hi...

      Can any one please suggest the best possible way of generating a 2.048MHz square signal with 50% duty cycle with a better accuracy, using PSOC3.........do i need to use any external hardware for getting good accuracy ?

How accuray you want it to be?

I think the psoc3 IMO is +-1%. You need to use ceramic resonator or external crystal. And crystal would be better than ceramic resonator.

 may i use XRT8000 device to generate an accurate 2.048 MHz signal

The XRT8000is a PLL, so it needs a stable source frequency. So you might be better of by generating this from the PSoC directly. Use a crystal as frequency source. Configure the PLL to the highest possible speed.Then, use a NCO (see http://www.cypress.com/?rID=39408&cache=0 for the component) and configure it to generate the frequency you need.

How exact and stable must the frequency be? 2.048MHz is somewhat unusual.

Another way might be to get a crystal which is a multiple of 2.048MHz, and use it as clock source in the PSoC. That way, you can spare the NCO, und use a simple divider to generate your output.

2.048 Mhz commonly used on GPS, with 10^-12 accuracy requirements. You cannot

achieve this with a simple xtal placed on PSOC.

Note the IMO is totally inadequate for these types of accuracy.

Regards, Dana.

 hi danaa

                i found that XRT8000 is a dual PLL chip that generate 1.544MHz (E1) or 2.048MHz (T1) if we configure it in Reverse slave mode with an input frequency of 8KHz.....

               so anyway,  we can generate 8 KHz signal accurately.........then ,why cant we use this chip to generate my 2.048MHz signal....... 

              does it have any problem in interfacing it with SPI of Psoc3 ?

here iam attaching the datasheet of XRT8000....you can go through it...... thank u.

 hi danaa

                i found that XRT8000 is a dual PLL chip that generate 1.544MHz (E1) or 2.048MHz (T1) if we configure it in Reverse slave mode with an input frequency of 8KHz.....

               so anyway,  we can generate 8 KHz signal accurately.........then ,why cant we use this chip to generate my 2.048MHz signal....... 

              does it have any problem in interfacing it with SPI of Psoc3 ?

here iam attaching the datasheet of XRT8000....you can go through it...... thank u.

 hi hli....

           other than 32.768MHz...i dont think the multiples of 2.048MHz crystals available in the market.......but ECO of PSOc3 will accept a maximum of 25 MHz.......so, unless the crystals less than or equal to 16.384Mz  available in the market, i cant proceed in this approach.....

     May i know, does any one from 4.096MHz, 8.192MHz, 16.384MHz crystals available in the market......

 If the accuracy of your 8khz signal is good enough, then the output from the XRT8000 should be Ok, but remember than there are phase noise/jitter after the PLL.

The SPI clock doesn't need to be very accurate, so there should not be problem between the Psoc3 communicating with the XRT8000.

The 8000 does not specifically state serial interface is any industry

standard, but looks like should work with SPI. You will have to

evaluate timing and API applicability.

Should there be some issue using SPI, simple bit banged GPIO pin

routine would suffice, given you are only sending simple commands

over the interface.

Regards, Dana.

You can check the page 14-16 of the XRt800 specification and compare it with that on the SPI data sheet of PSOC to get the correct mode.

 I think it should be mode 0, LSB first 8 bit.

But becarefull with the 1st byte as you need to send the R/W as bit0 and a0-a6 as bit1 to bit7.

 hi...

            i found 4.096MHz crystal available in market............i am looking to use it as XTAL, and with the help of frequency divider i hope i can generate 2.048MHz signal................

                                                                   what do you say guys?

That should be OK.

You don't need external divider, you can use the Filip Flop inside the PSoC3.