Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
It depends on the chip and some times your power supply if and how you need to do it.
For the chip, the manufacture should be able to answer that.
Usually it shouldn’t differ more than about a diode voltage drop.
If you have create your voltages with linear regulator, it’s probably not a problem, because a linear regulator follows the main fairly accurate.
With switching regulator one approach is with diodes, see for ex. Motorola app note **broken link removed** chapter 1.2.3.
The main problem is to select diodes with the right forward voltage drop.
An other approach is to control the feedback voltage to the switching regultors.
Their are some chip from I think Maxim or Linear that manages sequencing power supply voltages.
Another approach is to use FET switches. For example, if your chip uses 5V and 3.3V, and wants the 5V first, then you put a FET on the 3.3V wire, and you use a comparator (or power-good signal, or some other means) to determine that your 5V is O.K., and only then you turn ON your FET.
The disatvantage over using diodes is that it is more complex, and some times costs more. The advantage is that the power dissipated in the FET is small (especially when something goes wrong with one of the supplies), compared to the solution with the diode, and hence the solution may be smaller and more reliable.
OK, thanks for your thoughts. I have already used 'diodes' to prevent an excessive potential difference between AGND and DGND of the component and MOSFET based switch circuit to switch on voltages in intended order. In addition I implemented a circuit which controls a rise time of power voltages in situation when too fast switch caused large in-rush current I could not afford (related to capacitance which has to be charged).
I agree that information about proper sequence of power voltages should be given by the manufacturer, but the true is that it is usually omitted. Moreover, when I faced the problem and asked the manufacturer for such information I got in response that the particular component is not sensitive for power supply voltages sequence. It took a long time to find out that sequencing is an issue.
For that reason I wanted to ask what in your knowledge would be a good/acceptable/unacceptable sequence to power such components. Are there rules of thumb or general precautions?