Artificial Data found when reading AC voltage

[moonlotus]

I'm using card 6210 to scan AC voltage using NI sample code = ContAcq-IntClk.prj

This is the output when AC is switch on.

Afterwards, when I swith off immediately. I found the output becomes like this. Look at the plotting in red circle. Please advise what was happened.

[SteveD123]

Okay, some basic questions:
1. What are you measuring? I count 50 spikes per second, is this some mains-derived measurement?
2. What is the circuit configuration?

3. What, EXACTLY,do you mean by "when I swith off immediately" ? If there is a mechanical switch involved, it could be simply contact bounce.

4. Can you capture the signal with an oscilloscope?

[RobertoBozzolo]

Additionally, shown voltage is out of measutring range of the card (10V max): which is the level of the signal you are acquiring? I mean the peak level of the AC voltage, not the average or RMS. Beyond being unable to correctly measure the signal, you may damage the board if the voltage exceeds maximum safety voltage (±30V powered, ±20V unpowered).

[SteveD123]

Robert

How is that even possible? Surely, if the card is set to measure between -10 V and +10 V, does that not give ADC binary values between 0x8001 to  0xFFFF (two's complement)?

Whatever voltage is sent to the card, it cannot read above 10 V, 0xFFFF unless there is a problem in the subsequent floating point arithmetic. Otherwise, this is the equivalent of "My ADC goes up to 11"!.

Can an overloaded NI ADC break the confines of a short int somehow?

If you have come across a similar situation where there was not a problem in the scaling arithmetic, please let me know, because it would deeply challenge my understanding of what is going on!

Cheers

SteveD

[moonlotus]

1. I'm measuring power supply, 50Hz, we have some hardware to drop the voltage at ratio of 100, for example, to fit the card range of -10V to 10V.

2. I'll post another comparison pictures later. Our users measure the same input of Motor Starting / Transformer Switching using NI Card 6210 & another device.

[moonlotus]

Sorry I didn't realise that it was out of range. Will check this again later. We build a simple circuit to drop the voltage input.

[moonlotus]

This is the waveform I captured using NI card 6210.

Zoom in...

At the same time, I use another power meter to record the waveform, please compare the Current Waveform using both measurement tools.

[SteveD123]

Okay, this changes everything.

First: 11kV and 120 amp current spikes mean that you are in seriously lethal territory. Stop right now, and make sure that someone certified to do so ensures that your set-up is safe. This is high voltage, and must be certified by someone who fully understands what is going on.

The traces you post are of a switch-on event. How do these relate to your original question concerning a switch-off transient? What is your question now?

Switch noise: If this is modern equipment, we are talking about a vacuum circuit breaker (VCB) or air circuit breaker (VCB). The switching transient will be affected by the mechanical configuration of the breaker.

If there is an active element, e.g. a modern motor drive, it could be operating high-frequency PWM. This can feed noise everywhere, but also you will not get a true picture of what is going on unless your instrumentation can resolve it, which means sampling at a higher frequency.

If your instrumentation is competent, there probably will be more than 1 stage of isolation between the HV you want to measure and the DAC. Often these have limited frequency and phase response, and you might not get a true picture, unless you know the frequency response of your instrumentation. Do not attempt to measure it yourself.

However, I am concerned that, given your earlier instrumentation error, you do not understand what is going on, or the risks involved. I must repeat my advice that someone who understands 320kW high voltage systems is consulted in person on your site before you kill or maim yourself and others.

Remember that advice from the internet is worth exactly what you paid for it.

[moonlotus]

My company is certified to do the testing, don't worry.

My question is, I found a lot spike when we do testing on high voltage switching. (Especially when the voltage/current with a lot harmonics/noise)

In lab, I can simulate the similar pattern of spike (abormal point / artificial point) when I do switching on power supply. Thus, someone help me to design a drop down circuit which can transform 240V to 2.4V.

Engineer has confirmed that their circuit has transform any high voltage/currnet to range of -10V to 10V. I just rescale it back to corresponding voltage when plotting.

[moonlotus]

My question is, can NI card read a switching transient correctly like what the Power Meter read? It consist high frequency of noise/harmonics.