We think the MCS6A is well suited for your application. It has a large 1GB or 2GB FIFO that is permanently read out during an acquisition and is able to buffer count rates of up to 100 MHz.

We have now a solution to synchronize the MCS6A clock with a external 9.5 MHz clock that can be obtained from the 76 MHz clock by dividing by 8. It is also possible to connect the 76 MHz directly, but probably better results can be obtained by an external 9.5 MHz clock. You have a choice how many bins per period you want, between 131 and 135 in steps of a half bin. For example we could select 131.5 bins for a period, it means 100 psec per bin. We could then fold every two periods together and this way obtain  a time resolution of 50 psec, as two neighbored periods are shifted by a half bin versus each other. The folding means just a resorting of the bins:
bin 1, bin 133, bin 2, bin 134, ..., bin 131, bin 263, bin 132.  The next 263 bins can be respectively added to the histogram and so on. The MCS6A software supports now the folding, also interleaved 50 ps resolution with synchronization to a 80 MHz Laser frequency is now possible.

As the length of the folded spectra is not a binary number, endless mode is not applicable, but very long sweeps can be done. For example a sweep length of 65488 bins corresponding to 2 x 249 periods. This can be done with a data length of 4 bytes, so the maximum continuous count rate is limited by the USB bus to 30 Mbytes/sec / 4 bytes = 7.5 MHz. The maximum burst rate is 100 MHz.

Are you scanning a target to obtain images? This could be done  using sequential mode. Tag bits can be used to mark the data with the beam position.

A good method to increase the time resolution using a "constant fraction" method, is to use a TD2000 with a very low threshold, and then record as well the falling as the rising edges of the pulses, and use a "software constant fraction" method as described in the MCS6A manual on page 5-8, Fig. 5.9 . The modification of the MCS6A.INI file for the FLIM application is described in chapter 2.5.5.