Over-engineered 4.5v Replacement

  • Nominations will close March 25th. If you'd like to join the board and influence how TinkerDifferent runs in the next year, put your name in now!
  • Hey Guest, MARCHintosh 2026 is upon us. Check out community projects, join GlobalTalk, and have fun!

Stinkerton18

Moderator
Staff member
Aug 18, 2022
102
85
28
Howdy all,

Figured I should post this here in case anyone else wants to follow suit. So what is it I did exactly? Well, I took an 18650 (16650 works too for smaller areas), connected it to an adjustable output boost converter/lion cell charger circuit, and used this concoction to replace the leaking 4.5V alkaline CMOS battery in an AST Premia 66/d system. Power for the charger comes from a 4-pin molex connector so the battery gets recharged/topped off while the computer is in use, plus cut off for if the LION cell gets too low or gets to warm.

Here's the full build with the output measurement: (NOTE: Yes I used yellow for the positive from the 5v pin which is usually red. I did this intentionally to separate the Power-input DC lines from the Power-output/Battery DC lines)
IMG_5067.jpg


In short: 5V from Molex-> Battery Charger/Boost Converter
3.7V to/from 16650/18650 from/to Battery Charger/Boost Converter
~4.5V from Battery Charger/Boost Converter to CMOS/Battery Pin header of system

The specific module I chose does not have a minimum draw/auto power off function, so it is always drawing from either the battery or the incoming charge power to modulate to the configured voltage.

BOM/Links (Enough for 5 units)
Charger/Booster and 18650 holders: https://www.amazon.com/dp/B09X9VMGV4
4-pin Molex connectors: https://www.amazon.com/dp/B09ZXGXMDC
 

YMK

Active Tinkerer
Nov 8, 2021
414
347
63
Neat. Could it work without the boost converter at 3.7V?

In very low drain applications, I'd expect the converter to eat most of the power.

I did similar with Eneloops and no converter.
 

Stinkerton18

Moderator
Staff member
Aug 18, 2022
102
85
28
Neat. Could it work without the boost converter at 3.7V?

In very low drain applications, I'd expect the converter to eat most of the power.

I did similar with Eneloops and no converter.
It's possible but I do know the BIOS itself does measure the incoming voltage and will throw alarms if it dips too low. Alkaline cells tend to drop to around 1.2V when depleted, which comes out to about 3.6V, really close to the regular 3.7V of a lithium-ion cell. That's why I did the booster, just to avoid risking the low voltage issue.