BCC MC33771C measurement timing Hello Team, We are working with the MC33771C Battery Cell Controller (BCC) and have observed a difference in voltage/temperature measurement readout timing when using different read intervals. Test observations: Configuration: Same BCC configuration with 4 slaves. Read interval: 10 ms: Reading temperature data from all 4 slaves takes approximately 40 ms. Read interval: 100 ms: Reading the same temperature data from all 4 slaves takes approximately 7.7 ms. We observe that the total measurement readout time is significantly higher when the read interval is set to 10 ms, even though the configuration remains unchanged. Could you please help us understand the following? Why does the measurement readout time differ significantly between the 10 ms and 100 ms read intervals? Is there any dependency between the read interval and the ADC conversion, measurement update, or BCC communication timing? Is there any recommended configuration or procedure to achieve consistent and optimized voltage/temperature readout timing when using 4 BCC slaves? Are there any specific timing constraints or synchronization requirements that we should consider when reading measurement data from multiple MC33771C devices? Please let us know if any additional configuration details or measurement logs are required for further analysis. Re: BCC MC33771C measurement timing Hello Nirmal,
Thank you for confirming the ADC configuration. This is a critical clarification: with ADC_CFG[AVG] = 0 and 16-bit resolution, the hardware tEOC is only ~520 µs. The 40 ms readout time you observe at 10 ms intervals cannot be caused by the ADC conversion itself, it is almost certainly caused by the firmware measurement sequence.
With tEOC ≈ 520 µs, data is ready well before either interval expires. The difference in observed readout time therefore comes from how and when the firmware triggers SOC and reads results relative to the interval. The most common causes of this specific symptom are:
1. SYS_CFG1[CYCLIC_TIMER] is active at 10 ms If the cyclic timer is set to 10 ms, the device autonomously triggers a cyclic conversion every 10 ms. However, cyclic conversions do not update the MEAS_xxxx measurement registers — they only update OV/UV/OT/UT fault flags. If the MCU firmware waits for a cyclic trigger before issuing an on-demand SOC, the effective measurement cycle gets coupled to the cyclic period, adding up to 10 ms of latency before each read. At 100 ms cyclic period, this latency disappears because the MCU issues on-demand SOC independently. Please verify the value of SYS_CFG1[CYCLIC_TIMER] and, if active, consider disabling it or decoupling your MCU read sequence from it entirely.
2. Fixed interval delay inserted before reading (not waiting for EOC) If the firmware issues SOC, then waits the full read interval (10 ms or 100 ms) before reading — rather than polling the EOC flag — the behavior would be:
100 ms: SOC → wait 100 ms → read (7.7 ms comm) → measured ≈ 7.7 ms ✓ 10 ms: SOC → wait 10 ms → read (7.7 ms comm) → but then if the cycle repeats and timings compound, the effective measured duration grows.
3. SOC issued per slave rather than as a broadcast If the firmware triggers SOC individually to each slave and enforces the read interval between each, the total for 4 slaves compounds to a multiple of the interval.
The correct sequence for 4 slaves is:
- Issue a single broadcast SOC to all 4 slaves simultaneously. - Poll the EOC flag (or wait a minimum of 520 µs at 16-bit, AVG=0) — do not use a fixed delay equal to the read interval. - Read all 4 slaves sequentially after EOC is confirmed.
BRs, Tomas Repeat at your desired interval. Re: BCC MC33771C measurement timing Hello @TomasVaverka Thank you for the detailed explanation. In our configuration, we are using the following settings: ADC_CFG[AVG] = 0 ADC Resolution = 16-bit Number of BCC slaves = 4 We are using the same configuration for both test cases (10 ms and 100 ms read intervals). However, we observe the following timing difference: 10 ms read interval: Approximately 40 ms to read temperature data from all 4 slaves. 100 ms read interval: Approximately 7.7 ms to read temperature data from all 4 slaves. Since the ADC averaging and resolution settings are identical in both cases, could you please clarify why the conversion and measurement readout timing differs significantly? Is there any dependency on the SOC trigger interval, ADC conversion status, or EOC flag handling that could explain this behavior? Could you also suggest how we can achieve consistent measurement readout timing with a 10 ms measurement interval? Re: BCC MC33771C measurement timing Hello Nirmal,
The timing difference is most likely caused by your ADC_CFG[AVG] averaging setting.
When averaging is enabled, tEOC (time from SOC to data ready) scales proportionally with the sample count. For example, at 16-bit resolution with AVG=64, tEOC ≈ 32.5 ms — far exceeding your 10 ms read interval. This explains the observed behavior:
At 100 ms interval: Conversion is already complete when the MCU reads → only communication time is measured (~7.7 ms).
At 10 ms interval: The MCU must wait for conversion to finish → measured time = remaining tEOC wait + communication ≈ 40 ms.
A few additional points relevant to your questions:
1. Cyclic mode (SYS_CFG1[CYCLIC_TIMER]) does NOT update MEAS_xxxx registers — it only updates fault flags (OV/UV/OT/UT). For actual voltage/temperature readings, on-demand SOC is required. Averaging is also not available in cyclic mode.
2. Always wait for the EOC flag before reading results. Do not use a fixed delay equal to the read interval unless it is guaranteed to exceed tEOC.
3. For 4 slaves: broadcast a single SOC to all slaves simultaneously (all convert in parallel, completing at tEOC), then read sequentially. At 16-bit, AVG=0, the minimum total time is approximately 0.52 ms (tEOC) + 4 × 0.7 ms (TPL comm) ≈ 3.3 ms.
4. For a full timing optimization framework, refer to AN12637: MC33771C and MC33772C Based System Timing Optimization (available on nxp.com, login required).
BRs, Tomas Re: BCC MC33771C measurement timing Hello @TomasVaverka , I am using broadcast message for SOC, still i am facing timing difference.
View full article