output data type of a simple discrete TF and its VHDL code generation

Question

Safiya 2015-5-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/218713-output-data-type-of-a-simple-discrete-tf-and-its-vhdl-code-generation

编辑： Stephen23 2015-5-28

Hi, i want to design simple filter using discrete TF block TF=(1/(1+ 0.1z^-1 + 0.1z^-2)) and writing its VHDL code to be synthesized in cadence. i checked with input of const block=2 with default data type inherited and ufixdt(1,16), it works well.

but i dont know when it is synthsized in cadence will it provide me input port where 4bit ADC can be connected (ADC is driveing block). my target input to the discrete filter block which accepts 4bit signal from ADC. how can a 4bit filter be designed. does the input constant block data type in simulink affects the synthesized code. e.g in simulink run, scope give same result either inherited ... or ufixdt(1,16) ? thanks

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Stephen23 2015-5-28

编辑：Stephen23 2015-5-28

Here is the original question, just in case Safiya decides to delete the text of their question as they have with some other questions:

Hi, i want to design simple filter using discrete TF block TF=(1/(1+ 0.1z^-1 + 0.1z^-2)) and writing its VHDL code to be synthesized in cadence. i checked with input of const block=2 with default data type inherited and ufixdt(1,16), it works well.

but i dont know when it is synthsized in cadence will it provide me input port where 4bit ADC can be connected (ADC is driveing block). my target input to the discrete filter block which accepts 4bit signal from ADC. how can a 4bit filter be designed. does the input constant block data type in simulink affects the synthesized code. e.g in simulink run, scope give same result either inherited ... or ufixdt(1,16) ? thanks

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Tim McBrayer 2015-5-26

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/218713-output-data-type-of-a-simple-discrete-tf-and-its-vhdl-code-generation#answer_180476

HDL Coder will generate code based off your Simulink model. If you provide a 16-bit input (either implicitly or explicitly defined) via your Constant block, the Transfer Function implementation will have a 16 bit data path. If you need a 4 bit data path, I would suggest changing your input to provide a 4-bit data path.

Whether the input is signed or unsigned is also important; in your case this depends entirely upon the data type of the 4-bit ADC on your board. Since you have specific hardware you are targeting I would modify your model so that your target subsystem is accepting data that is identical to the hardware that you need to interface with.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Safiya 2015-5-28

Thankyou! for your reply, Initially Input signal varies from 0 to 1.5V and then stables after 60usec to 20mV. steady state level is 20mV. actually this comes from ADC to be fed to Type-III compensator. thank you

请先登录，再进行评论。