4
\$\begingroup\$

I'm new here, so forgive me if I breach any forum protocols in the asking of this question.

I have designed a voltage regulator scheme for a guidance system on an aerial drone. All components require either ~4V (they have an input range) or 3.3 V exactly. My solution is pictured here in a schematic and a board layout. I used the TPS53318 switcher to regulate my battery voltage down to 4 volts with roughly 3 mVpp noise, then used an LP38502 LDO to convert that into nice clean 3.3 V with as little wasted power as possible. Datasheets here: TI TPS53318 and TI LP38503.

Schematic:

enter image description here

enter image description here

And board layout(image isn't very good, but I figured I'd include it just in case it helps): enter image description here

So, my problem: It doesn't work. The output from the switching regulator is a very small oscillating negative voltage, which I assume is caused by some sort of LC oscillation in the absence of any real output.

  • I've isolated the LDO, and it does it's job fine when supplied with 4V from an external supply.
  • I have tried reflowing my solder, even replacing the switcher itself (I always order 3 times as many as I think I'll need) and I still see the same problem.
  • I've done my best digging around with a multimeter and a scope trying to find a resistor with the wrong value or something, but I'm coming up empty handed.

This is my first time trying to troubleshoot this type of circuit, or any electronics this small on a proper board, so I'm hoping that someone here might point out errors I might have made or possible bugs I can attempt to diagnose. The oscillating negative voltage output still seems odd to me, and I'm not entirely satisfied with my idea that it is caused by LC dynamics, so I'm also hoping that someone might be able to suggest a possible cause/solution for that effect.

\$\endgroup\$
  • \$\begingroup\$ What is your battery voltage, and what is your undervoltage lockout set to? \$\endgroup\$ – Dan Mills Oct 31 '17 at 23:49
  • \$\begingroup\$ Also I am a little suspicious of the R7,C7 pair values given the value of R10, see 8.3.13 of the datasheet for how to size this. It could be right, but is is setting my spider senses tingling. \$\endgroup\$ – Dan Mills Oct 31 '17 at 23:51
  • \$\begingroup\$ @DanMills The battery voltage is from a 3s lipo pack, so nominally 11.1V. I'm doing my testing with a 10V benchtop supply because 10 makes the math a little easier. The turn-on threshold for the EN pin is about 1.3V, and my enable divider has a ratio of about .17, so my low voltage shutdown will be around 7.5V if I'm not mistaken. I'll take another look at how I selected those C7 and R7 values, thanks for the tip. \$\endgroup\$ – phillipd94 Nov 1 '17 at 5:14
  • \$\begingroup\$ Reading the datasheet, I notice that the output capacitance (and ESR) is critical to loop stability; what type of capacitors are you using at the output (and are they in fact 33uF)? \$\endgroup\$ – Peter Smith Nov 1 '17 at 10:38
  • \$\begingroup\$ @PeterSmith I am using 33 uF low ESL ceramic capacitors, details here: digikey.com/product-detail/en/tdk-corporation/… \$\endgroup\$ – phillipd94 Nov 1 '17 at 16:05
2
\$\begingroup\$

The first thing I noticed when looking through the datasheet and your schematic was your output components. First of all, your output caps look rather low in value. You have 4x 33uF where the 2 examples on the datasheet give you 4x 100uF and 2x 330uF. These are usually a good guide as to what your output capacitor values should be. I'm unsure of your exact requirements but double check your calculations for those. Do the same with your input capacitors too. Yours again seem quite a bit lower than the ones in the datasheet. I apologise if you have calculated these and they are correct, but it usually works for me to use the datasheet examples as a good guide. If they ARE too low, perhaps that could be a cause of your output not working correctly.

The other thing is your PCB layout. With these kinds of IC's (with a power pad underneath) it usually indicates they need ALOT of solid grounding work on the PCB. You should at the very least have yourself a nice big ground plane on the bottom side of your PCB and some vias underneath the IC to allow it to dissipate heat properly. When I first designed circuits using IC's like this, I had a number of failed units due to insufficient grounding, this is really critical in these designs. Your input and output capacitors should be located as CLOSE to the IC as possible which it doesn't quite look like they are.

I know it is a frustrating thing to redesign a PCB that you have already had made, but sometimes you need to.

Just make sure first that every component is in the right place, the values are correct and orientation of components is correct (on those that require it) before thinking about my suggestions.

I think it could most likely be the PCB design over anything else though.

If anyone would like to disagree and correct me on something I may have missed, please feel free as I know what I suggested is a pain in the a** to do!

\$\endgroup\$
  • \$\begingroup\$ The image of my board layout shows only the top copper layer (an oversight on my part), the bottom copper layer is a continuous ground plane, and I placed 8 vias under the thermal ground pad of the switcher IC and on each grounded component. I agree that I could have placed the capacitors a little closer, but I had to work around a huge outline for my external inductor that isn't pictured either. It could be the PCB, but reordering would be a huge pain so I'm going to investigate all other options before accepting this answer. \$\endgroup\$ – phillipd94 Nov 1 '17 at 18:45
  • \$\begingroup\$ I feel like a real dummy, the problem was a cold solder joint on one of my tiny 0402 resistors. \$\endgroup\$ – phillipd94 Nov 2 '17 at 19:51
  • \$\begingroup\$ Ahh, well done for spotting it then! That is very lucky indeed! Sorry, I should have asked if you had a bottom layer before assuming the top was the only one. I am glad you managed to resolve your issue without having to reorder PCBs! \$\endgroup\$ – MCG Nov 2 '17 at 19:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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