EEVblog 1518 - Hoymiles Solar Microinverter TEARDOWN

[@EEVblog]

CORRECTIONS to the reverse engineering as pointed out by some viewers. The mains side isolation transformer is a common mode filter inductor: kzbin.info/www/bejne/r4CtaKuGqsymmsU

[@davidbeard7262]

Just about t say.... Way too small for a 50Hz part. The common mode filter almost certainly functions somewhat as a differential inductor, too. The isolation is done by the two-phase DC/DC boost stage transformers, which are modulated to produce the required wave shape; the H-bridge output transistors probably just serve as a polarity switch, as there's minimal capacitance on the output of the DC/DC boost circuit.

[@mpirron1]

@@davidbeard7262 the isolation is provided by the AC coupling capacitors, so it can't affect the bias voltage on the transistors' base

[@mpirron1]

At least that's what the book says

[@DonaldSleightholme]

what about using a solar panel and a motor to turn an induction motor, a mind mill thing.. ☀️🧲🤷‍♂️🙇‍♂️

[@mpirron1]

@@DonaldSleightholme why would you ever do a dumb thing like that?

[@jusaca01]

Really interesting! I would love a teardown of a different micro inverter! I always think you can learn quite a lot when you compare two different products and check what the two manufactures did similiar and where they differ. That's the interesting part! :D

[@Korni0816]

Those little transformers TR100 and TR101 are actually current transformers to measure the current in the respective switching MOSFET and winding. The current passes through the metal strip over the top of the ferrite core ( primary winding with 1 turn) and inside the core is the secondary winding with e.g. 100 turns. The sec winding is accessible on the two outermost pins on the left side. all the remaining Pins are NC and soldered just for mechanical stability.

[@EEVblog]

Ah, yes, that makes sense. Should have gone in for the microscope closeup.

[@Mr.Leeroy]

got P/N? Ah, NVM, found it: PA1005.100NL

[@ZomB1986]

I thought they were called "current transducers". (Still "TR")

[@johndododoe1411]

@@ZomB1986 At component level only current transformers exist.

[@quetzal4042]

I have 17 Enphase microinverters on my roof, installed 7 years ago. Over last 2 years, four of them have failed one by one, which is about 600 times their claimed failure rate of 50 per million per year. I suspect they would have lasted longer in Arizona or Australia, but in wet, humid and often overcast Southeast US, with all the thermal expansion and contraction, you just can't keep moisture out of those boxes no matter how hard you try. Theoretically they are under warranty but kind of a nightmare getting them replaced, because I'm too old to be getting up on the roof and the outfit that installed them is no longer in business. Enphase doesn't like dealing with homeowners anyway, so I have to find a certified Enphase installer willing to work on somebody else's system, not so easy. And the icing on the cake, so to speak - the warranty doesn't cover labor!

[@EEVblog]

Wow, four of them?

[@craigs5212]

Lots of hardware up on the roof, those primary DC caps are under a lot of stress, switch mod ripple current high temperatures (especially in AZ) and anything connected to the AC mains is subject to all sorts of nasty transients that are very hard to protect against. Would be very interesting to see the RCA (root cause analysis) of those micro inverters.

[@strehlow]

Yikes. We have 16 which have been up for two years. Fingers crossed.

[@SupremeRuleroftheWorld]

@@EEVblog enphase does not like humid weather. They are not known for their reliabillty in europe. Many people dont use micro inverters because they fail often and are a B to replace compared to a regular inverter in the attic

[@nickwallette6201]

It's tough to begin planning out a system like this. There are so many new names, and so many _no-names,_ that you can't just say "oh yeah, Eaton, they should know what they're doing..." I'm in the far north, with potentially a couple months of well-below-freezing temperatures, and lots and lots of precipitation. It would make me nervous to put _any_ active electronics outside. (Although baking in the sun can't be too easy on them either.) I want to DIY, because: cheaper, more fun; but also because if you pick an installer, they're just going to use what they carry -- which is either what's best for their profit margin, or what they can get, or simply just what they've always used. Learning curves always suck. What can you do.

[@RickS-zj2su]

Dave the TMS320 is a mainstay of solar (and motor control). Especially the Piccolo family. Why? They have a High Resolution PWM (HRPWM) that uses some magic technique to get PWM resolution greater than the clock for very precise mico adjustments. I think it's 150pico second steps (less than 200ps anyway)! I expected to see a TMS320 and lo and behold we have it.

[@GodmanchesterGoblin]

True. People tend to forget the TMS320 is a very loose branding that TI has applied to many families of DSPs over the last 30 plus years, with quite several completely disparate CPU and instruction set architectures (I can recall seven or eight different families at least without looking anything up). Some of the newer devices with dedicated power control peripherals certainly fit well here.

[@Flole3000]

It would be interesting to know *which* TMS320 is used in there

[@theshemullet]

Thanks! I liked the way you did the schematic over the parts. Makes it a lot easier to comprehend.

[@nissarin]

Regarding the AC plug - there are essentially two versions: daisy chain (double AC plug) or trunk like this one (you plug it into "T" connector on trunk AC cable)

[@McTroyd]

I had never heard of Hoymiles before the hardware appeared in your mailbag. Apparently we've all been missing out. Looks like they're ticking all the boxes for safety and reliability. Thanks for the teardown, despite the goop.

[@MegaCyklops]

one of the only micros that passed ÖVE and DVE Testing and are certified to run here in germany/austria on the grid.

[@MattyEngland]

I had to replace the relays on a full size inverter that weighed about 30kg. Some scary caps in there, I definitely triple checked they were discharged 👍 Was a very nicely designed piece of kit, very high quality boards in it.

[@onit996]

The electrolytic capacitors store energy to balance the 2x50Hz mains power pulsation, not clouds or birds covering the panel.

[@Tim_Small]

Yeah, I was going to say the same. Otherwise you drop efficiency a load - you have to draw a constant DC current from the panel at the MPPT to get max power output from the PV panel.

[@redsquirrelftw]

That was a cool tear down. A video on how these work would actually be really cool. Ex: how do they sync with the grid, and how do they detect if it goes down. Especially when there are many of these on the grid, how does one inverter not think another inverter is "the grid" and keep going, for example. I presume it goes by current draw, if grid goes down it will try to pull massive current?

[@DavePoo2]

Or even just how to DC to AC conversion works.

[@dogastus]

I was about to say a similar comment. How does the microinvertor keep in phase with the grid when it is supplying power to it?

[@AndreasDelleske]

@@dogastus Well they measure the voltage at every microsecond so they can follow.

[@AndreasDelleske]

The inverter turns off when voltages or currents go off limits. A blackout looks like an eventual voltage spike (depending on where in the phase the current stopped) followed by a shortcircuit = current too high on the grid side.

[@monad_tcp]

@@AndreasDelleske one would think that current would flow in the inverse direction and the solar panel would become an LED. just kidding

[@XynxNet]

Love your Video. Cried a bit though watching you tearing down a unit, which was very hard to get for month here due to supply chain issues. The educational aspect was worth it, though.

[@abisaiamatalo2769]

Great video. Watching all the way from Western Kenya.

[@AnatoFIN]

That goo looked nice, I’m ordering 10 of these. I would love a teardown of a different micro inverter!

[@timconnors]

Just don't put it down on the desk and let it pick up solder blobs and other detritus, and then let it flow back over high voltage high power electronics.

[@JFirn86Q]

Very interesting teardown and going through the path like that. I would like to see another to see how they compare to be honest.

[@todayonthebench]

To a degree I wouldn't be surprised if the "transformer" Dave suspects to be the main isolation transformer to simply be an inductor. And instead have the two DC-DC converters handle the isolation. And I suspect this to be the case given the two optocouplers top left at 21:30 going over to towards the "transformer". And there is more optocouplers on the back of the board for this region too. The advantage of isolating in the DC-DC ones is that one can make a compact design able to handle high power. And then use a class D output stage to generate the low frequency AC for the mains voltage. Ferrite transformers don't tend to like low frequency stuff, especially at high currents. Tends to over saturate and do a poor job. But this isn't a problem if one switches a class D amp at high frequencies to generate a low frequency signal. Efficiency is typically fairly decent too.

[@EEVblog]

Yes, correct, it's a common mode inductor. See the pinned comment. My brain was not engaged.

[@erlendse]

Totally, a iron transformer for mains frequency would be way bigger.

[@todayonthebench]

@@EEVblog It is easy to make these sorts of mistakes when faced with an unknown circuit. Educated guesses and a bit of fast analysis is the way to go. One can also jokingly say that a camera removes a good few IQ points to say the least. The spur of the moment reacting to something "live" isn't usually the best approach to also remain correct. But does give useful insight regardless and something for us commenters to ponder over as well. Overall, a decent video. Though, first thing I looked for where how the solar panel and main's were isolated, and where. Tracking down opto couplers, transformers and such clues to potential DC isolation. And generally following the copper-less valley across the board. Since large bulky inductors/transformers/chokes are usually a black box until one does a continuity check.

[@todayonthebench]

@@erlendse That were my main suspicion as to why the DC isolation must be elsewhere. Also helps that I for fun designed my own grid tie "transformerless" inverter a few months ago out of pure curiosity to how one would do such. (Designing a circuit based on pure "This is the goal" is more fun than looking up how others have actually done it already. Usually one stumbles down a few rabbit holes along the way, learn some tricks, and generally find a few alternate solutions to a problem.)

[@kurtturner9215]

Interesting! I was sold IQ7 Enphase from the solar salesman, but the installer wanted to install Hoymile inverters with 4 panels to each. They didn't look impressive from a general overview, and I didn't like their production dashboard. I insisted on getting the Enphase, but these Hoymiles don't look too bad. Definitely would be interested in seeing what the IQ7 inverters look like in comparison!

[@XFolf]

I say keep on opening them up until we see something weird/different. Bound to be some gems out there.

[@InssiAjaton]

Interesting black "goo". I have disassembled two IGBT modules at different times. In comparison to your semi rigid stuff, what I found inside was crystal clear stuff (except the black burnt traces indicating electrical tunneling). But, clear or already burnt, it still was sticky and stretchy beyond my belief. The modules were about 6" by 3" by 0.5". They contained a 6 diode 3-phase rectifier bridge and 4 IGBT H-bridge for variable frequency drive and a "bonus" IGBT for a brake. It took me over ten hours to pull most of the goo out with tweezers, forceps, small screw driver and exacto knife. Half of the time was spent on cleaning my tools with paper towels...

[@tanishqbhaiji103]

Thank you Dave, I really wanted to see a teardown of a microinverter.

[@AlexandreMejias]

Great tear down very useful for beginners like me, Cheers Brother.

[@TobyRobb]

I enjoyed this teardown. Pretty popular devices right now.

[@TobyRobb]

@offgridgarage

[@AndyFletcherX31]

Input fuses are not needed because the panels are current limited (Isc) so even if you reversed polarity they would not source any extra current to blow the fuses anyway.

[@EEVblog]

True.

[@russellashburn6268]

It would be cool to see some testing on these before you tear them down. Hooking them up to the lab power supplies and measuring output waveform and the thermal output at the case would be examples. Maybe also do some over driving of the inputs and see what breaks first?

[@teamsafa]

Nice video. Regarding no fuse at the solar-panel input. Not needed as the panels are current limited. So as long as the connected solar panel short-circuit current not exceeds the rating of the inverter there is no need for any fuse as the panel will not be able to generate enough current to blow it anyway.

[@serggorod1423]

Good job! Almost can( re) projecting my own device! It was wherry deep!

[@Randrew]

A few years ago I opened an Enphase microinverter and it was fully potted with the same kind of material. Yup, it liked to stick in components like inductors :) In my case, the potting material seemed to kind of flow back into place with a little bit of heat and time - warmed it over 30 minutes or so on low temp in my "projects oven" in the garage. Looking back in my circuit pics, the Enphase inverters (at least the model I have) don't have fusing in the PV input side either. And they have the same double-transformers that you showed here, so I'll guess the same push-pull "interleaved" drive that you pondered of them. I apparently didn't take pictures of the bottom side where the FETs live.

[@mpirron1]

Of course, their only copying enphase. They developed the "borrowed" idea In the 1st place, which is cool to me, I think IP Protection Laws are nothing but bull**** anyway. But anyway, I have a question for you. Do you have grid-connected solar? If so, Why isn't it necessary that your inverters must cease to generate electricity when the grid goes down, right when you could use it the most? And why are you not fighting this tyranny and theft you are subjected to?

[@Randrew]

@@mpirron1 Tyranny? There are sound technical and safety reasons to not having a random (and out of utility control) assortment of power sources pushing against a broken grid. At the most basic level, isn't the "grid-tie" concept orphaned if there's no grid to tie to? The "tyranny" of physics is always ready to get involved too. Also there are solutions to your no-power-when-there's-no-power problem. Of course they center around batteries, can involve transfer switches and cost more. Ah, the tyranny of No Free Lunch.

[@mpirron1]

@@Randrew but its okay for ice generators?? Why can't I install an automated transfer switch on my inverter?

[@mpirron1]

No, that's untrue. There is no code provision in the country for DC coupled Transfer switching. Even though it works perfectly fine. And the backup battery solutions that code allows is also tyrannical, I still can't use my solar panels for power. Only for charging the batteries. Look, here's the bottom line. If the grid operator is not going to be able to Monitor and siphon off your solar generation, then its not going to be used to generate AC power, period and its pathological.

[@mpirron1]

@@Randrew Tell me this, smart guy Why is it illegal for me to send my excess solar generation straight to ground if I am connected to the grid? Why am I forced to sell it to the grid operator? Why Is he paying me 1/6th of what I'm paying him? And only in the form of credits that cannot be accrued? Why can I not be paid in cash like I have to pay? Those provisions don't sound authoritarian in the least to you. I suppose.

[@kenwallace6493]

Great video. I do wonder if the drawing of the isolation transformer is right. It would have to output 240VAC at 50/60 Hz, rated for 400VA. It seems awfully small for this job.

[@andreasdill4329]

No. The mains voltage and waveform will be approximated through PWM modulating the Power at high frequency.

[@tomwimmenhove4652]

I was thinking the same thing. I think what Dave thought is the 'isolation transformer' is actually the output filter, and the flybacks are the ones that perform the isolation. That little thing will never do 400VA at 50Hz.

[@tomwimmenhove4652]

@@andreasdill4329 That PWM signal would still contain a very larege 50Hz component. (check the spectrum of a high frequency PWM 'approximation' of a 50Hz sine wave). You can't pass that through a high-frequency transformer.

[@EEVblog]

Yes, you are correct, I should have actually put my thinking cap on, it's a common mode choke.

[@craigs5212]

I believe your are correct, I think that is a common mode choke on the AC line side. From what I have seen the PWM is done in the unfolder after the interleaved flyback converter stage. Very clever, get your PWM 1/2 wave sine conversion and polarity swapping for the second 1/2 wave with the same components.

[@ToddDesiato]

Thanks for this. If you apply the panel backwards, the diodes in the FETS limits the voltage to -0.7V. I routinely get 356W out of my HM-350NT and they blow away my IQ8+. They're excellent, IMO.

[@Ferraday]

haha that’s not a good thing. a -0.7v “limit” is a short across the panels...

[@ToddDesiato]

@@Ferraday They are current limited to Isc, it doesn't hurt anything for a short period, and doesn't damage the inverter.

[@wasddasw6485]

@@Ferraday Across one panel - it doesn't matter since they'll only get as hot as they would with no load anyway. The panel will get the same amount of energy from the sun whether it's shorted, disconnected or properly connected. Only difference is where the heat gets dissipated. If the panel can survive sitting out in the sun, it can survive being shorted. I'm more concerned about the 0.7v drop across the FETs. At around 13 amps short circuit, they'll have to dissipate 9.1 watts worth of heating. Better hope that thermally conductive potting compound is up to the job since that's the only thing moving the heat from the PCB to the outer case.

[@maxtorque2277]

@@wasddasw6485 Clamped at around only 1v the panel will be so far from it's MPP that the output current will be nowhere near the rated value of that panel

[@benssolarandbattery]

That's the same goo that I pulled out of an Enphase IQ7+ microinverter. Sticks well when thrown at a wall. 🤪 I'm not great with electronics but great explanations of it all!

[@flymypg]

Assuming no ingress, what do you think would fail first? (My bet is a 65V max input rating could age those 63V input caps. I like to see 100% margins on input/output caps.)

[@EEVblog]

Residential panels are signifantly under that though.

[@strehlow]

@@EEVblog I could see someone putting a couple in series on each inverter, especially if they are hedging against a poor roof angle or something.

[@tactileslut]

Very disappointed to see a customer facing rating HIGHER than that on the caps. Negative headroom, and on something typically installed out in the heat.

[@timconnors]

@@tactileslut it's all a continuum curve anyway. It's just less time before it begins to fail.

[@EnricoConca]

@@tactileslut You have to consider that in normal operation, the panel operates at about 70% of rated open circuit voltage due to the MPPT. So, it would only ever experience 65 V in case of maximum solar irradiation during a grid failure. I would still not rate such an inverter for 65 V though.

[@adaminsanoff]

Yes. Please review also the other inverter.

[@J1mW357]

Wow thank you for this, I always wondered how that would look like from the inside.

[@DavidHembrow]

Nice to see. There's an HM400 in the very small system that I installed in my daughter's apartment. We have HM1500 on our second system on the garage roof. They seem well engineered.

[@getcartercarpark.]

I, too, would love to see teardowns of different micro inverters to see the differences in design and operation, please! Thanks, KW

[@demoncloud6147]

More please, I always dreamed of making these as DIY but it is beyond my skill level. This video gave me a rough understanding

[@serggorod1423]

I add it to my schematic playlist!

[@serggorod1423]

When I saw it! Hi resolution is good, but 2xxx is greater! Complete good job!

[@DantalionNl]

I would like to evaluate the micro inverters based on emc compliance and testing. In our country basically every imported micro inverter fails to meet regulation and its a big problem!

[@adhdengineer]

My solar installer wired one of my 9 panel strings reversed polarity into the hybrid inverter. Managed to handle it fine, i was impressed. Though it would have at least destroyed the mppt for that string, but no... It's running fine...

[@ThatEngineerGuy_]

Oooobaby can’t wait to watch this later

[@imark7777777]

I wish all potting compound would be like this it would be so much fun.

[@MeiklesAndDimes]

Definitely get the other micro inverter for comparison. How do these thing’s synchronize with grid frequency?

[@EEVblog]

Yes, they sync, that's the idea.

[@BrianHG.Ocean.Fitness]

Loose sync and KABOOM!

[@pedromorgan99]

@@EEVblog How ?? does it detect "grid" then switch on... but then how does it detect if grid is "disconnected"...?? ie islanding... cant figure that out in my little grey part...

[@jaro6985]

@@pedromorgan99 Simple solution would be to switch the output off for a cycle every 10s or so, and measure the voltage to see it collapses. Google says various methods are used, shifting the voltage or frequency and measuring for a response on the grid.

[@mymisc1896]

One would guess it uses PLL (Phase lock loop) type of circuit to achieve grid sync.

[@craigs5212]

Looks like its following a Microchip AN1338 app note but using a Ti processor. Basically the two large PQ cores are the inductors for two interleaved flyback boost converters. They are interleave for efficiency. These produce a high voltage DC that is galvanically isolated from the panel input DC by the inductor's secondary winding. The DC then runs to a combined PWM and un-folder stage to produce the AC output sign wave. The un-folder makes the opposite half phase. Micro inverters are current sources not voltage sources. You can't expect your 350W inverter to change the grid voltage but you can push current to the grid. I didn't go through the code but my impression is it adjusts the voltage phase relative to the grid phase to control the current supplied to the grid and to remain within the MPTT range of the panel. The DSP math is quite complex, with several servo loops, PLL's and error checking. All the while running the flyback converters, PWM and unfolder circuits. Lots of factors to consider. Every so often they turn off and check that the grid is still alive and shut down if the grid has failed. Don't want to fry the lineman up in the bucket truck.

[@EEVblog]

Nice find. Yeah, almost indential topology.

[@pranjal3727]

Always wanted to know how they know grid is there or not WITHOUT turning its output off and listening. And you said its that simple, they just turn off for a few cycles🙂 Apart from detecting an overload, any other way to detect grid is down or not without dropping output power?

[@LukaszTNT]

Thanks for pointing to interesting lecture 😊

[@jessicav2031]

Is that mains side correct? I thought you weren't supposed to put GDTs directly across high current protected lines because once they light, they may not extinguish. I always heard that this was a common failure so MOVs are put in series with GDTs. Maybe the GDT is in the center and goes to ground, while the other side goes through one MOV to live and another MOV to neutral?

[@andreasdill4329]

It will just blow the fuse on the mains.

[@jessicav2031]

@@andreasdill4329 Good point, I don't see the fuse but I'm sure it's there.

[@EEVblog]

I checked the traces, yep, it's in series with the MOV, nice call.

[@phester28]

yes please, i like knowing the failure points of these ☺

[@wouterke9871]

The cable glands and the seal on the lid that is to flexible with to little screws. I opened my DS3 AP-Systems; same goo and waterdrops in all 4 corners within 6 montths after install.

[@DanielR8392]

Just I recognize that is a HMS version. The models with HM and MI it exist already a opens source project in github (Ahoy and OpenDTU). With that you can read the inverter and control it. The HMS use a other frequency (868mhz) and it is in reingeneering process already.

[@robertpeters9438]

It would be great if you ranked them: for example: temp management via thermal camera, protection for reverse polarity or overvoltage

[@souhaibtouati9330]

Those Caps are called DC-Link capacitors, mainly energy buffer und reduction of current ripple, a common thing in Inverters 👍

[@souhaibtouati9330]

And regarding the TMS320, no suprise 🙂 the C2000 is still being used today, but mainly as a redundency/safety controller not as a main MCU

[@serggorod1423]

My congratulations from Krymia We are expecting for U! Th-x for all!

[@mcflapper7591]

Yeah! Please do the ZBeny review! :)

[@boomer9900]

My son's friend is a lineman. Protocol states that each primary phase must be grounded before any work can be done on the conductors. However, it is interesting to note that no such protocol exists for the stepped down 240V residential connections to the distribution transformer. As more and more solar go online I think these rules need to be changed.

[@randycarter2001]

The DC output of the solar panel is not as clean as you want it. They actually have a fairly high source resistance. I forgot how many cells you have to have in series to get 60 volts. Adds up to a lot of slop in the voltage hence a lot of beefy caps to smooth it out. What kind of current pulses do you have at 400 watts?

[@Mobile_Dom]

love everything about my 800w hoymiles microinverter, except for the factthe solar monitoring is only once every 15 minutes, so definitely not real time monitoring

[@flymypg]

One more inverter for comparison would be good.

[@MegaCyklops]

you can talk to them via an NRF and openDTU :)

[@jani140]

Prices have gone through the roof since they are controlable by opensource software

[@karlakunze9823]

Hi @Peter Kowalsky, MI- and HM-Models did use NRF24L01+ for communication via 2.4GHz Enhanced Shock Burst at 250kBps. But recent HMS- and HMT-Models instead employ an CMT2300A for communication via 868MHz / 915MHz at an expected 125kBps rate. OpenDTU and AhoyDTU do not support HMS-/HMT-Model communication yet. But we are working on that :D You may check our Wiki for the Protocol and some of the details for Germany and Austria, in case you want to keep the produced signal in line with the local Grid Profile(s).

[@MegaCyklops]

@@karlakunze9823 thanks for the info karla!

[@mikefochtman7164]

So gas discharge between the mains, and another between mains and ground. Good protection in case the mains are subject to something harsh, like a lightning strike?

[@piotrcurious1131]

nope. this will protect only from minor arcing, like f.e. ballast from fluorescent tube kicking in or disconnecting wall-wart . Lightning strike has current of about 10 000A. For 0.1ohm protection, this means 1000V spike . Also lightning strike generates huge EMP. IMHO best protection against lightning is insurance, as anything else just adds incredible complexity and cost. Another good protection are dpdt relays with at least 5KV isolation on both input and output. As storms usually do not happen in clear sky, if there is no sun unit should be disconnected and grounded by relays. Usually spikes coming from storm can be also detected on AC side. Most UPS devices have some kind of protection algorithm implemented basing on detecting voltage spikes in the grid, throwing protective relay and switching to battery mode.

[@hasenlamano]

I would love to see the switching topology of this inverter since me and a friend started our thesis making a 2kw inverter

[@TheOneJacinto]

Thanks for sharing!!!!

[@rcinfla9017]

Single phase AC has a sine^2 power profile where the PV panels need a constant DC load to provide maximum power. You don't want 2x AC line frequency power flow ripple current to show up on panels or you lose power from PV panel. That is the reason for all the electrolytics across PV inputs and the weak component reliability link in the micro-inverter scheme. Electrolytic caps MTBF is highly dependent on their operating temp. Micro-inverters operate in a very hot environment behind PV panels, likely a few inches from a hot roof. Not great for electrolytic caps longevity.

[@davidgunther8428]

ChargerLab channel just did a teardown of the Enphase IQ7+ micro inverter today. It must be a trend! 😆 Their layout seems a bit different, but they also use a lot of input capacitors.

[@matsv201]

There was a fireman that died in Germany due to DC arc over a few years back. Since then they have to check the solar panels before starting the fire fighting. Note, that even if its in nighttime. Just the back light from the fire, might charge up the panels to a considerate voltage, even if there is hardly any current. This is a problem, because if your house is on fire... for some other reason, and they can´t verify that your pannels are AC or low voltage DC, they will not save your house. (well.. they might if its just a minor fire, but if its a big fire and they have to spray the whole house)

[@toitmouton6285]

It is, in fact, a differential mode filter to filter out the switching harmonics.

[@donwald3436]

3:00 Dave playing with EE play-doh lol.

[@Citizen5000]

Theres an open source solution for the wireless protocol with an ESP32 and some wireless module as a bridge. Got it running, works like a charm and only costs cents!

[@eogmau]

Can you share the solution?

[@Ferraday]

The tiny transformers are CTs for the power switches

[@robertwhite4375]

The square device on top of the AC side MOV is most likely a thermal fuse in series with the MOV to guard against short circuit failure of the MOVs. Quite typical arangement for surge protection. Will isolate a MOV that has failed low impeadance. You loose the surge protection but it shouldn't catch fire. Safety feature rather than functional benefit.

[@DjResR]

I know someone had anti-islanding fail because two different inverters fed the grid keeping each other awake. Aishi caps are used in CFLs a lot._

[@wouterke9871]

I considered this to achieve island on a boat, haven't tested yet.

[@LutzSchafer]

Nice teardown Dave! I have a question though. The mains transformer looks so small for 400W and 50/60 Hz? Wonder what kind of core material is used to have the permeability and low Eddie currents required not to heat up like crazy. Also obviously the H bridge creates harsh rectangle waveform on the primary. There will be losses. Then again I am not an inverter person just wondering...

[@giangelov]

Nice catch. Dave got it wrong. Galvanic isolation is from the flybacks and the H bridge he is referring to is actually mains frequency unfolder. This is exactly the patent of Nphase inverter - flybacks operating with rectified AC reference and the H bridge converts it to AC. The magnetic components is a common mode filter with very high differential mode components as well.

[@LutzSchafer]

@giangelov thanks for the clarification. Though I don't see smoothing caps after rectification from the flybacks. So that means that flyback frequency halfwaves are used. Likely the filtering after the H bridge takes care of this.

[@giangelov]

@@LutzSchafer That is the reason they have used interleaved flybacks - less ripple. They do have filter caps and the DM inductance helps also. On the front section it is exact topological copy of Npahse. In their design I remember they used thyristors for the H bridge. I'm not sure if Dave did check these devices, but from the solder side of the PCB the "gate" driving circuit looks very simple for driving MOSFET.

[@richardspiek1401]

Please consider working another micro inverter for comparison. I'm using Enphase on my panels. Thanks for what YOU do. ...N3JLR

[@deterdamel7380]

I don't think, that the transformer on the top (AC-output) is an isolating one. It's too tiny for 400W @50Hz. I guess it's a filter in mH range.

[@vincenthamilton4792]

Does anybody know the name of the thermally conductive potting compound used in this device?

[@williamarmstrong7199]

Certianly. Some of them have a bit of an iffy rep. I was thinking of getting 3 of these for the 3 spare panels I have. However the cost!! So I bought a Solis 1 KW inverter 2nd hand. (As new under speced by an installer) so it has only a few hours of use. This will tie in with my 5 Kw Solis inverter. It wiĺl run 4x 260w panels so I can add another panel if I want. It cost less than one 300w micro inverter so all good.

[@ScramblerUSA]

Hmmm... It must disconnect if there's no mains voltage - got it. This means my neighbors who hope to enjoy electricity in case of power failure at least during the day are screwed. :) A question though - how does it know when to disconnect? I assume all those microinverters are connected in parallel at some point, so to each inverter there is no difference between mains and another inverter. How does it work?

[@mpirron1]

Each baby inverter's AC output all meet up in parallel in some kind of "combiner" box where they are AC- coupled to your smart meter in one of several possible schemes. And then comes in, like you so astutely grasped, their primary function. See we assume the primary function would be to generate electricity for the homeowner, but only the homeowner thinks that. You know how can you know? Because if the grid operator is down and can't take the excess at it's price, you don't get to generate as your inverters shut off.

[@TheHuesSciTech]

@@mpirron1 You've completely missed answering the OP's question. The inverter was sampling mains voltage and frequency, and output current accordingly. The the mains power goes out. But the inverter has been putting current on the mains line, which if it were balanced to the load, would create a sinusoidal voltage. V=IR. So how would the inverter distinguish between that power failure case, vs the mains still being on?

[@mpirron1]

@@TheHuesSciTech I didn't miss answering that question, Because it's not the one that anybody should be asking. But if I must... the inveteter is AC-Coupled to the mains, across the coupling capacitors And the frequency Is ... Oh, just read it in the other post below, or read the papers on it. However, there is a real question here. It's just not that one. The question is why do the inverters have to shut down?

[@mpirron1]

Hint.... It's got nothing to do with backfeeding the grid.

[@TheHuesSciTech]

@@mpirron1 That was explained in the video. It's so that electricians working on fixing the problem (e.g. getting the fallen tree branch off the power lines) don't get electrocuted by power being fed back onto the grid by solar inverters on nearby homes. To add more detail, you *could* have special circuitry to cut off the house from the grid so that the inverter could power the house without energizing the grid; but a) that would be an expensive addition for a rare circumstance, and b) there's no guarantee that the current level of light on the panels could power everything in the house anyway, so it'd be a really shitty and confusing experience for the average person. Meanwhile, a emergency diesel generator can always provide XXX kW on demand, and are selected and specced based on what the relevant business/hospital needs to keep going. And the whole intention of them is to deal with power blackouts, so the equipment to definitively and safely NOT back-flow the power onto the grid is worth the cost. So stop being a troll who dismisses perfectly reasonable and interesting questions as "not the one that anybody should be asking", giving BS answers, and thereby derailing their threads. You're perfectly entitled to start your own thread, so do that.

[@grapsorz]

more and more ppl is ordered to shut down there solar systems here if they use inverters of that type (one for every panel) they make so much radio noise together that they shut down airport radar and other communication systems. yes they complies by itself.. many of them together do not.

[@SiaVids]

It would be unusual to fit a spark gap over the top of an MOV, it is more likely to be a thermal sensing device to check if the MOV's are getting hot through over voltage.

[@nickyhathiramani7861]

Hi Dave what great timing , I was looking at these for my new installation. I have heard these units run bloody hot even when ambient is under 30 deg C , this worries me for longevity when compared to Enphase. What have you heard and your thoughts. I was also lucky enough to meet you at Norwest mall a year ago, keep up the good work. Cheers

[@DerekWoolverton]

Back when I worked on inverters the TI chips had much faster and more capable ADC which made the control software easier to write.

[@basaltnow]

I think they are using two 180 degree interleaved step up converters and than make the AC which they then step up to 230ac using that other output transformer. That could be, instead of using two fkyback with all their disadvantages vs step up in CCM.

[@Mike-gz4xz]

If installing one or two PV panels say just to supplement a few watts grid tied , dose Hoymiles Micro Inverters need the gateway to wake them up like Enphase micro inverters ?

[@danielheinrich8046]

Though I couldn't see the look on your face while you depotted this thing, I could practically *hear* the gigantic grin.

[@ovalwingnut]

Making the insaney more plainy. Thank you

[@wickedprotos1937]

That thermal 'putty' would be nice to identify and source.

[@petehiggins33]

Where are the output capacitors for the flyback converters? Surely they should be a similar size to the input capacitors.

[@deterdamel7380]

There're not needed in this design. The boost stage creating the sin-like-wave by PWM. A filter (small caps + Inductor > 1mH, on the upper part in the video) removing the HF. The DC-to-AC bridge switching to the phases with 50/60Hz (or 100/120Hz?).

[@petehiggins33]

@@deterdamel7380 So, are you saying that the big magnetic component at the top of the board is not a common mode choke as Dave has declared in his correction video but is in fact a differential mode choke?

[@deterdamel7380]

@@petehiggins33 I expecting an inductor with >= 1mH there, maybe common mode, or two on the same core or whatever. I said it's not an isolating configuration. But Dave mention this in the correction video already.

[@petehiggins33]

@@deterdamel7380 Someone on the correction video has pointed out that the input filter has to handle 100Hz ripple current and maintain the input voltage within the operating range and this is why the input capacitor is so much larger.

[@soniclab-cnc]

definitely more inverter teardowns

[@flymypg]

That potting compound looks like a goopy silpad!

[@WeBeGood06]

16:40 Mains Voltage, Mains Frequency but no discussion about Mains Phase. Is all power added to the grid as Voltage (Amplitude) or does it ever use leading Phase to increase frequency back towards 50/60 hz. Would be nice discussing how the Grid goes down and the response of microinverters as Voltage and Frequency drops.

[@petersage5157]

Those axial power resistors are pretty high spec for carbon film jobbies - 2% tolerance, 25 ppm/K. Very not bad.

[@pranjal3727]

Why not metal film or metal oxide? Metal oxide are flameproof and grunty.

[@jonathanfulcher602]

Please do that other inverter teardown.

[@mfessi]

By using the putty, no pressure equalization element (valve) is needed, since there is practically no air in the housing - right? Does this work well in practice? Might be also a good alternative (instead of controlled ventilation) for outdoor home projects (Wi-Fi APs...)?!

[@chaitanyasindagi1237]

The h-bridge drive after the flyback looks like it might be a full bridge driven LLC converter? - hence the two caps, one for DC, one for AC resonance. Did you check the pin connections there?

[@no_more_free_nicks]

oooo this is what I wanted to see: like + comment

[@PeaceChanel]

Thank You Everybody for All that you are doing for our Planet Earth.... Peace.. Shalom.. Salam.. Namaste .. 🙏🏻 😊 🌈 ✌ 🌷 ☮ ❤🕊

[@electrodacus]

The AC side is galvanically isolated from DC side so the clearance is irrelevant. There is no reason for current to flow between those two fully isolated sides. PV panel is also like a battery fully isolated (maybe frame will be connected to earth but that frame is still isolated from the PV cells).

[@EEVblog]

Not if you have a failure or leakage somewhere.

[@electrodacus]

@@EEVblog The point is that in normal operation there is no voltage potential between those two points thus even if there is water, there will be no current flow. Having water there is already considered a fault or abnormal operation and a second fault at the same time is almost never considered. The PV panels are a fully isolated source unless one of the PV wires has damaged insulation and maybe touches a grounded metal roof or the frame of the panel. Even so if the AC side near that DC is the neutral (as I guess it will be the case) there will still be no current flow.

[@mpirron1]

Galvanically isolated by those AC coupling capacitors, so mange voltage doesn't interfere with the bias voltage applied to the base of those transistors while it's getting frequency matched.

[@Piasecznik72]

Hi Dave, I wonder how mains sense works. How it can differentiate mains coming from grid from mains coming from another micro inverter? Can it be fooled by lets say UPS?

[@GTrainRx7]

That is a pretty cool (pun intended) solution for something that is not needed to be opened..

[@toomoii]

Can you do a video where you go into detail about circuit protection like with MOFs etc? I know little about what you can do and should do. There was sadly nothing lectured in university despite this maybe being the most important thing when designing an electric device… Also would be great if someone could recommend some existing tutorials.

[@alstare72gael88]

hello great video.. I have to repair the enphase microinverter which does not work because of 1 swollen capacitor.. I removed the gray paste.. but it must be replaced for the seal. Did you find the name of this thermal paste.. thank you

[@J_i_m_]

I'm still wondering how an inverter "knows" when the grid fails. Especially when production is in a perfect balance with self consumption. Also, how could an inverter measure the grid frequency (to stay in phase) when it's continuously generating it's own sine wave on the same wires.... It's like it cannot talk and listen at the same time. Maybe another idea for an episode ;)

[@danielaustin7643]

i was thinking the same thing

[@erlendse]

Except you can. With mains the voltage (and frequency) is given, the inverter just controls the current. Without mains connection the inverter controls the voltage too. Also, it can try to slightly push the frequency and if it works just shut down as soon as possible!

[@mpirron1]

@@erlendse it matches it's frequency to the input mains frequency. When there's no input mains voltage to sample, it shuts itself off because it has too by code. Now ask yourself why a backup ICE generator gets to work, but not a backup solar inverter. Let me know what you come up with, would ya?

[@J_i_m_]

@@mpirron1 Input=output. When a solar inverter puts it's power to the mains, it will also measure itself. My question was, how is the solar inverter able to distinct it's own sinewave with the grid's sinewave.... it's on the same wires! But @erlendse answer "it can try to slightly push the frequency and if it works just shut down as soon as possible!" is a brilliant answer!

[@brainthesizeofplanet]

It's usually not the inverter,.it's the hub it energy meter "listening" for the net 50hz frequency and then telling the inverter array what's up

[@happyhippr]

hell yeah, let's see a 3 phase inverter teardown, maybe motor controller teardown from a modern power tool or appliance

[@davidmcdonald3314]

Just thinking about what you said about the gap on the board and that you would have extended it down further... I certainly don't know the answers here But I do wonder if the re-entrant conformal coating might make it pointless Here's my thinking... Any moisture that gets to the board must surely have to first get through the conformal coating So moisture would be in the conformal coating anyway Wouldn't that provide a moist path that is able to bridge any gap??? Like I say, I don't know. Perhaps someone else here has knowledge and is able to share it