Switching solar panel strings between inverters/SCCs

[fafrd]

Yes.

Prior to the invention of switching power supplies, diodes and capacitors were used extensively in power supplies. The amount of ripple depends to a great extent on the load. For a high impedance comparator input the output voltage is very stable, unless the capacitor is leaky.

By using a string of series capacitors separated by inductors, you can get a very steady dc output voltage. It can be further stabilized by using a full wave bridge rectifier and a single transistor. For what I'm doing the single diode and a 1000 uF capacitor are more than adequate.

Yeah, this technique works nicely to detect import versus export.

My goal is a bit more complicated for two reasons:

I have an export limit of 3.5kW so I want to detect when power is approaching that level rather than 0W (& 0A) net.

Using an adjustable resistor and a comparator I should be able to find a bias voltage to correspond to approximately the correct current level, but at that point, I’m doing nothing more that an OTS current switch is found, so it hardly seems worth the effort.

The second issue I have with either a current switch or a modified version of the technique you are using is that currrnt is an accurate way to measure power levels approaching 0W (since AC Amps = 0 @ 0W with either 120 or 240VAC) but it’s not accurate at higher power levels.

I’ve measured my grid voltage varying between 248 and 251VAC and my Microinverters are specified to operate from grid voltages of 211 to 264VAC.

So if I want to be as conservative as possible, I’ve got to convert 3500W to 264VAC x 13.26AAC to determine when I am approaching my export limit, but that means I’m actually cutting off at only 3182W for the majority of the time grid voltage is only 240VAC.

I’m still on the fence about what approach I’m going to take for that but in the meantime, I ran into these is case they can be helpful to you:

https://www.amazon.com/dp/B098DPR9X9/ref=emc_b_5_mob_t?th=1&psc=1

$30 to switch 80A @ 28VDC (2240W).

My entire DC-coupled array is under 2240W, so I’m starting to think about using one of these to switch between a grid-tied inverter and an SCC once the AC-coupled array under NEM exceeds ~5A / 1.25kW.

I would just use a 10A relay on the output of the grid-tied inverter to cut off AC power generation and then use one of these 2.24kW relays to connect the DC-coupled array to the SCC once AC-coupled generation exceeds the limit I’ve set.

From what you stated earlier, there should be no difference between switching 56A @ 40VDC = 2240W (PV input) and 77.8A @ 28.8V = 2240W (SCC output), correct?

 

[McKravitts]

I have an export limit of 3.5kW so I want to detect when power is approaching that level rather than 0W (& 0A) net.

Why do you have that limit?

I’ve measured my grid voltage varying between 248 and 251VAC and my Microinverters are specified to operate from grid voltages of 211 to 264VAC.

I'm pretty sure that 264 VAC is the maximum the microinverters are capable of.

Because of the extremely low impedance of the grid, you should be able to use your measured value of max voltage of 251 VAC in your calculation of cutoff wattage. That would leave you with a worst case cutoff of 3347 watts.
Will you be restarting export when the power drops below 3500? If so can an occasional small error be tolerable?

TWTADE/JQX-62F-2Z​

That's a nice relay. I've used very similar relays before and the've lasted years.
The only issue is the high coil current.
A high current latching relay would be a better choice but then you have to pay more and develope the switching interface

From what you stated earlier, there should be no difference between switching 56A @ 40VDC = 2240W (PV input) and 77.8A @ 28.8V = 2240W (SCC output), correct?

That's the rule of thumb many engineers use, but I'd guess there is some minor difference in the cycle life but it wouldn't be enough to worry about.

 

[fafrd]

Why do you have that limit?

With my utility, after they have reviewed your grid-tied solar system and reviewed your hook up to the utility pole to handle the power export, they approve your grid-tie installation up to an an export limit which you are not able to exceed. My Microinverters total to a peak output of 3.6kW and have sustained output of 3.2kW, but the export limit in my agreement states 3.5kW.

I'm pretty sure that 264 VAC is the maximum the microinverters are capable of.

Yes. 211 to 264VAC is the Microinverter spec.

Because of the extremely low impedance of the grid, you should be able to use your measured value of max voltage of 251 VAC in your calculation of cutoff wattage. That would leave you with a worst case cutoff of 3347 watts.

Yes, or I could assume a worst-case of 255VAC to leave a bit of margin. I’ve even found programmable voltage-based relays that would allow me to activate a dump load on the very rare times grid voltage exceeds some limit I choose (such as 251VAC).

Will you be restarting export when the power drops below 3500? If so can an occasional small error be tolerable?

Let’s call it the ‘turbo export’ on ton of the base AC-coupled export that is always there, and yes, the idea is to be adding turbo export whenever the total of base+turbo is not above 3.5kW.

Short spikes (under 15 minutes) that are modestly higher (100W above) should not be an issue but I don’t want to discover what the limits are by discovering I’ve been flagged by my utility.

That's a nice relay. I've used very similar relays before and the've lasted years.
The only issue is the high coil current.
A high current latching relay would be a better choice but then you have to pay more and develope the switching interface

Lot’s of ways to skin the cat. Adding additional solar power that just gets wasted has to be an option to bracket cost - a solar array that just fills a battery and then gets throttled (and wasted) is nice and simple and fail-safe so that reliability also needs to be considered.

That's the rule of thumb many engineers use, but I'd guess there is some minor difference in the cycle life but it wouldn't be enough to worry about.

My ideal system would export all DC-coupled power on the morning since it wakes up earlier, then continue that ‘turbo’ export as the larger grid-tied array begins to wake up until the export limit if 3.5kW (or 13.9A) is reached, at which point I’d want to taper-back turbo export as base export increases further.

From 12-4pm when the base array is typically maxed out around 3.2kW, either stop all turbo export or ideally throttle-back to ~300W, and then after 4pm as base export starts to fade, scale turbo export back up to 100% (it fades earlier).

So I’m still toying with the idea of playing around with custom CT-sensor designs as you are doing.

Putting 3.5kW meaning 13.97A though a wire as a reference is impractical, but if I can put both base export and turbo export through a single CT sensor and find an effective way to ‘subtract’ the current corresponding to another CT sensor sensing 13.97A, it should allow me the use the GTILs to offset all consumption and maintain an export level which is as high as possible without exceeding 3.5kW throughout the peak part of the day.

This would widen my daily ‘plateau’ from 12-4pm to 10-5pm or even possibly 9-6pm.