The goal of this testing is to figure out why the MPPTs were not working at FSGP2019. There may have also been issues with the solar master box, but we are focusing on the MPPTs here.
The Nomura MPPTs have an SPI port that is broken out to pin headers. The SPV1020 datasheet mentions that important cell parameters can be measured via this port, however until now a serious attempt has not been made to get it working.
...
We did notice that once the modules go into their overvoltage state (reading 0x03FF), all 10 bit of voltage readings a 1, the chip had to be manually restarted - which could explain some of the issues that we were noticing on the previous car.
More testing was done on Nov 20, but no significant progress was made.
We noticed that the MPPTs never actually reached the maximum power point of the power supplies that were powering them (where the current and voltage limits were both hit). We believe that this was never actually reached due to the fact that the only thing that the MPPT does to change the input is the duty cycle of the switching, essentially changing the average current that is drawn from the input. Since the power supply does not react as fast as a solar panel to changing loads, nor does the output current drop sharply after a point (as shown in the below image), it is expected that the MPPTs were trying their best to track power point but could not due to the power supply limitations mentioned. These can be better addressed and tested when we actually have sun.
Solar panel voltage current curve from the SPV1020 application note. Also notice the location of the maximum power point.
This oscilloscope picture is measured on the input side of the MPPTs, with AC coupling on the scope to measure the voltage ripple. Notice the 10uS period of the ripples, and how there are roughly 4 spikes during each period. These are believed to be caused by the hard switching on of each phase (there are 4 interleaved phases). This confirms that the MPPTs are switching properly, and are attempting to find the Maximum power point.
Another reason that the MPPTs might not have been working during FSGP 2019 could be due to how they were configured and setup. The SPV1020 datasheet and application notes give lots of info on the operation of the MPPTs.
On the VOUT SNS pin, there are 2 voltage set points.
1.00V, where the converter goes in to somewhat of a constant voltage mode, still with all 4 phases firing. If the output still climbs higher, then it gradually turns off phases, reduces duty cycle, and eventually turns all phases off when it reaches 1.04V. At which point the output falls, and phases turn on again, and reaches 1.04V and the cycle repeats.
Setting the potentiometer (on the car for example) in no load (1.04V SNS) condition, we get 25.2V output voltage when vout SNS is 1.04V. However, when we then go to connect a load to the MPPTs, the voltage reference will maintain a 1.0V as there is now something to sink the output current into and pull the output voltage down. The MPPTs would be operating in the CV mode, where all the phases are running (maybe at reduced duty cycle) to keep the output SNS at 1.0V.Since the potentiometer divider when set to 25.2 VOUT, 1.04 VIN, we get a multiplier of about 24.2 for the output voltage in reference to VOUT SNS. So when we're in the CV mode when the load is attached, we have the voltage output at 24.2V, so we are 6V off of target output voltage for the whole string (possibly more or less depending if my quick voltage divider math was correct),and the battery would not be charged when it is almost full, and would charge less (MPPT CV mode) when it just reaches a low enough voltage to charge.