Digital battery power distribution system. is this idea safe?
rwgast_logicdesign
Posts: 1,464
Now that I am getting accustomed to working 70+ hours a week, I am starting to find bits of free time to work on my stingray based robot again. I would really like to implement a somewhat intelligent battery system that can outlast the rest of the robot, power wise that is. I am currently using two 8 cell NimH packs to power the bot. When the batterys go low I plug them in to a commercial charger capable of charging two 8 cell packs at one time.
Basically what I am looking to do is design a small board with a micro that can monitor and log the batterys voltages and there current draw then display the data on a 16x2 LCD among other things. The way I would like to wire it is like this, A main power switch starts the power monitoring system, the power system will then be able to start and shut down each 8 pack battery via a relay, the batterys can be switched on and off via a push button or software.
Now here is what i am a bit shaky on... I would like to connect the power monitoring set up to both 8 cell pack in parallel, then each batterys + and - branch off to its own system which is controlled by the relays. Now will paralleling the power board with each pack cause problem with charging? I.E as long as each packs individual + and - is connected two its own charger, the current wont try to flow back in to both batterys in parallel at once when charging right?
Basically what I am looking to do is design a small board with a micro that can monitor and log the batterys voltages and there current draw then display the data on a 16x2 LCD among other things. The way I would like to wire it is like this, A main power switch starts the power monitoring system, the power system will then be able to start and shut down each 8 pack battery via a relay, the batterys can be switched on and off via a push button or software.
Now here is what i am a bit shaky on... I would like to connect the power monitoring set up to both 8 cell pack in parallel, then each batterys + and - branch off to its own system which is controlled by the relays. Now will paralleling the power board with each pack cause problem with charging? I.E as long as each packs individual + and - is connected two its own charger, the current wont try to flow back in to both batterys in parallel at once when charging right?
Comments
Provided you avoid battery-battery shorts, you should be OK.
Using relays in a tree can ensure that can never occur .
There are some nice battery reading devices out there.
eg the TI INA226,
and Microchip have a good looking selection like
PAC1710 - single channel 1-25: $0.93
PAC1720 - Dual channel 1-25: $1.63
Package is a little small, but the price is good for what this does.
Avoid excessive temperature during charge and discharge.
Avoid excessive voltage during charge.
Avoid excessive low voltage during discharge.
Avoid short circuiting or reverse voltages applied to individual cells.
Test cells for short circuiting and exclude from use prior to any charging cycle.
If you can manage all that for each and every cell, you will do fine.
Of course the more cells you put in parallel and series, the harder it gets to do this. So you might add a liquid cooling system to moderate temperature.
Relays versus MosFets... hard to say, but I'd go with latching relays to reduce power consumption if I really needed them. Mechanical relays tend to fail in the off state; solid-state devices tend to fail in the on state.
I suspect the Propeller2, with all the i/o with indivdual ADC i/o will be a useful device for this.
What am I trying to say? This is a complex puzzle that many are looking at and if anyone has solved it, they are sitting on the knowledge as valuable intellectual property. They may have decided to keep their info as trade secrets and not patented any of it. There are lots of 'less than perfect' solution that are being touted on the internet and lots of abandoned one chip devices. There are lots of related chemistry issues.
The power board will basically be an AVR ran off a switching regulator that reads each of the battery packs voltages and current draw through its a to d using shunts, opamps, and resistor dividers. I elected to leave all that off the diagram for simplicity sake. As I said what im worried about is the connections to the power board causing the packs to be charged in parallel. The reason I want to run the power control system in parallel with both packs is to make sure it still has enough current going to it when both packs may be so dead they would not be able to power anything individually, this way the status of the batterys can always be seen.
Also is there a problem just using a regular relay and holding it's logic input hi for the duration the battery is needed to power the system? A latching relay would be the more ideal choice, but Im just trying to use what I have on hand. I have plenty of 30amp logic level fets but I like the clicking noise a relay makes
a. ones that toggle on and off from one input
b. ones that have two inputs - one for on and one for off.
The advantages of the type b are that the inputs are obviously going to do what you want with a pulse, no feedback is necessary. And you can make bigger more powerful latching relays.
++++++++++
There are a lot of aspects to integrating a multiple cell charging and discharging system. The more cells you put in series, the harder it gets - eight cells in series is quite ambitious as the variations in product of individual cells can be quite a bit. One cannot easily make assumptions about what is a standard internal resistance and rely on it. And in failure modes, internal resistances change dramatically and suddenly.
Added to all this is the fact that the more you do to actually balance the individual cells to be charged equally and to discharge equally, the more energy you waste in inter-cell transfers of capcity. You rob Peter to pay Paul with some loss at every transaction. So most battery packs have reverted to simpler designs and abandoned ambitious multi-cell control schemes. Electric cars have not given up the cause so far, but laptops have exploited less cells in series and operating at lower voltages. Cell phones geta away with quite a bit because they use just one cell configurations.
NiCd seems a lot more forgiving that NiMH or Lithium Ion. I actually like Lead acid gel cells because they are even more forgiving, can do long-term trickle charging with ease, and work better in sub-zero temperatures. Lithium ion is at its best when you want to recharge quickly and then use the device disconnected from a power source.
you have both + poles of the battery packs connected together, and that in turn means both chargers also have their outputs connected.
Nasty.
Drawing power from both packs at the same time when they're 'nearly dead' just means that you destroy 2 packs at the same time, not one.
And connecting up two packs is a pain in the ... ... no matter how you do it as you need to make certain the power can only go FROM a pack into the load, and not to the other pack. (this means playing around with Rectifier diodes)
And the best you can hope for without some really clever tech is that the most full pack is the one delivering the power.
By all means use a multiple pack design, but be prepared to only run from one at a time.
Perhaps some blocking diodes?
If you want to monitor both packs without having them connected in parallel as they are now you will need to have a monitoring circuit that has two separate inputs, or to switch the one input between the two packs using a relay, transistor, or ssr.
What do you want to do?