Cyclone IV goes into an undefined state

--- Quote Start ---

Do you want to say, the FPGA is working correctly if VCCINT isn't delayed? If it's working neither, I expect an elementary hardware problem, e.g. unconnected exposed pad.

--- Quote End ---

Sorry for my English:

The delay is only an test

MY PROBLEM IS that nCONFIG signal goes low. Not only, but other signals (i.e. TDI, TMS) go low when theoretically they should keep on staying high.

Can I get information on power supply?

I used these components:

"MICROCHIP - MCP1825S" for 1.2Volt.

"NATIONAL SEMICONDUCTOR - LM1117T-2.5" for 2.5Volt

"LM2937-3.3" for 3.3Volt

this component can be good?

--- Quote Start ---

Can I get information on power supply?

--- Quote End ---

As you're reporting "All these voltages are good", I don't see a need to discuss about the involved power supply. I have to assume, that the measurement is correct.

--- Quote Start ---

Not only, but other signals (i.e. TDI, TMS) go low when theoretically they should keep on staying high.

--- Quote End ---

You mean, they are forced low by the FPGA? NCONFIG, TDI and TMS should have pull-up resistors, so are you sure, that the resistors are still supplied with 2.5 respectively 3.3V?

YES, NCONFIG, TDI and TMS are forced low by the FPGA and All voltages are good.

Furthermore, even NCONFIG, TDI and TMS have pull-up resistors.

nCONFIG is connected via pull-up to 3.3V and TDI ,TMS are connected through pull-up to 2.5V