mugicha
Newbie level 2
I'm doing a design on a development board in Simulink. I have Realtime Workshop and all that fancy stuff so I can simulate my design in Simulink and then generate a bit file (with ISE working in background) and then actually download this design into the FPGA on the board and run it in hardware. This system works really great, its amazing actually how closely the hardware performance matches the Simulink simulation output. However, I am having a problem with the DAC on the board, but I'm not sure if my problem is just a lack of understanding of how DACs work or what.
Here's what I'm finding: when I have a stream of baseband symbols going into the DAC and put a scope on the output the waveforms I see are very similar to my Simulink simulations. The peak to peak voltage of the output uses pretty much the full range of the DAC's output, and my signed 16 bit fixed point numbers get turned into signed voltages centered around ground at the DAC's output. Perfect, just how I expect it and want it to operate.
HOWEVER, when I put a constant value into the DAC, no matter what that value is, I only get out 0 volts, or something close to it when I look at the actual DAC output waveform on a scope. In simulation the DAC behaves how I expect it to, in other words a constant value puts out a constant DC value. If I put in something close to the max input value into the DAC, then in simulation the DAC output is close to the positive rail. If I put in a max negative constant then I get something close to the negative rail. Also, these constants are identical in value and fixed point scaling to the min/max values of my baseband symbols, which as I said both simulate correctly as well as function correctly after getting turned into hardware. Therefore, it would seem safe to assume (to me at least) that if values such as the ones I'm putting in come out correctly when the input is time varying, they should come out correctly when the input is constant.
Am I doing something wrong? Is a DAC normally unable to output a constant DC value? Is there something that it likes about a time varying input such as my baseband symbol stream that it doesn't like about a constant? Maybe the Xilinx tools are optimizing out the constants? I'm confused....help anyone?
Here's what I'm finding: when I have a stream of baseband symbols going into the DAC and put a scope on the output the waveforms I see are very similar to my Simulink simulations. The peak to peak voltage of the output uses pretty much the full range of the DAC's output, and my signed 16 bit fixed point numbers get turned into signed voltages centered around ground at the DAC's output. Perfect, just how I expect it and want it to operate.
HOWEVER, when I put a constant value into the DAC, no matter what that value is, I only get out 0 volts, or something close to it when I look at the actual DAC output waveform on a scope. In simulation the DAC behaves how I expect it to, in other words a constant value puts out a constant DC value. If I put in something close to the max input value into the DAC, then in simulation the DAC output is close to the positive rail. If I put in a max negative constant then I get something close to the negative rail. Also, these constants are identical in value and fixed point scaling to the min/max values of my baseband symbols, which as I said both simulate correctly as well as function correctly after getting turned into hardware. Therefore, it would seem safe to assume (to me at least) that if values such as the ones I'm putting in come out correctly when the input is time varying, they should come out correctly when the input is constant.
Am I doing something wrong? Is a DAC normally unable to output a constant DC value? Is there something that it likes about a time varying input such as my baseband symbol stream that it doesn't like about a constant? Maybe the Xilinx tools are optimizing out the constants? I'm confused....help anyone?
