Buck converter not working as it should

[Diver300]

The solar panel will only give out 560 mA max. If you are using a buck converter you are expecting to take the full output current, which could be nearly 1.6 A, some of the time and zero current the rest of the time. You need a capacitor, most likely electrolytic or tantalum, that can supply the current for the time needed, and then charge up when the buck converter is in the off period.

If we assume 100% efficiency, the output current will be 34*0.56/12 = 1.586 A. If you are running at 20 kHz, the period is 50 μs. The time that the output is on will be 12/34 = 35% of the time. The capacitor has to support 1.586 - 0.56 A for 35% of 50 μs, so the charge is just over 6 μC. With the 22 μF capacitor suggested, the change in voltage will be just over 0.25 V, so that looks fine, but the ESR (equivalent series resistance) of the capacitor will also have to be considered. There is no disadvantage in having a much larger capacitor.

 

[spec]

Hy Driver,

I suggested a 22uF ceramic capacitor and a 2m2F (2,200 uF) capacitor in addition. But it could be that the solar panel assembly already has a large capacitor fitted.

spec

 

[spec]

I know you didn't ask for this, but here is a much simpler implementation of your solar battery charger. It also has a better performance.

spec

​

 

[ronsimpson]

Spec,
I like your P-FET idea. Watch out for 34 volts on the G-S of the MOSFET.

 

[spec]

Pravin,

Your inductor, is it really 40 mili Henrys or did you mean 40 micro Henrys.

spec

 

[spec]

Spec,
I like your P-FET idea. Watch out for 34 volts on the G-S of the MOSFET.

You are a star Ron. Thx again-will fix.

spec

PS: now fixed- I hope.

 

[spec]

Just done some calculations on the inductor:

(1) ASSUMPTIONS
(1.1) Inductor: 40mH
(1.2) Input voltage (from solar panel): 34V
(1.3) Battery Voltage: 12V
(1.4) Switching frequency (PWM): 20KHz
(1.5) Switching mark-to-space ratio: 1:1

(2) CALCULATIONS
(2.1) Switching period= 1/20Khz= 50uS. Thus half period = 25uS
(2.2) The voltage across the coil in the coil charge phase (MOSFET on) is 34V-12V= 22V
(2.3) Q= LI= VT. Thus maximum coil current (I) = VT/L = (22V*25uS )/40mH= 13.75 mA

(3) CONCLUSION
The inductor value is around two orders (x100) too high or the switching frequency is two orders too high.

spec

PS: Ron, please check this for gross errors

 

[dr pepper]

Have you designed the system to be continuous, 40mH sounds very high for the inductance.

 

[Pravin Gosavi]

The solar panel will only give out 560 mA max. If you are using a buck converter you are expecting to take the full output current, which could be nearly 1.6 A, some of the time and zero current the rest of the time. You need a capacitor, most likely electrolytic or tantalum, that can supply the current for the time needed, and then charge up when the buck converter is in the off period.

If we assume 100% efficiency, the output current will be 34*0.56/12 = 1.586 A. If you are running at 20 kHz, the period is 50 μs. The time that the output is on will be 12/34 = 35% of the time. The capacitor has to support 1.586 - 0.56 A for 35% of 50 μs, so the charge is just over 6 μC. With the 22 μF capacitor suggested, the change in voltage will be just over 0.25 V, so that looks fine, but the ESR (equivalent series resistance) of the capacitor will also have to be considered. There is no disadvantage in having a much larger capacitor.

Thanks. I need exactly what you said. I have used 470 uF cap.

 

[Pravin Gosavi]

Pravin,

Your inductor, is it really 40 mili Henrys or did you mean 40 micro Henrys.

spec

Yes. It is 40 mH. I got the value from an online calculator.

I know you didn't ask for this, but here is a much simpler implementation of your solar battery charger. It also has a better performance.

spec

View attachment 100164
​

Thanks. I will also consider this circuit.

 

[Pravin Gosavi]

Have you designed the system to be continuous, 40mH sounds very high for the inductance.

Yes. The system is to be cotinuous current.

 

[Pravin Gosavi]

Just done some calculations on the inductor:

(1) ASSUMPTIONS
(1.1) Inductor: 40mH
(1.2) Input voltage (from solar panel): 34V
(1.3) Battery Voltage: 12V
(1.4) Switching frequency (PWM): 20KHz
(1.5) Switching mark-to-space ratio: 1:1

(2) CALCULATIONS
(2.1) Switching period= 1/20Khz= 50uS. Thus half period = 25uS
(2.2) The voltage across the coil in the coil charge phase (MOSFET on) is 34V-12V= 22V
(2.3) Q= LI= VT. Thus maximum coil current (I) = VT/L = (22V*25uS )/40mH= 13.75 mA

(3) CONCLUSION
The inductor value is around two orders (x100) too high or the switching frequency is two orders too high.

spec

PS: Ron, please check this for gross errors

That value was calculated by an online calculator. Is it wrong?? Can you please calculate inductor value for me? I am totally lost. Should I increase frequency?

 

[Pravin Gosavi]

As per buck converter theory the current should be raised to about 1.5 A in this circuit. And the inductor should be able to withstand this much current. As per Spec's calculations the Imax through the inductor can be 13 mA only. How should the inductor be designed? Please help.

 

[Tony Stewart]

All problems are caused by designer not knowing the correct impedance Z(f) for every passive and active parts including lead wires and PCB layout capacitance where impedance is high or during Switch off.

Metrics such as SRF and ESR are critical for L and C. Rdson and Ciss are critical for FETs... etc.

Learn L/R ratios of inductor design and ESR*C choices. low ESR Caps are 1 to 10us, standard caps are much higher. Batteries can be 10k Farads and also have a higher ESR*C time constant than low ESR caps but since C is high in batteries, ESR is very low and your MOSFET RdsOn and DCR of L is probably way too high.

Examine all power losses in each part, especially if warm or hot.

 

[spec]

That value was calculated by an online calculator. Is it wrong?? Can you please calculate inductor value for me? I am totally lost. Should I increase frequency?

No increasing the frequency is the wrong direction. reduce the switching frequency to 2Khz and gradually reduce the frequency until the buck converter is working correctly (I hope). The correct frequency with that inductor is around 200Hz at a first approximation.

You say that you only have 470uF connected across the solar panel. This should be increased to 2m2F (2200uF) like I said.

spec

 

[Pravin Gosavi]

I have examined all the parts. Nothing gets warm neither mosfet nor inductor.
Thanks for help.

 

[Pravin Gosavi]

Ok I will try with these frequencies. Thanks for help.

 

[spec]

As per buck converter theory the current should be raised to about 1.5 A in this circuit. And the inductor should be able to withstand this much current. As per Spec's calculations the Imax through the inductor can be 13 mA only. How should the inductor be designed? Please help.

You can keep a switching frequency at 20KHz but then you would need to reduce the inductor to 400uH. That means that you would only need one tenth of the number of turn on the inductor that you have now.

By the way, the inductor wire needs to be substantial. It would be good to wind two wires side-by-side for the inductor winding.

Can you show a picture of your inductor with a ruler by it to indicate the size?

spec

 

[Pravin Gosavi]

What do you say about this calculator:

https://www.daycounter.com/Calculators/Switching-Converter-Calculator.phtml

 

[Pravin Gosavi]

You can keep a switching frequency at 20KHz but then you would need to reduce the inductor to 400uH. That means that you would only need one tenth of the number of turn on the inductor that you have now.

By the way, the inductor wire needs to be substantial. It would be good to wind two wires side-by-side for the inductor winding.

spec

For 40 mH I think I have 4- 500 turns. So now I need only 40- 50 turns. Right?

And the inductor is a transformer EE type which I found in an old smps.

Thanks.