Continue to Site

# Super Capacitor bank Charging (Buck-Boost)

Status
Not open for further replies.

Hi,

I need to charge a super-capacitor bank. The specs are as follows

C = 50F @ 13.5 V (5 caps in series)
Charging current 10A.
Input voltage range 9-18V.

What is the suitable topology for this purpose?

Best Regards

Your input voltage charges 5 series caps so each has 2 or 3 Volts. Is that your plan?
Do you plan to boost the input voltage? By capacitive multipliers? By switched-coil converter?

Multiple series caps may acquire unmatched voltages after a few charge/discharge cycles. Conceivably one cap might be at 13.5V, while the others take little burden. One or two components may become reverse-charged.

Therefore a chain of equalizing resistors could be a good idea.

Or do you wish to charge each cap one at a time separately? It's possible though not easy.

Your input voltage charges 5 series caps so each has 2 or 3 Volts. Is that your plan?

There are 5 caps in series and single capacitor voltage rating is 2.7V. So the maximum charging voltage is 13.5V

Do you plan to boost the input voltage? By capacitive multipliers? By switched-coil converter?

Initially the input voltage range was 15-20V. But now it is from 9-18. So i need buck-boost type circuit . The charging current is 10A. I want to keep the design simple.

Multiple series caps may acquire unmatched voltages after a few charge/discharge cycles. Conceivably one cap might be at 13.5V, while the others take little burden. One or two components may become reverse-charged. Therefore a chain of equalizing resistors could be a good idea.

Initially i don't want to use capacitor balancing.

Or do you wish to charge each cap one at a time separately? It's possible though not easy.

I want to keep the design simple but efficient.

Hi,
I want to keep the design simple but efficient.
Does this mean: although you know that the capacitor voltage will get out of balance, your primary goal is to "keep it simple"?

Klaus

I can focus on cap balancing once the basic circuit is finalized.

No matter if balancing is needed from the start, it will be usually implemented in the capacitor battery, not in the charge controller. In so far it's a separate topic.

I'd consider four switch buck-boost converter.

Sepic - easy choice ...

Is it possible to first boost the voltage to 24 V or close to it, then use a buck stage to charge the capacitor? I think the efficiency will be low as compared to the single stage4 switch buck-boost converter. However for a 4 switch buck boost convert i need to use specific pwm chip.

What is the complexity level with Sepci converter? loop stability etc

Regards
--- Updated ---

I found these to application notes for sepic converter. The design is built around UC1854 PFC regulator. Is it possibble to use UC3843 type IC?

#### Attachments

• slup103(2).pdf
1.4 MB · Views: 75
• slup104(1).pdf
776.2 KB · Views: 91
Last edited:

If you want a simple solution, use Sepic as suggested by Easy peasy. With output diode, you have only one controlled switch.

Easy peasy

Points: 2
Your application power is quite high >100W
So, I don't suggest SEPIC, which have lower efficiency, need high current inductor, larger PCB and hard to design stable control loop.
I designed digital control BUCK - BOOST as below:
My circuit had run with 5A@13.5V.
We can use MCU or controller UC3843.
MCU is smartter control. When input voltage < output voltage. Circuit run in BOOST mode.
When input voltage > output voltage, circuit run in BUCK mode.
The good result is I peak-peak of inductor is smaller, help better efficiency, cooler, lower cost, lower ripple.
If using UC3843, I peak-peak of inductor will be higher, because it always operate in BUCK-BOOST mode.

Points: 2
If using UC3843, I peak-peak of inductor will be higher, because it always operate in BUCK-BOOST mode.

Which MCU is suitable for this design. Did you implement the compensation loop in software?

Best Regards

currents in SEPIC chokes are no higher than in a buck / boost, and only one gnd referenced mosfet to drive - which makes total parts less - compensating is fairly straight forward, operating in DCM is possible if you want fast dynamics - but not needed for cap charging - slow dynamics are fine ....

Points: 2
UC3843 can simulate in LTspice as below:

Efficiency estimate is not so high as high switching current.

rename file uc3843-boost-buck-12V to13V8 CHARGER v3.jpg to .asc to simulate.
--- Updated ---

With MCU, need support these:

05 ADC input (04 also OK): Vin, Vout, output current, temperature, switching current.
02 independance PWM. For BUCK PWM and BOOST PWM.
For high speed MCU or DSP, I can use PWM trigger to ADC to measure switching current.

I used STM8S003 for the first prototype, HT32F5222 for the second prototype.
For charging application, don't need fast control loop as normal power adapter, because already know what is load - battery.
Only warning case is unplug, low speed MCU need a hardware circuit to trigger this event to shutdown PWM immidiately and reduce overshoot output voltage.

Compensation loop will combine base con Constant Current Mode (CCM) and Constant Voltage mode (CVM).
Infact, I have dip-switch to change charging current, floating voltage as 13.5V (12V lead acid) or 12.6V (03 cell LI-ion).

Firmware in MCU is not complex. It is my 12V DC UPS project.

Last edited:

Points: 2

I can not get the simulation file.

Regards

Status
Not open for further replies.