Measurement of the In-rush current

[Mr B]

Measurement of the in-rush current slope of a Switch Mode Power Supply when switched on:

Detail:

I am trying to measure the In-rush current slope of a Switch Mode Power Supply' Mains when switched-on. The current is measured on the mains live using a hall effect transformer (TELCON HTP25) and a DAQ (Labview U3-HV). The data are then passed on and processed in Labview.

The computation must be executed at high speed and provide a high level of precision. Although the in-rush current is extremely fast, it is not instantaneous and can still be measured. The attention of this experiment is to monitor the current demand and then decide whether to interrupt the mains or not.

Question:

- What kind of computation should I apply to the data acquired by the DAQ in order to obtain the relevant slope (dy/dx) of the In-rush current and further analyse it? I'm aware that I could simply use the min and max values and compute them but it is not a simple as it seems. If I use this method I will never be certain to obtain the min max at the correct time (or scan # here). The ideal analysis will happen as described in the attached picture. Here you can see where the most relevant info regarding the In-rush current is.

- I also would like to know, How I could obtain an X-axis related to Time rather than the scan number. Could I simply use the (1/scan rate) for the first sample and so on or is there a way of getting it done automatically?

- Additionally, would there be any methods to apply in Labview in order to process data in real time. I.e. As fast as possible? Is the real time module appropriate for this application? Even if the timing Vis, in the real-time module were used, I still don’t seem to be getting the loop speed I require for my application, which I estimated to be in the region of 100 μs.

Thank you very much for your time and help.

Regards,

Badis

[mikealder]

Why not post the vi or a screen capture of the code you are using to read the data from the card, what you are trying to do is quite easy but without seeing what you have done it is difficult to suggest options - Mike

[Henrik_Volkers]

To protect the fuse (line breaker) and/or the power supply  you should 'think' like a fuse: In the first interation it can be seen like a short time integral of the current, and the sum is represented in heat (there are much more effects, however this seems enough for this case)

If this is for a real world test (production line end test) I would not only go that way. I would use a hardware: R-CR as a short time integrator and a comparator that turn off the power. The trip voltage for the comparator could be set by the software..  and to slow down the process and prevent some board traces to change into smoke it's nice to add 10R+ in Series for the first start.....  OK nothing for higher volume, more for higher value equipment 😉 

[Mr B]

As you suggested, I've attached a copy of my VI. I’m feeling a bit relieved knowing that there’s way around my problem.

Well from now on I’m gonna try to think like a fuse !! J . My application is for a real world test, as you guessed. It is for a final test on a production line. I like you approach Henrik but I’m not too familiar with the tools you mentioned here. Could you please give me further details. For instance you mentioned a hardware R-CR, I don’t really know what this is! (short time integrator and a comparator that turn off the power).

Thank you

Regards,

Badis

[Mr B]

Hello everyone,

I've recently posted a subject in regard to the measurement of In-ruch current. I've had 2 replies but nothing since.  

Could anyone give me a tip or two, pleeease ?

Thank you

B

[Coq_Rouge]

Hi

Your application can not be run on a standard OS like windows. You need a real time target (with a real time OS)  for your application. Do you have access to a setup like this?

[johnsold]

Badis,

While it is possible to make something run that fast (FPGA or perhaps some of the real time modules), it is more common to use a hardware device to provide high speed protection and monitor the process from the software.  If you put a high speed comparator and power MOSFET switch of suitable rating in the circuit feeding your equipment being tested, you could interrupt the current within microseconds or less.  Monitoring of the voltages and currents could be done with standard data acquisition devices.  If the software took 10s or even 100s of milliseconds to detect the interruption and shut down the excitation, the device would still be protected. (You also do not need to worry about software bugs or priority timing issues using the combined hardware/software approach.)

Lynn 

[Coq_Rouge]

Im am not sure what you want to detect. From your picture, and description it seams to me that it is the current slew rate you want to act on. Not the absolute value of the current. Is this correct.

[Mr B]

Hi Lynn,

Thank you for your reply. Your approach seems interesting and similar to the one mentioned by Henrik's. I haven't designed something like this before but I will definitely give it a go. Would you agree that the use of a short time integrator is appropriate for my application?

Thank you.  

[Henrik_Volkers]

What do you want to protect: DUT or Fuse?

For analog circuit design: Horowitz:'Art of Electronic' 

It might be a good idea to simulate your protection.  Do you have access to PSPICE? ( PSST: wrong forum here, 'SwitcherCAD' from linear.com is fine for this task, runs from a USB stick, and costs 0$!)

You don't need the short time integrator, if your your break level has a good margin to the line breaker. Line breakers come in different tastes: The rated current and a break characteristik (that is a kind of integration).

Think about using a SSR to turn on and off the DUT. Worst case you have 10ms to break (since SSRs need a zero crossing of the AC current) but you can also turn on the DUT in alsways the same manner..  I assume AC Line testing...

And again If you put 10R in series (Increase R_i of the line source) you slow down and limit your inrush current, but you can correct that numerical...