bobc
10 kW
All good points. The problem I've encountered, working on the top side of the board, on the SMD part, it is SO easy for the side of the iron to melt through a relay case. I've ruined one relay this way and probably fatally damaged another (it's still working with some tape over the hole....)
Anyway, I did an update to the design the other week, to fix known issues & tart it up a bit.
1) changed all PTH circuitry to SMD. The relays are still PTH for cost reasons
2) put 4 screw holes in the corners. TBH this is not enough support - fully populated (>10s) the relays are pretty heavy & the board will need a ziptie over the middle in an EV
3) upgraded the on board switcher. Just swapped for an LM5161 in place of the LM5017. This delivers a full amp at 12V, enough for all the relays. The idea of beefing up the 12V by tapping off low down the battery string is an unscheduled imbalancing input - I've abandoned it.
4) various analog filters to improve fidelity
5) remove the short between 12V and B17 (balancing tap cell 17)
6) add catch diodes on the relay coils
While working on this project I find that I have damaged one cell on the scooter - I made the mistake of leaving the keys in it and switched on for about a week earlier in the winter when I'd just swapped some of the cells and I had not yet balanced the "new" pack (this thing wasn't working yet). It was a few days before I spotted this (that it had been left 'on') and the pack voltage had gone way too low. Because it had not been balanced there was one cell right down at 0.5V. I actually thought I'd got away with it, but it turns out that that cell is now probably 10% down on capacity & the result of that is that the whole array is 10% down; so I have swapped out the offending cell & am now discharge testing it. Still, fairly pleased with the LiFePo prismatic cells, more rugged than I expected.
Anyway, I did an update to the design the other week, to fix known issues & tart it up a bit.
1) changed all PTH circuitry to SMD. The relays are still PTH for cost reasons
2) put 4 screw holes in the corners. TBH this is not enough support - fully populated (>10s) the relays are pretty heavy & the board will need a ziptie over the middle in an EV
3) upgraded the on board switcher. Just swapped for an LM5161 in place of the LM5017. This delivers a full amp at 12V, enough for all the relays. The idea of beefing up the 12V by tapping off low down the battery string is an unscheduled imbalancing input - I've abandoned it.
4) various analog filters to improve fidelity
5) remove the short between 12V and B17 (balancing tap cell 17)
6) add catch diodes on the relay coils
While working on this project I find that I have damaged one cell on the scooter - I made the mistake of leaving the keys in it and switched on for about a week earlier in the winter when I'd just swapped some of the cells and I had not yet balanced the "new" pack (this thing wasn't working yet). It was a few days before I spotted this (that it had been left 'on') and the pack voltage had gone way too low. Because it had not been balanced there was one cell right down at 0.5V. I actually thought I'd got away with it, but it turns out that that cell is now probably 10% down on capacity & the result of that is that the whole array is 10% down; so I have swapped out the offending cell & am now discharge testing it. Still, fairly pleased with the LiFePo prismatic cells, more rugged than I expected.