LiPo not charging using MCP73831

[joeyjoejoe]

Can someone double-check for anything obvious I've missed?

5V USB is VDD.

I am pulling up the STAT as per the datasheet, and it's always reading HIGH. It should read LOW when charging (either fast or slow charge). (Page 13)





Datasheet : https://cdn.sparkfun.com/assets/learn_tutorials/6/9/5/MCP738312.pdf

[Mr.B]

I only have two observations from the datasheet.
You would need to confirm if your project has taken these into account:
Page 11.
3.1) A supply voltage of [VREG (typical) + 0.3V] to 6V is recommended. Bypass to VSS with a minimum of 4.7 µF (Your schematic does not appear to have this cap.)
3.6) An internal electrical connection exists between the Exposed Thermal Pad (EP) and the VSS pin. They must be connected to the same potential on the Printed Circuit Board (PCB) (Your schematic does not appear to show the connection of EP to Vss)

[joeyjoejoe]

You are correct, I did miss a bypass cap on VDD. I will try to bodge a 4.7uF on there.

As for the pad, that applies only to the DFN package, I am using the SOT package.

[joeyjoejoe]

Bodge complete - no change in behavior unfortunately.

[aiq25]

On your schematic it looks like STAT is pulled up to a 3.3V rail but on the app note it shows as connected to the input voltage (I assume in your case VBUS). Did you tie STAT to VBUS or is this schematic part of a larger circuit that has a 3.3V rail? Not sure if this matters but thought I would point it out.

[Mr.B]

Have you measured VBUS to ensure it is at least 0.3v above VREG when under load, or at least trying to draw the 450mA?
MCP73831-2 is 4.232v (max), so is VBUS above 4.532v when under load?

Other than that, there is not much else I can think of.

[Kasper]

Could be a few things. Some of these may seem obvious but they are also common.

Does the battery work? Is there a protection module (PCM) installed in the battery? Is its fuse blown?

What is the battery voltage?

Is the battery voltage too high or too low?

How do you know it's not charging? Is it just because of what stat pin is doing or have you verified this with measuring battery current? 

Is the fuse in your DMM blown?

[KL27x]

Is the battery charging?

If it is charging, but the pin1/STAT is not pulling low, then double-check your invoice and the label on your bag.

Microchip makes a 73811, which is the same package and nearly the same pinout. Except pin 1 would be an enable pin, which when pulled up will allow charging and when pulled low will disable charging. And pin 5 is a digital input pin used to select between 450mA charge rate or 85mA charge rate. This part has no STATUS output.

They also make a 73832, which has a tristate STATUS output pin, IIRC, but I don't remember what state should be active during charging. I have 4 different variants of this charger IC and I have mixed them up a couple times.

[joeyjoejoe]

Yeah STAT is pulled to 3.3v as the MCU is 3.3v logic.

Battery is 3.8V and works. VDD is at 5V, so defintiely getting the 300mV threshold.

I did order another 2 units from Arrow in case it's a fault or damaged IC.

I measured the power draw with a USB meter and the draw is identical to when the battery is disconnected.

[KL27x]

Another mistake I have made before is to look at the sample circuit in the diagram and wire the chip up as if that is the correct pin order. Esp in a case like this where they show the 5 connections with the space on one side, making it look just like the physical SOT 23-5 IC, in this case. It kinda makes me mad they do it like this. I understand it makes it easier to draw the schematic, but they could at least not put the pins in exactly the same arrangement as an IC.

[joeyjoejoe]

In this case it's a PCB, so no manual wiring

A head scratcher for sure.

[Kasper]

3.8V is probably low enough. You could drain it to about 3.4V to be sure.

If you post pictures of the PCB that might help.

[joeyjoejoe]

Follow up - ghost in the machine.

I got the replacement today and popped it on the PCB. Drawing almost 500mA now from USB

@KL27x, I'm thinking maybe I had the wrong version as you suggest.