I know how it used to work and I'm not quite sure how it works today. I am quite certain, thoguh, that there are still some sort of packets moving in cables. They are just different than they used to be.
They most definitely did have. E.g if I had a LV transformer pulling power from MFE and feeding into a furnace then the transformer pulled 128 EU from MFE, kept that 128 EU in internal storage and sent out small packets whenever furnace asked for them. When that 128 EU ran out it requested another 128 EU packet from MFE. Without internal storage in the transformer the older system would either had HUGE energy losses or there would have been way too many tiny packets moving in the line between MFE and transformer. As neither happened I feel quite safe when saying they did have internal capacity, generally equal to the higher input/output EU value (128 for LV transformer, 512 for MV etc)
It's WIP and FAR from finished. You can't really say it's needs to be rebuilt when it has many features missing still.
technically you still have one packet per tick, but you can't exceed it, now you just add to the EU/t which is like BC, but BC you can't blow stuff up nor is there a voltage level (BC is supposed to be pneumatic energy, but it does quite follow those rules)
It used to be parallel power connections now it is all serial.
That isn't an internal storage like an MFSU or batbox, the transformer is just maintaining a min packet size when stepping up. If the setup is one packet input it would output every 4 ticks for stepping up. For stepping down it would simply output 4 packets every tick. The old EU meter didn't read the number of packets per tick, just the average EU/t of all the packets going through, which is one reason it would get screwy readings if you gave it too short a sample time.
I wouldn't have put out this 1/4 finished product for public testing just yet. I would have gotten much farther along. Doing it this way makes cross mod compatibility a nightmare.