Further Nomura Testing

This has been a continued topic of discussion for the last term, and unfortunately we were not able to make a final decision on the Nomura vs Elmars yet due to not having enough information and confidence in the Nomuras, and not wanting to shell out money for the Elmars when the Nomuras might work. So, we're doing some more testing!

What to Test

There are a few things that we need to test in order to verify that the Nomuras will work properly.

• Recovery from input undervoltage condition
  • During some of our bench testing, we would need to turn off the power supply in order to get the MPPT to start working after an Input Undervoltage event.
• MPPT operation in series, at full load
  • We have been unable to get this to fail in our (limited) bench testing, but were not able to test at full power
• MPPT operation in parallel, at full load ideally
• Real MPPT algorithm function
  • We want to verify that the MPPTs can actually find a global max and not just a local max on the power curve
• Transition from pass-through to tracking mode (with both parallel and series arrangements)
  • This will happen with our batteries at low voltage, when charging through to the point of normal operation. We do not want to have to do any manual intervention here.

Notes

Expected power produced per panel: 28 cells * 3.72W (PMMP for Me1 Cells) = 104.16W

Test Setup

Given that the testing will be done in the winter (cold and not consistent sun) and that we need to be able to reproduce the testing results, we will be doing our tests inside, using artificial light. Power supplies also do not respond as quickly to the changing 'perturb and observe' algorithm that most MPPTs (including the Nomuras) use, and thus real MPP tracking is unachievable.

We will be using 2 of these lights (so 4x 500W halogen bulbs in total), pointed directly at the panels but not too close that the panels heat up (as the reviews mention that the lights get hot): https://www.canadiantire.ca/en/pdp/noma-1000w-twin-head-halogen-work-light-on-tripod-0524058p.html?rrec=true#spc

We will be using the array of MSXII for the testing, and the Nomuras that are attached to it. The modules on the rear trunk panel are identical, and we will use 2 of them (both 28 cells in series in a rectangular pattern).

Setup Procedures

1. Ideally, clean the solar panels
2. Panel Test Setup
   1. Get the Current-Voltage curve for the panels and light we will be using. This will give us a reference for where the MPP is, and what amount of power to expect from our setup.
      1. Connect the DC load directly to the panel in CV mode with remote voltage sense connected where the MPPT would be connected
      2. Do the test with 1-4 lights on each panel to see how effective 2 lights are to provide the required amount of light for the panels
      3. Sweep through the range of panel voltages (from 0V to 50V), recording the voltage and current at each step
      4. Creating a quick python script to run this test would be ideal, as this will allow testing of our new panels quickly
      5. Graph the results (for fun, and for proper data collection and visualization, and for calculation of the MPP).
3. 1-MPPT Test Setup
   1. For all MPPTs, connect panel, and adjust VR1 until P2 is 1.23V, relative to the negative input terminal.
      1. For ALL MPPT tests, connect the MPPT to the module, and then adjust the input potentiometer according to the 'Nomura-MPPT-Instruction-Manual', section 7.1: Connect solar panel, or DC power supply which adjusted as same as VOC of solar panel voltage. Check voltage between "P2" terminal and "GND PV_IN[-]". and adjust voltage to reach 1.25V (+/-0.02) by adjust screw "VR1".
      2. 1.25V is the reference voltage of the internal ADC, so adjusting to 1.25V at the Voc input condition will allow for a full range and highest precision in measurements. We could adjust to 1.23V is we wanted to allow some headroom for potentiometer vibrations changing the value during testing.
   2. Also adjust VR2 until MPPT shows 35V output voltage.
   3. Connect multimeter with PC logging capability to the panel input of the Nomura.
   4. Connect DC load power terminals to output of the Nomura, and the remote sense at the Nomura's output terminals.
   5. Connect clamp meter with logging capability to the panel input side of the Nomura.
   6. With these devices, we are measuring input current, output current, input voltage, and output voltage and will determine the efficiency of the MPPTs under various conditions.
4. 2-MPPT Parallel Test Setup
   1. Connect Nomura Ideal Diodes between the output of each MPPT and the paralleling point to avoid pushing current back in to the system.
      1. We should reverse engineer the Nomura Ideal diodes to fully understand them.
   2. DC Load and remote sense connected across the positive and negative terminals after the ideal diodes.
   3. The Current clamp will be moved around between the inputs and the outputs of the 2 MPPTs to check current sharing.
   4. Can be tested without the ideal diodes (use regular diodes, or no diodes since it should be built in to the Nomuras) if there are issues with the ideal diodes.
   5. An extra DC load or resistor stack will be required to achieve the full output power.
5. 2-MPPT Series Test Setup
   1. No ideal diodes needed (only one string)
   2. DC Load and remote sense connected across the positive and negative terminals of the stack.
   3. An extra DC load or resistor stack will be required to achieve the full output power.
   4. The Current clamp will be moved around between the inputs and the outputs of the 2 MPPTs to check voltage sharing.

Test Procedures

1. 1-MPPT Test
   1. Turn-on test
      1. With load off, turn off the lights, and cover the panel.
      2. Uncover panel and turn lights, and measure the output voltage to ensure the module turns on properly.
   2. Pass through to Tracking Test 
      1. Turn on the DC load (in CV mode) and swing the output from below the input voltage to above the max output voltage (10V - 50V)
         1. We could go down to 0V as well, and use the same DC load script used for the panel testing
         2. This will get us a graph of efficiency vs output voltage for a given input voltage, essentially efficiency vs boost ratio.
   3. UVLO Recover test (simulation 1 panel being shaded)
      1. Without turning the load off:
      2. turn off the lights and cover the panel (triggering the Under voltage lockout state)
      3. turn on the light and then uncover the panel (shade from panel removed) and verify that the MPPT turns on and produces power again.
         1. If it does not turn on, record what is necessary to get it to restart. We think it may be necessary to reset the whole chip by removing the load, and disconnecting the panel, then reconnecting.
   4. Temperature Test
      1. Connect panel and DC load, and all measurement equipment to measure efficiency. Close the MPPT lid, and monitor the temperature rise.
      2. If the temperature keeps climbing to the point where the module reaches thermal shutdown, then note an issue with cooling.
      3. We will get a graph of MPPT efficiency vs temperature, to see how much priority we need to put on cooling them.
2. 2-MPPT Parallel Test
   1. Do the same turn-on, pass-through, and UVLO recovery tests for each MPPT
3. 2-MPPT Series Test
   1. Do the same turn-on, pass-through, and UVLO recovery tests for each MPPT.
      1. For the series connection, when in UVLO (or shutdown), the MPPT will connect a diode across the output terminals, so we should see the stack function as 1 MPPT with a diode drop.

Nomura-MPPT-Instruction-Manual: explains more about the testing and tuning procedures for their MPPTs.

These python files can be used to plot the I-V curve of the panels and identify the operation modes of the MPPTs.

Testing Results

TEST SETUP:

(It was a lot brighter than the image shows. Each lamp is a 500W halogen lamp, for 2000W total)

Below are shown the graphs produced form the tests, the tests with single panels were completed with all 4 halogen lights shining on a single panel:

Panel Only

1 Panel, 1 MPPT

2 Panels, 2 MPPTs, Parallel (with regular diodes on outputs of individual MPPTs)

2 Panels, 2 MPPTs, Series

Notes - there was 'a little' more light for the 2 panel tests since the ambient lights were on both panels as well. During the single panel tests, all the light was concentrated on a single panel, thus giving higher per-panel output in the single panel tests.

SPI Testing

With all of our SPI testing with the power supply as the power source and not much success, I finally decided to just go for testing with the panels. And since we finally have lights that can at least give us some power, we were able to test it.

I swapped in a Nomura with the SPI-GND jumpers cut and probed the registers using Arshan's mppt-spi program. The resulting text file is added below.

With the panel, SPI worked amazing. The weird jumps in the input voltage values that we had seen with the power supply are no longer there. The values were changing almost continuously to maintain the maximum power point. Everything seems to be working well.

We also tested an undervoltage cutoff and restart - We loaded the output until the MPPT went in to pass through mode, then continued increasing the load until the panel voltage dropped below the UVLO voltage where the chip shuts off. At this point, we received all 0's over SPI. When decreasing the load, the MPPT turned back on, starting in pass-through mode, and went into MPPT mode when the input voltage was high enough. It worked flawlessly and went straight in to the MPPT algorithm. 

At this point, I don't see what we could have done wrong except for some weird wiring error that was never caught, and I'm starting to think more and more that this was actually the issue.