Inverter max VOC

[GoldDigger]

With the panels facing a cloudless night sky, the panel temp at first light could actually be colder than the coldest air temperature.

Tapatalk!

 

[electrofelon]

VOC is a real property of solar panels. The wiring between them must be rated for the VOC to prevent a fault, period. The only way to get a VMP that goes all the way up to the wiring rating would be to have cutoff controls that are within the listed solar module assembly. The code will never allow anything else. Give up on that. (Yes, there are safety factors that mean this is overkill, but safety factors exist for a reason and the CMP will never change their minds on that.)



No, it could actually be VOC, in case of system failure. I think that's what you're not getting.



The code needs to assure that installations fail safely, not just that they operate safely. The rarity of failure doesn't change this one iota. This is true across the board, not just in PV systems.






Generally, I feel that either
a) you are still not really getting how the technology and engineering really works, OR
b) you are trying to argue, without really saying so, that "the code should really allow us to ignore VOC because that is not the operating voltage." Regarding this, there isn't a chance in heaven that the Code Making Panel will ever agree.

Ok, I feel like you guys think this is the most ridiculous idea you have ever heard. I am proposing a small increased allowance in WORST CASE, RECORD LOW VOC. Say maybe we can go up to 650-675 volts. This would be in conjunction with inverter design that could handle the absolute worst case VOC and/or would hold the voltage at or below 600V. Yes an equipment failure, or opening the circuit would result in wire and equipment seeing 650 volts if all the stars lined up. Yes I am kind of arguing your proposition "b". I think it would be extraordinary rare that equipment would see this and even if it did, it is again very unlikey that any damage would occur. Design and codes are typically a compromise between safety, cost, and functionality. If you think an allowance to exceed 600V rated cable by 50-75 volts in certain very rare situations is an unacceptable increase in risk, than you are entitled to think that. Maybe you also have a problem with the "next size up" rule, or that you can size the SE conductors to the load rather than the sum of the rating of the 2-6 disconnecting means, or the "120%" rule, or the tap rules - or maybe not, maybe you think those are acceptable risks for the increase in flexibility and lower costs they allow. Maybe some of you would agree this would be fine, but are saying that the CMP's would never allow it - and I would agree with that. If they will never allow it than maybe you feel it is a waste of time to discuss it. We can complain about arc fault requirements all we want, as I do and many others on this forum do, but they are not going away.

Voltage has virtually nothing to do with irradiance. When the first rays of the sun hit the modules in the morning, the voltage goes to Voc for whatever temperature the modules are at the time, which is pretty much the low temp for that night. It's Voc because oc means open circuit, and when the inverter is off there is no current flowing through a series resistance to pull the voltage down. There is no "VOC" as you use it; it's either open circuit or it's not.

Despite that is what is taught in PV 101, that is not really true. Irradiance does effect VOC, sure not nearly as much as current but certainly enough to make a difference in the sunrise situation. Look at the charts, and note the voltage difference, consider there will be about 13 modules in series.

This system design you've described makes little sense and would work badly. One does not run an array directly to multiple loads where one of them is a MPPT CC and the others are water heating elements. The latter will mess with the former and cause it to work suboptimally. More than one MPPT device on an array will not work either.

I think if designed properly this will work fine. A series resistor will of course drop the voltage seen by the CC, but as long as this voltage is still within the mppt range of the CC, I see no reason the CC will not run the array at MPP. Why wouldn't it? Give it some thought and If you still think I am wrong and would like to explain further I am certainly open to learning what I am missing.

I believe the 2014 code would allow you to have up to a 1000VDC ground mounted array, as long you don't have more than 600VDC entering a single family home. (Disclaimer: haven't read the new code in all that much detail yet). In other words, perhaps you could have a relay cut power to the house if the temperature dropped so low that the VMP was over 600V. Now, this would mean losing power on the coldest days of the year, which seems ... dumb. But in the particular unusual case you have described above, it might be that code gives you the 50V of wiggle room you are asking for.

Yes I think 1000V PV is on the horizon and I will welcome it - obviously
http://solarprofessional.com/articles/design-installation/1000-vdc-utilization-voltages

 

[electrofelon]

Another comment on mppt and connecting a heating element to the line side of the charge controller: The only case I can think of where the the array will not be run at MPP is if the CC's MPPT algorithm has certain "expectations". For example say it open circuits the array to find the VOC and then figures that about 80% of that is my MPP. Then say it sweeps around that number but only in a rather small +/- range. Well in the case of a series resistor, it would of course get the correct VOC but under load the voltage would drop significantly thought that resistor thus dropping the voltage out of the range that it is playing with and if it is convinced that the MPP is within that range, it would not be working the array at its MPP. I would think mppt algorithms are more advanced than that but I do not know for sure.

 

[GoldDigger]

Another comment on mppt and connecting a heating element to the line side of the charge controller: The only case I can think of where the the array will not be run at MPP is if the CC's MPPT algorithm has certain "expectations". For example say it open circuits the array to find the VOC and then figures that about 80% of that is my MPP. Then say it sweeps around that number but only in a rather small +/- range. Well in the case of a series resistor, it would of course get the correct VOC but under load the voltage would drop significantly thought that resistor thus dropping the voltage out of the range that it is playing with and if it is convinced that the MPP is within that range, it would not be working the array at its MPP. I would think mppt algorithms are more advanced than that but I do not know for sure.

An initial panel voltage setting of 80% of V_OC will be a good starting point regardless of the nature of the load on the output of the CC.
What will make a difference is that the CC will try to hold the output voltage within a close range of the nominal "battery" voltage it is set for and may drop out or log a fault if there is not enough power from the panel to get the voltage that high.
The MPP will still be at about 80% of V_OC.
A "linear current booster", on the other hand, is an MPPT circuit which does not have any particular expectation of the output side voltage and is more likely to work with a heating element.

 

[electrofelon]

An initial panel voltage setting of 80% of V_OC will be a good starting point regardless of the nature of the load on the output of the CC.
What will make a difference is that the CC will try to hold the output voltage within a close range of the nominal "battery" voltage it is set for and may drop out or log a fault if there is not enough power from the panel to get the voltage that high.
The MPP will still be at about 80% of V_OC.
A "linear current booster", on the other hand, is an MPPT circuit which does not have any particular expectation of the output side voltage and is more likely to work with a heating element.

GD, I think you may have misread or misunderstood what I was proposing. I am planning to put a resistor in series between the array and the Input terminals of the CC. The load side of the CC would be connected to the batteries as normal. So in this case yes the array MPP doesnt ever change but the voltage that the CC "sees" after the voltage is dropped through that resistor will. So if the CC samples the VOC and looks for MPP close to 80% of that, it could miss the MPP of the array+resistor combo. I envision the algorithm to be smarter than that and do a broader sweep. Maybe I could talk to someone at Schneider to verify. The nice thing about a series resistor is it is self regulating on array current output, that is if clouds obscure the sun, the resistor will consume far less energy allowing most if it to go to the batteries which would certainly be the priority over heating water on a cloudy day.

 

[GoldDigger]

EF: When you talked about adding a heating resistor to the line side of the CC I did not think of it as a series resistor. That could confuse the initial guess of an MPPT algorithm alright, but I am not sure I see the purpose for putting it there. A series resistor will not, IMHO, do anything to mitigate high VOC problems

Tapatalk!

 

[electrofelon]

EF: When you talked about adding a heating resistor to the line side of the CC I did not think of it as a series resistor. That could confuse the initial guess of an MPPT algorithm alright, but I am not sure I see the purpose for putting it there. A series resistor will not, IMHO, do anything to mitigate high VOC problems

Tapatalk!

Right the conversation morphed beyond the high voc issue, sorry for the confusion. I was just talking about utilizing watts that are beyond the cc's capacity and discussing when or when not the array will be tracked at mpp

 

[GoldDigger]

When a significant amount of power is extracted from the array upstream of the MPPT device, and especially where the power going to that device is dependent on the MPPT devices input current, then the best you can say is that the MPPT controller is getting the maximum power it can for itself.
For any specific situation you would have to prove separately whether that would end extracting the maximum power possible from the array, adding up the two power numbers.
In that, you are quite correct.
If the parasitic load was in the form of a parallel resistor, I think we can say for sure that the total number would be maximized at the same point as the MPPT input.
For series I am less confident.

Tapatalk!

 

[iceworm]

I'm not a PV whiz. My closest association is 4 - 50W, 12V panels wired for 24V, I put up at my cabin 15 years ago to keep the batteries from freezing during the winter.

But still I'm interested and could use some education. I promise - I read all of the posts.

Concerning a typical electrofelon size system:
What is the magnitude of power? At full output, is this 600V at 10A, or 100A?

Say one had a temp corrected open circuit voltage of 675V - 700V, what would be a typical required current draw to load the string down to 600V? Is this like 1A or 10A?

just curious

ice

 

[jaggedben]

What is the magnitude of power? At full output, is this 600V at 10A, or 100A?

About 8-9 amps, regardless of how many panels are in series. "At full output" means none of them are shaded, etc..

Say one had a temp corrected open circuit voltage of 675V - 700V, what would be a typical required current draw to load the string down to 600V? Is this like 1A or 10A?

It could be 'like' 1A in low light or 10A in good light. An inverter's MPPT algorithim will probably find a Vmp below 600V if the VOC is 700, but maybe not.

I don't know what you mean by a 'temp corrected VOC', unless you mean the VOC at the lowest expected temperature.

All of the above numbers assume we are talking about crystalline silicon and not any other technology.

 

[iceworm]

... I don't know what you mean by a 'temp corrected VOC', unless you mean the VOC at the lowest expected temperature. ...QUOTE]

Yes that's what I meant. Poor choice of words on my part.

ice

 

[jaggedben]

Ok, I feel like you guys think this is the most ridiculous idea you have ever heard. I am proposing a small increased allowance in WORST CASE, RECORD LOW VOC. Say maybe we can go up to 650-675 volts.

Okay well... at least we know what we are talking about now.

I would point out that the code already contains some part of this allowance, because it doesn't require meet the 'worst case, record low', but rather the 'lowest expected ambient' temperature. It then refers to the ASHRAE handbook in the informational note.

This would be in conjunction with inverter design that could handle the absolute worst case VOC and/or would hold the voltage at or below 600V. Yes an equipment failure, or opening the circuit would result in wire and equipment seeing 650 volts if all the stars lined up. Yes I am kind of arguing your proposition "b". I think it would be extraordinary rare that equipment would see this and even if it did, it is again very unlikey that any damage would occur. Design and codes are typically a compromise between safety, cost, and functionality. If you think an allowance to exceed 600V rated cable by 50-75 volts in certain very rare situations is an unacceptable increase in risk, than you are entitled to think that. Maybe you also have a problem with the "next size up" rule, or that you can size the SE conductors to the load rather than the sum of the rating of the 2-6 disconnecting means, or the "120%" rule, or the tap rules -

Actually I do think that the SE conductors example makes not much sense. As for the others, I think they are more straightforward, or not analogous. For one thing, they all involve current. Can you point to an existing example in the code which allows fudging the safety factor on voltage? (not saying there isn't one)

or maybe not, maybe you think those are acceptable risks for the increase in flexibility and lower costs they allow. Maybe some of you would agree this would be fine, but are saying that the CMP's would never allow it - and I would agree with that. If they will never allow it than maybe you feel it is a waste of time to discuss it. We can complain about arc fault requirements all we want, as I do and many others on this forum do, but they are not going away.

I'm not merely saying that the CMP wouldn't allow it; I'm also yet to be convinced that your idea is safe. You keep saying that it would be rare or very unusual for the conductors or equipment to see the VOC, but I think that's a debatable point. We have systems that are left shut down waiting for PTO, systems that are offline for maintenance, systems whose inverters do fault testing at VOC when starting up in the morning...

 

[ggunn]

Despite that is what is taught in PV 101, that is not really true. Irradiance does effect VOC, sure not nearly as much as current but certainly enough to make a difference in the sunrise situation. Look at the charts, and note the voltage difference, consider there will be about 13 modules in series.

I am aware of that. Did you miss the "virtually" in my statement?

 

[electrofelon]

I am aware of that. Did you miss the "virtually" in my statement?

Well I wouldnt even agree with the term "virtually". I measured 2.5 volts difference from a 250 watt module facing the sun to pointing the opposite way, and that was on a bright day with snow on the ground. Multiply that by 13 modules, and consider how much lower the irradiance is likely to be at sunrise for many systems and its a difference worth taking into account.


Just to clarify since there are two discussions going on here. This system I am putting in at my place this summer is NOT incorporating any unorthodox high VOC stuff. At -25 and 1000W per irradiance, the VOC is still well below 600. I may in the future, if I get bored and need some more capacity, add some more modules and incorporate some controls to "clip" the voltage down and play with that theory a bit. I am off grid and there are no codes so I can do whatever I want. The array will be 3250 watts, 13 250 watt panels all in series for a MPP voltage and current of 400 and 8. The voltage drop in the array run is about 12% (PV is cheaper than wire!). I AM however going to incorporate some heating elements on the array side of the charge controller soon to use up some of that extra power that the charge controller cant utilize - its only good for 2200 watts. Note that doing this is pretty much a secondary consideration for the relative size of the array and charge controller: This array will give me significant output on cloudy days which is important for off griders, and full output on the hottest days so I can run power tools without using my batteries.

 

[electrofelon]

Okay well... at least we know what we are talking about now.

I would point out that the code already contains some part of this allowance, because it doesn't require meet the 'worst case, record low', but rather the 'lowest expected ambient' temperature. It then refers to the ASHRAE handbook in the informational note.

True, that gives some fudge room.

Actually I do think that the SE conductors example makes not much sense. As for the others, I think they are more straightforward, or not analogous. For one thing, they all involve current. Can you point to an existing example in the code which allows fudging the safety factor on voltage? (not saying there isn't one)

I cant other than the use of current transformers. Maybe another would be a 600 volt service - voltage could be higher at times of course.

I'm not merely saying that the CMP wouldn't allow it; I'm also yet to be convinced that your idea is safe. You keep saying that it would be rare or very unusual for the conductors or equipment to see the VOC, but I think that's a debatable point.

...You got that right

 

[winnie]

It seems to me that the idea of having a load not controlled by the MPP itself, connected to the solar array, is a bad idea. The job of the MPP is to control the load placed on the solar array to maximize output; anything outside of the MPP is going to 'dilute' this control capability.

Now the idea of an MPP that operates (say via a contactor) some direct loads on the solar array seems totally reasonable to me. In this way a smaller MPP could be efficiently used, operating some other 'loads of opportunity' when the power is available, without having to run the power through the MPP itself. In this case, however, the MPP algorithm and system would 'know' what is going on.

Regarding systems where the MOCV of the array exceeds the limits of attached components, but the normal operating point of the system is below the limits of the components: I think that this could be approached if you had additional safety components, at least as reliable as circuit breakers, which limit the voltage on any part of the system even with disconnects open. Because solar arrays have a well defined and limited short circuit current, you would need to have shorting crowbars rather than devices that open the circuit, but I don't see these as deal breakers. Possibly not worth the expense, but not unreasonable. If the duration of the DC over voltage is small enough, then it will put less stress on the insulation than the peak of the maximum allowed AC voltage.

-Jon

 

[iceworm]

... Regarding systems where the MOCV of the array exceeds the limits of attached components, but the normal operating point of the system is below the limits of the components: I think that this could be approached if you had additional safety components, at least as reliable as circuit breakers, which limit the voltage on any part of the system even with disconnects open. Because solar arrays have a well defined and limited short circuit current, you would need to have shorting crowbars rather than devices that open the circuit, but I don't see these as deal breakers. Possibly not worth the expense, but not unreasonable. If the duration of the DC over voltage is small enough, then it will put less stress on the insulation than the peak of the maximum allowed AC voltage.

Disclaimer - I am not a PV guy.
My reason for my previous post asking about the magnitude of the voltage over 600V and required current to pull the voltage down to 600V was likely following your reasoning. If is only took an amp or so:

Connect the output of the array to a 600V zener driving a 1000V, 1000W darlington - a 600V series voltage regulator. The heatsink would be the expensive part.

Add a fail safe crowbar - a current sensor such that if the current goes over say 1.5A, the darlington goes to full conduction - shorting out the PV array. Like any protective relay, have to be reset. I thinking that shouldn't be an issue if designed properly. The crowbar won't operate unless there is a failure.​

Just thinking, not my area, mfg not likely to incorporate for the limited market of those that want that last 50 V between a nominal 550V system and a 600V. For a 10A system, that is the difference between 5500W and 6000W?

So if one is spending $2W installed cost there is $1000 to spend to get the "free" 500W. So, the question is, can the mfg incorporate that for $1000?

Just thinking, nothing important.

ice

 

[ggunn]

Well I wouldnt even agree with the term "virtually". I measured 2.5 volts difference from a 250 watt module facing the sun to pointing the opposite way, and that was on a bright day with snow on the ground. Multiply that by 13 modules, and consider how much lower the irradiance is likely to be at sunrise for many systems and its a difference worth taking into account.

I'm not going to split hairs over the meaning of the word "virtual"; you can assume whatever risk you feel comfortable with when designing a system. Keep in mind, however, that some inverter companies maintain an internal black box register that keeps track of the highest voltage the inverter has ever been exposed to, and if you exceed their published maximum allowed voltage at any time your warranty is rendered void. In dealing with those inverters I give that maximum voltage a pretty wide berth when I am designing a system. Do whatever floats your boat.

There are, on the other hand, inverters which keep the DC inputs isolated from the rest of the circuitry when the inverter is off, and they do a voltage test before starting up. If they see too high a voltage they won't connect.

 

[GoldDigger]

If the object of the crowbar is to limit the voltage from reaching Voc, then you would have to reset it every time the sun comes up and also coordinate it with the operation of the inverter to somehow have the inverter apply load the moment the crowbar is released. Not going to happen.
You would have better luck crowbarring one or two panels in the string until the operating voltage of the remainder of the string went low enough that adding those panels back into the string would be safe.
You still have the problem of what to power the crowbar circuit from. It would be similar to a foldback current limiter only without the unregulated voltage supply to use to power the circuit....

Tapatalk!

 

[ggunn]

So, what's wrong with the design practice of sizing strings such that excessive/damaging/uncompliant Voc won't ever be reached?