MOSFET: preventing reverse bias

Guys and gals

I am putting a system together and was reviewing the schematics when I suddenly realised I had a reverse bias situation on my hands.
The attached pdf shows the charger control cct. This allows me to switch the battery charger to the 12v battery, if the solar panel doesn't supply enough power, during the winter, for example. The battery charger is auto sense so it selects 6v or 12v O/P as appropriate.

This means that U1 has 6v on the D and either 6v or 12v on the S, but U2 can have 12v on the drain and 6v on the S. BAD!!!

To isolate I am going to insert a Schottky diode into the D of U2.

Question: should I include a pulldown between the D and the Schottky?

Edit: Before any one jumps on my head, I know the Fan Control signals are inverted. I've amended it at my end.

Shed_Solar_System_Mega_Rev1.pdf (8.96 KB)

Oh dear. I don't think it will work as I want.

Just sitting going through the schematic, it sudenly dawned on me that the auto sensing battery charger probably won't be able to sense the voltage because of the MOSFETs.

I guess I will have to use relays, which I didn't really want to do or could I use additional MOSFETS as sense connectors?

I would be grateful for suggestions, but will draw it up when I get a chance, as I am going away tomorrow.

Lets step back a little.

I suggest you list the set of conditions and what should occur for each condition.
for instance:

  1. Bright day, Solar Cell providing more voltage than "________" we will run the LM2596 from the solar cell and also "___________________"

  2. etc

You mention a solar cell, however I cannot identify the solar cell in your diagram.

Is the "12v Battery Charger ON" connecting to U2 coming from an arduino? It doesn't have to but it would seem logical for this forum.

Suggestion for schematics that combine high and low power components. Start with a line near the bottom of the page and call that common (ground). Reference all components / modules to that.

The reason for this is, ground is not always ground. I normally say there is no such thing as ground. By that I mean high(er) current modules like the LM2596 can create voltage noise on the common connection. Controllers like the arduino do not link noise. The solution is often careful selection of where you connect the commons together.

Hope this helps.

JohnRob

@JohnRob

Thanks for the reply.

Yes, the signals "12v Battery Charger ON" and "6v Battery Charger ON" are both from the Mega. There will be a 2 second delay between one MOSFET being disabled and the other enabled.

The solar panel, sensors, and Mega ccts. are all on separate sheets.

The system is being bench tested, as I develop each part of the sketch, with LEDs to simulate the various loads.

The sketch, which is almost complete, works perfectly, on the bench. I only have a couple of sensors and RTC to be added

What I realised yesterday evening was that U2 would be reverse biased when the 6v charger was enabled, hence the need for a Schottky and pull-down? I have decided that a 10k pull-down had better be fitted. Agreed?

My doubts about the auto sense working is because the charger defaults to 24v with no battery connected but switches to the required O/P when connected to either a 2v cell, 6v, 12v or 24v battery. It this point I don't know if it needs to read the battery voltage prior to switching (MOSFET will block the battery being 'seen' by the charger), or if it O/Ps 24v at very low current which is pulled down by the battery, to tell it the voltage of the battery that is to be charged (Which will work). If the former then I will have to use relays, if the latter, then the MOSFETs will be the way to go. Tests are required but as I am going away, today, I don't have the time to run any tests.

Regards

Fof

Its impossible to follow your circuit diagram because you don't make it clear which components are power sources and which are loads.

You should as far as possible keep positive at top, negative at bottom and power/signal flow as left to right.
If you don't follow these rules you need to add clear voltage labels and current or signal flow arrows on the wires.

You also appear to connect 12V to 6V battery which doesn't make sense.

The answer is huh..... no. This configuration will not guard against reverse bias as the Arduino will not be able to turn U1 and U2 off.

To turn off the fets you will need to get their gates to within 1 volt of the source. The best you can do is 4.? from the arduino in if the battery is charging it is likely above 6.5V.

You'l have to rework the FETs or the drive to the FETs.

Still not sure where the solar cell is connected.

JohnRob

@JohnRob et al

Currently in Italy and while " 'er indoors" is out, with sister, visiting churches I got some time to sit back and review my schematics. :sunglasses:

DURRR! What a stuff-up. There is no way it would have done much, other then blowing the guts out of the MOSFET and the Mega control pin. :blush:

I promise I'll do more thorough reviews, before posting any queries, in future.

Fof

Hi,

  • What is you battery size, volts, Ah.
  • What is your PV size, Volts OC, Current SC and power Watts?
  • Why are you switching 6V charging and 12V charging?
  • What is the 6/12/24V charger?
  • What do you want the Arduino to do?

Thanks.. Tom.. :slight_smile:

@TomGeorge

This is just an exercise for me to assist my learning C. Prior to retiring, I promised myself I would learn embedded s/w, which I never had the time to do while working

The solar cell and 12v battery are just to supply power to LED strips in my shed and wood store, and are quite adequate in the summer. In the winter, with low light and low temps, they are not good enough. Hence the charger C/O.

I will be logging Battery temp, shed ambient temp, wood store temp and humidity with the ability to control the wood store exhaust fan (dependent on the humidity), rain fall and light intensity, all time stamped with a RTC.

I realised that none of the temp/humidity values will be precise, but they will all be relative to each other.

The data will be downloaded to Excel and graphed to give me a baseline for improving my 'man cave'.

As I say, this is just a programming exercise. Nothing more.

The schematic stuff up, should not have happened. It was done in a bit of a hurry, attempting to get to a particular point, prior to my taking my other half to Italy, to see her sister. The whole thing is now on hold till I get back to UK.

Having spent most of my life, as a technician, in electronic develoment labs, working on leading edge equipment, I am more than confident with hardware. I do admit, though, that, sometimes, I will see something that I may be not 100% sure of, so I throw out a query, and then stop and really think about it. One day I will learn to stop and think BEFORE posting a query. :astonished:

Fof

Not sure if it's been mentioned, but you need a level shifter for the FET gates. The gate voltage needs to be brought up all the way to the source voltage in order to turn the FET off. A Mega can't do that on its own when the source is connected to a voltage higher than 5V.

As for the actual problem you identified, put a second P-FET in series "backwards" on the 6V line, and tie both of the gates together. This is the way that Li-Ion protection circuits use to stop a battery from being overcharged or overdischarged. If you put the body diodes in opposite directions, when both FETs are off current is blocked in both directions.

Don't pay too much attention to the gates being connected to separate outputs on the Li-Ion example. That's because the protection circuit needs to stop one of overdischarging or overcharging at a time, but still allow the other thing to happen (I hope that made sense). Your application doesn't sound like it needs something that sophisticated, so controlling both gates from a single output is all you need.

The other option is to use relays. It doesn't sound like you need fast speeds since yo mentioned a 2 second crossover delay, so they might be worth a shot.