retepsnikrep wrote:Looking at a big board with 51 wires comming into it direct from the cell terminals for 50cells, the voltage at the board connector and fusing are a bit concerning.
1) Should we have any fusing in these wires?
Would not hurt if there are a short2) Where should it be? On the board? In the wire near the cell? Polyfuse or a proper holder onboard/inline
a holder, but where would depend on easy access, i would say on board and put the board a place where you can easy get to it3) These fuse will have to stand the 500ma max drain per wire ao can be quite a low value and the wires reasonably thin.
4) Should we go for a slightly smaller board (A5) with say 25 cells. these could then be stacked as reqd. They may be a bit easier to locate as well than a massive A4 thing? Hmm
A5 as they can be stacked, maybe even put them in a box with a sub-d connector? this can easy be removed if the car is sold or the unit has to be repaired etc.We can probably compensate for voltage drop in the cell/sensing wires with some dvm measuring and an extra line in the software to add 10mv or something when the load resistor is on.
D connectors won't give enough isolation from the pins to the shell for the high voltages ASFAIK
hmm... when the wires are soldered then maybe insulate them with heat shrink tube?So I'm thinking sil 0.1" pins on the board bottom edge for some sort of ribbon connector.
As there is no reset jumper for the Pic (we have removed both the jumpers) care will be needed during the programming cycle and a removable 50 way connector at least gives us a way of interrupting the supply if we have a reluctant pic which needs reseting. If the dig slave software is set to work at 4mhz anyway this probably won't be an issue. But that limits comms to 2400 baud scan rate of 50 cells/sec. Might be able to improve that with some more work on software.
Actually thinking about this we could just make up a 4.5v supply (3xAA bats) on a two pin jumper connector and then remove the HV connector and work down the sil 0.1" line with our simple supply plugging it onto the first two pins and them moving along a pin at a time, powering each pic in turn during the programming cycle. Much safer as well
![Very Happy :D](./images/smilies/icon_biggrin.gif)
The current Master software using the 28X1 only supports upto 128 cells, the new 28X2 would increase this to 1024 cells
![Shocked :shock:](./images/smilies/icon_eek.gif)
enough even for the biggest pack!! Should be a simple chip swap and a few constants to change in the Master program to deal with that.