How do Tigo optimizers work?

[fafrd]

I’ve been reading everything I can find about Tigo optimizers and it’s amazing how fully they have been able to obfuscate their actual function under the general umbrella of ‘impedance matching’.

So if anyone has actually tested one of these Tigo optimizers and/or understands j he is they function, I would greatly appreciate an explanation.

I’m especially interested in what they do with fully-shaded and partially-shaded panels:

Fully-shaded (no current output):
To ‘tunnel’ current means that the Tigo Optimizer will essentially bypass the shaded panel and deliver full current at 0V of current. A switch being closed, for example, would provide this function. Is that how Tigo’s optimizer deals with fully-shaded panels? Bypassing them without relying on bypass diodes and essentially turning an N-series string into an (N - 1)-series string?

Partially-shaded (1 or 2 sub-strings shaded):
Assume 1 or 2 substrings are unshaded and the remaining string or strings are shaded to the point that output power will be reduced unless the bypass diodes on the shaded substrings are activated).

The entire panel can be bypassed as on the fully-shaded case, but that’s not really ‘optimizing’ anything and wasting potential power.

So we have to assume that the Tigo optimizer is getting the potential power out of the panel and that means activating the one or two bypass diodes (there is no other way to get the current from the unshaded strings out).

So we have to assume that the Tigo optimizer is applying ~1/3rd or ~2/3rd Vmp_panel to get out a full ~Imp of current.

I don’t understand how this is any different than an MPPT controller dropping voltage on a string until bypass diodes kick-in to bypass shaded sun-strings - what am I missing?

In addition to the claims regarding optimized performance in the presence of shading, Tigo optimizers claim the following two features:

Compensate for up to 25% panel mismatch: the Tigo optimizers also claim that they can compensate for up to 25% panel mismatch (for example, one 150W panel in a string of 200W panels. In this case, the ‘impedance matching’ function seems pretty clearly defined: take 75%W and lower Vmp and/or Imp and boost to match 100% Imp of the larger panels @ 75% Vmp of the larger panels. This implies that the maximum amount by which a Tigo optimizer can decrease native incoming panel voltage is by 25% (maximum current boost of 33%).

Compensate for up to 25% in string-length mismatch: as long as a shorter string in no more than 75% the length of a longer string, Tigo optimizers can effectively compensate for the panel mismatch so that both strings operate optimally at the same Vmp_string. This might be by decreasing current of each panel in the shorter string by 25% and increasing voltage of each panel by 33% or this might be by increasing current of each panel in the longer string by 20% at 75% of native panel voltage level.

Any knowledge about how these Tigo optimizers do their thing appreciated.

 

[fafrd]

I found this excellent article which does a good job explaining what Tigo optimizers actually do:

http://www.diva-portal.se/smash/get/diva2:1249003/FULLTEXT01.pdf

They are buck converters that boost current to the current of the strongest cell in a string by reducing voltage accordingly.

It’s still unclear to me whether they activate bypass diodes in the presence of shading or not but they are certainly able to bypass an entire shaded panel without using any bypass diodes.

And I’m also not understanding what the limits of current increase possible by the buck converters in Tigo optimizers is.

Specifically, I have half-cell panels which will occassionally only be putting out 1/2 of Imp because one half of the panel will be so shaded that current at Vmp will be close to 0A while the unshaded half-panel will be putting out it’s full 1/2 Imp @ Vmp, and I am interested to understand whether Tigo Optimizers can handle a 50% current mismatch.

If the other panels in the string are unshaded and putting out a full Imp, the Tigo optimizer would need to convert 1/2Imp @ Vmp to Imp @ Vmp/2 in order to avoid choking off the string.

I see several references to Tigo’s optimizers being able to handle 25% mismatch or 25% difference on string length, but nothing indicating whether they can handle the 50% current mismatch that can arise from half-cut cells in the presence of some shading.

 

[fafrd]

Also found this on the Tigo website: https://support.tigoenergy.com/hc/en-us/articles/214947868-String-Output

“Current can go up to anywhere between Isc and 12.5A”

This implies an upper limit of 12.5A on current output but suggests no lower limit on the input current which can be scaled up to as much 12.5A or Isc.

In the case of half-cut-cell panels, some shade on 1/2 of the panel can easily reduce current to 50% of Imp @ Vmp, so the above statement implies Tigo’s optimizer would be capable of doubling current to 100% Imp @ Vmp / 2.

 

[newbostonconst]

With the added cost of optimizers I always questioned why people don't just do Microinverters then there is shading. Seems the better way to go on cost to system pay back.

Thanks for the info....

 

[ArthurEld]

With the added cost of optimizers I always questioned why people don't just do Microinverters then there is shading. Seems the better way to go on cost to system pay back.

Thanks for the info....

I don't think cost of microinverters vs optimizers is how people make their decision.
Some people prefer to bring DC to their inverter.
And if you want to use batteries I think it is cheaper to use string inverters instead of micro inverters.

I'm not very knowledgeable of this subject but it interests me. My PV doesn't use optimizers or micro inverters.
It just goes straight to my hybrid inverter. I can and probably will AC couple micro inverters to my inverter in the future because I want to add more panels that will be shaded sometimes.

 

[fafrd]

With the added cost of optimizers I always questioned why people don't just do Microinverters then there is shading. Seems the better way to go on cost to system pay back.

Thanks for the info....

I’d do microinverters in a heartbeat (and use them for my 4kW grid-tied system) but this is for an off-grid backup system being used to charge a backup battery, and microinverters do not self-limit like SCCs.

 

[fafrd]

I don't think cost of microinverters vs optimizers is how people make their decision.
Some people prefer to bring DC to their inverter.
And if you want to use batteries I think it is cheaper to use string inverters instead of micro inverters.

I'm not very knowledgeable of this subject but it interests me. My PV doesn't use optimizers or micro inverters.
It just goes straight to my hybrid inverter. I can and probably will AC couple micro inverters to my inverter in the future because I want to add more panels that will be shaded sometimes.

Pretty challenging to combine microinverters with a string inverter. There are two issues with microinverters:

1/ They need to detect a ‘grid’ signal (which can be faked with a PSW running off of a battery).

2/ They pretty much convert all available solar power to 120V or 240V A/C and once converted, that power needs somewhere to go. Much tougher to throttle-back microinverters compared to SCCs (at least with the crop of products available today).

For grid-tied, they are fantastic and simple.

For off-grid battery-based, you are pushing a square peg through a round hole.

Enphase is coming out with some innovative new offerings to address these challanges, but as usual, they will only work with an all-Enphase system.

 

[Hedges]

I’d do microinverters in a heartbeat (and use them for my 4kW grid-tied system) but this is for an off-grid backup system being used to charge a backup battery, and microinverters do not self-limit like SCCs.

I think they do when they are meant to play with batteries. Enphase combines their microinverters with a battery inverter. One of the guys here (Ampster?) says he uses their microinverters with Skybox.

I favor string inverters. Microinverters have to replicate all or part of the functionality in every inverter (part if another box performs other parts of UL1741 function.) I think efficiency can be higher with string inverter for this reason. But the difference between say 97% and 99% doesn't matter much in terms of power production (2% difference). It should help reliability - 3:1 difference in heating, and getting inverter off the roof to a shady location. I think Enphase is still able to deliver good MTBF by overdesigning, and of course a couple failed is just a percentage reduction in output while waiting to replace.

You can get a string inverter with 3 MPPT, so it can have 3 different length strings. However, I like to have parallel strings of multiple orientations and the inverter I'm thinking of (Sunny Boy 7.7 kW -41 model) has limited current per MPPT. I would have to parallel all MPPT (which is allowed) to over-panel with 1.5x or 2x rating using the latest high wattage panels. So for me, a single MPPT works as well. But it won't do well with two strings where one is 50% shaded, something that separate strings would handle.

It looks like Solar Edge with its optimizers supports a single string of multiple angles and a battery backup is available. This may support your > 2:1 current output.

 

[ArthurEld]

Pretty challenging to combine microinverters with a string inverter. There are two issues with microinverters:

1/ They need to detect a ‘grid’ signal (which can be faked with a PSW running off of a battery).

2/ They pretty much convert all available solar power to 120V or 240V A/C and once converted, that power needs somewhere to go. Much tougher to throttle-back microinverters compared to SCCs (at least with the crop of products available today).

For grid-tied, they are fantastic and simple.

For off-grid battery-based, you are pushing a square peg through a round hole.

Enphase is coming out with some innovative new offerings to address these challanges, but as usual, they will only work with an all-Enphase system.

As far as I know it costs more to use Enphase with a battery. Hopefully that will change.

My Solark inverter is a string inverter that can AC couple micro inverters. I hope to try the AC coupling someday.
But for now I'm putting up 10K of PV in 4 strings with no optimizers. I can add optimizers later if needed. Or cut down some trees.

 

[fafrd]

I just got off the phone with Tigo and confirmed their optimizers are limited to 25% buck compensation.

What this means is, if you have a string of 2 half-cut-cell panels and one panel is half shaded (1/2Imp @ Vmp), the optimizer will be limited to boosting current to 133% of that 50% Imp, meaning to 67%Imp at a reduced voltage of 75% Vmp.

That means the unshaded panel will be operating at 67% Imp or ~67% of full power.

Without an optimizer, voltage will be driven above Vmp on the unshaded panel so that current drops to 1/2 Imp and the string will put out 1/2Imp at total voltage of >Vmp x 2, or total power of >50%.

(Alternatively, if all 3 bypass diodes on the shaded panel are activated to bypass it, current would get back to full Imp, but voltage would drop to Vmp -3Vdiode, meaning <50% total power, so running the shaded panel at half current is better).

Since the string with half-shaded panel will get above 50% power with no optimizer and 75% power is the theoretical maximum, I’m not sure the cost of an optimizer is justified.

Once you are putting in rapid disconnect MLPE anyway, the incremental cost to also get optimizer capability is modest, so it might make sense.

But it’s not the no-brainer it would be if Tigo’s optimizers were designed to handle buck compensation to a full 50%...

 

[fafrd]

As far as I know it costs more to use Enphase with a battery. Hopefully that will change.

My Solark inverter is a string inverter that can AC couple micro inverters. I hope to try the AC coupling someday.
But for now I'm putting up 10K of PV in 4 strings with no optimizers. I can add optimizers later if needed. Or cut down some trees.

How is your Solark throttling back your microinverters?

An AC-coupled SCC that could throttle-back Microinverters and charge a battery would be interesting to me if it;

1/ did not cost a fortune (I’m spending ~$500 on 4.5kW of DC-coupled SCCs

2/ worked with any 24/48V battery (especially DIY LiFePO4)

3/ had a high charge efficiency (I’ve seen 85% specified like that’s something to brag about - I’d need another 0.8kW of panels to make of for that loss)

 

[ArthurEld]

How is your Solark throttling back your microinverters?

I have no idea. I think it uses the AC first to cover house loads then to grid if there's excess. And uses the DC to charge the batteries.
That is what it will do first because it is most efficient. It will only convert AC to DC or DC to AC if necessary.
Some people use the Solark with only ac coupled micro inverters.

An AC-coupled SCC that could throttle-back Microinverters and charge a battery would be interesting to me if it;

1/ did not cost a fortune (I’m spending ~$500 on 4.5kW of DC-coupled SCCs

2/ worked with any 24/48V battery (especially DIY LiFePO4)

3/ had a high charge efficiency (I’ve seen 85% specified like that’s something to brag about - I’d need another 0.8kW of panels to make of for that loss)

 

[fafrd]

I think they do when they are meant to play with batteries. Enphase combines their microinverters with a battery inverter. One of the guys here (Ampster?) says he uses their microinverters with Skybox.

We’re going pretty far OT, but what the heck, it’s my thread, right?

The DC-coupled stuff has the benefit of being pretty-much open-standard. Any MPPT SCC can charge a battery alongside any other MPPT SCC and any battery-powered inverter or inverters can convert that stored battery energy to AC power. No walled gardens.

As a general rule, the AC-coupled world is more proprietary.

So I’m interested in any non-proprietary solution to AC-coupled PV charging, but in particular what I’d be interested in is an efficient AC-coupled battery charger with sensors which supports zero-export by:

1/ monitoring PV production and self-consumption so that any production in excess of consumption is used to charge battery as long as battery is not full (charger needs to be bigger than PV array). This means that during the day when PV is active, all self-consumption is supplied by the Microinverters or the grid as long as the battery is still charging.

2/ if battery is full and PV production exceeds self-consumption, either throttles-back Microinverter generation (back to needing open communication standards) or shuts down Microinverter generation by blocking grid signal. DC-coupled SCCs throttle very easily by increasing string voltage above Vmp and all the way to Voc if current needs to drop to zero. Throttling is tougher and not yet as mature with microinverters.

If you or anyone knows of anybody working on non-proprietary smart AC-coupled battery chargers like this, please let me know.

I favor string inverters. Microinverters have to replicate all or part of the functionality in every inverter (part if another box performs other parts of UL1741 function.) I think efficiency can be higher with string inverter for this reason. But the difference between say 97% and 99% doesn't matter much in terms of power production (2% difference). It should help reliability - 3:1 difference in heating, and getting inverter off the roof to a shady location. I think Enphase is still able to deliver good MTBF by overdesigning, and of course a couple failed is just a percentage reduction in output while waiting to replace.

I’ve got NEP dual-microinverters which have run for 6 years now without a hiccup. Exceedingly effective at dealing with any shading (MPPT per panel) and 96% conversion efficiency. The only issue with Microinverters is when you need to throttle them back (brute-force shut down works but probably not good for lifetime, smoother/gentler throttling is still evolving).

I paid under $1000 for 3kW of microinverters which include Rapid Disconnect functionality, so lower cost than all of

You can get a string inverter with 3 MPPT, so it can have 3 different length strings. However, I like to have parallel strings of multiple orientations and the inverter I'm thinking of (Sunny Boy 7.7 kW -41 model) has limited current per MPPT. I would have to parallel all MPPT (which is allowed) to over-panel with 1.5x or 2x rating using the latest high wattage panels. So for me, a single MPPT works as well. But it won't do well with two strings where one is 50% shaded, something that separate strings would handle.

Yeah, that’s the challenge of my new array. 10 panels with a shifting 5-panel shadow progressing it’s way across the array all morning-long. 60% production is easy because the shade is gone 40% into the day.

But getting back as much of that list 40% as makes sense is the challenge.

Half the panels are unshaded at any point in time, so a full parallel string would get me back at least half of the lost 40% (which was my original plan).

Parallel arrays require a great deal of current (lots of money for wires and wiring) so I’m looking for a solution with 3 2S or 3S strings that can get me close to the same ~20% loss from shading (without costing more that it would be worth).

2 more panels in another part of my roof gets me another ~20% power, so anything I spend to squeeze 20% more power out of the partially-shaded section of roof needs to cost less than that...

It looks like Solar Edge with its optimizers supports a single string of multiple angles and a battery backup is available. This may support your > 2:1 current output.

SolarEdge only works with SolarEdge, so that’s a non-starter for me. In addition, both the optimizers and the string inverter are expensive and U’m just looking to charge a 24V/48V battery, not invert power.

But if solar edge offered a cost-effective smart PV battery charger with optimizers, I’d consider it...

 

[fafrd]

I have no idea. I think it uses the AC first to cover house loads then to grid if there's excess.

Yeah, that’s the usual solution with Microinverter-based arrays. I need zero-export, so if it can be programmed to consume any excess AC power to drive smart battery-charger (and converting to DC), I could be interested. But I suspect not since all of this PV+storage stuff is evolving so much right now...

And uses the DC to charge the batteries.
That is what it will do first because it is most efficient. It will only convert AC to DC or DC to AC if necessary.
Some people use the Solark with only ac coupled micro inverters.

Can it be programmed for zero-export and will the AC-to-DC battery charger adapt to consume only available excess AC power?

 

[ArthurEld]

Yeah, that’s the usual solution with Microinverter-based arrays. I need zero-export, so if it can be programmed to consume any excess AC power to drive smart battery-charger (and converting to DC), I could be interested. But I suspect not since all of this PV+storage stuff is evolving so much right now...


Can it be programmed for zero-export and will the AC-to-DC battery charger adapt to consume only available excess AC power?

I'll let you know when I find out. People seem very happy with Solark so I assume it does what most people want.
I have net zero grid usage so I only plan to use my batteries for uninterruptable power and outages.

I think there is a good chance that net zero deal will go away some day.

And power might go out from a hurricane passing by.

 

[fafrd]

I'll let you know when I find out. People seem very happy with Solark so I assume it does what most people want.
I have net zero grid usage so I only plan to use my batteries for uninterruptable power and outages.

I think there is a good chance that net zero deal will go away some day.

And power might go out from a hurricane passing by.

Backup power in the case of a hurricane or outage is pretty straightforward.

Throttling-back microinverter output so you don’t export is not.

I’m sure good solutions will materialize - I’m just not sure if they are here yet...

The ‘standard’ approach is to pass-through any AC power (including that generated by microinverters) and to control both AC-to-DC conversion for charging batteries as well as DC-to-AC conversion for powering loads.

 

[dakine]

I bought a few dozen of the Tigo Optimizers since they weren't that much more than the Tigo Rapid Shutdown units, and in California you're required to have a rapid shutdown system for all roof solar installs.

I did hook up a string of panels with the optimizers installed in a temporary setup, and they seem to do fine, but I don't have any baseline data to compare them to. The nice thing is that the optimizer function doesn't require the Tigo Access Point (TAP) set up to function, and am using it with a Victron MPPT.

Once the weather gets better, I'm going to compare a set of panels with and without optimizers to see if there's much of a difference, and if my Growatt or Sol-Ark systems have any difference on results.

As for using micro inverters with a Sol-Ark, I don't think that is possible since it's a DC based system. @ArthurEld, have you experienced otherwise?

 

[newbostonconst]

I have both Enphase IQ7 Micro's and APsystems QS1 Quad Micro's on a AC coupled system with a Hybrid Inverter - Outback GS4048a and 20 Kwh of batteries(18 cells of 280 in series for a 63 volt pack then paralleled to that with two Battle Born 24 volt batteries and two 100 ah cells all together run off one 5 amp active balancer - guess you would call that a franken pack). Originally bought the Battle Born and they wouldn't even power up the inverter without going into fault mode - junk....so had to add the 280Ah cells.

The GS4048a throttles back the micro inverters by changing the ac frequency on the power side of the micro from the standard 60 Hz. I think it degreases frequency by stepping it down to 58.5 or increases it to 60.5 can't remember. The micro inverter sees the frequency change and shuts down because the feed is out of NEC? The specs for the power line requirements of grid power.

If the GS4048a sees the micros are suppling to much power to the system the frequency is shifted and the APsystems Micros shut right down and then when the frequency is back in spec they power back up 5 minutes later. They don't throttle back they just shut off.... The Enphase micro's are said to throttle back but I have not had a chance to witness that with my eyes...I have run the Enphase micros on the GS4048a and know they shutdown when I was testing all my options but just haven't looked to see if they throttle back or just shut right down. The system works fine for me so it doesn't really matter.

I have 3 APsystems QS1 Quad inverters that have 4 panels on each of them. And 24 IQ7's with one panel on each.

 

[ArthurEld]

I bought a few dozen of the Tigo Optimizers since they weren't that much more than the Tigo Rapid Shutdown units, and in California you're required to have a rapid shutdown system for all roof solar installs.

I did hook up a string of panels with the optimizers installed in a temporary setup, and they seem to do fine, but I don't have any baseline data to compare them to. The nice thing is that the optimizer function doesn't require the Tigo Access Point (TAP) set up to function, and am using it with a Victron MPPT.

Once the weather gets better, I'm going to compare a set of panels with and without optimizers to see if there's much of a difference, and if my Growatt or Sol-Ark systems have any difference on results.

As for using micro inverters with a Sol-Ark, I don't think that is possible since it's a DC based system. @ArthurEld, have you experienced otherwise?

Solark can AC couple. It's combined input with DC can be more than 20k PV

 

[fafrd]

I have both Enphase IQ7 Micro's and APsystems QS1 Quad Micro's on a AC coupled system with a Hybrid Inverter - Outback GS4048a and 20 Kwh of batteries(18 cells of 280 in series for a 63 volt pack then paralleled to that with two Battle Born 24 volt batteries and two 100 ah cells all together run off one 5 amp active balancer - guess you would call that a franken pack). Originally bought the Battle Born and they wouldn't even power up the inverter without going into fault mode - junk....so had to add the 280Ah cells.

The GS4048a throttles back the micro inverters by changing the ac frequency on the power side of the micro from the standard 60 Hz. I think it degreases frequency by stepping it down to 58.5 or increases it to 60.5 can't remember. The micro inverter sees the frequency change and shuts down because the feed is out of NEC? The specs for the power line requirements of grid power.

Newer-generation microinverters throttle-back smoothly through frequency-shift (ie: slightly increased frequency results in output reduction from 100% to 50%).

An older-generation micro does dot have that capability and so it just shuts down once the frequency increases past the limit.

You don’t want to let your hybrid inverter brutally shut off and on your Microinverters - that ‘thrashing’ will greatly reduce the lifetime of your microinverters.

If the GS4048a sees the micros are suppling to much power to the system the frequency is shifted and the APsystems Micros shut right down and then when the frequency is back in spec they power back up 5 minutes later. They don't throttle back they just shut off.... The Enphase micro's are said to throttle back but I have not had a chance to witness that with my eyes...I have run the Enphase micros on the GS4048a and know they shutdown when I was testing all my options but just haven't looked to see if they throttle back or just shut right down. The system works fine for me so it doesn't really matter.

I have 3 APsystems QS1 Quad inverters that have 4 panels on each of them. And 24 IQ7's with one panel on each.

I’m pretty sure newer-generation QS-1s support the power-throttling-through-smooth-frequency-shift capability, so I suspect it’s your Outback GS5048 that may not know how to control them correctly.

I have an older-generation microinverter-based grid tie system that I have considered modifying to support off-grid backup.

If I was to do so, I would assure the microinverters wake up with the sun and are only shut off once per day (once the battery is charged).