S32K3 LPCMP module Latency issues

[yinq]

Hi,

I have encountered some problem of S32K3 , LPCMP module.

Which position may Tper appear in A, B, C, D, and indicate the number of Tper in this position?

[zhanguiyang]

How to handle the cross clock domain transition from window function to filter when using mode # 4A? Because in # 4A mode, the signal input to the filter is asynchronous with the filter clock window/sample.

[zhanguiyang]

How to handle the cross clock domain transition from window function to filter when using mode # 4A? Because in # 4A mode, the signal input to the filter is asynchronous with the filter clock window/sample.

[PetrS]

Hi,

If CCR1[SAMPLE_EN] is set to 1, the low-pass filter samples COUTA on each positive transition of the sample/window input. Window signal must be of minimum 4 bus cycle pulse for window mode to function. The output state of the filter changes when all the consecutive CCR1[FILT_CNT] samples agree that the output value has changed.

BR, Petr

[zhanguiyang]

When CCR1[SAMPLE_EN]=1, the window/sample signal acts as the clock of the filter, asynchronously with the bus clock?

If they are asynchronous, can window/sample signals count against COUTA directly?

[PetrS]

Hi,

If CCR1[SAMPLE_EN] is set to 1, at positive transition of the sample/window signal COUTA is sampled and filter counts specifies the number of consecutive samples that must agree before the comparator output filter accepts the sample as a new valid output state.

BR, Petr

[zhanguiyang]

So,is the working clock of the filter a bus clock? window_sample signal is not the working clock of the filter?

[PetrS]

Hi,

this is a feedback I got from Apps team.

When CCR1[SAMPLE_EN]=1, the window/sample signal acts as the clock of the filter, asynchronously with the bus clock? 

Yes

If they are asynchronous, can window/sample signals count against COUTA directly?

I can understand your mean. But you can refer the attachments which include the detail of filter, sample and window.

So, is the working clock of the filter a bus clock? Window_sample signal is not the working clock of the filter?

The LPCMP working clock of the filter is bus clock. Window_sample signal is the sample input and is not the working clock of the filter.

BR, Petr

[zhanguiyang]

In 4A mode, such as filt_ cnt=5,couta is output by the bus clock.but,when couta go in filter, the clock of the filter at this time is asynchronous window/sample signal. The first sampling count of couta by window/sample may have metastable state,How to solve this problem？

[PetrS]

Hi,

in this mode, the path from the analog inputs to COUTA is combinational (unclocked). COUTA is then input to Filter and COUT is sampled signal. A first sample will always happen at uncertain time and can be from 0 to period of sampling signal. This cannot be influenced anyhow.

BR, Petr

[PetrS]

Hi,

there will be 2 Tper clock synchronization in the Window function (point B) and ((CCR1[FILT_CNT] x CCR1[FILT_PER]) + 1) x Tper clock for the Filter function (point D).

BR, Petr

[zhanguiyang]

How to handle the cross clock domain transition from window function to filter when using mode # 7?