Comments by "SeanBZA" (@SeanBZA) on "DENKI OTAKU" channel.

  1. Heat transfer compound is a 2 component mix, that is poured in at final assembly, to fill all the voids. Stays flexible and somewhat sticky, as it will slowly flow with heat to help ensure all the voids are filled. Similar to the thick thermal pads used in equipment, in that it ensures better heat transfer from hot areas to the case, though it is not as good as a proper aluminium heatsink, but as it is cheap, insulating, and also helps with minimising creepage on the board, it is used.
    38
  2. I would say the steps are to keep the power input constant during starting, as the power needs of the motor change because of the need for increased slip during the initial power application, and the steps are there to reduce the CPU need for a smooth sweep, as the original designs used lower speed processors, so offloading most of the power control into discrete power steps was needed, so as to not cause excess slip in the motor and thus poor accelleration, and too fast a ramp would result in a motor stalling. thus the constant steps, so that a driver could keep a constant pull away and not run the risk of a stall, which would mean a stop and a slower ramp up again, and a possibility of the motor generating an overcurrent or tacho feedback error. remember this was originally running likely on aZ80 or similar processor, with large parts being done by other processors as well, and the designers had both limited memory and limited performance. 400Hz was running the GTO devices at the high end of the range, as they, for power devices at the time, needed 2 or more same size power switches for each of the 6 legs of the drive, and GTO devices need you to have a very beefy current source and sink, as you need to turn it on fast, so high current at a relatively low voltage of around 5V on the gate, and turn off you need a current about what the device is conducting, and around -10V, to pull the gate off hard and fast enough to switch off. At least this was better than earlier thyristors, where turn off was done using LCR circuits, and a second equally rated thyristor to pull the anode negative for long enough that the first one would be able to turn off, and then the second one, due to the high power resistor providing it with power, would stay on till the first fired again to turn it off. High static current, high power loss, the GTO made this power use so much lower, at the expense of needing 4 isolated relatively high power supplies, 3 for the upper bridge, and the lower one having a common cathode connection. Instead of 500W of power dissipation now down to under 100W at idle. Add this up per axle and the power use becomes considerable, plus higher frequency, and the ability to do regenerative braking with much less processing power, and you can see why they, and now the IGBT, won out.
    2