Just finished doing the physical install for a sun2000-10ktl-usl0. In general it was rather straight forward, and since I had most of the components for the install ready, it went rather fast.
1) Getting dc cables into the inverter. Huawei suggests using 600V cables. Maybe a bit old fashioned, and wanting some safety margin, I used the same pv cables to wire the inverter. These are 1000V, and unfortunately a lot thicker than regular thw or simlar cables. The dc holes are a bit tight (3/4"), but a full inch would had been nicer. I used both the PV and battery inputs. I grouped the positive cables and negative cables by themselves.
Also don't know why the terminals are arranged by arrays, rather than polarity, considering that they all are tied together. This might had allowed for better isolation between them.
2) Making the connection. You really need to apply some force to open the slots. If you are as concerned as I was, watch the install video. The terminal gets distorted a bit to open the clamp. However, I'm wonder that if they heat up, for then the metal would loose some of the springiness. It is only 10A, but if you don’t make a good connection, these can heat up.
3) The Power Compartment isn’t just that, it is a combiner. It connects all 3 arrays in parallel, so you might have problems testing your wiring after you plug it in. The diagram is quite accurate, and the DC Switch is the big one at the bottom of the unit. So you parallel PV array 1 and 2 after connecting them to the inverter, even with the dc switch disconnected (took me a few minutes to find the “short”). Once you turn the dc switch on at the inverter, all 3 strings are connected together. So if you are planning your cabling runs, the positive wires are all tied together, and the negative also get interconnected.
4) The optimizers can be a bit misleading. Until you run the inverter for the first time, you won’t get a voltage out of them. You just need to be careful when wiring, since you don’t have a chance to check the voltages beforehand. On the plus side, when the inverter is scanning for the optimizers, it only activates a few at a time, so you shouldn’t have issues with connecting two strings backwards.
5) The optimizers aren’t just blackboxes. When planning my strings, it was hard to figure how big to make them, since I couldn’t find enough information on how the optimizer works. Now from what I saw, it determines what voltage a string should be, and the other strings try to adjust to the same voltage (it is never good to connect sources at different voltages, since you induces a current to flow among them).
At first I thought the optimizers were buck converters, but turns out these can adjust the voltage up or down. They aren’t limited to the pv panel's voltage, but can go all the way up to 80V. So you can connect a shorter string to larger ones, as long as you can provide enough voltage to match the others. Otherwise, I guess, the longer strings will have to reduce their voltage and you might lose power doing so.
My guess is that the optimizers can adjust their voltage anywhere from 4 to 80 volts, which allows you to string more panels than the Voc (or even the Vmpp) would allow. You can also use less panels than the minimum operating voltage or than the other strings.
The congress of optimizers and inverter, determines a working voltage so that all the strings can be within the operating voltage (300–500 V) of the inverter. On the other hand the optimizers would try to adjust the voltage so that the current is within their limits (15A). The relation of the output should be close to the input Vin * Iin = Vout * Iout. The current limit only becomes an issue when the output voltage is less than the input.
6) Power Meter. I used regular category 5 STP to wire the power meter to the inverter, as the RS485 cables are too expensive. The distance was probably about 50–60’ away, and not sure if termination was needed. Since I connected two meters (the inverter only works with one, but maybe in a future update), the Gavazzi meter has a built in termination terminal. As far as meter selection, I probably would had chosen a different. I went with the CCS-WNC, but might had used the EM340 or DTSU666-H. These have a display, so it gives you an idea of what is going on. The EM340 has the ports for the RS-385 over the screws to the output power cables. So be careful when connecting it.
