TPS63900: Are the TPS63900/TPS63901 recommended for Coin Cell battery regulation?

Hi  Michael,

i think TPS63900 is more suitable than others.

Regards

Tao

Hi Tao,

Thanks for getting back. My concern with the TPS63900 is its high peak current which doesn't make it coin cell friendly. Does this mean I need to place more capacitance on the coin cell to provide for the higher current demands?

Other adjustable output TI regulators I've researched specifically call out use with coin cells: TPS61094, TPS61099x, TLV61220, TPS6122x. Also looked at a few coin cell regulators that output specific voltages. They all do have higher quiescent current.

The reason for all this is that the TPS63900 is not available presently in the WSON(10) package. We were told it is available in a 0.35mm BGA package that requires a via in pad on one of the balls which of course adds to our costs. Is this true?

Have a Very Happy New Year!

The TPS63901 which is available comes in a WCSP(12) package that requires the via in pad.

Hi  Michael,

Happy New Year too.

about the via, i am afraid that is correct as shown the picture below. Pin CFG1 need a via to be connected.

Yes, more caps at input will help if you have a short pulsed load. an additional resister between coin cell and input cap may help too. But this resiter will bring a poorer efficiency and higher voltage drop at input when transient load happan.

besides, seems there is High Drain series of coin cell which has a lower impedence internal as below. is it will help?

Regards

Tao

I’m adding caps in parallel with the coin cell based on your recommendation plus other research I’ve done. We can consider the LR44R for quick turn-On and discharge scenarios.

We won't use the TPS63901 because it requires a laser drilled via-in-pad, adding ~30% to our PCB cost. Considering we're looking at over 100,000+ units per year that adds up in a hurry.

As mentioned above I have also been looking at other higher quiescent-current controller options from TI and Analog Devices/Maxim. Meaning of course the coin cell won’t last as long. I'm avoiding ICs that don't specify batteries. But even "battery" often implies something larger than a "coin cell." So my 1st preference is a controller IC that mentions "coin cells," specifically Li-Ion.

We need 3.3VDC @ 200mA, powered in short duration bursts with lots of dead time, by a 625mAhr Li-Ion coin cell. An output option requires 4VDC @ 200mA. Prefer an ENABLE control. Quiescent current < 25uA.

TI's ICs I'm reviewing are:

• Outputs too low
  • TPS613221A (3.3V fixed, ~114mA – too low)
  • TPS610986 (3.3V, 100mA – too low (Main Boost))
• TPS610994 / TPS610996 (3.3V / 4.5V fixed, ~0.2A)

Is this how you read the specs for these ICs? Do you recommend other ICs?

Hi  Michael,

I would like to do a summary of your case as below, please help to check if they are correct.

• a device designed especially for coin cells is better
• input---3.3VDC @ 200mA  output---4VDC @ 200mA
• device has a lower quiescent current and lower peak current limit is better. 
• could you let me know how long the bursts time will last?

Regards

Tao

Hello Tao,

Input:

• Li-Ion coin cell, 625mAHr
• The coin cell will operate from approximately 3.0V to 3.1V when new, down to almost 2.0V before considered dead.

Output

• The regulated output is 3.3VDC @ 200mA
• 300mA output current would be preferred and 400mA would be even better.
• Power supply quiescent current less than 35uA total worst-case

Additional Control

• An ENABLE hardware control that turns the supply On and Off is highly preferred

Operation

• With a hardware controlled feature selecting a 4.0VDC output @ 200mA (preferred, not required)
• The external circuitry will draw upwards of 100mA to 150mA from the controller for 4-5 seconds then be Off for 15 to 30 minutes, possibly longer.

I ran WEBENCH Power Designer to find designs that met the following criteria:

• Would generate these outputs with the inputs listed above
• Be high-efficiency, efficiency >= 75%
• BOOST or BUCK-BOOST
• Output ripple <= 59mV

The issue is finding ICs that can run off a coin cell, not a large battery or unlimited power source.

WEBENCH came up with several designs using the parameters above. Looking thru the list for max quiescent current (Iq) I came up with the following. The output currents are based on going thru the individual specs and seeing what the parts did.

1. TPS613221A (3.3V fixed, ~114mA out – too low, Iq = 10uA)
2. TPS610986 (3.3V, 100mA out – too low (Main Boost), Iq = 23uA) - no stock
3. TPS610994 / TPS610996 (3.3V / 4.5V fixed, ~200mA out, Iq = 1.6uA) - very few available

Do you agree with the numbers above?

Do you recommend any other low quiescent current DC-DC controllers that will work off a Li-Ion coin cell?

The resistor-divider from output pin to feedback pin, adjustable regulators had much higher quiescent currents. Upwards of 50uA which is too high. I would prefer this type since I can use a MOSFET switch to select a different resistor divider ratio to switch from 3.3VDC to 4VDC.

====================================================================================================

Why did TI effectively discontinue the TPS63900? Not available until 4/7/2024 per Digi-Key. It offered a package we could easily use, had an adjustable input current limit, adjustable output, capable of 400mA output - meaning I'll need to add more capacitors. The TPS63901 package is unusable for us.

Thanks for all your help on this. I need to resolve this right away and get a dev board into fab and assembly with a low Iq part that will output 3.3V at 100-150mA off a Li-ion coin cell.