The 65A rating is for #8 wires, sheesh! It's only rated at 15 and 20A for #14 and #12 wire. A string of 300W panels can't create more than about 9 amps.
--
Thanks,
- Win
Seems like big arrays can be series-parallel wired. I guess that saves on inverters, one big inverter instead of several small one.
Copper is expensive!
Probably Tesla wants to blame Amphenol.
The inverters I've looked at had multiple inputs to handle many strings. Tying strings in parallel doesn't work, they must be individually optimized. But once they start combining string outputs, they supposedly go through a junction box, and shift to metal conduit and a totally different kind of wiring. That's what happens on my roof.
Actually, not. Recently I was shocked at Alpha and Belden's high prices, and did some research: raw copper is a tiny fraction of their price.
--
Thanks,
- Win
Somewhere I had read that the arrays were "de-energized".... Doesn't make much sense that a connector's series resistance will cause any heat if the series circuit is opened.
There is more going on here that they are not talking about. Could even be bad bypass diodes in the modules. Don't know where their modules come from but maybe China ?
And combiner boxes with series fuses or breakers are used when three or more strings are combined. Series fuse ratings are usually 15 amps but some more than that. Maybe
20A which is why combiners are required so that a bad (shorted) module cannot be "ganged up" on by the other strings.Also, DC arc fault protection and ground fault detection/protection are required these days as well as RSS Rapid Shutdown Systems.
But last November, the roof of a Walmart in Yuba City, Calif., ignited. When the blaze was discovered, wires on the roof were still sparking, the complaint said, even though the solar panels had been disconnected for five months.
Either:
If "de-energized" means tripping the RSD (rapid shut down) switch, then yes, there is no load on the panel, and therefore no current to cause heating. However, if it was de-engergized by shutting down the micro-inverters or a central grid-tie inverter, these inverters, or the associated wiring, might still present a load.
I had originally thought that because of the lack of similar solar panel fires at locations other than Walmart, that it must be either something unique in the components used by Tesla, or something in the manner they were installed. However, that just changed as Amazon is now indicating that they had a similar problem: The article claims that Tesla knew about the problem and is blaming the Amphenol connectors. Specifically, Tesla was replacing connectors and optimizers, parts that are meant to regulate the amount of energy flowing to a solar panel. Too much energy can cause a fire.
The connector in question is on the SolarEdge optimizers: "Optimizer" seems to be the same as "micro-inverter". The modules are mounted on the back of the panels and are far away from the roof. In my never humble opinion, a burning module is not likely to start a fire at that distance.
The cables, with connectors at midpoint, are a different story. If installed in the manner of the SolarEdge photo, they might be dragging on the roof and start a fire. However, that installation not to code according to current standards (NEC Article
690). The DC wiring is suppose to be enclosed in metallic (RMC, GRC, IMC, or EMT) conduit or in a raceway or cable tray. The following is form before the 2017 changes to Article 690, but does cover most of the applicable and proposed wiring requirements: If a connector caught fire, the fire would be contained to the conduit, raceway, or cable tray. The enclosed DC wiring (part of RSD) became effective on Jan 1, 2019, so it is likely that the Walmart and Amazon systems were installed to an earlier NEC standard and therefore might lack enclosed DC wiring. I can't tell from the burned out roof photos if the DC wiring was protected by conduit, was dangling wire dragging on the roof.-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Did one of these articles say that Tesla was using Solar Edge for these installs ? I know they were going to use them a while back...
Raw in what sense? There's a lot of processing between spot price and drawn wire, and more still to make that into stranded insulated wire.
Also, were those brand names priced manufacturer-direct, or via distributor?
The overhead from Chinese production -- if that's what you're comparing with -- is indeed very small, something like one third.
(It may be even more, if you don't notice or mind that you actually got copper plated/clad steel or aluminum!)
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website: https://www.seventransistorlabs.com/
Sorta. Tesla does use SolarEdge, but allegedly may not have used them on the Walmart rooftops that caught fire: The faulty parts in question were connectors - Amphenol H4 connectors - and SolarEdge optimizers. (...) SolarEdge responded to Business Insider noting that it was not in any way involved in the lawsuit between Walmart and Tesla, nor are SolarEdge products on the rooftops cited by Walmart in the lawsuit. (...) We can affirm that to our knowledge, our power optimizers have never been the source of any fires.
At this time, I'm not sure whom to believe.
SolarEdge CEO & Founder Guy Sella Dies, Aged 54
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Not that it's relevant to your point, but an "optimizer" and a "micro-inverter" are two completely different beasts. An "optimizer" individually takes each panel's output and DC-DC converts it to an optimum series voltage so all its available power is delivered to the series string, which must have equal current through each panel's optimizer.
OTOH, there's a "micro-inverter" on each panel, taking its output and converting it to Vac line voltage, delivering as much AC current as it can for its instantaneous power level.
E.g., my Enphase micro-inverter roof puts out 60Hz 230Vac, up to 12 amps per section, and 37A max total. But if the 230Vac line source is removed, all microinverters cease to function, and the system is OFF. It's possible an optimizer system will still try to maintain 600 to 800 V at its output.
--
Thanks,
- Win
Notice that I said "seems" which is my way of saying that I'm not quite sure. Both Enphase and SolarEdge have some kind of MPPT (Maximum Power Point Tracking) circuit, which what you're describing. However, the SolarEdge device outputs DC and requires an external DC to AC inverter, while the Enphase outputs AC directly. I would guess this was done to avoid patent debates. Since both have MPPT circuitry, I would consider both to be "optimizers". Since the inverter is external in the SolarEdge system, they could not be called "micro-inverters". I guess "optimizer" is an adequate description.
I don't know which model SolarEdge "optimizer" Tesla may or may not be using at Walmart and Amazon. Skimming the data sheets, I selected: as a likely device. "OUTPUT DURING OPERATION (POWER OPTIMIZER CONNECTED TO OPERATING SOLAREDGE INVERTER) 85V or 15A DC"
Also from the data sheet, the SolarEdge devices go to 1V output with the "POWER OPTIMIZER DISCONNECTED FROM SOLAREDGE INVERTER OR SOLAREDGE INVERTER OFF".
MC4 connectors:
MC4 Y-branch connector failure video:
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
The M4-type connectors generally come pre-molded from the PV panel companies and the optimizer / micro-inverter companies, so installers simply plug them together. A special tool is required to unplug them. They look and feel quite rugged. But images of the contacts look very wimpy.
Supposedly optimizers shut off panel output, if the inverter is off or disconnected. However there are stories of shut-off cables sparking.
A series string of panels with M4 connectors is simply one wire, carrying DC current, yet supposedly the inverter somehow controls all the optimizers. I couldn't find out how it does that, or how reliable that scheme is. There may be an issue of shorted MOSFETs.
In the case of the micro-inverters on my roof, one per panel, they require 230Vac to run, so disconnection is a good on/off control method. But micro-inverter systems cost about $50 to $100 more per panel than optimizer systems.
--
Thanks,
- Win
Perhaps you can explain what you think optimizers do and how they are used in the system. I'm not following what you seem to be saying. Are you sugg esting there is more than one optimizer per inverter? I'm surprised there are separate optimizers and inverters.
-- Rick C. - Get 1,000 miles of free Supercharging - Tesla referral code - https://ts.la/richard11209
Every panel has its own optimizer (OK, now they have dual optimizer boxes, four wires, two panels each.) Current must be the same in all optimizers in series. The optimizers control their output voltage, to match their panel's available MPPT power: low power low voltage, high power, higher voltage. An additional trick is that SolarEdge optimizers somehow control all the optimizers, so total series voltage is fixed, e.g., 400V. I wonder a) how they do that, and b) failure modes.
--
Thanks,
- Win
If the current in all panels in series is equal, how can each panel be "opt imized" separately??? I guess I've been thinking the current through the s tring to the inverter would pass through all the cells. But the reality is the optimizer isolates the cells of a panel from the output current so pan els in series run at the same current, but the cells of each panel run at d ifferent current from other panels?
-- Rick C. + Get 1,000 miles of free Supercharging + Tesla referral code - https://ts.la/richard11209
Originally, per module MPPT was to mitigate against partial shading where a partially shaded module would go completely out of circuit through its bypass diode(s). National Semiconductor had its "Solar Magic" units that it tried to sell for like, $60 each per module. It worked and they had a neat golf cart like example of it but it was too expensive.
Then others like TIGO (partially owned by SMA) got into it. It is very arguable that these optimizers actually help much in that regard but they also tout that they help with unmatched solar panels. The real way to mitigate partial shading is with a chain-saw and not to have any, but there are places you just have shading at least part of the day or in winter.
Solar Edge's modules evidently also limit the highest voltage that the PV string can see.
Another thing that can be done wihen multiple PV strings are in parallel and there is partial shading on say, one panel in one string is to just boost the voltage of the shaded string up a bit so that both strings are contributing their available power at the same voltage....
A shaded module in a string of several panels can really reduce the power output of the combined system because the entire PV array's MPPT voltage that the inverter finds is a compromise. It's a better compromise if both or more strings are all putting out maximum power at the same voltage.
Of course, an MPPT box (optimizer) on every module can typically also double as a rapid shutdown switch. BTW it is the SunSpec specification that tries to require that 1V residual across a panel that is OFF so that they can supposedly tell that it is working or not.
So, just as solar PV was starting to get inexpensive, along comes some fireman with their great ideas (and some with self interests) PLUS tariffs to bring the price of solar back up again.
It's the optimizer outputs that are in series, and of course each MUST be operating at the same current, regulating its series voltage to determine how much power it's contributing. The solar panel is the input to the optimizer, which works as a buck-boost converter between the panel and the optimizer output to the string. The ratio setting of the buck-boost converter determines its MPPT characteristic, where the panel sees a current load that's just right to pull its voltage down, from its open-circuit value, to the maximum output power point. It determines the optimum load current and voltage by continually moving the buck-boost ratio up and down, calculating the power output as it does so. It's fair to assert optimizers earn their keep, making a huge improvement over a simple series string of solar panels.
SolarEdge claims their optimizers run with over 98% efficiency, which is something one can achieve if a buck-boost ratio is close to unity. Think of the job the buck-boost is doing as adding or subtracting a little bit between the input and the output. The little bit comes from the input voltage, stepped down and applied as needed. If the little bit is 10%, and that aspect is 90% efficient, that's only a 1% loss.
--
Thanks,
- Win
The part I'm not getting is that the optimizer is designed to set the curre nt drawn from the solar panel to set the voltage so the power is at a maxim um. It is then not clear to me what happens at the output.
In a typical buck voltage regulator, when the input switch is closed the in put current is pretty close to the output current. The only variation is c urrent from the output cap which is there to smooth the voltage and the cir cuit will still operate without it. When the switch opens the input curren t stops, but the coil supplies current to the output. The on/off time rati o defines the ratio of the input to output voltage. It also defines the in put to output current ratio even though the currents are the same, but beca use the input current is duty cycle modulated.
Applying this same circuit to the solar panel optimizer, now there has to b e a large input cap so the input current is relatively constant. The switc h current still pulses and the duty cycle is managed to adjust the current flowing from the panel to find the MPP. The duty cycle will set the ratio of the input current and output current. The input voltage will vary with the panel illumination and the output voltage will vary along, but also fac toring in the duty cycle multiplier.
This is an easy circuit to imagine when driving current into a battery. Th e battery is nearly a constant voltage load so all the output variation wil l be in current. How does the optimizer work when the output is in a strin g of equal current optimizers? Do all the voltages just float? Seems to m e the multiple optimizers would interact and could be unstable. What actua lly sets the output current? Is that the inverter?
-- Rick C. -- Get 1,000 miles of free Supercharging -- Tesla referral code - https://ts.la/richard11209
If you are going to connect multiple MPPTs in parallel e.g. for charging a battery, each MPPT must have a constant voltage but variable current output.
If you are going to connect multiple MPPTs in series, each MPPT must be floating constant current but variable output voltage type.
Of course, the string voltage varies which may cause problems with some loads. You can't directly connect multiple strings in parallel. If you are going to feed the mains, you need a variable DC / constant AC voltage inverter. This should not be a problem, as long as the string minimum voltage is greater than the mains peak-to-peak voltage, just a simple PWM.
It would also be beneficial to have a parallel control voltage to set the constant current for each series connected MPPT, if the average output power may drop significantly (> 50 %) below peak output power.
Suppose the inverter is a 400V voltage sink... and the optimisers just put as much current on the line as they can. That'd work wouldn't it.
-- When I tried casting out nines I made a hash of it.
Assuming 1 m^2 panels with individual constant current MPPT, the maximum output power is 150 W. A constant loop current of 10 A could be used, thus the maximum output voltage from each panel could be 15 V. Make two strings with 50 panels each in series. Connect these strings in series and connect the midpoint to neutral/ground, thus the maximum output voltage between the ends of the strings is +/- 750 V relatively to neutral, still fitting inside the Low Voltage Directive (LVD 1500 V).
From this it easy to make AC with PWMing without needing any transformers.
With reduced solar radiation, the string voltage can drop to +/-350 V (7 V/panel) and it is still possible to produce clean sine wave for
230/400 V three phase. If modified sine wave or square wave is allowed, the string voltage could drop to +/- 250 V (3 V/50 W per panel).For US, the string voltage could drop to +/-200 V and still deliver clean 2 x 120 Vac.
Switching all MPPTs to a lower constant loop current, say 3 or 5 A would make sense for early morning and late evening and also during partly cloudy conditions.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.