Help with question about wire derating

[fafrd]

Key assumption is that bypass diodes work, don't fail if continuously carrying Imp. Some brands/models can't handle that.

What happens when… bypass diodes fail?

Series troubleshooting: Bypass diodes fail regularly, either because they do not have a high enough power rating or because they are overloaded due to nearby lightning strikes. With the following hypothetical, but realistic case pv magazine starts a series which aims to make the estimation of...

www.pv-magazine.com

Issues with bypass diodes

As the latest in its investigation into quality issues at pv plants, Heliolytics looks at failure modes in bypass diodes - and what can be done about them.

pv-magazine-usa.com

Fig. 6. The root cause of top #1 failure mode (burnt bypass diode)

Download scientific diagram | The root cause of top #1 failure mode (burnt bypass diode) from publication: The Reliability Investigation of PV Junction Box based on 1GW Worldwide Field Database | The reliability of junction box plays the critical characteristic in PV development. We perform the...

www.researchgate.net

Just when I was thinking I finally see light at the end of the tunnel...

These are Trina panels, and 72-cell so aimed at the insustrial market, so lower-quality diodes are not a huge concern.

But because of my shading, bypass diodes will be exercised until about noon daily (if I allow it), so if even high-quality diodes can be a point of failure from regular use, maybe I should start thinking about a parallel array again.

With a parallel array, shaded half-panels will just not contribute to the current being output and bypass diodes will pretty much never be activated (SCC will be locked at Vmp_panel from first light to last light).

I suppose I could always rely on my warranty and switch to a parallel string if I get repeated failures...

 

[Hedges]

RTFW

Apparently some panels have instructions that say partial shading during full sun is improper use. So would violate warranty terms.
Other brands are fine.
The diode would (almost always) be spec'd to carry that much current, question is whether the module design provide adequate heatsinking.

I don't have info on which ones have this problem.

 

[fafrd]

RTFW

Apparently some panels have instructions that say partial shading during full sun is improper use. So would violate warranty terms.
Other brands are fine.
The diode would (almost always) be spec'd to carry that much current, question is whether the module design provide adequate heatsinking.

I don't have info on which ones have this problem.

I guess I better do more research on the panels I’m considering before finalizing my plan.

Also just learned that the Vmppt range of the Epever SCC I was considering doesn’t extend all the way to the Max voltage of 200V (180V @ 25C).

Since Vmppt max is < 3 x Voc, I think that means the SCC has no way to cut off a 3-panel string when the battery is fully-charged and if that’s correct, I may be limited to 2S strings.

 

[Hedges]

SCC should have no problem deciding not to switch transistors when battery is charged.
There tends to be a max voltage input spec which is higher than top end of MPPT range.
I suppose that could mean when the current it wants to draw results in voltage exceeding its MPPT range it turns off. That might mean fails to keep delivering an absorption charge, not sure.

My Sunny Boy 5000US specs include:
Maximum MPP voltage 250 V … 480 V
Range of input operating voltage 250 V … 600 V
PV start voltage 300 V

In grid-tie operation it would always deliver maximum, it wouldn't try to reduce power output, so would run up to 480V but not higher if Vmp was higher.
If used off-grid and requested power output is reduced, then something funny might happen.

My PV strings are whatever number of panels I would put in series without exceeding 600V in cold weather (+15% voltage adjustment)
Nominal conditions, I see 480 Voc and 380 Vmp

 

[fafrd]

SCC should have no problem deciding not to switch transistors when battery is charged.
There tends to be a max voltage input spec which is higher than top end of MPPT range.
I suppose that could mean when the current it wants to draw results in voltage exceeding its MPPT range it turns off. That might mean fails to keep delivering an absorption charge, not sure.

My Sunny Boy 5000US specs include:
Maximum MPP voltage 250 V … 480 V
Range of input operating voltage 250 V … 600 V
PV start voltage 300 V

In grid-tie operation it would always deliver maximum, it wouldn't try to reduce power output, so would run up to 480V but not higher if Vmp was higher.
If used off-grid and requested power output is reduced, then something funny might happen.

My PV strings are whatever number of panels I would put in series without exceeding 600V in cold weather (+15% voltage adjustment)
Nominal conditions, I see 480 Voc and 380 Vmp

Yeah, just opening a switch would be an easy way to have an upper Vmp range lower than rated maximum voltage.

Strange though, why make a 200V MPPT that only covers Vmp up to 144V? I’m not sure it’s suitable for 3S strings with 50Voc panels, compared to the WZRELB 170V MPPT which has a Vmp range which extends all the was to 170V.

 

[Hedges]

Possibly its a holdover from how SCC or inverters were rated to tolerate Voc cold.
Usually temperature rises and Voc drops as the sun comes up.
Power drops off faster at voltage higher than Vmp, as compared to voltage lower.

In the design of a buck converter, most of the power dissipation occurs during transition between transistor on and off. Larger voltage swing dissipates more power. But that should be managed by reducing power output when hot.
I could see where the inverter might not be able to deliver full output power above 144V input. We would like it to still operate continuously but at reduced power above that. Don't know if it does or doesn't. I believe it will pass through the range between 200V and 144V, gradually ramping up current, so it may well function.

Your 3s x 50 Voc would get pulled down to 144V by MPPT algorithm, still not producing much power. Probably works fine.

My 120W panels have spec 21 Voc. Wires 24s, that would be 504 Voc, and with +15% for cold, 580 Voc
Actual Voc on a typical day is 480V, and Vmp 380V

Scaling to your situation, 480/504 x 150 = 143V (which is within 144V max MPPT range)
I expect you to do fine with 3s. But yes, 144V max MPPT is a bit low compared to 200V maximum, would have MPPT range issue if Voc near 174V was used.

We can hope, but haven't confirmed, that the MPPT continues to operate but can't deliver specified power above max MPPT range.

 

[fafrd]

Possibly its a holdover from how SCC or inverters were rated to tolerate Voc cold.
Usually temperature rises and Voc drops as the sun comes up.
Power drops off faster at voltage higher than Vmp, as compared to voltage lower.

In the design of a buck converter, most of the power dissipation occurs during transition between transistor on and off. Larger voltage swing dissipates more power. But that should be managed by reducing power output when hot.
I could see where the inverter might not be able to deliver full output power above 144V input. We would like it to still operate continuously but at reduced power above that. Don't know if it does or doesn't. I believe it will pass through the range between 200V and 144V, gradually ramping up current, so it may well function.

Your 3s x 50 Voc would get pulled down to 144V by MPPT algorithm, still not producing much power. Probably works fine.

My 120W panels have spec 21 Voc. Wires 24s, that would be 504 Voc, and with +15% for cold, 580 Voc
Actual Voc on a typical day is 480V, and Vmp 380V

Scaling to your situation, 480/504 x 150 = 143V (which is within 144V max MPPT range)
I expect you to do fine with 3s. But yes, 144V max MPPT is a bit low compared to 200V maximum, would have MPPT range issue if Voc near 174V was used.

We can hope, but haven't confirmed, that the MPPT continues to operate but can't deliver specified power above max MPPT range.

If that’s all it means (reduced maximum power above Vmp_max), it will be fine.

I’ll have to find the manual and query some owners here in DIYSF before finalizing my choice.

Support for 2S with the panel class I am interested in seems certain; support for 3S is closer to the bleeding edge (today).

And then there is the issue about whether I want to enable bypass diodes in the hunt for a few extra % or discourage them by going full parallel with an MPPT controller whose Vmp_min is > 2/3 Vmp_panel - 1.4V.

The WZRELB MPPT has a minimum Vmp of 34V, so used with 450W panels with Vmp of 41V, I should never get any single-column bypass diodes activated and only half-panel-level bypass diodes when a half-panel is shaded.

 

[Hedges]

If Vmp decreases 15% at 65C vs. 25C, then 41V goes down to 35V, just above MPPT minimum Vmp.
A question would be how much Vmp decreases at less than full sun. We know Voc is hardly affected but Isc is proportional to illumination so probably not much.

I don't think bypass diodes for one side of your (internal configuration) 2s2p (or is that 3s2p, 2p2s, or 2p3s?) panels is a thing. If a half-cell is shaded, that parallel string of half-cells can't deliver any current at Vmp of the array. It also doesn't get backfed, just sits there around zero current. The other string of half cells delivers its normal current.

This example shows 2p3s

Physical models used > PV Module - Standard one-diode-model > Twin half-cut cells modules

Twin half-cut cells modules

www.pvsyst.com

 

[fafrd]

If Vmp decreases 15% at 65C vs. 25C, then 41V goes down to 35V, just above MPPT minimum Vmp.
A question would be how much Vmp decreases at less than full sun. We know Voc is hardly affected but Isc is proportional to illumination so probably not much.

In general, I think Vmp stays pretty constant as radiance increases, but something I saw indicated that the ‘sharpness of the elbow’ when a panel goes from being current-limited to being voltage-limited ‘softeners’ / ‘rounds’ as the panel ages. So Isc remains roughly constant and Voc remains roughly constant, but Vmp starts to reduce as the transition becomes less and less sharp (there is less current at the original Vmp and the voltage to achieve the same Imp is reduced versus the original Vmp).

But in general, you are right. I’m most concerned about cutting off charging in the summer when daily output levels are high and ambient temps are above 68C (20C). Voc of the panels I’m looking at will be 50.2V at ambient temps of 20C and will be below 50.0V as long as panels temp is 3F above that (71F) which seems certain.

I’m not understanding why you are referring to 65C - that’s 149F and I don’t think my panels get that warm (ever, let alone late in the day).

I don't think bypass diodes for one side of your (internal configuration) 2s2p (or is that 3s2p, 2p2s, or 2p3s?) panels is a thing. If a half-cell is shaded, that parallel string of half-cells can't deliver any current at Vmp of the array. It also doesn't get backfed, just sits there around zero current. The other string of half cells delivers its normal current.


This example shows 2p3s

Physical models used > PV Module - Standard one-diode-model > Twin half-cut cells modules

Twin half-cut cells modules

www.pvsyst.com

Didn’t realize that - if it’s only a single bypass diode for both upper and lower columns (and no panel-wide bypass diode) I’ll never need to worry about any bypass diodes being activated once a full half-panel is out of the shade...

Need to redo my analysis.

 

[Maast]

Lot of very interesting discussion and links in here so far - but I havent seen (or I missed it) any mention of using optimizers like Tigo's. It would remove any shading and voltage variability issues. And it's rapid shutdown capable and easy to implement. Each Tigo TS4-O optimizer runs about $50 US and will work with any inverter or MPPT controller.
Also, it gives you panel-level monitoring so if you ever lose a diode or a panel goes dead you'll know about it immediately.

 

[Hedges]

Lot of very interesting discussion and links in here so far - but I havent seen (or I missed it) any mention of using optimizers like Tigo's. It would remove any shading and voltage variability issues. And it's rapid shutdown capable and easy to implement. Each Tigo TS4-O optimizer runs about $50 US and will work with any inverter or MPPT controller.
Also, it gives you panel-level monitoring so if you ever lose a diode or a panel goes dead you'll know about it immediately.

Yes, we could put a $50 optimizer on our 250W $35 sale-priced SanTan PV panel!




Once we need RSD, then optimizers come along with that.
If I ever move my 120W 12V panels to a new permitted rooftop installation, I'll look into 4-panel RSD. I'll connect three of my MC3 panels in series, then adapt to MC4 and have 63 Voc 360W connected to each of four inputs. Two RSD units covers one 24 panel string.

One of these days I'll get a clamp DC ammeter. With that I could check my many PV strings, spot any out of family.

 

[fafrd]

Lot of very interesting discussion and links in here so far - but I havent seen (or I missed it) any mention of using optimizers like Tigo's. It would remove any shading and voltage variability issues. And it's rapid shutdown capable and easy to implement. Each Tigo TS4-O optimizer runs about $50 US and will work with any inverter or MPPT controller.
Also, it gives you panel-level monitoring so if you ever lose a diode or a panel goes dead you'll know about it immediately.

Welcome to the thread - I didn’t realize anyone was lurking and have more or less been treating this as a private exchange with Hedges.

I did mention optimizers in one of the early posts when I was fixated in all-parallel and first mentioned the possibility of considering a series string.

Optimizers are nice, but pricey. Rapid disconnect lowers the bar a bit but they still add a significant premium (and the RD market continues to evolve and cost-optimize).

More importantly, optimizers will only help with a 1P string and their usefulness is greatly reduced with half-cut panels.

I understand a lot more about how bypass diodes work now and am also understanding more about the potential risks of using them, so reconsidering optimizers in order to prevent bypass diodes from ever activating is something I should do...

 

[Hedges]

I’m not understanding why you are referring to 65C - that’s 149F and I don’t think my panels get that warm (ever, let alone late in the day).

Typical assumed temperature. Either sun heats panel 40 degrees above ambient 25C, or maybe that's a hot day with less rise in panel temperature.

How Hot Do Solar Panels Get? PV Temperature Explained

In the summertime, solar panels are exposed to high amounts of heat. Learn about the effect of temperature on solar panel efficiency.

news.energysage.com

I've seen safety specs in the range of 65C. I have found items in the sun (e.g. windshield) to hot to keep my hand on.

How Hot is too Hot? The Real Maximum Temperature of Your Product

www.boydcorp.com

It could be that 65C is above operating temperature in most applications. I think I heard of designing PV around that temperature so that's what I assume. (When I do electronics I uses 85C at the PCB unless otherwise specified.)

 

[fafrd]

Yes, we could put a $50 optimizer on our 250W $35 sale-priced SanTan PV panel!

Wow, and I thought I got a good deal at $0.39/W!

Were those decommissioned solar power array panels?

Knowing their are viable options at $0.14/W, if I wasn’t roof-space constrained, I’d jump on them (though you need to add another $25-50 per panel for BOM into the inverter or SCC...).

Once we need RSD, then optimizers come along with that.
If I ever move my 120W 12V panels to a new permitted rooftop installation, I'll look into 4-panel RSD. I'll connect three of my MC3 panels in series, then adapt to MC4 and have 63 Voc 360W connected to each of four inputs. Two RSD units covers one 24 panel string.

One of these days I'll get a clamp DC ammeter. With that I could check my many PV strings, spot any out of family.

Also, at least in today’s market, the Togo products don’t fit the capability of 72-cell high-power panels well (hopefully that will evolve with time).

 

[fafrd]

Typical assumed temperature. Either sun heats panel 40 degrees above ambient 25C, or maybe that's a hot day with less rise in panel temperature.

How Hot Do Solar Panels Get? PV Temperature Explained

In the summertime, solar panels are exposed to high amounts of heat. Learn about the effect of temperature on solar panel efficiency.

news.energysage.com

I've seen safety specs in the range of 65C. I have found items in the sun (e.g. windshield) to hot to keep my hand on.

How Hot is too Hot? The Real Maximum Temperature of Your Product

www.boydcorp.com

It could be that 65C is above operating temperature in most applications. I think I heard of designing PV around that temperature so that's what I assume. (When I do electronics I uses 85C at the PCB unless otherwise specified.)

Safe to assume as a hottest temperature, perhaps, but probably not as a coolest temperature (which was the exercise we were engaged in to determine maximum Vmp).

There is probably a simple model based on Watts of output to determine degrees C above ambient.

 

[Hedges]

Yes, I paid around $0.35/W for SunPower, wanted to favor quality and W/m^2 over price. Then some other guy bought same panel for half that price.

I think SanTan had some sales at that price, but for panels with visible degradation and no UL sticker. Presently listed for $45, or $50 without "Snail Trails"

Mounting hardware is a growing percentage of total these days. When I started, I paid $6/W or $5/W for panels.
Getting 50% more watts from the same area was attractive, not needing to extend mounting hardware.

 

[Hedges]

Safe to assume as a hottest temperature, perhaps, but probably not as a coolest temperature (which was the exercise we were engaged in to determine maximum Vmp).

There is probably a simple model based on Watts of output to determine degrees C above ambient.

Any watts that are output are watts not dissipated in the panel, so doesn't heat it.
Current flow will convert some available watts to heat. But leakage through the diode at Voc (that's 100% of Isc leakaing) is converted to heat.
When a 250W panel is open circuit, there is an extra 250W heating the panel.

Breeze cools the panel. That's part of PTC ratings. In still air and on a rooftop is where panels get hottest. Backsheet and some other failure mechanisms will be higher there.

 

[fafrd]

Any watts that are output are watts not dissipated in the panel, so doesn't heat it.

Good point.

Current flow will convert some available watts to heat. But leakage through the diode at Voc (that's 100% of Isc leakaing) is converted to heat.
When a 250W panel is open circuit, there is an extra 250W heating the panel.

So late in the day, a panel will be hottest if it is not outputting (which will not be my case since shade will have passed).

This means Voc will be highest for the panels passing current to the SCC (not good for assuring that SCC can choke off current by raising voltage all the way to Voc).

Breeze cools the panel. That's part of PTC ratings. In still air and on a rooftop is where panels get hottest. Backsheet and some other failure mechanisms will be higher there.

I’m still hung up on what MPPTs do when they cannot increase voltage all the way to Voc and want to stop current flow.

A switch is a good solution but I feel like this is something that should be known rather than us having to speculate.

Perhaps Will Prowse (sp?) would know...

 

[Hedges]

Voc is likely to be highest some time on a cool morning, unless a cold front arrives later.

I think an MPPT SCC can stop switching, draw zero current, let voltage reach Voc.
I'm pretty sure it just won't exercise its MPPT algorithm trying to find a maximum power point (Vmp) above the top of its range.
It probably just ramps voltage up or down as needed, but won't make excursions there looking for more power and probably would go into derating if while operating there it found too much power.

But yes, good project for Will. Easiest to implement with an SCC that has a large voltage spread between those parameters.

 

[fafrd]

Voc is likely to be highest some time on a cool morning, unless a cold front arrives later.

I think an MPPT SCC can stop switching, draw zero current, let voltage reach Voc.
I'm pretty sure it just won't exercise its MPPT algorithm trying to find a maximum power point (Vmp) above the top of its range.
It probably just ramps voltage up or down as needed, but won't make excursions there looking for more power and probably would go into derating if while operating there it found too much power.

So essentially an MPPT can continue working in ‘PWM mode’ beyond it’s Vmp limit...

But yes, good project for Will. Easiest to implement with an SCC that has a large voltage spread between those parameters.

Do you know Will well enough to ask him about this?