Any Advantage to Isolating a Battery During Charge

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
Background:
As a hobbyist, every circuit I have designed was either powered by a 5V source, e.g., a USB jack, or by a battery. For battery power, I use a switching jack to isolate the battery during charging, which is a common practice the the RC equipment with which I am familiar. With my present project, I am considering allowing USB charging without use of an switching jack. There are two possibilities I have considered: 1) Use a mosfet switch to isolate the device while under charge; or 2) Simply put the system voltage regulator in series and allow charging and operation during charging. EDIT: The P-mosfet and direct connection to V-USB would be deleted from the circuit without isolation.

Details:
The system will be powered by a single lipo battery and operating voltage will be 3.3V. The battery charger will be a Microchip MCP73831/2. The system regulator (VR) will be an LDO MIC5319 or similar with an enable function. On/Off will be a single push button switch with "soft power." That is, first push is on, and second push is off, which will be done by controlling the enable function of the VR. I haven't decided whether to make the soft-power using discrete components (e.g., a pair of NAND gates) or a single chip (e.g.,MAX16054) that also provides debouncing of the on/off switch push button.

Questions:
1) Any opinions on not isolating the battery during charging?
2) Opinions on the following circuit outlined with dashes to provide isolation?
NB: That is just a random P-MOSFET form Eagle's library and is not the one that will be used.


upload_2019-1-28_15-3-10.png
 
Last edited:

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
Can you clarify "stand alone" or give a link? I have seen designs in which it charges batteries installed in circuits without isolation.
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
I re-read the datasheet and believe the "standalone" comment is an alternative. That is, it can be used standalone or with an MCU. There is no statement that I can find to the effect that it cannot be used as suggested here.
 

sghioto

Joined Dec 31, 2017
5,376
In your circuit you need to add a diode from 5v to +Vbatt. Anode on 5volts. This will reverse bias Q1 when USB is connected. Powering load and isolating charging
SG
 

BobTPH

Joined Jun 5, 2013
8,804
What kind of battery are you using? There are tons of dirt cheap battery management boards for Lithium Ion batteries available. These have the further advantage of preventing over discharge.

Bob
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
Thanks. I had not really thought about that as charging and using the battery at the same time was just a possibility, but not likely to happen in actual use.

With USB power attached, that would also apply full USB voltage minus 2 diode drops to the battery. That is intended to be a P-mosfet. Some people put an N-mosfet with a much smaller resistor in the P-mosfet's gate to turn it on or off. I have edited the original post with: The P-mosfet and direct connection to V-USB would be deleted from the circuit without isolation.
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
What kind of battery are you using? There are tons of dirt cheap battery management boards for Lithium Ion batteries available. These have the further advantage of preventing over discharge.

Bob
I am using a rectangular Lipo, and the MCP73831/2 is only $0.61 at Digikey in single units. That is close to dirt cheap considering how many I anticipate using.
 

Hymie

Joined Mar 30, 2018
1,277
I see many lithium battery charging circuits similar to this.

Should IC1 fail short circuit (pin 3 to 4) or the FET Q1 (gate to drain), then the USB 5V in will be applied directly across the battery; and if IC1 fails short circuit (pin 3 to 2) the battery will suffer a direct short circuit (anode to cathode).

And some people wonder why these batteries catch fire or explode.
 

sghioto

Joined Dec 31, 2017
5,376
Thanks. I had not really thought about that as charging and using the battery at the same time was just a possibility, but not likely to happen in actual use.
That is not what I am suggesting. The USB powers the circuit and charges the battery which is isolated during charging. That's what you want ,correct?
SG
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
Maybe I viewed your suggestion wrong. VBATT goes to the VR, but the mosfet's body diode goes directly to the battery under charge.
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
I see many lithium battery charging circuits similar to this.

Should IC1 fail short circuit (pin 3 to 4) or the FET Q1 (gate to drain), then the USB 5V in will be applied directly across the battery; and if IC1 fails short circuit (pin 3 to 2) the battery will suffer a direct short circuit (anode to cathode).

And some people wonder why these batteries catch fire or explode.
I have not seen reports of such failures. If the Microchip BMS failed in that manner, I suspect the behavior of the two diodes would be affected (e.g., both on or both off) and the user should be aware. No Lipo should be charged unattended. Fortunately, this is for only 4 or 5 people -- all but one of them follows directions well, and that one has a teenage daughter who is as compulsive as I am. That's my "failsafe." :)

EDIT: The MPP73833 is more sophisticated and is also available in a hand-solderable package at $0.88. Maybe I will switch. My main question is whether isolation is needed. My perspective is decades old and as said, based on switching power plug (round) inputs.
 
Last edited:

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
I modeled (LTSpice XVII) the boxed part of the original circuit with the P-Mosfet. It works as expected. With power from the USB at 5 volts to the gate and charging voltage less, the voltage on the output is zero.

1) Is that adequate isolation?

Back to my original question, why should the rest of the circuit, which consists of a VR with enable be isolated? If the enable is off, I don't see any reason as quiescent current of that and many regulators is <100 µA.

If someone were to attempt to enable the VR with USB connected and without the mosfet for isolation, I can see that the charging program might be affected and possibly result in overcharging. The circuit should draw about 100 mA or less; most of that due to an LCD backlight. But that effect should not be immediately disastrous. If mistakenly turned on, it could be turned off. In fact, voltage on the USB line could be detected with the MCU, and that could force a shutdown if the MAX16054 on/off controller is used.

2) Is inadvertent enabling the VR the only reason for isolation?

I do not need operation while charging.
 

sghioto

Joined Dec 31, 2017
5,376
I had not considered using an N-channel to source current like that
It's not a N channel mosfet. The FDN360P is a P channel mosfet you selected in your edited schematic, however you have the the source and drain reversed in your schematic.
I do not need operation while charging
OK, that's what I wasn't sure of. Then your schematic is all you need, just remember the mosfet is shown in reverse.
SG
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
Hi sghioto,

Here is a snippet from the On-semi datasheet for that device:
upload_2019-1-29_11-48-53.png

I believe I had the Source properly connected such that the intrinsic diode was reverse biased from the voltage supply. Your revised version shows an N-channel mosfet:

upload_2019-1-29_11-51-54.png

It is more than a little confusing, as Wikipedia (and some datasheets) show mosfets without a bulk connection, for example a P-chennel without bulk connection from Wikipedia is shown as:
upload_2019-1-29_12-0-31.png

and the datasheet for the MIC5318 VR shows the output P-channel mosfet the same way:

upload_2019-1-29_12-3-36.png

That's why I keep a crib sheet handy.

Regards,



John
 
Top