential of the circuit; if this is high like in offline SMPS, an enormous amount of gases will be generated which may be too much for the safety vent and cause an explosion. High current can also set the paper aflame. If a false capacitor with an insufficient voltage rating was installed, which is easy because they all have the same pinout, it will take some time before it fails, hence, if the SMPS seems to operate, the joy will be premature, the overstressed capacitor will soon fail. Experienced engineers will therefore switch off after a short time and test the temperature of the electrolytics and semiconductors, if one is already warm, it is probably overstressed.

Electrolytics function because the aluminum foil carries a thin oxide film built up during manufacturing. In operation, a very small leakage current flows and keeps the oxide film intact. During storage, the oxide will deteriorate, so that the capacitor partly loses its voltage blocking capability. If it is suddenly connected to its nominal voltage or even less, a high leakage current with a positive TC will flow, it will quickly heat up the capacitor, thereby increasing exponentially. It is a race: depending on the quality of the capacitor, the length of storage, the ratio of applied voltage to the nominal one and the current capability of the source: the capacitor either survives, because the oxide film restored itself quickly enough so that the leakage current decreases, or it fails within seconds. This danger is the greatest for high voltage electrolytics.

The experienced engineer will avoid this danger, because he knows that the “new” electrolytic may have been in stock for years. In order to prevent such unpleasant surprises, he will test each major capacitor before installing it. This is done by connecting the capacitor to a voltage source via a current limiting resistor of say 10 K; a voltmeter is connected across the capacitor. The voltage is increased in steps, and the rise of voltage across the capacitor watched. With a good capacitor, the voltage will rise quickly to just below the value set, then the voltage will be again increased and so on, until the capacitor voltage has risen close to its nominal one. It will then be discharged via a resistor, never by shorting it! This procedure is also necessary if an electronics gear has been stored for years. Some electrolytics will need several minutes, some can not be restored.

A word of caution: Most SMPS fail due to defective electrolytics! There is a lot of know-how in the manufacturing of electrolytics, their quality and the load in the application determine when they fail. The author has instruments with electrolytics made 50 years ago which still function. Most renowned European manufacturers had to give up, so their know-how was lost. There are many new manufacturers on the market; it is risky to trust their know-how. Also, due to the ubiquitous cost and price squeeze, design margins have become small. During manufacturing, the oxyde forming voltage used to be quite a bit higher than the nominal one, this, however, costs more and increases the size, hence, many manufacturers’ products barely meet their specs. The standard tolerance is - 20 + 50 %, but capacitors are hard to find which have more than - 20 %! Modern electrolytics contain very fragile aluminum foils, the leading Japanese manufacturers request not to use a capacitor which fell to the floor!

Far and above, it is the temperature which influences life and reliability most! When, after some time of operation with full load, the electrolytics of a SMPS are too hot to the touch: forget it, unless short life is expressly accepted. In the EU, manufacturers have to honor a 2 year warranty by law. On SMD boards the electrolytics are often heated by nearby diodes soldered to the board. Inexperienced designers place them close to semiconductor heat sinks or hot transformers. When testing SMPS, the temperatures of the electrolytics belong to the most important parameters, they are measured on top, “ambient” temperatures are immaterial. N.B.: Specs have to be carefully read, often “life” is specified only without any AC current load which is worthless. The maximum operating temperatures do not imply longer life: Life is cut by half for every 9 degrees C, this is valid up to appr. 85 C, above halving occurs already for every 4 to 5 C more! The electrolyte dries out with time, this raises the ESR which in turn causes higher temperature, then thermal run-away will set in. There is no way around: long life and high reliability sternly require high quality electrolytics, low AC loading and placement far from heat sources. Each degree less counts, it is an exponential relationship!

Foil Capacitors

The popular foils (PE, PP etc.) burn. Since many decades, the so-called X capacitors which are connected across the mains in offline SMPS have set many power supplies and then whole apartments aflame. Metal-paper capacitors are safe but too bulky. PE was abandoned, today’s X capacitors are almost exclusively made of PP and specified for a nominal voltage of 305 Vrms, although no higher mains voltage than 254 V is allowed. The mains carry overvoltages far into the KVp range which can ignite a X capacitor. Of course, test voltages and surges are standardized, but the mains may be disobedient to the standards committee and deliver still higher surges. They should be protected by VDR’s. Normal foil capacitors which are used in other parts of the circuit can also burn, this is less likely, because they are not subjected to unknown surges.

Ceramic Capacitors

Today’s SMD capacitors are predominantly MLCC’s (multi-layer ceramic capacitor). The renowned Japanese manufacturers state that no higher voltage than half the nominal one may be applied. The large-sized MLCC’s break often, if the voltage applied is high enough they will burn with fireworks. Another problem is the use of an inappropriate dielectric material. There are ceramic materials like Z5U, Y5V etc. which lose almost their whole capacitance when warm, also lose it when the voltage applied comes close to the nominal one, are highly nonlinear and exhibit very high losses at high frequencies. If they are used as filter capacitors in SMPS, they can become so hot that they unsolder and fall off the e.c. board. Ceramic capacitors are not marked, hence a wrong material will not show, in order to measure the capacitance, they must be unsoldered. Ceramic capacitors take substantial test overvoltage, hence it is impossible to determine their nominal voltage. By warming them up while they are hooked up to a measuring instrument one can find out whether it is class 1 material (COG) or class 2 or 3 material: class 1 will not change, classes 2 and 3 will decrease; if only by 10 %, it is likely X7R or X5R, if by 80 %, it is Z5U, Y5V or the like. SMPS which contain ceramics of the latter type should not be considered.

Power Semiconductors

If power semiconductors fail, they will usually short, but they can also burn red-hot inside until the current is turned off, the plastic will go up in smoke which should not be inhaled. SMD transistors or diodes can fall off the board.

In offline SMPS one can not rely on fuses for disconnecting in case of failure. According to the norms, a fuse must withstand 1.5 times its nominal current for at least one hour. If a power transistor or the main electrolytic is a dead short, it will blow the fuse, but it is also possible that another component in that circuit will burn such as common mode chokes or NTC’s because the current is not high enough to blow the fuse. There are no electrical means to protect all SMPS components from burning. Therefore one should always provide a metal housing which also provides shielding.

Visual Inspection

Experience showed that a thorough visual inspection is the most valuable and effective means to detect faults prior to turning on an unproven circuit. Semiconductors, resistors and the larger components are marked, SMD ceramic capacitors and glass diodes are not marked. As mentioned, it is especially necessary to check whether electrolytics and diodes are not installed backwards. The unmarked components pose a veritable problem in case the circuit malfunctions, especially the ceramic capacitors. In a worst case scenario one would have to unsolder and test all, one by one. Example: Assumed that instead of a 0.1 uF ceramic capacitor across the power supply pins of an ic, a 1 nF one was installed; the ic will probably malfunction. How long will it take, before this is detected? Probably, the ic will be first suspected and exchanged. It would be still worse if the ic functions, but is in fact on the edge of malfunction!

It is advisable that, for the first engineering board, the designer himself places all components, this is the best way to ensure that the correct components are on the correct places. However, many components have become so small that manual installation is only possible under a microscope.

Test Set-Up for Offline SMPS

As mentioned, offline circuits are only operated via an adjustable certified isolation transformer! A Power Analyzer which displays: line voltage, current, active power, apparent power, power factor lambda is inserted between transformer and test object. On the output side variable resistive test loads and a DVM per output are connected. The resistors are initially set to emulate the nominal loads. For this purpose ceramic wirewound power potentiometers are ideal which are available up to more than 100 W and with values from a few ohms to several kiloohms.

It is not advisable to start without loads because there may be problems in the regulation circuit which would cause the output voltages to go high. Most SMPS are designed for regulation from the secondary side; this implies that the SMPS must be able to start with full loads plus a reserve. The regulation circuit in the secondary is inoperative during start-up because there is no supply voltage yet. As soon as the reference output reaches its nominal voltage, the regulation sets in by reducing the power in the primary. Consequently, a broken loop will always cause overvoltage! (There are SMPS where the output voltage is sensed on the primary side which suffices for many applications, however, without any load, the output voltage can still rise above nominal because voltage spikes from the transformer will be rectified.)

A well designed SMPS incorporates transil (transzorb) zener diodes across each output, they conduct when their specified voltage is transgressed and limit the voltage, if the current rises too high, they will short and thus protect the load in the application. If transils are provided, start-up without loads is possible, it is easy to identify and exchange a shorted one, but it is necessary to be on the quick switching off, because the shorted transil will cause overstressing of the diode and the transformer winding belonging to this output.

The most important converters are flybacks, which are superior to > 250 W, these are “constant power” converters, i.e. it is the regulation loop which changes the behaviour to “constant voltage”. If they are not loaded, and the regulation loop is broken, the voltage will rise, theoretically, towards infinity, until some component will short; as the secondary voltage rise will be reflected into the primary, mostly components on both sides will be destroyed. It is not advisable to load only the reference output, because the unloaded outputs may rise so far that their capacitors are overstressed and fail. With other converters, the output voltage will rise, but not necessarily to destructive limits. NB: If a main power transistor in an offline SMPS shorts, a direct connection between the high voltage (rectified line resp. PFC output) and the low voltage control circuitry, typically 12 V, will be established so that the complete control circuit will be destroyed; depending on the size and worth of the SMPS, it will mostly not pay off to exchange all defective components, also for the reason that those components which seem to have survived may be already partly damaged!

The scope is the most important instrument, during the first test phase, the primary and secondary grounds can be connected, so that primary and secondary signals can be looked at by the same scope. It is highly recommended to use a > 100 MHz analog scope for that purpose, only top models of DSOs may be used instead, no middle- or low-priced ones. See the first article in this series for the explanation. For all high voltage measurements on drains etc. use only 100:1 passive probes; 10:1 probes are only good for 400 to 600 Vp! Their input impedance decreases at high frequencies which effectively dampens voltage peaks, so those are measured too low. The probes must be carefully adjusted, including their high frequency adjustments.

The most important signal to be measured at start-up is the drain voltage of the switching transistor(s). The second most important signal is the switching transistor current which is measured with a DC/AC current probe. The drain pin of the transistor has to be lifted off the board; one uses a small loop made of Teflon litz wire, the loop just large enough for the probe, the wires twisted for appr. 5 .. 7 cm, soldered between drain and board.

Turning on Offline SMPS

Of course, it matters whether it is a 10 or a 1000 W unit. With low power SMPS, one can be a lot bolder, because the possible damage will be small. Before turn-on check the fuse, only ceramic sand-filled heavy-duty fuses should be used on the mains! Many SMPS on the market contain cheap glass fuses and also such with a much too high rating. This is irresponsible and does not recommend such products. If the line is switched on, there is a high current transient which charges the main electrolytic; if that is not limited by a NTC or a resistor, it may blow a correctly rated fuse. The solution is not to increase the rating, but to use a slow blow (T) or very slow blow (TT) fuse. The normal procedure is to slowly increase the isolation transformer voltage while watching the Power Analyzer displays, the output voltage displays and the scope. A word of caution: a professionally designed SMPS can be turned on like this, however, there are SMPS which are not “brown-out proof”! If the 230 V are switched on, they will start, but if the voltage decreases much below the 207 V, they will break into wild oscillations which can lead to self-destruction. Caution: There may be defective transistor or capacitor, placing a short across the input, in such a case the line current will immediately rise drastically, hence increase cautiously at the beginning and watch the current display.

Normally, the converter(s) will already start switching, before the lower line voltage design limit is reached; the SMPS will deliver output power, in this voltage range between zero and the lower design limit the input current can rise above the value at the lower design limit, this can overstress components, hence the input voltage should not stay too long in this range, just long enough to take a look at the displays mentioned and then increased to the lower design limit. (With PFC in general 105 or 85 V, without PFC 207 V) The most important parameter is the active power display on the Power Analyzer! If these values are close to the correct ones at the respective mains voltages, it can be rather safely assumed that the whole SMPS is all right and functioning properly. The active power at 230 V and full load should be indicated in the specs, if not, it can be estimated by adding all loads connected and assume an efficiency of 85 %. After the full test of the first sample, the testing of more samples can be shortened to: Input voltage cranked up to the lower design limit, a look at the active power display and all output displays, then the input voltage increased to 230 V, then to 254 V, same checks.

With PFC

The PFC output is disconnected from the input of the main converter, a test load equivalent to the input power to the main converter is connected to its output. With rising mains voltage the output voltage will already follow before the PFC converter starts switching, because the current flows through the bypass diode (or flyback diode). The mains voltage must not stay too long at voltages below the PFC lower design limit, because the bypass diode, although husky, may be overstressed (without a bypass diode and with a marginal choke there is danger of destruction of the diode and the transistor). At 85 or 105 V the PFC regulation loop should lock, and the output voltage have reached its nominal design value between 360 and 420 V. 360 to 380 V: good design, above: poor design. It is assumed that the PFC is of the high quality “continuous” type where the fixed frequency converter current rides on the 100 Hz half sine and typically amounts to 20 %pp of the line current.

The scope is connected to the drain with the 100:1 probe. The DC/ AC current probe is clamped around the “cold” conductor of the choke which will also require a test loop. The drain voltage is clamped by the diode at the output voltage level, the rise time should be around some ten ns. NB: The scope display will depend on the scope used. If this is the very first turn-on, the choke must be checked, if the design data are not available. The absolute current peak consists of the 100 Hz current at low line plus the 10 % of the e.g. 125 KHz saw tooth. This peak current must not drive the ferrite into saturation at the highest temperature which the choke core can attain under wc circumstances. First the PFC is operated with full load for e.g. 1 h at low line, then the core temperature is measured with a contact thermometer, not with an infrared instrument. The winding temperature is either measured by measuring the resistances at 25 C and hot or by a sensor placed inside the winding. Because these measurements were at ambient room temperature, the difference to the maximum ambient around the choke expected in actual operation must be added. The result should not exceed 100 C except for special SMPS using special ferrites, see the part Transformer Saturation in the following text. Then the choke is heated with the use of a hair dryer to 100 C and the current watched. Impending saturation will be visible: the linear current rise of the sawtooth will change into a nonlinearly steep upward rising one. As soon as this appears, switch off! If this occurs at full load, low line, the choke must be redesigned and the procedure repeated. If the SMPS was bought or a test sample, it should be rejected. If it was of one’s own design, the choke must be redesigned. An acceptable choke will allow + 10 to 25 % overload without signs of saturation.

At last, the line voltage is increased to 230 and then to 254 V and the active power and the output voltage checked; because this is a booster, the losses will decrease with rising line voltage hence also the active power.

After the test of the PFC, its output can be reconnected to the input of the main converter. A boost converter cannot be protected against overload due to the direct connection from input to output via the diodes. If the main converter is defective, it may draw an overcurrent from the PFC. It is recommended, for this test, to insert a fast or medium-speed fuse with a rating equivalent to the maximum design input current of the main converter. Use only sand-filled ceramic fuses, because this is DC, a standard fuse could ignite an arc rather than disconnect!

The start-up of the main converter is different with and without PFC: a PFC will deliver its full output voltage in the moment the line reaches its lower design limit; this means that the input voltage to the main converter will at first follow the increasing line voltage and then suddenly jump to the PFC design value of between 360 and 420 V. This should not surprise! If problems in the main converter are to be expected, this sudden application of the full voltage is discouraged. It is then better to leave PFC and main converters disconnected, and to install a connection between the output of the bridge rectifier and the input of the main converter, thus bypassing the PFC and follow the procedure below.

Without PFC

The main converters in offline SMPS with and without a PFC differ hardly: the range for 230 V SMPS extends from 207 to 254 V, hence the converter must function from less than the rectified 207 V onwards, i.e. appr. 250 V DC, minus the peak ripple voltage which is often forgotten up to 310 V DC with load plus the peak ripple; at low load and high line the input voltage will rise to 360 V. If a PFC is on board, one might think that the input to the main converter is a stable e.g. 360 V; this is wishful thinking, because the standard PFC voltage regulation loop is very slow, hence load changes cause wide fluctuations of the output which the main converter has to digest without losing regulation.

Not all types of converters can be treated here, the flyback is chosen, because the bulk of offline SMPS are < 250 W which is flyback territory. The flyback is both the least expensive and the best converter, it not only delivers an excellently regulated reference output voltage, but in addition several output voltages which are well stabilized so that mostly post-regulators can be dispensed with. Their voltages depend on the loads on all outputs. The transformer (the term is false but customary) determines the whole performance and especially the cross-regulation.

There are two modes of operation: voltage-mode and current-mode control (invented by the author at Tektronix). A current-mode controlled flyback will operate by nature over a wide range of input voltage and features the fastest response of all converters. At voltages below its lower limit, the switching transistor (and transformer) current will be limited by the primary inductance resp. the maximum turn on time resp. duty cycle of the transistor, i.e., the circuit is low-voltage resp. brown-out proof. As soon as the control circuit starts operating, the flyback will deliver, so that all secondary voltages start to rise, as soon as the lower input voltage limit is reached, the voltage regulation loop should lock, and the output voltages reach their nominal values.

Hence, with this type of converter, it is safe to slowly increase the line voltage from the isolation transformer and watch the active power, the scope drain voltage and current displays and the output voltage displays. Normally, the flyback will function below its lower design limit and already deliver full power; however, in this range the line current will be higher than designed for; hence one should not leave the SMPS in this range for extended periods of time.

The temperature measurements on all vital components must be performed at low and high line, because some losses are more current, others more voltage dependent. Basically, these are the transformer, the transistor, the diode(s) and the electrolytics. Because it is so easy to choose the false type of ceramic, the MLCCs’ temperatures should also be checked.. NB: For measurements of temperatures at electrically hot points like transistor cases = drains, the SMPS must be switched off, even if handheld isolated instruments are used! The fast high amplitude pulses will disturb the measurements.

Output ripple measurements require the use of the special Tektronix probe connectors into which the probe is inserted.

Test Set-Up for Low Voltage SMPS

There are marked differences between offline and low voltage SMPS. A major difference is the existence of a negative input resistance with all converters which deliver a constant output power. If the input voltage rises, the current will decrease, this constitutes a negative resistance. The author refers to his article in “Bodo’s Power August 2012”. The design of the SMPS input circuit, the voltage source and the connections come into play. The whole input can oscillate wildly which has nothing to do with the regulation loop, which, however, can contribute.

As an example, here a buck converter is chosen. These are very popular in the shape of small modules. When testing (or selecting) it is important to check whether a module is a complete SMPS or whether the small size is misleading, and a whole lot of external components have to be added in order to complete a functional SMPS! As a rule, there are only rather small MLCC’s at the input, and this is why the input circuitry external to the module is critical and determines whether wild oscillations will occur or not. A proper test set-up must therefore come as close to the set-up in the application as possible, or the test results will be of doubtful value. The “application circuits” provided by the manufacturers cannot be trusted, to the contrary: it happens that they produce oscillations which lead to massively increased p-p and rms input currents requiring a big capacitor which is superfluous if the input is properly damped with just a small RC. The ESR of the output capacitor is vital for the stability of the regulation loop, hence no MLCC’s should be added there.

The scope is hooked up to the source of the transistor, the current probe is clamped around the cold conductor of the choke, if accessible, later it is clamped around the input conductor. The input current of a buck is a square wave, depending on the amount of capacitance at the input inside the module, at the input pin it will be filtered to some extent.

Turn-On of Low Voltage SMPS

The adjustable load should first be set to perhaps half load. With this type of converter the worst which can happen is a short of the transistor: then the input voltage will appear at the output, if the short was caused by a shorted diode, the input will present a short. In proper operation the output voltage is equal to the input voltage multiplied by the duty cycle. Most SMPS of this kind are not overload and short circuit proof. While the input voltage is increased, the output will first follow, as soon as the input voltage exceeds the programmed output voltage, the regulation loop should lock. Further increases of the input voltage will just reduce the duty cycle. After ascertaining that the supply functions, and the transistor source voltage, the choke current and the output ripple voltages on the scope are all right, and there are no output oscillations, the input voltage and current should be measured in order to verify that there are no input oscillations and no excessive p-p and rms currents. Last, the temperature measurements on the transistor, diode, choke and output capacitor follow. With modules, in order to check whether the incorporated choke is adequate, the full load should be connected, then the module should be heated to its maximum specified case temperature and left there for ½ to 1 h. Because the choke current is not directly accessible, output and input ripple voltages should be observed with the scope. The transistor is in danger, as soon as the choke approaches saturation, however, if the purpose is to select such modules, it is better to destroy some than to choose an inadequate product.

Some Special Problems

Transformer Saturation

The design of flyback transformers requires enormous know-how and experience as well as knowledge of ferrites and winding materials. Whereas true transformers like those in forward converters are comparatively simple, the flux density in their cores low, the flyback transformer is none, but a choke, and the flux density very high. Ferrites have a maximum operating temperature above which the losses rise steeply towards the Curie point; this minimum marks the highest temperature which the core may reach, for the customary power ferrites this is around 100°C. The saturation flux density is specified at 25 and 100°C, only the latter value is of interest; no specs are available for higher temperatures, hence it is wise to design the transformer for a 100°C maximum core temperature under wc circumstances. SMPS transformers are at least partly wound with TexE, TexELZ or TexF for class F; if the core is at 100°C, the winding will also be at or close to its maximum permissible temperature where its life is already shortened. The usable flux density is, of course, far below saturation, because near saturation the core loses its permeability and thus the transformer its primary inductance. The maximum flux density translates into a maximum current. The test whether the transformer is adequate as regards saturation is done as decribed for the PFC. If this is not performed with thoroughness, the following can happen: The SMPS functions until it warms up to a temperature at which the core saturates; due to the loss of inductance, the switching transistor current rises extremely sharply. In current-mode control, it depends on the speed of turn-off whether the transistor will survive. As a rule, it will be destroyed. Most probably, a bad transistor will be suspected, it will be replaced, but also the new one will be destroyed. It may take a lot of time, before the real cause, the transformer, will be found out!

Special Load Problems

No SMPS should be designed without information about the exact type of load, this applies also to selecting SMPS, otherwise serious disappointments may be programmed. A typical case are motor loads. Few users of “DC” - motors realize that there are no “DC” - motors; this designation misleads, a DC field cannot cause rotation. Each “DC” - motor draws a high peak-to-peak AC current the average of which is the DC current! If a SMPS for motors is selected on the base of the nominal specified DC current of a motor, this will not work for 3 reasons: 1. The SMPS must be able to deliver the positive current peaks which can be several times the DC current, 2. It must be able to soak up the negative current peaks, 3. The AC current stresses the output capacitors which have to carry the rms sum of the SMPS and the motor currents. In addition, it is also often forgotten, that a motor draws a start-up current which can be several times the steady state DC current; if the peak start-up current can not be delivered, the motor will not start but stall.

Another typical case is incandescent lamp loads which draw typically 15 times as much current during the start phase; if the SMPS cannot deliver that much, it will “hang itself up”, i.e. the load lines of the SMPS and the lamp intersect at a stable operating point at a low voltage, the lamp will glow dimly and never light up.

Regulation Problems

SMPS regulation loops are tested with so called electronic loads, those are resistive loads which are switched by a transistor, the frequency and the duty cycle can be selected. The scope will show the reaction of the loop. The problem is that there are loads which do not load the supply periodically but stochastically, that is more severe and can cause hiccups of the regulation loop by overdriving the amplifier, because output voltage excursions overlap. The effect is that suddenly much larger excursions in both directions occur.

Temperature effects on components

In general, the permissible loads on components decrease with rising temperature. Life will be impaired with all. “Cold electronics live longer.”

Many vital components change their properties more or less drastically over temperature, this is one reason why SMPS must be tested over the full range of internal ambient temperature. Examples: The ESR of electrolytics decreases with rising temperature and vice versa, this has several effects: the ripple decreases, a minimum amount of ESR is often necessary for the stability of regulation loops, i.e. a SMPS may be quiet as long as it is cold, but start to oscillate wildly when it warms up!

Ceramic Y capacitors are made of the worst material Z5U, when warming up, they will lose almost their whole capacitance with the effect that conducted emi norms are violated! This is one reason why SMPS manufacturers mostly specify 25 C as the measuring temperature for emi. Ferrites were already mentioned. Semiconductor losses mostly increase, one big advantage of modern SiC components is the fact that losses do not increase.


VN:F [1.9.17_1161]
Rating: 6.0/6 (1 vote cast)
What to Watch for when Turning on a SMPS for the First Time, 6.0 out of 6 based on 1 rating

This post was written by:

- who has written 791 posts on PowerGuru – Power Electronics Information Portal.

Contact the author

One Response

  1. avatar DhananjoyGhosh says:

    Sir, when i pass current through the ups a small burst made in smps. My ups voltage was in good condition. What i do sir

    VN:F [1.9.17_1161]
    Rating: 0.0/5 (0 votes cast)

Leave a Response

You must be logged in to post a comment.