PSU Back EMF Option, for GlobTek EL6 design series products


When a large inductive load such as a DC motor or a solenoid controlled valve is placed on the output of the power supply, at certain points of the load operation current/energy may be returned from the load to the PSU, which may cause the output voltage to surge higher, and in certain situations, could conceivably damage the PSU.

All GlobTek GTM96 SERIES power supplies, are relatively rugged with respect to back EMF damage to the power supply.

This ruggedness is due to the basic primary-side output overvoltage protection design feature. To understand this concept, it is necessary to first understand typical back EMF events, and OVP events.

Back EMF events which are the result of inductive type devices utilizing magnetic field energy storage, operating in the energy releasing mode, can typically result in the output voltage of the power supply increasing anywhere from 1.05X to 1.50X (5% to 50% increase). After the back flow energy is transferred to the PSU output capacitor, then it will slowly drain back to the nominal output voltage, in a period based upon the typical output current draw of the load, or in some situations, the drain from bias current the power supply itself, may return the output capacitors to the nominal output voltage level.

Output OVP protection circuitry can be designed in several different ways:

  1. Historically, SCR crowbars with zener diode triggers, could prevent the output voltage from exceeding a certain level. At present extremely rare, due to the higher cost, and larger component sizes. This technology would trigger the latching SCR if the back EMF event exceeds the OVP trip level. Not good if periodic back EMF events are expected.
  2. With smaller power supplies, often times parallel zener diodes are placed in parallel with the output. This may or may not be an acceptable OVP circuit design technology, depending on the back EMF energy level which needs to be absorbed. If the energy exceeds the zener diode energy absorption capability, the zener diode could fail to a shorted state.
  3. An optocoupler, may be placed across the isolation barrier, with the specific purpose of providing a OVP protection signal pathway to the primary controller. The OVP signal to the primary controller may activate a latching or non-latching protection scheme, dependent on the PSU circuit designer, or product specification requirement. Typically this solution is not acceptable for the OVP protection when a back EMF state is expected, since the zener diode in series with the optocoupler, could experience a huge current impulse from the back EMF event, and damage either the zener diode or the OVP optocoupler.
  4. Primary-side OVP protection, is the apex of OVP product design, it is low cost, highly reliable, and is fundamentally immune to back EMF induced PSU damage, with a few provisos. Primary-side OVP protection may be configured as latching or as non-latching. Obviously non-latching would appear to be best in the given scenario, however, since the back EMF event is not created on the primary side, a back EMF event is not “seen” by the OVP controller on the primary. In other words, to the “controller”, the output voltage appears to be normal during the back EMF event.

In the exposition of OVP methods a)-d) above, it should be explicitly understood, that there is a safety standard control aspect requiring component analysis of OVP containment componentry, and this influences the required circuit design, such as the requirement for a second optocoupler feedback path, when method c) is provided. Method d) is utilized on all GlobTek GTM96 SERIES products, as well as many of the previously designed products in the past. Although there are few limitations, some are worth examining, in order to be aware of any potential pitfalls, when evaluating new system configurations.

The below “cut” from a typical GlobTek schematic (Fig 1) is used to illustrate these provisions.

FIG 1. Fig 1

Capacitors C49,50 and 51 will be exposed to whatever the overvoltage stress caused by the back EMF event. Therefore if back EMF will cause 30V once every 10 minutes, then it may be worthwhile changing to a 35V capacitor. Alternately, most aluminum electrolytic output capacitors have a short term “surge rating”, which is typically around 10% to 20% higher than the continuous capacitor voltage rating. If the back EMF voltage surge event is very infrequent, this could perhaps be an allowable option.

Additionally, when the back EMF event occurs the 3 pin voltage reference / voltage error amp component U7 will instantly turn on hard, after it “sees” that the output voltage is too high. This can cause excessive heat dissipation in resistor R68 for certain product designs. In this particular product design, the resistance is relatively high at 2.7K ohms, but let us suppose the circuit designer had used a 270 ohm resistance instead, and let us analyze the result of a 1 second high voltage exposure at 30Vdc, assuming a square rise and fall time, surely this pulse shape would not exist in the real universe, but good for analysis purposes.

Case 1, R68 = 2.7K

U7 will drop about 2V in saturation, optocoupler LED U6 will drop about 1V, therefore 27V will be across R68. The heat dissipation on R68 will be V2/R = 0.27 watts. R68 is a 0805 size SMT resistor with a typical power dissipation rating of 125mW, therefore it will be overstressed by a factor of 270/125, for a 1 second pulse it ought to be OK. Most resistors can handle a 6X over-power stress for 1 second. Not sure about this particular resistor, but you get the point, it ought to survive a 1 second pulse without long term damage.

Case 2, R68 = 270 ohms

Again 27V is across R68, however now the power dissipation will be 10X higher, at 2.7 watts. When a 1/8W resistor is dissipating 22X it’s nominal power rating, damage can rapidly occur, survival without damage would probably be limited to about 1/10 of a second or less. Additionally, at 100mA of current the U7 IC is at the limit of it’s current capability, therefore it is also possible to damage it.

Thus although method d) is superior to other methods in withstanding voltage stress resulting from a back EMF event, an analysis of:

  1. Level of induced voltage, required to assess if output E-caps are adequate
  2. Duration of back EMF event, to determine is heat dissipation damage may occur in the output regulation circuit may occur, is necessary.

When output regulation circuit issues appear to be a concern, it is usually possible to modify the output regulation circuitry to reduce the power dissipation level during an back EMF event to acceptable levels. Bearing all the above in mind, if your system being developed has a known back EMF occurrence into the power supply, please determine the voltage level and time duration of the event, oscilloscope screen captures are the best way to convey this information to the power supply design engineer at GlobTek.

Performance estimation of interior permanent-magnet brushless motors using the voltage-driven Flux-MMF diagram - Scientific Figure on ResearchGate. Available from: [accessed 13 Jun, 2019]

Request Info