MM9Z1_638 LOW PASS FILTER

[a852013777]

Hi Nxp Partner

I don't quite understand how to write the LPF_Ax (low pass filter coefficient Ax) of MM9Z1_638.
Suppose I want to write A0=0.25, A1=0.5, A2=0.25, what should I write in LPF_Ax?

Is there a ratio relationship or a conversion formula?
Low pass filter A0 should be a floating point number that cannot be filled in integer format LPF_AX

影像 (3).png

Br,

Mike Liang

[TomasVaverka · Accepted solution]

Hi Mike,

The use of the FIR LPF is optional and allows to fine tune the overall frequency characteristics of the acquisition. In most cases this FIR filter is not used/bypassed.

One big tradeoff of the FIR low pass filter is that it will significantly increases the latency of the acquisition. This is relevant if different VSENSE channels are used/switched.

See attached PDF for:

- Digital signal processing/filtering p. 107
- Latency p. 115

The FIR LPF can be enabled/disabled (bypassed).

#define ACQLPFEnable()   (B_ACQ_ACC1 = B_ACQ_ACC1_LPFENM_MASK| B_ACQ_ACC1_LPFEN_MASK)     //!<  enable low pass filter for CSENSE & VSENSE channel

#define ACQLPFDisable()    (B_ACQ_ACC1 = B_ACQ_ACC1_LPFENM_MASK| 0)    //!<  disable low pass filter for CSENSE & VSENSE channel

If enabled the FIR coefficients and number of used coefficients can be programmed. The device already has a default set of coefficients programmed.

The code below is representing the default values (code is actually not required, as reset values of the respective registers).

B_ACQ_LPFC = 15;
B_LPF_A0 = 0x03DB;
B_LPF_A1 = 0x87B4;
B_LPF_A2 = 0x86B1;
B_LPF_A3 = 0x1580;
B_LPF_A4 = 0x04DF;
B_LPF_A5 = 0x9F70;
B_LPF_A6 = 0x81F5;
B_LPF_A7 = 0x5195;
B_LPF_A8 = 0x5195;
B_LPF_A9 = 0x81F5;
B_LPF_A10= 0x9F70;
B_LPF_A11= 0x04DF;
B_LPF_A12= 0x1580;
B_LPF_A13= 0x86B1;
B_LPF_A14= 0x87B4;
B_LPF_A15= 0x03DB;

I attached an Excel calculator which shows the impact of the individual filter stages and allows to evaluate own values. I guess it is what you are looking for.

BRs, Tomas