Setting up:
It’s been a while since I have worked with the board powered up. I hooked up the lithium cell and USB cable to my Computer, then the charger started doing it’s job. I then started the serial terminal on my computer. I hooked up my Oscilloscope to watch Vpp live during testing.
The code still in the board tries to regulate Vpp according to the reading it gets from the ADC. This isn’t working well. Since It’s been a while, I built and re-uploaded the project without any changes. This reminded me how I had the tools set up. Of course, the behavior of the board didn’t change with this upload. Next, I disabled the feedback loop to do my testing.
Voltage verses duty cycle:
I created and filled out the table below. As I tested and filled in the values, I added the filtered ΣΔ columns to try to get a better understanding of what is going on. The numbers didn’t make sense to me. I measured the voltage at the base of the transistor and as expected it was consistent value(0.64V) when the duty was 76 or higher. For most of the operational range the feedback voltage to the FAN5331 is 0.24 volts. I was expecting it to be 1.23 volts(page 6 of On Semiconductor FAN5331 datasheet).
ΣΔ Duty | Filtered ΣΔ Voltage (V) | Measured Vpp (V) |
---|---|---|
25 | 0.40 | 4.48 |
51 | 0.720 | 4.48 |
76 | 0.96 | 6.48 |
102 | 1.28 | 8.64 |
128 | 1.52 | 7.92 |
153 | 1.76 | 7.52 |
179 | 2.08 | 8.08 |
204 | 2.32 | 8.48 |
230 | 2.56 | 9.04 |
I went back and started analyzing the design. I noted in the data sheet that the recommended inductor should be around 10μH. I used the 2.2μH that was already on the board. I changed it and re-tested.
ΣΔ Duty | Filtered ΣΔ Voltage (V) | Measured Vpp (V) |
---|---|---|
25 | 0.56 | 4.56 |
51 | 0.88 | 4.56 |
76 | 1.12 | 13.5 |
102 | 1.44 | 25.2/9.4 |
128 | ||
153 | ||
179 | ||
204 | ||
230 |
Clearly the inductor was the problem!
Setting duty cycle to 102, I ended up with a output voltage of 25.2 which then dropped down to 9.4 volts. 25.2 is well over what I think I will ever need. I could put a resistor between the emitter of the transistor and GND. This should give me more consistent gain for different circuit conditions. Going from 4.5 volts at 51 to 25 volts at 102 seems like a narrow range. If I assume it’s linear from 76 to 102 duty cycle, I get 12 Volts/ 26 counts = 0.5 volts per step. I would like to have 0.1 volts per step. That’s a 5 to 1 ratio from where I am now. I think I would like to try changing the 100 K resistor to 10K to see what happens. That makes me a little nervous, I will have to recalculate the gains to make sure I can get to 20 volts.
Feedback:
As always I am looking for your feedback. Have you cut in a circuit and forgot to change a critical component like the inductor? Do you think I need to make the volts per step smaller? Do you have any similar stories?