48 V PV panel compatibility – MPPT controllers

DIY
,
1

Hello everyone.

This concerns a domestic installation with two solar arrays: one rated at 3 kWp (EAST-facing) and one at 6 kWp (WEST-facing), charging a 48 V, 15 kWh LiFePO battery with a nominal voltage of 51.2 V.

I had initially planned to install 15 DualSun 600 Wc, 48 V, 144-cell panels – Vco = 52.4 V, Vmpp = 45 V, Impp = 13.35 A.

The PV panel-to-battery interface was to consist of three MPPT 150/100 charge controllers, each drawing current from five PV panels connected in parallel at the input.

1 - After reading the MPPT charge controller manual, I have realised that this configuration will not work: The MPPT input voltage must be at least the battery voltage + 5 V, i.e. 56.2 V (or even 63.4 V). I therefore need to connect the panels in series in pairs to reach a maximum voltage of 2 × Vco = 104.8 V, which is acceptable for a 150-xxx charge controller. HOWEVER, section 4.3 on page 12 of the manual states that these charge controllers cannot support more than 216 PV cells in series. My two panels form a series string of 288 cells, which would mean switching to a 250-xxx charge controller.

2 - Since I am obliged to connect the PV panels in series pairs, I have a problem with configuring the EAST array (five 600 Wc panels), as this is an odd number! A simple solution could be to use six 500 Wc panels and configure them as three branches, each consisting of two panels in series. BUT there is an issue: the 500 Wc panels have a Vmpp voltage of 33.5 V, i.e. 67 V per series string. Can I connect my three MPPT charge controllers via a VE-CAN cable to synchronise them, even though the two on the WEST side will have 105 V at the input and the one on the EAST side will have 67 V? Is this useful, given that the sunlight on each side will occur at different times?

3 - Page 20 of the manual specifies the MPPT charge controllers’ output voltage levels to the LiFePO batteries: 56.8 to 54 V, depending on the charging phases. This is incompatible with high-capacity domestic batteries (51.2 V – 16S), for which the maximum charging voltage can reach 58.4 V, the float voltage is 51.2 V, and the end of deep discharge is around 42 V. Can these levels be adjusted via VictronConnect in expert mode?

Is my reasoning at the end of §1 correct? Could you tell me whether the configurations described at the end of §§2 and 3 are feasible?

Many thanks for your assistance. Jean

2

Please use the MPPT Calculator for your system sizing.

And solar panels are almost always connected in series to form strings in larger installations.

3

Hello DirkW

Series connection: I definitely want to avoid it because of the shade cast by the adjacent trees, depending on the season. The panels will never all be active at the same time, but they will all generate power during the day. So I’m going to limit myself to strings of no more than two panels.

4

You’d need one MPPT per panel/string in your setup.

Alternatively, you could use microinverters and AC coupling.

5

That’s not right! Although each module has 144 cells, they are connected in parallel in two blocks of 72 cells. You can connect those in series without any concern. Only two modules can ever be connected in series to one controller. I would set up several Smarsolar 150/45 controllers, each with two series strings connected in parallel. I would connect the 4 mm² cables from the series strings in parallel only at the controller.

6

With three or more strings connected in parallel to the same MPPT, you need blocking diodes so that the two less shaded strings don’t damage the weaker one through reverse current.

I have the same problem, which is why I connected most of the panels using Hoymiles microinverters. That’s also why I have three MP2 5000s, so I can charge at 7 kW. In addition, I have an MPPT 150/45 with 2s2p for black start.

Where possible, you should build a system with 2/3 MPPT and 1/3 AC PV. You’ll have to see what works best for you.

7

Diodes aren’t always necessary.

I have, among other things, a 2s4p configuration, which is why I’ve spent quite a bit of time looking into the subject.

Even when two modules are fully shaded and the other modules connected in parallel are in full sunlight, the voltage difference is so minimal that no reverse current flows at all.

I therefore opted for DC circuit breakers rated between Isc and the maximum permissible reverse current.

However, I also have a few “special modules” from DAH Solar in use, each of which has around 115 V. So they can be connected individually or in parallel to an MPPT and used to charge the 55.2 V LFP batteries.

I don’t know whether they’re still available, though…

IMG_20241010_172456
IMG_20241010_1724561920×2560 488 KB

8

Let me respond to your suggestions:

Dirk: I’m going to avoid AC coupling, and therefore microinverters (Bluetooth/Wi-Fi/maintenance on the roof, etc.). One MPPT per string: some controllers have multiple inputs, so this is permitted by the manufacturers.

RL: That’s very interesting information, which I had already read somewhere, but which isn’t documented by the manufacturer. What a shame. Your observation is confirmed by Victron’s MPPT calculator, which allows this configuration with the 150-xxx models. The basic PV cells would therefore produce 0.73 V open circuit, whereas I had read that they generated more like 0.6 V. Regarding the connection: you’re right, and I had planned to connect each PV panel to a copper busbar in the photovoltaic switchboard. From now on, it will be the eight “series strings” that are connected to three copper busbars, each connected to one of the three MPPT controllers. Each controller would then receive 6 × 600 Wp, i.e. 3,600 Wp, which is too much for a SmartSolar 150/60 (3,440 W maximum). I’m therefore going to use SmartSolar 150/70 units, which allow 4,000 W (I want to have identical SmartSolars in case of a fault).

Björn: Noted regarding the need for additional diodes in addition to those in the panels. I’ll study the reverse characteristics of my PV panels. However, I don’t want AC coupling in the current setup, even though, if properly designed, the principle may offer some advantages. Perhaps if the system’s power is increased …

Jetlag: Yes, that’s my view too. I’ll study the reverse characteristics of my PV panels. However, using 115 V panels raises the issue of changing electrical class (in France): a Vco of 115 V at 20°C rises to 122 V at 10°C and means that the regulatory measures for voltages above 120 V must be taken (change from SELV to LV, meaning one pole earthed and installation of a permanent insulation monitoring device (GFDI in English)).

However, I would like my three SmartSolars to be able to synchronise in order to protect the battery. The solution is to connect them with a VE-Can cable. But is that relevant, given that the three controllers will never have the same sunlight at the same time?

Thank you for your comments and clarifications.

9

You can oversize the panels on the Victron MPPTs by up to 130%. This is the default setting in the MPPT Calculator.
The loss in output is small, as the maximum output is only reached rarely.
If you combine several strings in parallel, I would install the surge protection between the combiner box and the MPPT to save money.

Depending on how many MPPTs you need, consider whether VE.Direct will suffice or whether CAN would be better.

The CAN models are usually larger and more expensive, but with daisy chaining you can essentially connect as many as you like. The smallest is the 150/70.

Depending on the Cerbo GX, only three VE.Direct units can be connected. There are USB VE.Direct four-way adapters. In addition, the expensive VE.Direct cables have to be purchased. All things considered, CAN bus could be cheaper and would certainly be a cleaner solution.

The MPPTs should be connected to a Cerbo GX or similar device either via VE.Direct or CAN bus. They then switch to “external control”. Venus OS takes over DVCC control (distributed voltage and current control) based on the data from the battery’s battery management system (BMS). Venus OS thus controls all the devices, including the MP2/Multi RS.
When the battery is full, the MPPTs are throttled so that the battery is not overcharged. There are many more scenarios, but this should be sufficient for illustration.

10

Jetlag,

the reverse-current protection must always be taken into consideration with three strings or more!
You can use diodes, but special high-voltage PV fuses are better.
In the case of your DAH modules, these would be 20 A (FUSE RATING).

Otherwise, unprotected reverse current can cause damage in the module array (hotspots), which could potentially trigger fires.

11

Hello Karl-Heinz,

Yes, it’s something to consider, but diodes in particular aren’t the best choice. They always cause a voltage drop and power loss. There was a video by Energie&Hobby on YouTube that showed how these diodes can burn out…

That’s why I also installed DC fuses. The modules I connected in parallel have an Isc of around 11.5 A and a reverse current of 20 A. So I installed a 16 A DC fuse.

IMG_20221115_125650
IMG_20221115_1256501920×1440 320 KB

But as I said, I monitored and measured it carefully at the time, and the voltage difference between the shaded and sunlit modules was so small that no reverse current occurred at all.

IMG_20221211_111756(1)
IMG_20221211_111756(1)1920×1371 500 KB

12

The reverse current only occurs in the event of a fault, but then it is uncontrolled and at full strength, fed from all the connected strings into the faulty string.

13

Hello, and a thousand apologies for my absence over the past few days.

Communication between modules: I’m not going to connect the MPPT controllers to one another … for the time being. I’m already using the VE.Bus for another system, with communications centralised and managed by an STM32 microcontroller equipped with several serial inputs (and therefore able to handle several VE.Direct devices). The STM32F407 model also includes a CAN controller, which I have already used with NMEA2000 devices (marine navigation). If the Victron CAN protocol is well documented, I should be able to use it for a DIY solution, as I’m not particularly fond of “proprietary” systems.

Protection of the branches against reverse currents: the DUALSUN 600Wc modules deliver a short-circuit current (Icc) of 14.12 A and can tolerate a reverse current of 30 A. In a parallel arrangement of three strings, if one fails, the combined current from the other two strings (28.24 A) should not damage the installation. But let’s be cautious! Two types of protection are therefore possible: either a diode, or a fuse connected in series on each branch. “Diodes” based on MOS transistors produce virtually no voltage drop or power loss. However, they are electronic components and can of course burn out. I admit that I prefer fuse protection (bidirectional!), and the fuse rating needs to be determined. In my case, either 16 A or 20 A: I’m going to look into this based on the fuse-blowing curves. I had overlooked this point, but you’re right—it deserves a great deal of attention.

Jetlag: a fine production unit!

14

I’d also advise you to use fuses, even if your calculation is correct and there shouldn’t really be a problem.

I’d go with 20 A fuses.

Active diodes are great too, but the connections alone—in other words, wiring them up—are considerably more difficult, because you usually have to cobble something together yourself.

You could measure the parallel strings and let us know what the voltages are. :wink:

15

I expressed myself badly earlier: I will not be installing any MOS “diodes” (a Zener diode and two resistors to be added to the transistor—ready-to-use modules are available), but only one fuse on each string. I am thinking of installing 20 A gI-type fuses, as the panels can withstand 30 A reverse current and produce an Icc of 14.12 A.

Another point: distance between the panels and the controllers = 30 m—10 mm² cables will be needed … a very significant cost. But so far, so good!

For the moment, I am looking for a system for mounting the panels → combination screws with a 130 mm-long, 10 mm-diameter lag screw (for wood) + a metal screw. It is difficult to find the right equipment … and a supplier. The provisions have already been made on the roof. As soon as the fixings have been found, installation could progress quickly as supplies arrive. The inverter will be a 48 V MultiPlus II. In my view, the next tricky part will be studying the LiFePO battery bank: there are a huge number of 51.2 V, 300 Ah modules available on the market … but how good are they? Is the communication between the inverter and the battery bank compatible? That will be the job of the next two months, after studying the datasheets!

As for the monitoring, I am currently at the design and layout stage for the technical cabinet. So those measurements will not be available for another six months!