I have my own design (see attached), and I'm pretty new at all this. I designed it in kicad and then tested it pretty extensively using LTSpice. Upon having a dev board made when I plug in BAT+ I see GND throughout the circuit go to VCC - 0.7 for a brief period and then drop to 0V. One it's at 0V the EN pin stays at VCC - 0.7 (also unexpected). I attached two multimeter traces; one trace is of the EN pin and the other of the GND connection (as measured from the GND side of R16).

Here are a few things I think I know?
* The pre-startup is really the part of interest, so I don't think the GPIO connections are meaningful at all. For example a diode test setting of the multimeter in the unpowered circuit shows no connection between GPIO4 and GPIO14 in either direction.
* There is a 0.7V drop between ground and BAT+ in the unpowered circuit
* When the ESP-12F powers on (when EN rises high enough which doesn't happen immediately) the GND drops to 0V and stays there...

How can I troubleshoot this design in a non-destructive way with my multimeter and/or oscilloscope? The hardest part for me is that the circuit is actually on a PCB so it's all connected unless I start destructively removing parts of it. The only thing I can really think of is to get a breakout board and start to build up the relevant parts on a breakout by hand...

 

If you have an unpopulated PCB that will help tremendously in the trouble shooting process.
Are you willing to post your CAD schematic files and gerber files?

There are other ways to find a short on the board.
Here are some techniques that can be useful.
1) IR camera.
2) Freeze mist
3) Current tracer

Edit: Sorry. I just saw your oscilloscope screen shots. I was thinking of a different problem, i.e. a short from Vcc to GND.
I have to review your situation again.

 

Where is the oscilloscope probe GND clip connected?

 

LED strips require a non battery power supply. 100 LEDs x 2ma = 200mA constant current, the battery will fall for 5 hours if its 1A/h. In practice you need to limit the current or the LEDs will burn.

Something seems to be burning in your PCB.

TP4056
·Charges Single Cell Li-Ion Batteries Directly from USB Port ·Preset 4.2V Charge Voltage with 1.5% Accuracy

The schotky diode D1 is not correctly connected, your battery current flows towards VBUS with it. The anode and cathode are switched on Schotky diodes.

What is the end purpose of this circuit? If its to power LED strips, its better to use LM2576 or LM2596, you don't need the switching between VBUS and the battery. Its better to just put the battery at the end of the TP4056 pin 5 and if the VBUS is on, that will both charge the battery and supply voltage to your circuit(less current for the battery is better, as your circuit will draw current too). This way your power supply becomes 4.2V, the charging voltage of the 18650 battery.

 

Thank you both for the replies!

The GND clip on both oscilloscope probes is attached to BAT-. I actually do have several unpopulated PCBs! Maybe I should start running similar continuity tests on those unpopulated boards...

What do you mean something is burning? I agree it's not right, but I'm not sure what to do about the burning theory.

The Schotky diode info is interesting, maybe that's part of my problem? During my testing I modified on of my assembled boards to remove the schotky diode's connection to VBUS so I don't think it's now the cause of the GND voltage I'm exploring.

Annnny thoughts and ideas are welcome. Thank you!

 

I believe the voltage spice on ground is actually due to me plugging the battery in and not having both terminals connect at exactly the same moment. If I carefully connect the ground terminal first and then BAT+ I get no spike on the GND node of the circuit.

When I do connect BAT+ without BAT- connecting to GND I do see the +3V on the GND node, but through two different tools I see 0 mA flowing so I think it's just charged due to high resistance connections to ground with no current flowing.

Take together there doesn't seem to be a problem here. So the real mystery I'm chasing is why is the EN pin going high when I first connect BAT- to GND and then BAT+ to BAT+ on the scematic? I really didn't expect that and it's causing incorrect operation for me.