SMF - Just Installed!

Main Menu

Recent posts

Green Energy / How I Connected An External Fa...
Last post by electron - Mar 29, 2024, 12:16 AM
If you are going to run a Power Jack type Grid Tie Inverter at full output, it's a good idea to move some extra air over the case.

A lot of people have put small 80 mm PC power supply type fans outside of the GTI and it seems to work good.

The problem is, how do you make them turn on and off only when needed like the little one inside does?

A quick calculation showed me that running the fan 10 hours a day at about 2W would be 20 Wh a day wasted (sort of). Plus wearing out the fan quicker.

I did some measurements and decided to use a 80 mm PC style fan instead of the one supplied and also to run the GTI with the cover off and blow air directly at the circuit board and transformer.

I can run it with the cover off because I am the only one with access here and it's in a place where it's hard to put your hand in there. It's also sitting flat on top of a big piece of metal and away from anything that could catch fire.

FIRE RISK IS SOMETHING YOU NEED TO THINK ABOUT - If this thing decides to go poof, normally it should be mostly contained inside the metal case, but when you remove the cover like I did then you are taking a risk of fire.

You have been warned. I am just showing you what I did.

I may change this set up later and put things back to normal and run an external fan blowing on the outside of the GTI.

The internal fan runs off of the 10V inside the GTI, that comes from a LM317 regulation circuit. The LM317 has it's own heat sink. See below for technical details.

This GTI is the OLDER style, with the flat top.

The pics below show you the adapter cable I made for this from an old dead portable phone battery. Example batteries are shown too. You could simply cut the fan wires and splice them to the connector cable then tape them up. I added the extra pins because I am always changing things and had a hard drive power connector adapter running a different fan.

I took pics of some old PC power supply cases, you can see how you end up with enough extra metal to cut and make a nice fan holder. The metal is easily bent to shape to what you want. You already need to screw down the GTI so you can use those mounting screws to hold your fan in place.

Technical Details:

The internal fan runs off of the 10V inside the GTI, this comes from a LM317 regulation circuit. The LM317 has a heat sink on it and is rated for only 1.5A max.

The LM317 is powering other things and gets pretty hot already so you don't want to load it much more.


When I measured the fans directly, the small internal fan draws about 80 mA and the new bigger 80 mm fan draws 120 mA. The LM317 inside the GTI seems to be OK with that little extra 40 mA amount but I wouldn't push it too much more.

You can make your own external LM317 circuit and make it adjustable so the fan isn't too loud. The only problem is that it won't go on and off only when the GTI needs it. I may make a post later about a way to do that.

Here's a simple LM317 circuit and calculator for figuring out resistor values. It only needs 2 resistors for a fixed voltage or put a pot in series with R2 (see below link) to be adjustable.

The GTI connector is a .1" center connector, so are most of these portable phone battery connectors. The fan connector is also a .1" center type, but has 3 pins typically. It is possible to take a razor blade and carefully cut the fan connector to fit the GTI connector.
I put a dallas temperature sensor on the Grid Tie Inverter and logged it. It has been doing this same thing for over a year now. Full power output to the grid at 250W sometimes I see 260W, and you can see when it goes to constant output at about 9:30AM.

This is a cheap Power Jack 300W Grid Tie Inverter. The old model with the flat top.

This one has the 80mm fan blowing into the inside as described in the previous post. The small stock internal fan is disconnected.

Some people have said you can't run these at full output. Since the max temperature here is only about 79F, I think you can run them full out with the normal fan but mount them to a piece of metal.

Most MOSFETs are rated to operate up to 300F !

I have measured the efficiency of this one, direct amp measurement, and it's about 78% efficient at full output.

I'm also running this off of two big deep cycle 12V batteries (GTI at 24V), but that's another story.
See previous posts below for info on the external fan.

I put a dallas temperature sensor on a NEWER style Grid Tie Inverter and logged it for an hour. Ran it on 24V batteries.

This GTI was mounted to a vertical wood board, fan facing up. The sensor was clamped to the side using a plastic clamp so it wouldn't skew the results.

This GTI has no extra fan blowing on it and is not mounted on metal.

At one point in the chart you can see how the GTI shut down for some reason, it was only at 95.6F. One green light was on and did the quick short blink thing, fan was on but no output power for one minute. Like all machines, it would probably do that again over and over but this was only a hour test. I guess that is how they protect it.

Room temp was 65F and fan output temp was hanging around 86F (using a different sensor).

I also took voltage and amp readings every 10 minutes. Watts is output to the grid, read on the Kill-A-Watt meter.

Time Amps Volts Watts Temp Efficiency
 00 12.7A 23.5V 251W 65.0F 84.1%
 10 11.9A 23.2V 230W 88.0F 83.3%
 20 11.8A 22.9V 224W 92.7F 82.8%
 30 11.9A 22.6V 222W 94.9F 82.5%
 40 12.0A 22.4V 221W 95.6F 82.2%
 (shutdown happens in here for 1 min, fan output is at 86.5F)
 50 12.3A 22.2V 224W 94.0F 82.0%
 60 12.4A 21.8V 221W 96.0F 81.7%

Kill-A-Watt reads 200 Wh total to the grid. Should be 220 Wh but that's what it actually read.

Conclusion? Needs a extra fan blowing on the outside, even a slow movement of air would help. 96F is really not that hot for most electronics and mosfets can go much higher but they have programmed this device very conservatively, or it's possible that this particular one is a little sensitive. Consumer grade thermistors and resistors have a wide tolerance so it's possible that is the problem.

I have added a fan from a power unit that was burn after 9 years i hooked it up with a simple bracket and i control it with aux. controller thermostat switch for cars or trucks that use it for trans coolers or help out a/c systems run better when u add a aux. fan to the cars radiatror or trans. real well. These are fully self contained units u can get them at most auto parts stores for about $40.00. They made to handle big power like 40 amps or more. I have them in all my Rides. U can see this set up on Utube. Its good to cool these things off and its great when u can run off battery bank. See my video below starting at time index 1:06 to see the thermostat.
[media][/media] to all

Green Energy / 60A EPpsolar charge controller...
Last post by electron - Mar 29, 2024, 12:12 AM
Hello guys,
Please I have installed a 60A/24v MPPT charge controller, however I noticed that when the current from the panels (6 300w mono panels)reads between 0.5A to 9A, the charge displays on the charge controller, however when the value of the current rises to about 30A it doesn't read on the LCD on the charge controller, it reads 0.00A, but a multimeter reading from the panels shows it reads 32A. Please any one with experience on epsolar charge controller and any setting I need to adjust on controller?Have you been to the EPsolar site? They have some software updates that might fix it, the code on your unit may be old or may have become glitched. Their support dept has been helpful to me so maybe contact them after trying to update the software.

See Products --> Technical --> Download
Green Energy / California Public Utilities Co...
Last post by electron - Mar 29, 2024, 12:08 AM
"FAT CAT IS OUT OF THE BAG: Evidence has now been made public of illegal actions and collusion between former California Public Utilities Commission president Michael Peevey and utility PG&E, as criminal investigations continue."

"Former California Public Utilities Commission president Michael Peevey and PG&E are now subjects of federal and state investigations, as reported by the LA Times and other media."

"Within a release of 65,000 emails is proof that Peevey knew � since 2010 � that "smart" meters can cause physical harm, and said his own bill more than doubled after an install on his vacation home. Read our feature article and watch the shocking video of his "stage-managed" farewell:"


Green Energy / Free And Virtually Unlimited F...
Last post by electron - Mar 29, 2024, 12:04 AM
YouTube user "Tiny House Listings" shows us how to easily make junk mail / newspaper fireplace burn bricks with a cheap plastic bucket and a hand drill.

Free And Virtually Unlimited Fuel To Heat Your Home

"Every day we have tons of paper sent to us in the mail, boxes that our food is packaged in and in my case, lots of beer packaging. In this video I show you how to turn all that paper into something very useful, a way to heat your home for free. "
goodsaving beer packing now

IUPUI's solar power installation harvests the sun
How I wish more and more universities would be doing something like this, not only does it keeps cost of education down but would also make students more aware of renewable energy.

Two Small diy Solar Panels Made Out of Glass and Aluminun Make 55 Watts
I only had some small pieces of glass so I made two separate panels out of glass, aluminum and silicone encapsulate. Tied the two panels in a series pos into neg then it plugs into the inverter. They are making about 55 watts. I think at solar noon it will be more wattage. the voltage open circuit is 18.7 volts. And the two panels are making 14.8 volts with a load. I hope to make another small panel of only 6 cells to attach to it. to make the 21 volts short circuit it needs on cloudy days. :D

Green Energy / Grid Tie Inverter Inside Look ...
Last post by electron - Mar 28, 2024, 11:57 PM
This is a list of YouTube Grid Tie Inverter technical / repair / inside look videos I listed as "Favorites" so that they would be easily accessible for reference.

The new YouTube format seems to be leaning toward not having "Favorites" shown anymore, so I am going to list them here. They are not in any particular order.

Inside sun 300 grid tie

PowerJack 300W Grid Tie Inverter Modifications

PowerJack 300W Grid Tie Inverter Component Upgrades

Update 2010-10-16 - Inside the Sun 500G.avi

Gti replacement of capacitor 1

Gti replacement of capacitor 2

Gti replacement of capacitor 3

Gti replacement of capacitor 4

Gti replacement of capacitor 5 final

INSIDE the SUN 250 and Power Jack 300

Power Jack 300 and SUN 250 Grid Tie Inverter REPAIR

PowerJack Inverter Repair... (j/k)

Other related videos are listed in the "Favorites" bar to the right on my YouTube channel, for as long as that feature lasts:

GTI power jack inverter solar panel MPPT electronic repair replace capacitor resistor transformer MOSFET transistor diode IC 14V 28V TO-220 heat sink opened up look inside internals

Green Energy / Help a new user understand... ...
Last post by electron - Mar 28, 2024, 11:54 PM
Wanted to get started with solar for a cabin. I bought a 50 watt Renogy panel and the classic cheap solar charge controller with an MPPT unit on my wishlist further down the road.

I have my panel outside in direct sunlight, not a cloudy day.

Here's the setup: You see two batteries but I'm only connected to one. I am concerned about several things. The "load" light on the controller, should that be on? I'm only getting 5 watts & .38 amps? I didn't expect 50, but I didn't expect 5 either...

Different configuration just for my own knowledge. I've got the meter inline with the battery charge.

Any thoughts are appreciated.


I am concerned about several things. The "load" light on the controller, should that be on? I'm only getting 5 watts & .38 amps? I didn't expect 50, but I didn't expect 5 either...

My load light is on when it's charging, I think it only goes off if the battery is low, in your case on a 12V system probably 10.5V maybe?

Here's some video, I use the same kind CMP 12/24 but put a small extra heat sink so it can handle a little more amps. You can see 14A in the second video, I have had seen them handle 20A.


It looks like the panel is getting some shade in that picture, is it behind a fence? Any shading will cut back the power, even a small corner shaded. It also needs to be pointed at the sun angle as best as you can get it, generally face it south and adjust it at least twice a year for tilt.

On that CC, generally it starts backing off at 14.25V (12V system) but if you watch the second video you can see how the panel's amps go down as the volts go up, that's just the way it is.

When you see the "charge" LED start blinking on your CMP 12/24 that is the controller backing off, meaning your batteries are charged or getting there.

The controller I am testing on the second video seems to have a lower on/off point so I prefer the others. The blinking you see is slower than the CMP 12/24 you will see, it's pretty fast.

And you can see the added heat sink better on the second video. It can't be too heavy of a heat sink, it's only held up by the mosfet leads, this is a small old 386 PC heat sink from very old computers.

See this thread for info on how I beefed it up:
Thanks for your reply! I saw that article before I bought my kit. I have also heard that these chargers are not PWM, they're just voltage regulators... I'm looking at getting a better controller, so might not bother modding this one.

I had some time to move the panel around and get full sunlight around 6pm today. I started getting about 25 watts and 1.5 amps from the panel. I noticed if I cover just the corner of the panel with my fingers the voltage output drops to 4 or 5 watts. I REALLY didn't expect that. I've got my homework to understand how that works!
Green Energy / LTO Lithium Titanate Batteries...
Last post by electron - Mar 28, 2024, 11:48 PM
If you get the flat pack cells made by Kokam, the tabs/terminals at the top are a very soft aluminum and are long and straight.

(2024: You should look up the info for the NEW "Salt" batteries, their price is going down)

For my test pack I used a Leather Punch Tool that you can get at Walmart in the craft section. I think the biggest hole size is 3/16" or .188" on the tool.

I punched two holes to match a metal piece from Home Depot in the cabinet hardware section called "Mending Plates" the 2 in. long ones with two holes are perfect for this, EVERBILT P/N:339482 or SKU 030699152995, zinc plated, I used 2 plates per terminal.

I used short 10-24 3/8" phillips pan head screws and nuts with lock washers, internal style. Longer ones were put on the end plates for connection to wire lugs.

I had to hold the plate with my thumb, then used a scribe tool inside the holes to make a mark, then used those marks to punch.

While punching I squeezed hard and then gently twisted sideways only slightly back and forth and I could feel when the hole was done. After one hole was done through TWO terminals, I would insert a screw and nut temporarily and check everything else for alignment, then punched the other hole.

I think the punch is better than trying to drill holes.

I had to be SUPER CAREFUL to insulate the terminals and operated on only 2 cells at a time basically. I worked backwards and numbered the cells as I went making sure they are aligned on top of each other for each one, then putting them aside in a safe way so the terminals don't touch anything. 2 cells at a time slowly and carefully, put one aside and move to the next one in order.

If the punch tool ever shorted, these batteries will put out a lot of power really fast and so I was super careful and used cardboard to insulate other terminals and made sure I could concentrate the whole time and not be interrupted. No distractions (good idea to use packaging tape to insulate all tools).

As I added in the screws to finish the stack after all the cells were correctly punched 2 at a time, I also put on a piece of black tubing used for "drip irrigation" with a slice down the length on one half, covering the entire terminal and screws, it just slips on from the side and that insulates the terminals and is cheap. This helps when using tools so I put them on as each connection was finished. I taped the cells together as a stack using some strong packaging tape when done because they will slip around and they are very heavy once assembled. I used them stacked on top of each other, they could probably be used sideways or with the terminals at the top with the proper container.

Well, that's what I did and it worked. I don't think this is something just anyone can do, especially the - impatient in a hurry types. One "24V" 10S stack took probably 2 or 3 hours of total concentration.

Below are some pics [sorry. not here anymore] of OTHER PEOPLE'S BATTERIES for 12V, not mine. But the cell pic is the same type I used for 24V (10 pcs in my case). This single cell is about 10x10x0.5", about 55 to 60 Ah and gives me about 1 KWh out of the inverter (20V cut off) for 10 cells.
I never heard about "super capacitor balancing", do you know how this works or do you have a link to this balancer?
about the 30s BMS: have you concidered about building 2 batterie packs each with a BMS and wire it up in series later? 2 x 15s or one 14s + one 16s BMS should be plenty of choices.
Green Energy / Re: How to: Fix a solar panel ...
Last post by electron - Mar 28, 2024, 11:42 PM
I know this thread is old, but seeing as how panels have become cheaper these days, would it be more practical to fix broken panels or just go buy new ones? Some of mine badly need repairs but I�m yet to calculate how much it�s gonna cost. Buying new ones might be cheaper and save me the hassle of repairing.I suggest to you that you should buy another one. If you will just fix it or repair it I think your expenses would be bigger than having new solar panels. Its now your choice I am just giving suggestion.Just wanted to say, thanks to everyone for their kind comments on this fix.

It shows it's possible and you could fix a panel from the back side in some cases, most panels have a plastic sheet on the back that you could cut into if you had to.

If you can fix a panel by bypassing one cell and it's not to hard to do I would do it. If there is a bad connection, same thing.

If the panel is in really bad shape like this one and you got your use out of it (I used it for years) then maybe it's time to get a new one since prices are getting so cheap (as you all said).

This panel was putting out very little power and someone local recently sold me a 75W panel for $35 but it had no frame at all, just the glass and a connection box on the back side.

So I simply put it on top of the broken panel now, the connection box sticks out above the top of the old panel so the whole glass panel lays flat pretty much, and used some small 26 ga wire across in 3 places to hold it in place and now the broken panel has made a nice mount and I am getting 75W in the same place for only $35 more!

The 26 ga wire is thin enough that it doesn't seem to effect the 75W panel's output power at all but it's enough to keep the wind from moving it.

So I found a new use for the old panel and don't have to fix it anymore!
Green Energy / How to: Fix a solar panel that...
Last post by electron - Mar 28, 2024, 11:38 PM
I have a factory made panel I've had to fix over and over, the company went out of business of course because these panels suck!

[sorry the pics are not here but the text is still valid]

The back side is one piece of aluminum and gets WAY, WAY more hot than the other normal plastic backed panels I have.

This type of repair should be possible on other more normal panels, except you would cut into the plastic backing near the tab wire positions. You could even add a wire back there to bypass a dead cell.

Most likely it's a connection problem and not the cell going bad.

Whats nice on a plastic backed panel is when you are done with the repair, it won't block any light coming into the cells.

The front of my panel is encapsulate, then a thin clear plastic of some type, maybe a special mylar. I can fix it by cutting into the front.

You can see the thin clear plastic if you look close at where I made the cuts in the picture below. The encapsulate is still mushy, a little more than dried silicone sealer.

You can see on the left what their connection "tab wire" looks like, not very wide, that's the first problem, but the bend was very sharp to try to get the cells closer together.

Here's what happens. It gets hot and the cells expand away from each other because the encapsulate stretches, The metal wires bend. It cools down at night, wires bend back. After some days that action over and over breaks the connections.

You can find the broken connections using a volt meter that has sharp probe points and with the panel under a light source, I poked through the plastic to the tab wires.

The expansion is so severe that it has actually caused my solder joints to break! I have had to re-solder joints several times so far.

I fill in the hole I made with some clear silicone sealer RTV. And that attracts dirt so I try to make the cuts as small as possible.

Followup on the fixes I did.

In the pics below you can see where I did the fix a long time ago and now what's happening is the "mylar" top thin film plastic (I don't know what it really is) has kept on ripping from where I made the cuts.

It sort of makes a bubble and then dirt gets in there and blocks the light.

I was able to remove most of the dirt using a vacuum and then used Super Glue to glue it down against the encapsulate layer, using a metal block to keep it down until it dried. You can't really get under there and clean it perfectly, it will have to be good enough.

This pic is the worst two spots I had, it's possible they are the very first two I did.

I think the temperature swings from hot to cold every day cause the thin film to rip.

Now I get to see how long the Super Glue holds up to the daily sun.

The second pic is the same spot after I glued it down and cleaned it out a little more.
Another followup on the fixed panels, the super glued parts came up and had to be glued again and now I see one part that is turning dark for some reason.

Super glue was best because it's like water and flows under the plastic film very easily. I thought it would be OK but I may have to find another type of glue.

I looked at the stuff you glue onto windshield glass to put a rear view mirror back on, but it's usually a gel.

The liquid worked really well.

If this turns really dark I may just strip off the plastic film and do something else, I don't know yet.

Pic below.... [NOT SORRY] Followup on the fixes I did.

In the pics below you can see where I did the fix a long time ago and now what's happening is the "mylar" top thin film plastic (I don't know what it really is) has kept on ripping from where I made the cuts.

It sort of makes a bubble and then dirt gets in there and blocks the light.

I was able to remove most of the dirt using a vacuum and then used Super Glue to glue it down against the encapsulate layer, using a metal block to keep it down until it dried. You can't really get under there and clean it perfectly, it will have to be good enough.

This pic is the worst two spots I had, it's possible they are the very first two I did.

I think the temperature swings from hot to cold every day cause the thin film to rip.

Now I get to see how long the Super Glue holds up to the daily sun.

The second pic is the same spot after I glued it down and cleaned it out a little more.
Green Energy / Grid Tie Inverter Repair - Tra...
Last post by electron - Mar 28, 2024, 11:32 PM
If your Grid Tie Inverter has a bad transformer it's not going to be a simple fix.

The GTI shown here is a 300W Power Jack style one, also the "sun" GTIs have the same board type.

If you have measured all the mosfets already and found that none of them are shorted, then you can move on to the next thing and check the transformer.

Measure using the low ohms setting on your meter, mine has a 200 ohm position.

Short your leads to see what a "zero" reading really is, mine reads 1 ohm or so, yea it's cheap but it doesn't matter, you now know what "zero" (a complete short circuit) looks like.

The DC side of the transformer here as "A" and "B", when you measure across the "A" pins, you will see "zero" ohms, same across the "B" pins. The "A" pins are one side of a center tapped transformer coil, "B" is the other.

On the high voltage output side of the transformer you will measure low ohms, maybe a couple ohms across the "C" pins and then measure across the "D" pins. The "C" pins and the "D" pins are separate transformer coils.

If you measure high ohms (more than say 100) or open on any of these measurements then the transformer is open and bad.

You can remove the transformer and see if you can repair it. It is possible that a wire came off a pin or they didn't solder it right at the factory, check all that.

We are talking about simple wire wound around inside the transformer, so if you can find the break and fix it without causing any short circuits, then you may be able to save it.

Other than that you will most likely not be able to find a replacement, but you might find a motor rewinding place that would be willing to rewind it, maybe. If they have the equipment.

If only one of the output coils "C" or "D" are bad then you can still use the GTI but at lower output power and it will probably lose some efficiency.

On the newer models of this GTI, they have the output coils connected in parallel, so you will have to lift one of the jumpers on the bottom of the circuit board to measure the coils separately. A low output on a newer model would indicate that one of the output coils has gone open.

Hi Sir , I hope you can help me out.
I got an almost exact version of GIT.
I checked transformers as you describe , seems ok.

Few months ago , my GIT stopped workin.
I ordered exact replacement parts for the MOSFETS.
Following instructions found in Youtube I changed 2 of the AC side Mosfets for new ones. Although I was not absolutely shure they where bad.
The reparation did not work out.
I found out later that LM317 regulator was not working ok. I replaced it and the unit starded working again.
It did so on the bench with a Pc power unit. Lights blinking and power converted ok.

As I go to connect to my solar panels , the unit worked a short time ( minutes ) and goes off again.

Back to the bench I found out there now is a short on DC main side, the LM317 does not get the DC input voltage on corresponding pin.
I have checked almost everything , to no avail.

Please, if so kind, how I can further check this DC input side short.

Thank You very much

Please, if so kind, how I can further check this DC input side short. On the DC side the mosfets are in parallel, one set on each side of the transformer, also the transformer looks like a short so it's hard to figure it out easily.

You should be able to tell which side is really shorted.

Generally, but not every time, the G - Gate pin is shorted too, and the gates are not completely paralleled with the other one so there is a small resistance, so you can check that and maybe get lucky and tell which one is gone.

The other way is to pull both mosfets and then check each of them.

Thank you very much for prompt answer.

I will try check this, I am not shure if I did understand well.

I will start a new threat titled "Grid Tie Inverter Repair - I have 6+4 = 10 Mosfets on one side and 4 diodes on he other.

Maybe you are in the mood and make an analysys about this.

I would very much appreciate it.

I am shure there are many trobled DIY users out there stuck with broken cheap GIT's.

Thank you again. :)

Greetings from Chile

I will start a new threat titled "Grid Tie Inverter Repair -. I have 6+4 = 10 Mosfets on one side and 4 diodes on he other. Sorry, but I can't teach a class in electronics repair here, there is way too much information out there on the net for me to spend time on that.

If you don't know how to use a ohm meter, I suggest you practice and mess around with it on a old radio or something or go on youtube and learn about it.

If you can't trace down which mosfets are on the DC side you are in trouble already, the traces are right there.

Sorry, but I already said what to do.

Like I said, just remove the DC side mosfets and test them one at a time, that is the easy way if you don't know much else.

Fearless has some good videos on youtube about GTI repairs, maybe you can figure it out that way. You could have found them by searching on youtube or even google.

I have a powerjack 1200 with a short dc
I removed the 8x dc mosfet pin "s" from all, and they are ok
With diodetesting they show 600

I removed the elkos to

BUT without the mosfet there is still a short between "s" and "d"

Can the transformers be the cause of the shortcut ???You mean you cut the S pin and are going to solder it back?

Interesting idea. That might work well on a Sun GTI since the mosfets are bent over from the board and laid flat. Actually on the Sun's you can leave long leads to solder on a new mosfet and not have to take the whole thing apart.

It's not easy to fix the DC side since things are in parallel.

But if the S pin is disconnected, the transformer coil should be "disconnected" basically.

Maybe the capacitors are what you mean by "elkos" ? The caps are across the DC too so that is something to check but not likely unless they look bad.

Is it possible one of the S pins is still making connection?

I am not sure what is left to short out the DC side except maybe a bad voltage regulator?

Could your meter DVM be tricking you somehow with caps charging?

I am assuming that you mean the short is still across the PCB pads marked S and D. Solder bridge somehow?

Get a good light on it and look around.

My GTI is labelled "smart grid tie microinverter". It comes with 15 years limited warranty, but the small print says the customer has to send the unit to Taiwan at his expense and risk.
So when it stopped working last week I decided to do surgery myself and transformed a small  torxx tool bit to fit the triangular holes of the Chinese screws. After removal of the circuit board I saw in the center an SMT diode marked M7 that was cracked. This is an 1N4007 in a very small package, but there is plenty of space there, so I replaced it with an old fashioned one. Above the diode there are two wires going to the on/off switch.
I connected DC, switched the unit on and immediately saw smoke coming from a capacitor next to the diode. That part (and all other SMT caps) is unmarked, so you can only judge the capacity by the package size. This one is 0402, so probably a ceramic one, around .01-.022 uF and 25V.
I fitted a small .01/50V ordinary cap there, applied DC and saw that the red and green LEDs started to blink.

The inverted is back on the wall, shows a green light and fan just started to work, so I guess my repair was successful!