6
\$\begingroup\$

On a PCB of mine I'm employing two ADP2504 Switching DC-to-DC Converters (one fixed version outputting 3.5V and one adjustable version set to output 4V) which have their inputs connected to the output of a Texas Instruments bq24032A (a power management IC that switches between USB 5V and a Li-ion battery). The 4V ADP2504 IC is only used to power one SIM7600 cellular module whereas the fixed 3.5V ADP2504 is used to power all other circuitry (microcontroller, LEDs, small LDO (3.5V -> 3V, few mA), etc.).

To power my PCB, I plug a USB cable (that is supplying 5V from a bench-top current limiting power supply) into a USB receptacle on my board that feeds the bq24032A's input (note that no Li-ion battery is connected to the bq24032A). Doing this, the PCB draws an excessive amount of current (the 0.3A limit I set on the power supply gets tripped. This also happens with limits set to 0.5A and as high as 0.9A). Additionally, the ADP2504 outputs are not 3.5V and 4V as expected (checking with a multimeter, the 4V output drops to ~2.5V in this state for example. The 3.5V output drops similarly). Very rarely does this problem NOT occur (such as when I assembled my first board prototype (layout remains unchanged) or one time when I soldered a cut trace back together).

To troubleshoot this, I:

  • Disconnected/cut the input trace of the 3.5V ADP2504 from the bq24032A's output and I disconnected/cut the trace connecting the SIM7600 from the 4V ADP2504's output. Plugging the USB cable in again did not result in excessive current draw.
  • Soldered/connected the 3.5V ADP2504's input back to the bq24032A's output. Plugging the cable in again did not result in excessive current draw and the output was indeed 3.5V where expected.
  • Soldered/connected the SIM7600 back to the 4V ADP2504's output again. This caused the excessive current problem to come back. Note: one time when I soldered this trace back together, there was no excessive current upon plugging in the USB cable and the 4V ADP2504 output was 4V indeed. However, the problem came back shortly after ("Problems that go away by themselves come back by themselves").
  • Disconnected the 4V ADP2504 from the SIM7600 once again. Powered the SIM7600 module independent of the ADP2504 by applying 4V from the bench-top power supply directly to the module's power pins. There was no excessive current nor drop in voltage from the power supply, indicating that there is no short circuit associated with the SIM7600. Checked the voltage of the 4V ADP2504 (disconnected from the SIM7600) and it was indeed 4V. Key point: when the 4V ADP2504 and SIM7600 are disconnected, they seem to operate without issue independently.

The above points and looking at other questions on the website led to suspicions that there is a problem with the 4V ADP2504's load (or that the ADP2504 may be oscillating or that there is an issue with its feedback). A comment on this question (which deals with a linear regulator however) by Peter Smith states: "I would try removing the output cap to start with. The datasheet states 'no external components required' (perhaps...); as John D notes, the output may be oscillating, and an incorrect capacitive load can cause this in some regulators."

On the input of my SIM7600 (which is the 4V ADP2504's output on which there is a 22uF ceramic capacitor (datasheet recommendation)), I have a 1500uF tantalum polymer capacitor (T545H158M006ATE035) (see Edit/Update at bottom of post for explanation why this is used, thanks Tim Williams for the suggestion). I removed this 1500uF capacitor and lo and behold, there is no excessive current draw and the 4V ADP2504 indeed outputs 4V.

Now, the issue is I'm not exactly sure why removing this 1500uF capacitor seems to fix the problem and more importantly, I'm not sure how to fix the problem without removing the 1500uF capacitor.

Some things that I've tried (with the 1500uF capacitor still in the circuit):

  • Placing a series inductor (1.5uH) between the 1500uF capacitor and 22uF capacitor (Bimpelrekkie's suggestion to avoid the large capacitance affecting the feedback loop). The problem still occurred. I think 1.5uH may also be (way) too small of an inductor here.

  • Placing a series resistor (tried with a 5 and then a 10 ohm resistor of appropriate power rating) between the 4V ADP2504 and the SIM7600 in hopes of reducing inrush current from the large capacitor that may be "throwing" the ADP2504 "off". The problem still occurred.

  • Adding up to 20uF of capacitance to the ADP2504 input/bq24032A output (as I've seen sources suggesting that the input of the switching regulator may be dropping too low such as here and here (" Do not attempt to put a large value filter capacitor (C3) on the output of the regulator."). I also added 10uF capacitors alongside the minimum recommended 0.1uF caps for the bq24032A input/output. I understand this may very well be (way) too small of an amount compared to the 1500uF. I also added a diode (Schottky) across the input and output of the regulator as the second link suggests in the case of this input voltage drop (to avoid backfeeding) but to no avail.

Are there any thoughts here as to what is going on, if it can be remedied and how? Looking around at other questions and answers (the 2nd paragraph here as an example), I think it's very plausible that I am having issues with oscillation/stability/feedback of my DC-to-DC switching regulator due to a large capacative load. Am I on the right track? I want to take an oscilloscope to the inputs/outputs of my power ICs and I'll be getting my hands on one in the next few days but until then, I've been trying to do what I can and I wanted to make sure I'm not overlooking something simple.

If anything else is needed please let me know.

Schematic: enter image description here

Board Image: enter image description here

Board Layout: enter image description here

Board 3D Render: enter image description here

EDIT/UPDATE: Tim Williams brings up a good point (why the 1500uF capacitor and how did I get there)? As Tim noted, the SIM7600's peak current draw is 2A due to GSM emission bursts with the following recommendation for capacitor sizes. enter image description here The ADP2504 can supply up to 1A which is < the 2A bursts. I decided that I would go with a compromise by using the ADP2504 (which looked very promising and easy to use) despite this current capacity discrepancy and that'd I compensate for this with capacitance to avoid a voltage drop as suggested in the SIM7600 datasheet. I feel I perhaps could have gone with a regulator that could supply 2A but these were difficult to find and I felt I would hit a limit before the regulator itself: I felt 2A may be too much for a standard USB wall wart one may find laying around and I didn't exactly want to draw 2A bursts from a 1800mAh Li-ion battery I'll be using in fears that it would be bad for it. Additionally, the specific "1500uF" value came about by looking at the value of large capacitors I could buy and seeing which of these capacitances would satisfy ("satisfy" meaning the ADP2504 output/SIM7600 input would not drop below 3.4V) a small LTspice simulation I made using an ADP2504 model and a current source drawing 2A bursts periodically. 1500uF fit the bill.

Potentially off-topic but this 1500uF cap may not even be necessary (?). 2G networks have been shutdown in my country so I'll be using the SIM7600 for LTE and I'm not sure if LTE needs the same (or similar) 2A bursts of current. If it doesn't, then I think I can use a smaller capacitor (these bursts of current are pretty much the only thing the 1500uF cap is for). This would be the most practical scenario (get rid of the 1500uF capacitor) however I need to read up more on LTE and I'd still like to know/learn what's happening.

EDIT/UPDATE #2: With RemyHx's comment about inrush current and John Birckhead's answer regarding the soft-start (and 30A of current!), I think it's very likely that this inrush current is the problem. I went out to a nearby electronics store and bought a few NTC thermistors to see if putting them in series between the ADP2504 and SIM7600 would do/fix anything (as there are NTC thermistors used to limit inrush current, however I don't think I bought any designed with this purpose in mind). Putting a thermistor in series (I bought different thermistors each specified as 1k ohm, 4.7k ohm and 100 ohm, I bet this is the room temperature resistance) and powering the board, the excessive current problem did not occur. However, the problem is the thermistors I tried are acting more like resistors as they don't heat up/lower their resistance (between the thermistor and the ADP2504 I indeed have 4V as desired but after the thermistor I don't get more than 3V). The next step is to look to buy a specific thermistor made for this purpose or explore other means of limiting inrush current (RemyHx mentioned using a MOSFET). Will definitely update/report back.

Also, it is possible that the series resistors I used before (5 and then 10 ohm) in hopes of limiting the inrush current were too small and didn't limit the current enough (which allowed the problem to keep occurring).

\$\endgroup\$
9
  • 1
    \$\begingroup\$ I think the problem is the very high inrush current because of the initial load of the capacitor (the 1500uF one). If you really need this one, you should reduce the inflow when the circuit is starting up. Like with a mosfet. Or smaller cap. \$\endgroup\$
    – RemyHx
    Commented Dec 20, 2022 at 5:51
  • 1
    \$\begingroup\$ Welcome! I see explanations of most things, but not the titular 1500uF capacitor. Why was this chosen? I don't see any more than 330uF in the radio datasheet, at least at a glance, and the ADP definitely doesn't need it. Also the radio's peak current draw is "2A typ", while the ADP's output is 1A. Why the discrepancy? \$\endgroup\$ Commented Dec 20, 2022 at 6:01
  • \$\begingroup\$ Hi @RemyHx , that is a good point and those are good suggestions (that I need to read up on). My 5 and 10 ohm series resistors likely weren't enough... I've also heard of using NTC thermistors in order to reduce inrush current which may be useful... Do you have any thoughts on that? \$\endgroup\$
    – entrée
    Commented Dec 20, 2022 at 6:42
  • \$\begingroup\$ @TimWilliams , Thank you for the welcome :) and thanks for the suggestion, those are important details. I went ahead and added an explanation for current capability discrepancy and the 1500uF capacitor at the bottom of the post. I hope that helps \$\endgroup\$
    – entrée
    Commented Dec 20, 2022 at 6:44
  • \$\begingroup\$ Ok. Then it sounds like you didn't expect the ADP to draw ~1.5A input? \$\endgroup\$ Commented Dec 20, 2022 at 7:00

1 Answer 1

3
\$\begingroup\$

Your problem is the soft start. From the ADP2503 data sheet: enter image description here

Since you have only 200 usec to ramp up to 4 volts, your rate of rise (dv/dt) would need to be 4 volts/200 usec or 20,000 volts/sec. In a capacitor, current is C*dv/dt so for 1500 uF this is 30 amps. Per the data sheet, 20 uF is recommended.

You could try replacing the 1500 uF capacitor with a circuit like the one below, depending on how often you need to handle a burst.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.