Calculating Average Power in a HV Pulsed system

[DejanC]

Hi All,

I am looking to validate some measurements I did. I have a 400 kHz signal with about 1% duty cycle, producing a roughly 400 Vpp input on and 270 Vpp on the output. My current probe is rated for upto 800 kHz so I am assuming that signal should be fine for accuracy. However, my 270 Vpp shows 30 V RMS. Seems like a discrepancy there. Is it valid to calculate the average power using V_rms X I_rms?

What can I do to more accurately measuring the output power if I am worried about the accuracy of the voltage readings without buying an expensive HF power analyzer?

[Terry Bites]

masteringelectronicsdesign.com/how-to-derive-the-rms-value-of-pulse-and-square-waveforms/

[wasedadoc]

Square root of (270 x 270 x 0.01) is ?

[DejanC]

Thank you.

So V_rms = V_pp *sqrt(D)

If my duty cycle is 1% and Vpp = 240 V, then the scope showing V_rms = 30 V is not too far off. CorrecT?

[wasedadoc]

Thank you.

So V_rms = V_pp *sqrt(D)

If my duty cycle is 1% and Vpp = 240 V, then the scope showing V_rms = 30 V is not too far off. CorrecT?

No. Yes

[DejanC]

I calculated 27 V from your comment and my formula. Why is it not correct?

[gf]

Assuming that you mean a pulse waveform with 1% duty cycle and 270V_pp, any value >= 26.865 V_RMS could be correct, since the DC offset matters as well, and you did not specify it.

[wasedadoc]

I calculated 27 V from your comment and my formula. Why is it not correct?

Sory, my previous answer was a rushed single word.  Should have been 'yes'.

[CaptDon]

You are giving voltage readings and talking about a current probe. Also you state two different voltages peak to peak 400 in and 270 out. What sort of circuit? How much power are you looking at? Is it in the current probes range? This reminds me of radar where you calculate the power during the pulse and multiply it by the duty cycle which in pulsed radar is typically .001% to .005% which is why a 4KW pulse radar only has an average power of 4 watts and isn't particularly dangerous to humans at close range.

[jbb]

Hold on…

For a really-not-sinewave signal, multiplying V RMS and A RMS won’t make a lot of sense.

Ideally, you’d calculate instantaneous power p = v * i and then calculate average power P = avg(p). But this is much easier said than done.

You mentioned 400kHz and 1% duty cycle. That’s a 25 ns on time, which is pretty short. Direct measurements are likely to be difficult…

You also mentioned 400 V p-p input and 270 V p-p output; what are you measuring with? I’m guessing a voltage probe and oscilloscope (that’s what I’d use!) - but are they fast enough to measure what’s really going on?

[bdunham7]

I have a 400 kHz signal with about 1% duty cycle, producing a roughly 400 Vpp input on and 270 Vpp on the output. My current probe is rated for upto 800 kHz so I am assuming that signal should be fine for accuracy. However, my 270 Vpp shows 30 V RMS. Seems like a discrepancy there. Is it valid to calculate the average power using V_rms X I_rms?

You have asked various aspects of your question multiple times and you aren't going to get good specific answers without being clear about the details of your circuit and signal.

Are you saying that your signal has a repetition rate of 400kHz and a duty cycle of 1%?  So it is a 25ns pulse every 2.5µs?  And 400V_p-p input and 270V_p-p output (whatever input and output are)?

If so, your current probe is not even close to being appropriate for the task and if you are using standard 10X oscilloscope probes, those are only hanging together because of the low duty cycle.  What kind of current is involved?  This sounds like a pretty high energy circuit with a BW of ~50MHz or so. 

[zapta]

800Khz bandwidth for 400Khz 1% pulses doesn't sound sufficient.

[zapta]

If your load is resistive than power = vrms * irms is right. Otherwise you measure VA and not power.