Chaining two batteries to achieve a specific voltage range under their maximum

Question

I'm building a portable 8x8 LED matrix powered off of two CR2450 coin cell batteries. Most of my components ( Attiny85 and LEDs ) can run easily off of the supply voltage range of the CR2450, but the shift registers aren't being as conducive...

There are three shift registers in my design: two DM7495N and one NTE74LS164. All three shift registers have a recommended voltage range of 4.75 to 5.25 in order to operate. If I chained the two batteries, they would supply a maximum voltage output of 6V and a minimum of maybe a little over 5.2V, before the other components become under-powered and the batteries would need to be replaced. That is not within the range associated with the shift registers..;(

Is there a really simple way to step down the voltage to fit the ratings of the shift registers without using a voltage divider? Like maybe with a couple of resistors or some other common electronics component? Something tells me that there is in this case, since I am using two separate batteries... If I have to, I could ignore the recommended ratings, as both registers have a maximum input voltage larger than 6V, but I would like to know if there is an efficient and easy way to step the voltage down in this case. Thanks!

Answer 1

I would recommend changing the parts in your design to use parts such as the 74HC195 and 74HC164, or change the design slightly to use other variants of CMOS technology.

The 74HC parts operate over the range of 2-6V and consume hardly any power, much less than those you are currently using - they could operate without any regulator. They can interface with the ATtiny the same as the TTL and LS parts and are pin compatible.

The static power consumption of CMOS devices is close to zero - for example the 74HC164 has a typical value of only 8uA at 25°C.

When the device is operating there will be additional current required to charge the internal capacitances that depend upon both the power supply voltage and the operating frequency.

The data sheet usually has a value called "Power Dissipation Capacitance" or similar - this is 40pF for the 74HC164. This can be used to calculate the dynamic power consumption. (See page 9, note 3 of the datasheet for the 74HC164).

If the device is operating at low frequencies this will be very low. For example if the 74HC164 is being clocked at 1MHz with a 5v supply and the input data is a constant zero or 1 the resulting supply current will be 1mA, much less than the 30mA of the 74LS164.

40*(5)^2 * 10^6 * 10^-12 = 10^-3 = 1mA.

I would expect the average frequency when used for driving a display to be in the kHz range with a correspondingly low current consumption.

Any current required to be sourced from the output pins would be added to this currents. Sinking current does not require any additional supply current.

Answer 2

This arrangement will not work at all, not even for a short time.

The CR2450 battery (per Sony, Panasonic datasheets) has a drop of ~100 mV with a load of 7.5k, or 0.4 mA. This means that the internal battery resistance is about 250 Ohms. Even if you try 4 LEDs at 2 mA each, the voltage will drop by 2-3 V. Your Attiny will eat another 1 mA even if you run it at 1MHz. The main killler will be DM7495 with typical consumption of 50 (!!!) mA. Two of them will take 100 mA, plus 17 mA for NTE74LS164. There is no way for these batteries to deliver 120mA, the short current (voltage drop to zero) will be ~12-15 mA.

So scrap your project and start over, with CMOS chips as Kevin suggested, and a different battery.

Answer 3

Try searching up voltage regulators. Linear voltage regulators can do the job of stepping down the voltage but the voltage difference between the input voltage and output voltage of the linear regulator gets dissipated as heat. If you want an efficient voltage regulator, try switching voltage regulators like the Buck Converter.