Micro Inverters on Ground Mounted Array

[pv_n00b]

Tigo makes an MLPE that provides monitoring only, the TS4-A-M.

 

[winnie]

And for those of us who are paranoid, are there any module level power electronics systems that don't require reporting to the 'cloud'??

-Jon

 

[pv_n00b]

And for those of us who are paranoid, are there any module level power electronics systems that don't require reporting to the 'cloud'??

-Jon

Not that I know of. That's how they keep that reoccurring revenue coming in. Assuming the system is monitored off-site by a 3rd party uploading the data is going to be required. I remember in the old days when the monitoring system had a little web interface built in that someone could go to on their LAN.

 

[GoldDigger]

If by that you mean some strings would have more of the western panels, and some would have more of the eastern panels, I would think that would only make sense if each string is on a separate input on the inverter. Otherwise, if you need to parallel strings on one input, you'd want equivalent shading on each paralleled string, so for an E-W moving shadow line, you'd do the opposite.

Cheers, Wayne

As long as there is no differential shading within a string (e.g two panels in shade, four still in sun) there is no significant harm in having otherwise identical shaded and unshaded strings on one MPPT input. The current from the shaded string will be lower, but the Vmp values will be almost identical.

 

[wwhitney]

As long as there is no differential shading within a string (e.g two panels in shade, four still in sun) there is no significant harm in having otherwise identical shaded and unshaded strings on one MPPT input. The current from the shaded string will be lower, but the Vmp values will be almost identical.

I agree for one string unshaded, and one string with cells equally partially shaded.

So I guess the question is if one string is fully shaded by an obstruction say 50 ft away (the context of the earlier discussion), is there enough diffuse light from the atmosphere to still illuminate its cells 5%, 10%, or whatever the minimum is to get a string whose V_oc is close to the V_mp of the unshaded string. Because if one string is completely unilluminated, its V_oc must be 0, right?

Cheers, Wayne

 

[ggunn]

I agree for one string unshaded, and one string with cells equally partially shaded.

So I guess the question is if one string is fully shaded by an obstruction say 50 ft away (the context of the earlier discussion), is there enough diffuse light from the atmosphere to still illuminate its cells 5%, 10%, or whatever the minimum is to get a string whose V_oc is close to the V_mp of the unshaded string. Because if one string is completely unilluminated, its V_oc must be 0, right?

Cheers, Wayne

Voltage is changed very little by differences in insolation. Current is the major dependent variable with respect to sunlight.

 

[wwhitney]

Voltage is changed very little by differences in insolation. Current is the major dependent variable with respect to sunlight.

Agreed, but at the limiting case of no illumination, the voltage is 0. What level of illumination is required for V_oc to increase to near V_mp of the fully illuminated string, so that string can produce power efficiently? And how does that compare to the ambient diffuse illumination during the day, for panels in direct shade?

Cheers, Wayne

 

[jaggedben]

Agreed, but at the limiting case of no illumination, the voltage is 0. What level of illumination is required for V_oc to increase to near V_mp of the fully illuminated string, so that string can produce power efficiently? And how does that compare to the ambient diffuse illumination during the day, for panels in direct shade?

Cheers, Wayne

Just a guess from experience, but probably only a handfull of watts per meter squared. Maybe 1-2% of typical mid-day illumination.

 

[electrofelon]

Voltage is changed very little by differences in insolation. Current is the major dependent variable with respect to sunlight.

This has always been a pet peeves of mine. Big solar pedagogy pushes this idea and Certainly current changes "more" than voltage with differences in insolation, but voltage changes "quite a bit" with insolation too. I admit it's all in how you interpret certain adjectives.

 

[GoldDigger]

This has always been a pet peeves of mine. Big solar pedagogy pushes this idea and Certainly current changes "more" than voltage with differences in insolation, but voltage changes "quite a bit" with insolation too. I admit it's all in how you interpret certain adjectives.

I think that some people may use different references for the applying a value judgement to a measurement.
In this case there are two plausible ways to go.
1. Look at the change in current and say that, for example, a 5% change is significant. Same for a 5% change in voltage.
2. Look at the loss in energy production resulting from an deviation from Vmp and now a 5% change in voltage may correspond only to a 2% reduction in power since the MPP is by definition a point in the power curve whose tangent is horizontal.

 

[ggunn]

This has always been a pet peeves of mine. Big solar pedagogy pushes this idea and Certainly current changes "more" than voltage with differences in insolation, but voltage changes "quite a bit" with insolation too. I admit it's all in how you interpret certain adjectives.

Once I got the #### shocked out of me by some modules that were illuminated only by a few fluorescent tubes on the warehouse ceiling 20+ feet away. Current varies fairly linearly with insolation but voltage jumps up pretty quickly at low light levels and flattens out. "Big solar" has nothing to do with it. Look at any family of IV curves for a module; the X intercept (Voc) changes very little for differing insolation levels.

 

[electrofelon]

I think that some people may use different references for the applying a value judgement to a measurement.
In this case there are two plausible ways to go.
1. Look at the change in current and say that, for example, a 5% change is significant. Same for a 5% change in voltage.
2. Look at the loss in energy production resulting from an deviation from Vmp and now a 5% change in voltage may correspond only to a 2% reduction in power since the MPP is by definition a point in the power curve whose tangent is horizontal.

What bugs me I guess is that temperature changes voltage too, and this change is small, yet the code takes this onto account, but ignores irradiance (IIRC, newer code cycles allow the modeling of voc for larger systems and the irradiance can be taken into account).

 

[electrofelon]

Once I got the #### shocked out of me by some modules that were illuminated only by a few fluorescent tubes on the warehouse ceiling 20+ feet away. Current varies fairly linearly with insolation but voltage jumps up pretty quickly at low light levels and flattens out. "Big solar" has nothing to do with it. Look at any family of IV curves for a module; the X intercept (Voc) changes very little for differing insolation levels.

I didn't really articulate my complaint well initially. Let me rephrase. I have an issue with big solar pedagogy saying irrradiance is irrelevant but temperature is relevant. from a code standpoint I suppose that is true, but from a practical standpoint these effects are about equal.

 

[winnie]

I think that it might be more that there are some very good rules of thumb that get mistaken for physical laws/gospel truth, rather than useful approximations with limited scope.

Jon

 

[wwhitney]

Just a guess from experience, but probably only a handfull of watts per meter squared. Maybe 1-2% of typical mid-day illumination.

OK, so a random reference I located stated that diffuse insolation for most of the day is around 30 - 50 watts/meter². The graphs I've seen of I vs V for various insolation levels haven't gone below 200 watts/meter². But you're saying a graph of V_oc vs insolation, which has to go through the origin, would rise very rapidly well before 30 watts/meter²?

In other words, if you have two identical strings in parallel, one being fully illuminated, and you cover the other with an opaque layer on the panel, you're going to kill the power generation from both. But if you instead erect an opaque shade 50' away that directly shades one string without shading the other, the diffuse insolation on the shaded string is sufficient to allow the unshaded string to produce power.

Good to know.

Thanks, Wayne

 

[ggunn]

I didn't really articulate my complaint well initially. Let me rephrase. I have an issue with big solar pedagogy saying irrradiance is irrelevant but temperature is relevant. from a code standpoint I suppose that is true, but from a practical standpoint these effects are about equal.

Irradiance isn't considered irrelevant; Voc is measured at 1000W/m^2. Voltage doesn't change much with irradiance and irradiance doesn't get much more than 1000W/m^2. Voltage dependence on irradiance is already measured very near the top of its range.

Under very cold conditions voltage can change enough to bust the max voltage limit of an inverter. I am working with a system at the moment where the string Voc increases by 94V at minimum temperature.

This is all per the NEC. I guess you're just going to have to be peeved at "big solar pedagogy", whatever that is.

 

[pv_n00b]

A cell reaches about 80% Voc at 20% STC insolation. It levels off pretty flat after that. If you add 80% more insolation you only get 20% more Voc. Voltage is a big hockey stick on the graph, while current is linear with insolation.

 

[jaggedben]

OK, so a random reference I located stated that diffuse insolation for most of the day is around 30 - 50 watts/meter². The graphs I've seen of I vs V for various insolation levels haven't gone below 200 watts/meter². But you're saying a graph of V_oc vs insolation, which has to go through the origin, would rise very rapidly well before 30 watts/meter²?

In other words, if you have two identical strings in parallel, one being fully illuminated, and you cover the other with an opaque layer on the panel, you're going to kill the power generation from both. But if you instead erect an opaque shade 50' away that directly shades one string without shading the other, the diffuse insolation on the shaded string is sufficient to allow the unshaded string to produce power.

Good to know.

Thanks, Wayne

This week my microinverters (same town you reside in) wake up around sunrise, or maybe 1/2 hour later on a cloudy day like this morning. Their first DC voltagr report is about 11volts. (This is apparently not Voc, since the the micros have enough power at this point to wake up and communicate but not output AC power.) Next report 15min later DC voltage is around 23 and they are maybe outputting 1-2 watts per panel already. My panels have a nameplate Voc of 37 and Vmp of 30, although real world seems to be lower, probably because they have aged. Anyway...Yes, Voc rapidly rises with very low insolation. Power production starts with ambient light around dawn.

Speaking to your second paragraph, yes, shaded panels still produce power. Recently had a situation where a client had tarped an array to protect panels during unrelated work. At mid-day the panels were will producing half a dozen watts each directly under the tarp. Now, I don't know how translucent that tarp might have been, but still.

 

[GoldDigger]

In other words, if you have two identical strings in parallel, one being fully illuminated, and you cover the other with an opaque layer on the panel, you're going to kill the power generation from both. But if you instead erect an opaque shade 50' away that directly shades one string without shading the other, the diffuse insolation on the shaded string is sufficient to allow the unshaded string to produce power.

Good to know.

Thanks, Wayne

As I understand the voltage/current/insolation curve family for a solar cell, the bold statement is true as long as there is no blocking diode in the shaded string.
So, yes, you want to avoid zero insolation, but during daylight hours the indirect insolation, even when direct sunlight is blocked, is sufficient to bring the voltage of the shaded string up to a level which will allow the unshaded string to deliver significant power.

But in the absence of bypass diodes, even partially shading one cell in the series string will limit the current (Imp, Isc) for that string and may cause harmful power dissipation in the shaded cell(s)