Optimizers... what are they good for?

[nwdiver]

Hard to imagine a bigger waste of money and resources than all the MLPEs out there....

The 'shade fix' thing is unfortunate. This would occur with any reasonable MPPT. But you need a pithy name to counter nonsense.

 

[arnolddeleon]

Hard to imagine a bigger waste of money and resources than all the MLPEs out there....

The 'shade fix' thing is unfortunate. This would occur with any reasonable MPPT. But you need a pithy name to counter nonsense.

Wait, making sure I understand correctly, this is SMA just branding MPPT with rapid shutdown?

 

[nwdiver]

Wait, making sure I understand correctly, this is SMA just branding MPPT with rapid shutdown?

No. Has nothing to do with rapid shutdown. It's how string level optimization out-performs module level the vast majority of the time.

 

[arnolddeleon]

Ok, it is even simpler, it's giving a brand name to string level MPPT.

I believe the new Tesla solar inverters also don't use module level optimizers

 

[nwdiver]

Ok, it is even simpler, it's giving a brand name to string level MPPT.

I believe the new Tesla solar inverters also don't use module level optimizers

Yep. String level is all you need. Generally speaking MPPT is MPPT There are some differences in the trackers 'intelligence' but the principle is the same.

 

[mspohr]

Hard to imagine a bigger waste of money and resources than all the MLPEs out there....

The 'shade fix' thing is unfortunate. This would occur with any reasonable MPPT. But you need a pithy name to counter nonsense.

Wow! What a bunch of BS!

 

[Ampster]

I saw a reference to RSD. How do SMA and Tesla accomplish panel level RSD?

 

[nwdiver]

I saw a reference to RSD. How do SMA and Tesla accomplish panel level RSD?

I've read that Tesla has string level somehow. Not sure how that complies with NEC 2019. SMA uses a sunspec module.

I hope the whole RSD thing gets re-evaluated like Australia is doing with their insane roof-top DC switch. The cons are >10x the pros on RSD.... The premise is wrong. I was terminating some ~500v PV lines and I could touch the conductor without harm. Why? Because it's an ungrounded system and I wasn't also holding the (-) end. I wasn't completing the circuit. I've never heard of a single firefighter being harmed by DC voltage on roof. It's nearly impossible but we're spending hundreds of millions soon to be billions of dollars a year to prevent it.....

One problem is that most installers have merrily gone along with the insanity. Tesla clearly isn't buying in. Hopefully they're powerful enough to get some change.

 

[uskalu]

@nwdiver Thanks for providing valuable info and educating us. I'm also planning to install solar on my roof, located in Hillsboro, OR. Tesla 12kw system predicting yearly energy harvesting about 11000kWh. A few local companies offer ~10kW system with enphase micro-inverter system, which predicts ~10500kWh yearly. In terms of cost, Both Tesla and other companies quoted $25000. In this situation price per watt and price per Energy are batching between Tesla and others. So do you think it is better to go with non-tesla since they offer micro inverters ?

 

[jrweiss98020]

I have seen so many complaints about installation and service problems with Tesla, that I HIGHLY recommend going with a local installer when practical.

I went with A&R Solar, who also services the Portland area. I am VERY happy with their service, and was well worth a few more $$.

There are several people here who are much more technically knowledgeable than I am regarding solar power, but I am a fan of the microinverters. If you can get them for the same price as a string inverter, I think it's a win-win situation.

 

[nwdiver]

So do you think it is better to go with non-tesla since they offer micro inverters ?

No. It might be better to go with non-Tesla for better service.

I would be suspicious of the discrepancy between system size and yield. A 20% larger system is generally going to yield 20% more kWh/yr regardless of the inverter type used. The difference between inverters is <2%. I would get exactly what's being proposed and run your own estimate via PV Watts to get an apples-apples comparison.

 

[uskalu]

No. It might be better to go with non-Tesla for better service.

I would be suspicious of the discrepancy between system size and yield. A 20% larger system is generally going to yield 20% more kWh/yr regardless of the inverter type used. The difference between inverters is <2%. I would get exactly what's being proposed and run your own estimate via PV Watts to get an apples-apples comparison.

That is another question I had. Different companies showed me different yields. So basically when I devide the yield by system size I get number of hours per year. So I can see that they have used different hours per year in their own calculations. When I average those numbers, it gave me about 1000 solar hours per year of 2.75 per day. I too agree that there shouldn't be such a significant yield difference. But once they use different solar hours they it is hard to compare them. Is there a better place to find solar hours for a given location.

Cheers

 

[nwdiver]

That is another question I had. Different companies showed me different yields. So basically when I devide the yield by system size I get number of hours per year. So I can see that they have used different hours per year in their own calculations. When I average those numbers, it gave me about 1000 solar hours per year of 2.75 per day. I too agree that there shouldn't be such a significant yield difference. But once they use different solar hours they it is hard to compare them. Is there a better place to find solar hours for a given location.

Cheers

Your best bet is to input system specs into PV Watts for your location.

 

[pdx_m3s]

Another thing to note is that I’ve seen multiple examples now of Tesla “jumping” two strings together at the inverter as opposed to running a pair of wires for each string. They’ve done this for strings in different orientations too. My understanding is that this will negatively impact performance, all for trying to save a few bucks. Maybe @nwdiver can chime in.

 

[nwdiver]

My understanding is that this will negatively impact performance, all for trying to save a few bucks.

A little. Really the biggest loss IMHO is no longer having that diagnostic tool. It's a lot easier to spot a blown bypass diode on identical strings when String A is 370v and String B is 360v than when String A/B is 365v. But to run them separately is so cheap it's absurd to run them together just to save ~$50....

 

[pdx_m3s]

A little. Really the biggest loss IMHO is no longer having that diagnostic tool. It's a lot easier to spot a blown bypass diode on identical strings when String A is 370v and String B is 360v than when String A/B is 365v. But to run them separately is so cheap it's absurd to run them together just to save ~$50....

Interesting. Thanks. So to confirm, combining two strings at the inverter with a jumper, say one East, one West, has little impact on generation?

What about E/W on the same string? I’ve seen Tesla do that too.

 

[nwdiver]

Interesting. Thanks. So to confirm, combining two strings at the inverter with a jumper, say one East, one West, has little impact on generation?

What about E/W on the same string? I’ve seen Tesla do that too.

It's not too bad unless you get shade. If a few panels in one string get shaded it effectively makes that string shorter and you lose the unshaded panels.

 

[mspohr]

It's not too bad unless you get shade. If a few panels in one string get shaded it effectively makes that string shorter and you lose the unshaded panels.

In this situation wouldn't microinverters be an advantage?

 

[nwdiver]

In this situation wouldn't microinverters be an advantage?

Nope. Just need to run each string to its own MPPT channel.

 

[nwdiver]

I thought this was another good example of why optimizing at the panel level is expensive nonsense. We just completed an install of some newer 365w panels. Unlike previous panels I've worked with these actually have a MPP current of ~10.7A. On hot days that creeps a little higher. Unfortunately the latest generation of SMA inverters haven't caught up to the newer panels so their MPP channels are limited to 10A. I was kind of expecting this but each string hit that limit at ~11:30 and stayed limited until 2:30. That's ~3 hours of lost production. Since a partial solution was fairly simple I thought I'd try it.

The SMA inverter has 3 MPP channels and this install has 2 strings. The inverter has the option of 'bridging' 2 channels so I split one string between A & B and put the 2nd string on C. The results were interesting. It went from 10A @ 377v on channel A to 10.6A @357v on channels A&B. For a gain ~14w or 0.4%. The lesson here is the gain from precisely hitting that MPP is INCREDIBLY small. Voltage and current were off the MPP by ~6% due to the current being limited but power loss from that was only ~0.4%. AND optimization isn't free. They're ~99% efficient. So you're likely spending 1% to gain ~0.4%.... a net loss.