Guru
Joined: 08/02/2015
Location: New ZealandPosts: 1077
| Posted: 02:40am 24 Nov 2022 |
Its good that others are interested in having a go designing a mppt charger. I have a few general observations about that design for use at higher voltages.
1: Those panels are a high voltage 90voc, with 2 in series giving 180v. This presents a few issues when charging a 12V battery.
The pwm buck duty cycle will be very small, so the mosfets will have high peak currents leading to inefficient operation. High voltage mosfets have quite high on resistance compared to the lower voltage types, thus necessitating multiple parallel devices. I would start by placing your panels in parallel to get the PV voltage down, so you can use lower voltage mosfets 150v, for 60 amps charge current you will still require 2 devices in parallel for efficient operation.
2: Your 3.3 and 12V regulators wont run off higher input voltages, I would power them from a 12V buck\boost switch supply running off the 12V battery AND replace the low 3.3v switching supply with an analog regulator. This will get the EMI noise down especially for the ADC circuitry.
3: The circuit has no means to measure the battery charge current, there is no need to measure the input PV current, after all it is max battery charge amps that is required when in MPPT phase; also to prevent the synchronous buck stage from acting as a reverse boost converter during very low buck duty cycles its important to know when the battery current could start reversing, when this condition starts and detected by the very low charge or reverse current then the synchronous rectifier mosfet can be disabled so preventing this disastrous condition. You will require a different mosfet driver to achieve this.
4: Power the buck driver from a 1212-1W isolated psu block rather than using the D3 bootstrap diode, with batteries the charge currents may be halted, thus your driver chip will have no volts available.
5: PV and Battery voltage sensing; I have found simple 3rd order active filters work much better than using just a large capacitor alone, the response is quicker and noise level lower. I now use an analog mux for selection of all analog signals with the mux output going via a single 3rd order filter then the CPU ADC.
Cheers
Mike