sorry there was a trade show here last week and I've had some visitors so my emails have slipped, I am now trying to catch up as best I can but do have some backlog to get through.
Sorry to hear you are having some problems, (it is a pack from 20Ah ouch cells BTW, not 26650s) I have a few comments:
The voltage measurements you last took do suggest 1 cell is at a lower state of charge than the others, that could be a bad cell, it could just be slightly higher self discharge on 1 cell that is not being addressed when charging, due to the charge voltage not being high enough for the BMS to do anything. All other cells would need to be at 3.65V plus before a low cell is brought back into balance. The charger you are using only brings the high cells up to approx 3.55V, that is not enough. A charge voltage a little higher than the standard 87.6V will help help to bring any low cells up to 3.65V. If you can charge the low cell using a single cell charger or power supply that would be good. Regarding Headway BMS, Headway do not and never did make their own BMS, they have bought from different companies over the years and their latest supplier are much better than some of the earlier BMS they sourced so my contact in Headway tells me. They had a lot of BMS issues in the past with units bought from other suppliers, so no I do not buy anything from Headway and the BMS I supply are 1 of the most reliable products I do supply, very few issues have been reported, so there is nothing so far that suggests a BMS issue, so there is no need to do any mods to it or the wiring of the pack.
Also, you really need to have some way to measure the current draw accurately, if Lyen didn't measure the CA shunt value and as far as I know he doesn't, that number is no more than a guess. You need to either accurately measure the R Shunt value and also ensure the CA wires have been fitted to the controller PCB correctly, or get an independent way to measure the current, such as a watt meter and use that to discharge the pack. Watt meters are generally 60V max, so you can only measure the discharge current, you cannot measure the total pack voltage or you will fry it. Bottom line, the Ah figures mean nothing right now, you need a reasonably accurate way to measure the current and Ah used, preferably at home with a test rig. There are just too many unknowns and variables in your present setup.
If you can charge the low cell up to 3.65V and then discharge the pack into a reasonably accurate test system, preferably at home in a controlled way then we can be sure if there is a bad cell or not. Once the BMS trips, all cells should be checked to determine the cell voltages of course. The CA should also be ok, but it would be good if at least the current measurement has been verified to be reasonably accurate.
I have replied to your email, can we please continue there if that is ok.