Whenever an electronic circuit is first energized, transients occur in current and voltage waveforms. These start-up transients can affect the electrical and thermal behavior of components and circuits with serious reliability, EMI, and random effects. Try to characterize how your circuits start and stop.
Have you detected anomalous behavior in your design, especially during the start-up of the system? Unintended resets? Failure in some components? Random events?
If you find those situations try to measure how your circuit wake-up during the start-up of the power supply. Perhaps you will discover hidden events. Let me show you one example with a small 19V power supply part of a more complex design in a consumer product.
The start-up of the product will be analyzed using a 1GHz scope (Agilent DSO7104B). A current probe (TCP202) is used to measure input current in channel CH1. A voltage probe is used to measure output voltage in channel CH2. The complete setup is shown in Figure 1.
The screen in Figure 1 is a typical view when triggering with channel CH1 (current) in AUTO MODE with 57mA approximately in trigger level. Input current is a quasi-sinusoidal 50Hz signal with 313mA peak to peak in amplitude. Output voltage is 18.4V DC with some ripple.
How is the start-up for this product?
To characterize the start-up for this system, the probes are fixed in place but TRIGGER is configured for SINGLE mode: only one acquisition when triggering conditions are satisfied and then screen is stopped.
With the power supply in OFF condition, the scope screen is empty and the trigger circuit is waiting for some event crossing the desired level (approximately 360mA in our example).
When the power supply is energized, input current has a transient and the captured data during that time is shown in the scope screen (Figure 2).
Note the fast transient in current at the moment of start-up in region (A). It looks like a near vertical line similar to one of the vertical gridlines.
This is a very short transient with high di/dt with great conditions to create radiated EMI, crosstalk, immunity problems, and perhaps failure in some components
(e.g. semiconductors).
If the transient is really fast, you will need a high bandwidth scope to be able to see it.
The measured peak current is out of the screen, more than 4A in our scope, and far from the nominal values in our design. Sometimes the peak of current will be really big, for example, if your input circuit includes a big capacitor (i.e. electrolytic).
In regions (B) and (C) we identify the start-up of the system (approximately 100ms in duration). Then, in region (D) the nominal current and voltages are measured as in Figure 1.
Some interesting information is found with a zoom in region A as shown in Figure 3.
The short transient close to an ideal vertical line in Figure 2 top trace (region (A), offers a strong HF activity associated to parasitic in layout, cabling, components, and the topology of your circuit.
You can try to correlate failures in your system with these kinds of transients. If you are able to see them, you will be able to try fixes. There are a number of techniques to significantly reduce the amplitude and duration of start-up transients: dimension of components, filters, inrush current limiters, intelligent switching strategies, etc. Learn about those methods and your products will be more reliable and successful.
My final advice: do not forget to characterize the start-up operation of your designs to identify sources of problems.
