Author Topic: 5V step up burns Li-ion charger IC  (Read 844 times)

0 Members and 1 Guest are viewing this topic.

Offline deltoroTopic starter

  • Newbie
  • Posts: 2
  • Country: us
5V step up burns Li-ion charger IC
« on: December 03, 2019, 09:21:09 pm »
tl;dr - After powering a 1A load with Li-ion and 5V step up converter (U2), connecting to USB causes the Li-ion charger (U1) to overheat and burn

In the attached schematic, U1 is the single cell Li-ion charger. U2 is the 5V step-up converter. 5VON is hooked to a microcontroller digital pin that is driven high during radio transmission.


Background
---------------
This circuit charges a single Li-ion cell and powers a LoRa radio that is 3V-5V tolerant and changes in demand from 20mA during reception to 1A during transmission (it can receive at <5V but in order to achieve max power during transmission it needs to operate at 5V). A 5V step up converter is turned on during transmission but turned off at other times to conserve battery.

Design requirements
-----------------------
1. When connected to USB, use USB 5V to both charge the Li-ion and power the load (Li-ion is disconnected from the load)
2. When disconnected from USB, power the load with Li-ion
3. Power the load with Li-ion stepped up to 5V if 5VON is set high by a microcontroller
4. There is the case when USB is connected and 5VON is set high - let's assume for now that never happens

Steps to reproduce the problem
------------------------------------
1. Charge the battery with USB. The radio receives on USB power during charging.
2. Disconnect USB. The radio switches seamlessly to using Li-ion for power.
3. Turn on the 5V step-up converter (U2) by driving 5VON high during transmission
4. Turn off the 5V step-up converter (U2) by driving 5VON low after transmission
5. Reconnect USB
6. Watch U1 produce magic smoke :(

Why is this happening ?
 

Offline jaromir

  • Supporter
  • ****
  • Posts: 344
  • Country: sk
Re: 5V step up burns Li-ion charger IC
« Reply #1 on: December 04, 2019, 12:28:48 pm »
After swapping D and S leads of Q1 the circuit may behave more reasonable.
For more details, see http://ww1.microchip.com/downloads/en/appnotes/01149c.pdf , chapter "CONNECTING THE SYSTEM LOAD TO THE BATTERY".
 
The following users thanked this post: deltoro

Offline jhpadjustable

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: us
  • Salt 'n' pepper beard
Re: 5V step up burns Li-ion charger IC
« Reply #2 on: December 04, 2019, 12:35:19 pm »
I notice that your schematic shows an NMOS symbol while the value calls out a PMOS device. After you have corrected the discrepancy and cleared the smoke, :) take another look at your circuit and consider what the body diode of the MOSFETs are doing for you (or doing to you). I speculate that they are back-feeding power into the 73831, destroying its battery output pin, setting the stage for spectacular failure when VBUS is subsequently connected.

A higher-level idea for your consideration: how about a low-cost, simple boost converter IC that is non-synchronous and does not have true load disconnect capability? When they are shut down, the input passes through a diode and inductor to the output without boost, which absorbs at least half of the power-path circuitry you have added, at a minor efficiency cost due to the diode voltage drop. With that in place, you can treat the power-path problem separately using any means you like, from two Schottky diodes, to a charge controller with integrated load sharing such as the MCP73871.
"There are more things in heaven and earth, Arduino, than are dreamt of in your philosophy."
 
The following users thanked this post: deltoro

Offline deltoroTopic starter

  • Newbie
  • Posts: 2
  • Country: us
Re: 5V step up burns Li-ion charger IC
« Reply #3 on: December 05, 2019, 05:04:15 pm »
 Argh! Thanks for pointing out the reversed part. I was confused from the start, like you said the schematic symbol was wrong, and I figured the body diode is pointing the wrong way too.

Do you have a part suggestion for a low cost boost converter that is non-synchronous without true load disconnect ?
I am not sure how to search for the pass through on shut down function.
 

Offline jhpadjustable

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: us
  • Salt 'n' pepper beard
Re: 5V step up burns Li-ion charger IC
« Reply #4 on: December 05, 2019, 10:47:56 pm »
Look for something where the input inductor connects to a diode which connects in turn to the output, in the traditional fashion. You can be sure these will simply pass current through to the load because they have no way to stand in the way of it.

The 3608 step-up is a decent choice. Respectably high switching frequency, 2A switch current rating, SOT23-6 package, under 10 yankee cents each in lots of 5, and broadly sourced by Asian semiconductor manufacturers (Aerosemi, Hengxiang, Seaward, Fuman).

If you prefer something a little more Western, the Monolithic Power Systems MP3221 does provide enough power for this application and its input disconnect is optional.
"There are more things in heaven and earth, Arduino, than are dreamt of in your philosophy."
 
The following users thanked this post: deltoro


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf