Hi,
I have an application in which I need to power an LCD backlight controller chip. This chip is a switching regulator designed with a current drive on the output for the backlight. The power supply range on this chip is 3-10VDC. The circuit I'm working on is powered from a 12V wall supply, or an 8V battery. I could just put in a 5V regulator to supply power to the backlight chip, but that would be a waste of power and would create a bunch of heat. The backlight chip will draw less current when the supply voltage is higher, so I'm thinking about a zener regulator with emmitter follower transistor circuit to power it. At 12V supply, the zener (10V) would be in regulation and the output from the transistors emmitter would be around 9.3V. Current draw at this point would be around 70mA. When the wall supply is unplugged, battery takes over. Input drops to 8V, less as battery drains. It could go as low as 6V. As the supply drops, the backlight driver will require more current... around 130mA at 5V input. I want to minimize the voltage drop across the transistor for increased battery life. My question is, is this a good approach? Is there a better approach? How about a linear LDO regulator, such as LP2952, set to 10V and running in drop-out when on battery power?

 

I think either the zener/emitter follower or the LDO would be good choices. The former may have a slight advantage unless you find a LDO that doesn't need an output cap for stability, I think TI has some of those.

 

The LDO's have around .5V of dropout spec. Don't know what the input/output difference is at lower voltages. I have tried a couple transistors on the bench... MPS2222 and 2N3904. These are just what I had on hand. I set up an experiment with the supply voltage set to 6V and the load on the emmitter of 50 ohms. I noticed that the drop across the 2N3904 was about 1.25V with a 300 ohm resistor from collector to base. With the MPS2222, the drop was about .93V with the same 300 ohm resistor collector to base. Looking at the datasheets for these devices, it's not immediately obvious to me what specs are responsible for the difference. Beta's are similar. Therefore, I'm not sure how to spec a transistor that will minimize the drop. I need the drop as small as possible to maximize battery life. On the other hand, if the drop on the LDO is constant around .5V, that would be good, but the supply current increases to around 6mA in dropout... not good. Any ideas on how to spec the transistor to minimize drop across it?

 

Any ideas out there?

 

This portion of your initial post:

<snip>The backlight chip will draw less current when the supply voltage is higher<snip>

...seems to indicate that your "backlight chip" is a switching supply, which is about the only thing I could think of that explains less of a current draw when the supply voltage is higher. This is due to the power conversion that occurs within the switching supply itself.

If this is indeed the case, attempting to add linear regulators into the mix will simply dissipate power unnecessarily, both in the linear regulator AND in the switching supply.

Without more information about your existing backlight supply, we really can't give you any better suggestions.

 

Try using a PNP transistor instead.

 

The backlight driver is a switching regulator. I'm just trying to limit the voltage into it to 9-10V. I don't want the expense of building a second switcher to run from a large input range when it should be relatively easy to clamp. Again, the supply will see aroun 8V with battery power and 12V when charging. As for using a PNP, I know LDO regulators use them, but the control loop is more difficult to manage than with an NPN type. This is why they have specs on the output caps. Anyway, I don't see how to use a PNP without getting complex in the control circuitry. With an NPN, I can just use a resistor and zener diode.