How to test GC4016

Hello,

The GC4016 is a Digital Down Converter (DDC).  All of the functions are not testable just from a DSP. 

   - there are 4 digital input ports, normally connected to a 2s complement ADC MSB justified - this is not tested directly, can be tested with a signal processing test

   - there is a sync input that is used to synchronize the DDC internals, either the Sync Output is connected back to the Sync Input, or the device is internally synced, this is also tested with the signal processing test   

   - clock a 2.5v LVTTL clock has to be provided to the GC4016, this is tested by the device working, and the checksum test

   - the JTAG must be written to, to initialize the GC4016 out of JTAG, this is t ested by the device working

   - the MPU interface an 8 bit Asynchronous bus interface - this interface is testable over the DSP MPU bus, this is tested by the device working, and the checksum test

  - DDC output - there are several different modes for the I and Q DDC outputs.  These can be tested with a constant input, diagnostic tone, or a known input, tuned, filtered, and decimated, this is only tested by the device working

Assuming you have an external sync input, logic low for one GC4016 clock based on external data pattern.  The external data pattern can be input digitally, or through an A/D converter.  The clock to the GC4016 meets the clocking requirements, and the data and SyncIn meet the Tsu and Thi for digital data.

The JTAG interface is only used for interface test.   It is not used for register programming.

The TCK is connected to the WRB and the TMS, TDI are pulled high.  Normally the power on reset, for both the JTAG and core, that comes out of reset, without additional logic.    You would write at least 32 times to register 1.  This is used to bring the GC4016 out of JTAG mode.  (There is no TRSTB signal to force-reset the JTAG Tap controller).

The GC4016 program is used to configure the DDC in a specific mode, and the MPU registers are written over the 8Data, 5 address, WE#. (RD# is held low), and 8 data.  Look at Figure 1-2 the last 2 sections, the CE# is connected to DSP Chip Enable - active low.  The addressing in the GC4016 has a general page of registers, and then several pages of registers. 

Most of the registers are read write, following the address tables on page 49 of the datasheet.  On the TI website, one of the checksum tests can be loaded into the MPU registers using the DSP.  The registers can be read back from the GC4016 to the DSP.

Once the power supply is OK, GC4016 clock is running, the DSP loads the GC4016 registers for one of the checksum tests,   The section 5.14 diagnostic test is then followed to read the checksum registers, and see if they match the table 5-14 values for the expected checksum. 

Once the checksum test passes, we build a cmd4016 program for the customer configuration,

     - clock rate

     - which input port

     - CIC, CFIR, PFIR, Resampler rate - specific channel BW (using the existing 68% standard configuation filters, and bypass resampler ratio)

     - once the project is built, the output file is then loaded into the GC4016.  this can be tested either by loading a test-constant and setting the mixer frequencies to an inband  tone, or with an external signal, at the desired tuning frequency.

If you have an existing board, use the diagnostic program (only one of 4 are needed), it can be loaded from the register set on the GC4016 Developer's toolkit on the GC4016 portion of the TI website.

Regards,

Radio Joe