Before I get into the details below, since this is not a controller-specific problem, is it ok if I split your posts and my replies off to a separate troubleshooting thread? (best to move them to your build thread if you have one).
1.4v is VERY low, below typical minimums, for a LiFePO4 cell. 2.5v is typically basically empty, and setting above that point, more like 2.8v or higher, is a better idea as it will have less stress on it. 2.1v or so is the lowest I would use if I absolutely had to...but I would use 3v myself to improve pack longevity by reducing stress.
Similarly, 3.4v is really "full" for LFP cells; the capacity between there and 3.6v that they are usually charged to (for balancing) is so small it would be better to leave that and charge lower, for less cell stress and greater longevity. So 90V as overvoltage is high; even at 3.65v/cell (typical balancing / full voltage) for 24 cells, that's only 87.6v.
Less stress means less change of each cell vs other cells over time, and they'll stay balanced better over time, too. (the further they are discharged, the more stressed and unbalanced they will be each time, and the longer a charge cycle will take trying to fix that).
I recommend looking at the specification sheet for the cells you used; this is available from the manufacturer of the cells. (if they don't have these, then you have no certain data about / specifications of the cells and will have to determine all their limits experimentally to be able to properly setup your BMS and controller for them--it also probably means they don't hold their cells to any standard and so every cell could be radically different from every other, so any pack of cells won't be matched in characteristics to each other, and the pack won't perform as well as it would if they were all the same (unless the pack builder tests every cell and discards all cells that don't match each other, which can mean buying many times the amount of cells the pack actually needs).
For a 24s pack, at 2.8v per cell, that gives a pack-level LVC of 67.2v (at the BMS); I recommend something higher at the controller (2.9v/cell 69.6v, or 3v/cell, 72v), but the 67.2v at lowest.
If the controller is set for 67.2v LVC, then set the BMS for a tiny bit lower, like 66.5v or something (whatever the per-cell voltage is for that), so that the BMS won't shut off before the controller has the chance to just release the load. Otherwise you're going to have power cuts on every throttle up when the battery is near empty, because the BMS will cut all power.
If you can provide details about the cells, pack build, BMS, etc., it may help diagnose the problem and suggested settings or repairs to correct it.
If there's no voltage out of the BMS to the controller (actually measured with a meter external to the battery, not just displayed on the BMS program), then the BMS has shut off to protect the cells.
That means that some parameter in the BMS was exceeded (or there is a fault in the BMS itself). If the BMS software doesn't show you what that is, you'd have to measure things yourself during these events, using multimeters or cell monitors, and watch it realtime under load to see what happens. If the event that cuases the shutoff is short enough, you may not see it realtime because of the way meters read and display their data. It might require a reasonably fast multicell logging device to catch, and then see in the logs what happened. This isnt' normally needed; there are other ways to check things.
You'd need to look at any logging info the BMS takes. If it doesn't log the maximum cell voltage difference, or the minimum cell voltages, etc., you'll have to use other methods to find the problem. These smart BMSs generally have some form of logging, even if it is only a simple record on screen of what the highest and lowest cell voltages were, which cell that was, what the highest and lowest currents were, etc.
But first I'd recommend setting the cell limits (votlage, current, etc) to something closer to what such cells are normally designed to handle.
The display you see is probably massively delayed from realtime, and updated very few times per second (or even slower). First the BMS has to read the cells, then process teh data, then encode the data for sending over BT to your phone, then your phone processes that incoming BT stream, sends it to the app it's for, the app has to decode the data stream, process it, and then display it. So it's not unusual to not see voltage sag in some cases like this, especially if the sag is short. (always if the sag is shorter than the display update interval).
Where is the 78v measured, and how?
If it is just displayed on the BMS app, that means the BMS reads that voltage. It doesn't mean it's actually present on it's output.
If it's measured at the controller input, then the controller *is* turned back on (or rather, it has power and could be turned on, if whatever turns it on is engaged).
Also, when disconnected from a load, a BMS's FETs can show a voltage on their output as a "ghost" or "leakage" but when the load is applied, this voltage drops to usually nothing.
If the BMS won't turn it's output back on until it's placed on a charger, it probalby means a cell or cells dropped so low in voltage that the BMS's settings indicate the cells would be damaged if further discharge was allowed without recharging them first.
At some point those cells will probably drop far enough that the BMS will even refuse to charge them, because a cell that goes too low is permanently damaged and can be a fire hazard.
