Hello,

the picture is normally not understandable without the textual description.

Anyway, I'd guess it is a test specification describing the response to an instantaneous change-of-load.

The graph refers to voltage deviations as a full step load change occurs (from open circuit to rated load). The real voltage deviates from the reference one (dotted line in the picture) as a change in the load (linear or non-linear) occurs. The standard IEC 62040 imposes the envelope of the instantaneous voltage deviation, which should be within +/- 10% at steady-state.

I agree with both aforementioned answers.

However, I would also like to add one more thing. When looking at transient responses we do not only look at the amplitude of the waveform but also at the time constants for rise time and decay time. In the above picture you can see that any transient anomaly should be cleared within 10ms and within 1 second everything should be back in steady state.

Hello Majidreja Salehi!

What we are talking here is an international standard developed and issued by a reputed and expert organization. The information cannot be "misleading".

Let's try to understand what the diagram says or reveal information to us.

The first plot (left hand side) shows a short transient riding on the fundamental sine wave of the UPS output. The next plot (right hand side) states the acceptable limits or limits as defined by the standard. As Stelios has summarized in his response above, the timings state the duration for which the deviation should stay or acceptable and the time duration within which the output from the system should be "normal"

I shall try to elaborate it a bit more -

The transient can range for a duration ranging from 100us (microsecond or 0.1ms) till 5ms with a defined tolerance of +/-30% of the rated nominal output. The next 5ms to 10ms window is where any transient should try to recover or diminish toward the normal range and from 10ms to 1s period the acceptable variation is within +/-10% of the nominal value.

Therefore, if the UPS output is exceeding these limits of falls outside the specified range (say +/- 10% of the nominal values), the said product "DOES NOT" comply to the "STANDARD".

I hope that I have taken a small effort to explain what that figure might mean, forgive me for any mistakes.

Thanks with regards,

-PRASANNA WAICHAL

Many thanks to all of you for your answers, 

but the question is still remaining! As you mentioned, the picture shows deviation of "instantaneous voltage" but if you read this standard, you can conclude from some sentences that the criteria +/-30% for 5ms, +/-10% steady state is defined for "RMS voltage"!

and I've seen in many papers that RMS voltage is checked according to this criteria. DO you have any idea about this? Isn't it a contrast in comparison to the picture?

Hello,

I haven't read the standard itself. Anyway, looking at the picture it would make sense to specify the tolerance band as 30/10 % "of the NOMINAL RMS voltage". (Thus 30/10 % from around 230 V.)

The reason is that the load change may occur at any point of the sine. As you can see, the right half of the picture shows the x-axis to be the reference line. This is the pure sine from the left half - here "flattened" as only the deviation is relevant, not the absolute value.

The standard specifies nominal voltage NOT rms. A nominal voltage of 230Vrms is the same as a nominal voltage of a sinusoid of 325Vpeak.

https://www.energystar.gov/sites/default/files/specs//private/UPS_Final_Draft_Test_Method_Appendix_B_IEC62040_FDIS_Excerpts.pdf

From my experience when dealing with surges/transient voltages it is better to work with peak values because surges have very small duration. Also most surge protection devices like the Metal Oxide Varistors (MOVs) work with peak voltages; Surge protection devices will clamp the surge at a peak value not rms.

For a single phase system: RMS = Peak/sqrt(2). Therefore, for 230Vac rms =  325V peak. Now applying the tolerance this deviation ranges from 161 to 299 rms  which is the same as 227 to 421 Vpeak.

That being said,  I see that you are given an envelope based on your system. IF your nominal voltage is 230Vrms=325Vpeak then the allowed surge/transient can be kept within an envelope of 227Vpeak to 421Vpeak for 5ms and then within 1 second it will be within the envelope of +/-10% which translates to  292Vpeak and 357Vpeak.

For a 120Vrms system = 169Vpeak then the envelope for 5ms will be 118Vpeak and 220Vpeak and within 1 second it should decay to a limit between 152Vpeak and 186Vpeak.

You can also check the following literature:

1.      Check CBEMA curves: http://www.powerqualityworld.com/2011/04/cbema-curve-power-quality-standard.html

2.      A thesis about Lightning/Transients and MOVs: https://www.researchgate.net/publication/281202416_Comparative_Study_of_Metal_Oxide_Varistors_MOVs_for_Failure_Mode_Identification

3.      A paper on Surge Protection Devices: https://www.researchgate.net/publication/3183116_New_MOV_Failure_Mode_Identification_Invention

Thesis Comparative Study of Metal Oxide Varistors (MOVs) for Failur...

Article New MOV Failure Mode Identification Invention

I appreciate your attention but I think it's too conservative to keep Instantaneous voltage in the above acceptable curve (+/-30% threshold below 5ms) for transient events and it is not even possible for RC loads. 

I must mention that this standard introduces reference RC, RL and nonlinear loads for checking the UPS output at the moment  of connecting these loads to UPS. For a discharged RC load, the instantaneous output voltage will become zero at the moment of connection. So, the +/-30% criteria will be IMPOSSIBLE for Instantaneous output voltage. 

In addition, the following text is written in IEC62040-3 (1999):

"UPS designed for both types of loads (linear and nonlinear) have to maintain the RMS values within the acceptable curves"

The text is a bit changed in IEC62040-3 (2011) and the above picture is added for indicating instantaneous voltage variation. That's why I can't make a clear conclusion.

When Line power is lost then the UPS will kick in. According to this standard the switching transfer time, from main power to backup power, and the switching transient caused during this transfer should be done within 5ms. Also, the UPS output power should be within +/-10% withing 1 second.

Now, for a 50Hz signal the period is 20ms. This means that a settling time of 1 second (1000ms) is 50 * period, so it is safe to say that the settling voltage will be +/-10% of the RMS value.

In the case of 5ms it is 1/4 period.

So now we have 2 signals. V1=VmSin(w1t) and V2=0.3VmSin(w2t)

V1RMS=0.707Vm and V2RMS=0.707*0.3*Vm

Total RMS = sqrt( (V1RMS)^2+(V2RMS)^2)

 = sqrt( (0.707Vm)^2+(0.212Vm)^2)

= 0.738Vm

Therefore, for 5ms the RMS value will change from 0.707Vm to 0.738Vm which in my opinion this is about 5%.

If you assume that v2=0.707*1.3*Vm then the Total RMS = 1.15Vm which is indicates an increase of 64'% (from 0.707Vm to 1.15Vm).

Electronic equipment can handle surges (overvoltage) for a few microseconds to ms in the range of kilovolts BUT when the overvoltage is in the range of several cycles then the power problem is no longer call Transient but it is called a Swell. Transient overvoltages are important for the dielectric strength of equipment whereas Swells are important because of the extra amount of energy that might exceed the rated power of devices.

Lightning is a different story. Lightning is a Current injector and due to this high amount of current (KiloAmperes) even low impedance systems will generate high voltages (Ohms Law V=IR). But this is a different topic.