How do Tigo optimizers work?

[fafrd]

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.

Especially if you’re a fellow Californian, I’m very interested to understand what your Tigo Optimizers do for you. The cost over an RSD MLPE is modest enough to make it a no-braineif they actually perform as advertised.

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.

The test I’d be interested in is to see how they handle one partially-shaded panel in a string of 2 or 3 panels:

If you completely block one cell in a 2-panel string, current and power should plummet until the MPPT drops voltage enough to activate one bypass diode in the shaded panel (~1/6th voltage drop) at which point current should fully recover, but power is only ~83% (5/6ths) of what it was without that shaded cell.

I don’t believe the Tigo optimizer is going to be able to improve upon this since but it would be interesting to measure as a baseline.

Where things get interesting is if that shaded cell is only shaded over 1/4 of its area. In that situation, current should drop to 75% and power along with it. A good MPPt will discover that dropping voltage to activate a bypass diode drops voltage by 1/6th and increases power to 83%.

But with the Tigo optimizer in place, it should be able to increase the current out of the shaded cell from 75% to 100% (without activating a bypass diode) while decreasing voltage out of that panel from Vmp to 3/4Vmp, meaning string voltage should be 1.75Vmp and power should be (75% + 100%)/2 = 87.5%.

I’m pretty sure a +33% buck is the limit of what these Tigo Optimizers can deliver, so as the shade of the1/4 shaded cell is slowly increased, power should drop from 87.5% to 83% where the MPPT drops voltage enough to activate a bypass diode and the optimizer stops bucking.

The other thing that would be worth checking is what the optimizer does in the case of a mostly-shaded panel.

Sn MPPT will need to bypass a shaded panel by activating all 3 bypass diodes, which will drop another ~1.5V of of the Vmp of the unshaded panel (so string voltage will be Vmp - 1.5V rather than Vmp).

If the optimizer is able to internally bypass entire panels with a single diode bypass diode rather than 3, that should mean voltage of Vmp -0.5V rather than Vmp -1.5V without the optimizer.

And if the optimizer uses a switch to bypass panels, that would result in a string voltage of a full Vmp (half the string voltage of the unshaded string).

If Vmp = ~30V, avoiding 1V of diode drop amounts to 3.45% more power and avoiding the full 1.5V drop from 3 bypass diodes amounts to 5.3% more p

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.

If/when you do that testing, please plan for some shading. If you have the ability to make longer strings, Tigo claims that their optimizers allow a 4S string to be put in parallel with a 3S string, which is another thing you could try.

Unequal string length results in a midpoint Vmppt which is too high for the shorter string (reducing it’s current) and too low for the longer string (decreasing it’s voltage and power along with it).

With Tigo optimizers, buck compensation should allow the panels in the longer string to increase current by up to 33% while decreasing voltage up to 25% and the net result should be both strings operating at ideal Vmppt of the shorter string while the panels of the longer string is internally operating at Vmp even though each optimizer is only being given a voltage of Vmp * (#short/#long) bolts from the MPPT controller.

 

[Maast]

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....

Because if you want to use batteries you'll need another inverter that can AC couple anyway - and when you run through the cost of a inverter (that has to be sized for your load) + the microinverters its less expensive to go with optimizers, inverter, MPPT charge controllers. Though once you throw in the cost of high-end MPPT controllers (Victron, Midnite, Outback, etc) the cost savings isnt as much. If you're using a cheapie like a EPever then the cost savings is considerable.

Plus I didnt want to have all my eggs in one inverter/charger basket and if my central inverter died I wanted to be able to still charge the batteries and run off a backup inverter until I could repair or replace the main inverter. I've got three parallel Midnites so even if I lose one I'm not dead in the water.

If you DONT want batteries then the optimization and redundancy of microinverters is definately the way to go.

@fafrd I dont know if you've seen this yet, but this guy put the Tigos through its paces and they performed very well with shading: https://www.mcelectrical.com.au/tigo-energy-solar-panel-optimisers/

 

[fafrd]

@fafrd I dont know if you've seen this yet, but this guy put the Tigos through its paces and they performed very well with shading: https://www.mcelectrical.com.au/tigo-energy-solar-panel-optimisers/

Yes, I’ve seen that (may have even linked to it earlier in this thread).

If you have the kind of ‘partial-shading’ that reduces panel output by 25% or less, for sure Tigo Optimizers can deliver the goods.

When you have the kind of partial shading that results in bypass diodes being activated with standard (non-cut-cell) panels, they actually do very little - it’s the bypass diodes that do all the work (of getting current up to ~Imp) and the optimizers just pass on the results.

I could live with that but the issue I have is with how they perform with half-cut cells (which that video did not test).

A half-cut-cell panel is essentially 2 half-panels in parallel (with shared bypass diodes). The beauty of half cut cell panels is that with shading on only 1/2 of the panel, you are guaranteed to get at least half-current (Imp/2) without any bypass diodes being activated.

That’s in a 1S parallel string.

If you put two half-cut panels in a 2S string, shading on half of one panel drops the output of the entire string to just over 50% (from 75% available energy).

If Tigo’s optimizer could deliver 100% buck, that Imp/2 could be boosted to full Imp (at Vmp/2) and the string would put out 100% of available energy in the form of Imp @ string voltage of 3/2Vmp_panel.

Unfortunately, Tigo’s optimizers are limited to 33% buck, so Imp/2 can be bucked to at most 67% Imp and the second (unshaded) panel in the string will be limited to 67% Imp @ > Vmp = >67% Imp x Vmp meaning the whole string will be delivering just over (50% + 67%)/2 = 58% of maximum output rather than 75%.

I’ll have a sizable portion of the morning with half-cut panels half-shaded, and a parallel string will give me ~30% more output from those panels than if I connect them in series with unshaded panels.

I’d love it if anyone has tested Tigo optimizers with half-cut panels with partial shading, but I haven’t found any such thing...

 

[Maast]

Yes, I’ve seen that (may have even linked to it earlier in this thread).

If you have the kind of ‘partial-shading’ that reduces panel output by 25% or less, for sure Tigo Optimizers can deliver the goods.

When you have the kind of partial shading that results in bypass diodes being activated with standard (non-cut-cell) panels, they actually do very little - it’s the bypass diodes that do all the work (of getting current up to ~Imp) and the optimizers just pass on the results.

I could live with that but the issue I have is with how they perform with half-cut cells (which that video did not test).

A half-cut-cell panel is essentially 2 half-panels in parallel (with shared bypass diodes). The beauty of half cut cell panels is that with shading on only 1/2 of the panel, you are guaranteed to get at least half-current (Imp/2) without any bypass diodes being activated.

That’s in a 1S parallel string.

If you put two half-cut panels in a 2S string, shading on half of one panel drops the output of the entire string to just over 50% (from 75% available energy).

If Tigo’s optimizer could deliver 100% buck, that Imp/2 could be boosted to full Imp (at Vmp/2) and the string would put out 100% of available energy in the form of Imp @ string voltage of 3/2Vmp_panel.

Unfortunately, Tigo’s optimizers are limited to 33% buck, so Imp/2 can be bucked to at most 67% Imp and the second (unshaded) panel in the string will be limited to 67% Imp @ > Vmp = >67% Imp x Vmp meaning the whole string will be delivering just over (50% + 67%)/2 = 58% of maximum output rather than 75%.

I’ll have a sizable portion of the morning with half-cut panels half-shaded, and a parallel string will give me ~30% more output from those panels than if I connect them in series with unshaded panels.

I’d love it if anyone has tested Tigo optimizers with half-cut panels with partial shading, but I haven’t found any such thing...

I'm unclear why 33% buck (convert voltage downward) would be a limitation as the primary function of an optimizer is as a boost (convert voltage upward) converter when the panel is shaded to match the other unshaded panels in the string voltage so that one panel doesnt drag the entire string down.

 

[newbostonconst]

Fafrd, your plan to shut them off only once is unrealistic... Batteries do not come up to full smoothly and neither does the amount of sun coming to the panels... It is going to cycle unless you make a circuit to latch the charger off the first time and are happy with partially charged batteries... Partially charged is fine with me because I don't charge all the way anyway...

Pretty sure I remember seeing the unit step frequency smoothly when I had an oscilloscope on it but will check again another day and report back.

 

[fafrd]

I'm unclear why 33% buck (convert voltage downward) would be a limitation as the primary function of an optimizer is as a boost (convert voltage upward) converter when the panel is shaded to match the other unshaded panels in the string voltage so that one panel doesnt drag the entire string down.

That’s not what buck converters do - they increase current at the expense of voltage.

So 75% Imp @ Vmp can be converted to 100% Imp @ 75% Vmp (33% buck).

It’s the current of panels in a series that needs to be matched, not the voltage.

Current is common/shared, do it’s a low-current panel that ‘drags-down’ the other panels in a series string, not a low voltage.

Voltage drops by ~Vmp/4 for the full string (using this example) and string output goes to [Vmp x Isc x (n - 1/4)] which is the maximum theoretical power given that shading and is far better than the [Vmp x 0.75 x Isc x n] that would result without optimizer / buck conversion.

 

[Maast]

That’s not what buck converters do - they increase current at the expense of voltage.

So 75% Imp @ Vmp can be converted to 100% Imp @ 75% Vmp (33% buck).

It’s the current of panels in a series that needs to be matched, not the voltage.

Current is common/shared, do it’s a low-current panel that ‘drags-down’ the other panels in a series string, not a low voltage.

Voltage drops by ~Vmp/4 for the full string (using this example) and string output goes to [Vmp x Isc x (n - 1/4)] which is the maximum theoretical power given that shading and is far better than the [Vmp x 0.75 x Isc x n] that would result without optimizer / buck conversion.

Thats not quite the way they work. Buck/boost converters (which is what a optimizer is) are voltage devices not current devices. What the 'buck/boost' refers to is voltage and the current adjusting follows that. There's only one 'tool' electronics have to vary current - like a power supply, and thats almost always voltage, there are simple devices that use resistors to reduce current and voltage but thats very lossy and not used very much.

An optimizer buck/boosts the voltage output to the maximum power VMP of the panel if it's by itself, or to match its neighbors if its series/parallel. Since it has to send its own panels power along it can't do that if the panel output voltage is less than the other panels in the string - otherwise the entire strings voltage will go down.

Watts is watts so while its true that when a optimizer downregulates voltage to VMP (voltage maximum power) like on a sunny cold day the current goes up, however when the optimizer has to upregulate the panel output to its VMP the current goes down, but it's not 'bucking' the current its 'boosting' the voltage and the current goes down as a consequence.

If you have any documentation that says otherwise I'd very much like to read it.

 

[fafrd]

Thats not quite the way they work. Buck/boost converters (which is what a optimizer is) are voltage devices not current devices. What the 'buck/boost' refers to is voltage and the current adjusting follows that. There's only one 'tool' electronics have to vary current - like a power supply, and thats almost always voltage, there are simple devices that use resistors to reduce current and voltage but thats very lossy and not used very much.

An optimizer buck/boosts the voltage output to the maximum power VMP of the panel if it's by itself, or to match its neighbors if its series/parallel. Since it has to send its own panels power along it can't do that if the panel output voltage is less than the other panels in the string - otherwise the entire strings voltage will go down.

If you know of an optimizer that does buck/boost, please provide a link. I spoke directly to Tigo and they confirmed to me that their optimizer only does buck compensation and only to a maximum of 33%.

Watts is watts so while its true that when a optimizer downregulates voltage to VMP (voltage maximum power) like on a sunny cold day the current goes up, however when the optimizer has to upregulate the panel output to its VMP the current goes down, but it's not 'bucking' the current its 'boosting' the voltage and the current goes down as a consequence.

This sound great but is not what Tigo’s ‘Impedance matching’ optimizer does. It takes a panel with suboptimal current from a series string and bucks the current up to a maximum of 33% higher at the expense of lowering voltage by a maximum of 25%.

They may also have a bypass mode that fully bypasses a shaded panel using only one bypass diode or a switch instead of 3 bypass diodes, but I have not been able to confirm that yet.

If you have any documentation that says otherwise I'd very much like to read it.

I linked to this article earlier: http://www.diva-portal.se/smash/get/diva2:1249003/FULLTEXT01.pdf

Look at figure 3.18 - this is apparently the only compensation circuit integrated in Tigo optimizers.

 

[fafrd]

Thats not quite the way they work. Buck/boost converters (which is what a optimizer is) are voltage devices not current devices. What the 'buck/boost' refers to is voltage and the current adjusting follows that. There's only one 'tool' electronics have to vary current - like a power supply, and thats almost always voltage, there are simple devices that use resistors to reduce current and voltage but thats very lossy and not used very much.

An optimizer buck/boosts the voltage output to the maximum power VMP of the panel if it's by itself, or to match its neighbors if its series/parallel. Since it has to send its own panels power along it can't do that if the panel output voltage is less than the other panels in the string - otherwise the entire strings voltage will go down.

Watts is watts so while its true that when a optimizer downregulates voltage to VMP (voltage maximum power) like on a sunny cold day the current goes up, however when the optimizer has to upregulate the panel output to its VMP the current goes down, but it's not 'bucking' the current its 'boosting' the voltage and the current goes down as a consequence.

If you have any documentation that says otherwise I'd very much like to read it.

And here you can see it stated by Tigo directly: https://support.tigoenergy.com/hc/e...rences-Between-Tigo-and-SolarEdge-Optimizers-

SolarEdge optimizers are Buck/Boost while Tigo are Buck-only...

 

[fafrd]

I found this article which provides a nice visual to see the difference between SolarEdge (Full module-level optimization), Tigo (Buck-only module-level optimization), and Maxim (Buck-only substring-level optimization): https://help.valentin-software.com/pvsol/en/calculation/power-optimizer/

The Tigo buck-only module-level optimization extends the optimal Pmp power peak to voltage below Vmp by bucking so that current is increased and voltage is decreased (and current x voltage = power remains the same).

The other thing this graph makes clear is that Tigo’s optimizers (and all optimizers) make use of a panels bypass diodes. I see lots of marketing misinformation to the effect that Tigo optimizers get maximum power out of a panel without using bypass diodes, but that is nonsense (only true if there is no partial shading).

 

[fafrd]

I found this article which provides a nice visual to see the difference between SolarEdge (Full module-level optimization), Tigo (Buck-only module-level optimization), and Maxim (Buck-only substring-level optimization): https://help.valentin-software.com/pvsol/en/calculation/power-optimizer/

The Tigo buck-only module-level optimization extends the optimal Pmp power peak to voltage below Vmp by bucking so that current is increased and voltage is decreased (and current x voltage = power remains the same).

The other thing this graph makes clear is that Tigo’s optimizers (and all optimizers) make use of a panels bypass diodes. I see lots of marketing misinformation to the effect that Tigo optimizers get maximum power out of a panel without using bypass diodes, but that is nonsense (only true if there is no partial shading).

I’m still coming to grips with what these Tigo optimizers do and here is a bit of analysis for a 2S string with shading causing one substring of the shaded panel to drop to 75% Imp:

75% Imp @ Vmp with no bypass diode activated and no optimizer will constrain the second panel to 75% Imp @ > Vmp + (Voc - Vmp)/4 resulting in string power of > 75% of max efficiency for the full 2S string).

75% Imp @ Vmp, activating a bypass diode will deliver Imp @ 2/3Vmp or 66.7% efficiency while second panel will operate at full efficiency (so 83.3% of max output for the full 2S string).

A Tigo optimizer can boost that 75% Imp to 100% Imp @ 75% Vmp resulting in 87.5% of maximum efficiency for the full 2S string (75% + 100%)/2, the maximum output possible given that 25% shading situation.

Repeating this exercise with 30% shading on one substring, the bypass diode delivers the same 83.3% of maximum output while the Tigo Optimizer boosts the shaded panels output to 93.3%, meaning the unshaded panel is constrained to >93.3% of max output and the 2S string delivers >81.7% of maximum output (meaning the MPPT will likely prefer to activate the bypass diode and the optimizer didn’t add much of anything).

So in the case of partial shade limited to a single substring, the Tigo optimizers are effective up to about 30% output reduction from shading on that substring.

Beyond that, they are really not adding anything to what the bypass diodes are able to deliver on their own (with or without an optimizer).

 

[Runolfsson35]

This sound great but is not what Tigo’s ‘Impedance matching’ optimizer does. It takes a panel with suboptimal current from a series string and bucks the current up to a maximum of 33% higher at the expense of lowering voltage by a maximum of 25%.

I have tigo on panel which gets occasional shading from chimney and when the shade appears the current of whole string goes down by voltage goes up. I think it matches behaviour described by @Maast

 

[timselectric]

In my opinion, from everything I have read. DC string inverters are more efficient for charging batteries. Because, they are DC to DC chargers. They don't have to invert to AC, and then convert back to DC.

 

[fafrd]

I have tigo on panel which gets occasional shading from chimney and when the shade appears the current of whole string goes down by voltage goes up. I think it matches behaviour described by @Maast

I’ll take your word for it that I’ve had the wrong understanding then.

So the Tigo optimizers will reduce current until available power is being supplied at target voltage.

Now that I think of it, that makes more sense.

If two panels are in parallel and one panel has partial shading causing a single bypass diode to be activated, the Tigo Optimizer operating in the manner you are describing would boost that panels voltage from ~2/3Vmp to 88% Vmp meaning lost power from the unshaded panel would be reduced from ~33% to only ~12%…

I wonder if this might also allow 1S of a 48V LiFePO4 battery with the new 150-cell panels Trina has started producing.

Those panels have a Vmp of 43.0V which is too low to charge a 16S LiFePO4 battery, but if that string voltage can be boosted to 57.3V using Tigo Optimizers, it might be just enough to get a 48V LiFePO4 battery charged…

 

[Runolfsson35]

To be fair there is a lot of confusion about how tigo should work especially when large batches were found faulty across many years, see https://www.mcelectrical.com.au/tigo-optimiser-recall/

It's not unlikely that my tigo are simply broken and the data misleading see my other thread

 

[Runolfsson35]

I found another confirmation from tigo employee about that it's current that is boosted, not the voltage which goes down. So you may be right here and my tigo are just broken.

 

[Strousy]

Optimisers will never switch on bypass diodes as the optimisers get their energy from the panel. They cannot switch individual panel bypass diodes either as there are only two cables going to the panel. The minimum operating voltage of a Tigo unit I believe to be 18V, read that somewhere. Operating 3 bypass diodes accross a panel would imply a supply V to the Tigo of 3x0.6=1.8V. This will be impossibly low for a Tigo. Tigo will come with limitations which are in true capitalist fashion NOT HIGHLIGHTED IN THE SALES BLURB.

of interest might be: https://www.mcelectrical.com.au/tigo-optimiser-recall/

TIGO Energy Solar Optimisers. Better than solar panel bypass diodes?

Why do we need Tigo Energy solar panel optimisers when solar panel bypass diodes are doing the job? This blog shows how optimisers do a far better job than bypass diodes, and it explains why Tigo Optimisers are the preferred optimiser solution.

www.mcelectrical.com.au

 

[Strousy]

Sorry guys, but it appears that the first question is not: "How do Tigo optimizers work?" but "Do Tigo optimizers work?" We have 3 people on this forum currently questioning if they work at all. A fourth one is an electrician putting his name and reputation on the line:

Tigo Optimiser Recall?

Tigo optimisers seemed like a good idea at the time. In the past, I’ve spruiked Tigo as an affordable shade solution. But my recent Tigo testing has uncovered two major issues that render Tigo…

www.mcelectrical.com.au