What TI DDCs interface directly to a TI DSP?

About GC4016, pass-band bandwidth expressed as a percentage of the DDC-channel output sample rate.

the maximum percentage for PFIR and CFIR are 150%, we want higher bandwidth about 160%. Can you tell us how to do this? Currently we are using ICS-554 data acquisition card, it uses GC4016. http://www.gefanuc.com/products/2052

Regards,

Prasanna.

I'd like to use a C5505 DSP chip with a DDC.  Can the GC5016 be used with that DSP?  The spec sheet doesn't say anything about the GC5016 supporting a bit serial output interface, but can it be interfaced to the DSP through some other way?  Would another DDC be a better choice?  Thank you very much for any help you can provide.

Hello,

The GC5016 has a 4, 8, 16, and 32bit parallel interface.  There is a GC5016 IO Application note on the TI website.  Another approach is to read the IQ data from the DDC output, or send to DUC input from logic that contains a FIFO, and the Frame, and IQ clock interface. 

The GC4016, GC5018, AFE8406 (GC5018 with integrated dual ADC) have bit serial outputs that work with the McBSP interface. 

The GC4016 also has MPU registers that can read back the IQ value for a specific channel. 

Regards,

radio Joe

Hey Radio Joe,

So you mention that the GC4016 has MPU registers that can be read for the IQ value.  I'm assuming that you're talking about the microprocessor output mode, right?  Just to confirm, the registers can be simply read through a DSP's EMIF port, right?  You would just have to make sure that the microprocessor mode (normal or wide mode) matches the number of EMIF data lines supported by the DSP.

Thanks for all your help.

Marc

Hello Marc,

The GC4016 output mode can read the DDC outputs to a microprocessor register.   Section 1.2.8 in the datasheet discusses the different modes:

   a) as an 8bit peripheral (standard uP mode) section 1.2.8.1 in the datasheet discusses the standard reads, and Output modes setup needed.

   This method is used by some customere as a self test output.

   b) as a 32bit peripheral (wide uP mode) Section 1.2.8.2  in the datasheet discusses the wide up mode. 

Regards,

Radio Joe

So for example, I'd like to use the newly released C5505 DSP with the GC4016.  The C5505 supports 8 or 16-bit EMIF, but does not have the McBSP.  Therefore, can I use the C5505 8-bit EMIF with the GC4016 in microprocessor mode 8-bit output?  Thanks again.

Hello Marc,

If the RDY pin is used as an interrupt, (check output control page, 98 - registers 16,17 - The RDY pin can be programmed to be 16 GC4016 clock cycles Wide.  You need to check if the DSP external interrupt can respond to a interrupt pulse that is (given a 10Mhz clock, 1.6ms in width).  The EMIF reading of the channels is based on Table 1-4, Table 1.5 .  If the GC4016 output is 20Khz, and we have 24bit DDC output data, for four channels, you would set page 96, read the 16 values, read page 97 and read the 16 values.  You would then assemble them into your complex buffer output.

Write page

read 4 8bit registes for each channel. 

write page

read 4 8bit registes for each channel. 

I think this is do-able, but you need to make sure you can read the IQ output fast enough over the EMIF port.

Assuming the DSP to GC4016 bus cycle takes 200nsec (Setup-Strobe-Hold)  per cycle. 

If we clear RDY, write PAGE, read 16 register, write PAGE, read 16 registers - 35 bus cycles, this is needed for updating 1IQ value for 4 channels.

1/(7us for HW + ?us for SW) would be the maximum DDC output rate.

AS long as you can setup the DMA buffers, you will need to select the GC4016 mode that results in the least copying of buffer data to the internal 32bit IQ DSP format.

You might post in the DSP user group to see if there are special conditions for using the DMA for 16 addresses 16-31.

Regards,

Radio Joe