DW3000 TOF very large

When I tested DW3000 on STM32, the official process of DW3000 I transplanted, the TOF time obtained when using ds_twr_responder was very huge, and sometimes negative numbers appeared. What is the reason?

poll_rx_ts:761720209479
resp_tx_ts:761777733633
final_rx_ts:761822478535

poll_rx_ts:762614756463
resp_tx_ts:762672280577
final_rx_ts:762717025479

poll_rx_ts:763509316743
resp_tx_ts:763566840833
final_rx_ts:763611585790

poll_rx_ts:764403876487
resp_tx_ts:764461400577
final_rx_ts:764506145558

poll_rx_ts:765298422975
resp_tx_ts:765355947521
final_rx_ts:765400692551
This is part of the data.

The numbers are in local responder time domain. They can be from 0 to 40-bit numbers.

They do not makes sense on their own, but what makes sense is the difference in between them and math equation of diff of TS of initiator and responder where ToF will be calculated from.

you need to see the full ToF formula and understand that initiator and responder runs in different clock domains.

Thanks for your answer, I know what it is.