A-to-D sampling of the Nissan leaf battery.

yarisboy

The Nissan leaf batteries have three terminals. Referenced from the center terminal the voltages when fully charged read +4.2 volts and -4.2 volts respectively. I want to read both cells and feed the results into a Propeller. I'm aware that a MCP3208 could read the +4.2 volts but what would be the easiest way to get the -4.02 into the uP.

RDL2004

Use four different wires (1 positive, 2 center, 1 negative) and reverse the connections to the ADC for the negative/center pair.

altosack

You can use a voltage divider to get the voltage less than (-)0.3 volts, and then sample using differential mode. The MCP3208 is rated for -0.3 V to Vdd + 0.3 V.

I've used this method to read voltage from a shunt that can be +/- with an MCP3304. While it does reduce your resolution, you can raise it again with oversampling/decimation if need be.

yarisboy

RDL2004 wrote: »

Use four different wires (1 positive, 2 center, 1 negative) and reverse the connections to the ADC for the negative/center pair.

I've over simplified this. With 12 cells I'll need 24 channels of sampling. I notice the MCP3208 has a single digital ground for each of my three chips. I don't have experience with the MCP3304 but suspect it may be a better candidate for the application. During pack bottom balancing resolution is not an issue. During long term monitoring after to detect soft shorts resolution might be more of an issue.

altosack

Both the 3208 and 3304 support differential mode, which is the only way you can read a negative voltage (down to -0.3 V). This means you would need 5 chips (two for the 12 positive inputs and three for the 12 negative inputs), not 3. Note that you can use common MOSI, MISO, and SCK pins, with a unique _CS pin for each chip. This is all easily doable in one cog.

As far as I know, the only difference between the chips is one bit of resolution, but I haven't used the 3208. Plenty of others have, and there is extensive support for it in the OBEX; while I don't think there are drivers there for the 3304, I have a simple 30-line subroutine in PASM that reads it, including oversampling/decimation. It requires 88+ MHz to sample at 100ksps (2 MHz SPI); since I use it at 3.3V, at which it's only good for 50ksps (1 MHz SPI), it can do that at 44 MHz or greater.

As long as you don't need speed, the resolution thing is pretty solvable; by oversampling 1024 times, I get 16 bits of resolution with the 3304 from a shunt that only reads +/- 0.1 V, so you should be able to get 16+ bits with the 3208 reading at +/- 0.3V. Using a 0.5 V voltage reference (the minimum that the 3304 supports) helps.

yarisboy

I've got an Arduino Due bolted down to a really big breadboard and I just ordered a Propstick USB from Parallax. I've always wanted to do a head-to-head application for comparison. If I wanted to make it a Nuts&Volts piece I'd just forward it to JonyMac of Spin Zone fame. My target date for presentation is spring break.