I am working on a project where I want to have 3 batteries create 3.3V with a step down converter / switched mode power supply. I designed a circuit where I use the TPS560430X3FDBVR as a step-down converter to generate 3.3V from 4.7V. The TPS560430 does not step down the voltage properly and I see my input voltage at the output.
According to the datasheet:
The allowed input voltage range is 4…36V in the recommended operating conditions. For a 3.3V output, the feedback resistors should be chosen as 51kOhms and 22.1kOhms. The inductor should be 12uH.
I used a slightly larger inductor but from what I understand from peak current mode control (which is what this IC uses), this should not be an issue. As input and output capacitors, I used ceramic ones. For the layout, I used 0603 components (except the 1210 inductor) and I have a GND-Plane on the bottom side.
For testing, I only have a rather cheap multimeter which should be fine for most measurements. I double checked all soldering connections with the multimeter, they are all fine. I measured the feedback resistor values, 50.6kOhms and 21.9kOhms. When measuring the resistance, it took a few seconds to reach a stable value with my multimeter. Not sure if this tells me anything.
I also tried to use 5V as my input voltage and I also saw 5V at the output. Then I tried to solder a 100nF capacitor in parallel with R8, just to ensure that the reference for the feedback loop is stable. I also soldered electrolytic caps in parallel with the input and output capacitors (100uF for the input and 10uF for the output is what I had laying around). I did all my voltage measurements either with no load or with a 33Ohm load, which would be a 100mA current at 3.3V or a slightly higher current for the higher output voltage that I am getting. (No load current should be no problem according to the recommended operating conditions: 0…600mA.)
I soldered the circuit on two different PCBs and see the same problem for both PCBs. Also, at first I only saw an output voltage of 2V for an input of 4.7V on one PCB. Then I found that the bootstrapcap (C12) was not connected properly. I resoldered it and got 4.7V at the output again. This should tell me, that the TPS560430 should be working correctly, because this is exactly the behavior that I would expect if the bootstrapcap is missing.
I am all out of ideas now, what else I could try. Has anyone encountered similar problems with those ICs?
I would start with buying a (decent) oscilloscope. Then at least you can see whether the thing is for example oscillating in the first place. It also shows you noise on nodes which is a great help in diagnozes fault causes.
That is a horrible idea. This thing is supposed to switch at 1MHz, and then you create a time constant of
1/(1/51000 + 1/22100)*100e-9 = 0.00154 = 15ms
in the feedback loop.
Layout is also bad. The GND from the voltage feedback should have an independent return path from the power path. This is a synchronous chip, so there will be high currents both through it’s power and GND pins. The datasheet layout suggestion also shows vias to GND both directly under the IC, and on the left side of the GND pin just above the voltage feedback resistors.
Even worse is the layout of C16 and C19. These feed any output ripple voltage directly into the GND reference of the voltage feedback. Such things are completely unacceptable with an SMPS working at 1MHz.
I would not be surprised if this bad layout would be the sole cause of the thing not working at all. With an oscilloscope you would be able to see if this thing attempts to oscillate at all. I would also be curious at the noise level on R8, pad 1.
Overall this forum is intended for discussing and problem solving about KiCad itself, and not for circuit design. There are better forums for that, and I expect this thread will be closed soon.
You need to tie the FB pin to the OUTPUT directly as it has an internal divider for regulation and if you add an external divider it it change teh regulation point and/or follow the input voltage.
I don’t have a problem with somebody designing electronics without an oscilloscope but it’s absolutely required for debugging a board and characterizing it’s operation along with ensuring that it’s working correctly.
It is NECESSARY for working on most electronic circuits. Switching power converters are among those that require it. But skill and knowledge is needed to interpret what the oscilloscope shows you.
