Teardown and investigation of "LP" 150W sinewave inverter

[Th3_uN1Qu3]

I bought a new multimeter today (Uni-T UT70A), it's an average and kinda overpriced meter, as the "A" is the lowest end model in a series, but it's beefy and it's got inductance measurement (that's why i bought it). I guess i could've done better but oh well. The UT60E will still be my primary meter but i really really really needed inductance measurement and this was the best meter available locally that doesn't cost a fortune.

While i was at it, another device caught my eye - a 150W sinewave inverter branded "LP", made for a Polish company, for $54.49. It will enable me to scope the primary of SMPS supplies and design snubbers, and it sure is a heck of a lot smaller than an isolation transformer. A quick look thru the vents revealed nothing terribly wrong, and Chinese for Poland is still better than noname Chinese so i bought it.

I wired the inverter to an UPS battery, hit the power switch, and.... NOTHING. Not even the power LED coming up.

Of course, i have a warranty. But like i know those guys, they're gonna take a month to order another one. So screw the warranty. The teardown shall begin!

Edit: The only thing that was bad was the cigarette lighter plug thing, i took that apart and wired the battery straight to the wires and hey presto. Attached are a few pics of the guts, i'll comment them later. The third IC is a KA3525 (Fairchild equivalent of SG3525). Other than that, just the usual. 7500 (TL494) and a LM324 quad opamp. I'm curious on how the sinewave part is done tho, i'll take a more in-depth look the next few days. And i bet those are some cap brands you've never heard of.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

I guess somebody got their terms mixed up - this isn't even modified sinewave, it's pure 100% square wave. And it doesn't run at 50Hz, it runs at 55Hz... They probably figured, hey, there's 50Hz and there's 60Hz, let's do something in between so we can sell it on both markets. Not that i'd care for my application, but lying in your marketing blurb isn't considered good practice.

[kaboom]

Re: Teardown and investigation of "LP" 150W sinewave inverter

Hope those MOSFETs have really low Rds- all that cast aluminum of the housing and those little heatsinks! Inside the unit besides, where the heat will only build up.

My inverter, "modified sine," aka stepped square, has a small micro for the line frequency, full bridge switching, current limit, both input and DC bus voltage monitoring, UVLO. The "power" switch applies battery voltage to a 78L05, which then powers the micro. When the analog input of the micro for battery voltage sensing gets a sample (thru resistive divider), an 'enable' signal is sent to the 494. The 494-MOSFET-transformer converter boosts battery voltage to 165V. The DC bus voltage feedback is taken directly back to the 494, through another voltage divider. Once the DC bus is at 165V, the micro starts the full bridge, converting the DC bus voltage to AC.

There actually was a jumper wire in there to select 50/60Hz. I just removed and did away with that, saving on space. Of course, a different transformer ratio would be required for 220-240V out.

Which of the two PWM ICs control the four MOSFET full bridge output? Increase it's timing cap to lower the output frequency.

One of them will drive the two MOSFET pair that drives the transformer and will have its feedback coming from the DC bus. The other will be 'locked' at 50% and driving the full bridge at line frequency. That one will run open loop.

There's probably a neat little set of gate drive ckts for the full bridge, probably with 2N3904s and 2N3906s.

[japlytic]

Re: Teardown and investigation of "LP" 150W sinewave inverter

I don't know how a pure sinewave inverter is made from a TL494, SG3525 and an LM324 (or their equivalents).
One possibility is the LM324 being used as a sinewave generator.
Anyone want to make a circuit of this unit?

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

Hope those MOSFETs have really low Rds- all that cast aluminum of the housing and those little heatsinks! Inside the unit besides, where the heat will only build up.

It did its rated load short-term while staying fairly cool. My 7.2Ah battery was fairly weak so it didn't last much, and unfortunately i don't have a 12v heavy PSU to spare - but i bought a brand new 9.5Ah SLA today, so i'll see how it does on this.

On this topic, the battery low alert is triggered when battery voltage reaches 10.9v. And that buzzer is really freakin' annoying. If you continue running it, it'll start cutting off the output, turning it back on, cutting it off again... at a ~2Hz rate. Not my idea of protection but oh well. I think some hysteresis is called for in that comparator.

My inverter, "modified sine," aka stepped square, has a small micro for the line frequency, full bridge switching, current limit, both input and DC bus voltage monitoring, UVLO.

A micro? That's too expensive.

Which of the two PWM ICs control the four MOSFET full bridge output? Increase it's timing cap to lower the output frequency.

A cap? Bah... i was expecting something a bit more advanced.

One of them will drive the two MOSFET pair that drives the transformer and will have its feedback coming from the DC bus. The other will be 'locked' at 50% and driving the full bridge at line frequency.

IIRC i only saw two parts on the HV side heatsink... so mine's gotta be half not full bridge. On the positive side, i get a free power upgrade - there's room left in the transformer core for more (bigger) wire.

I don't know how a pure sinewave inverter is made from a TL494, SG3525 and an LM324 (or their equivalents).
One possibility is the LM324 being used as a sinewave generator.
Anyone want to make a circuit of this unit?

Well, that's the point - it ain't pure sinewave output. It ain't even modified sinewave. It's straight plain old square wave. And i believe it is monitored to some degree because there is some dead time between the two halves, which decreases slightly under load.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

Well, apparently whoever design this pile of rubbish, erm, marvel of engineering, didn't even figure out how to make a proper 50Hz oscillator. When i viewed it on the scope the squarewave would jump every now and then. I thought it was just the trigger circuit on my scope getting confused. It wasn't. The output is actually "hiccuping" - it isn't continuous. I plugged a CFL desklamp into the inverter and after it lights up the starter flashes it again every few seconds. Looks like this is going to need some significant rework. Or better, keep the case and heatsinks and build some new guts from scratch.

I have a big drawing to make and the only place that fits the paper size is the kitchen table. I need to run the laptop and a desklamp. There is only 1 free wall socket in the kitchen and i don't have any spare extension cords. So i thought i'd run the desklamp on the inverter. Well, that didn't end up so well... Luckily my laptop battery lasts 3 hours.

Lesson learned - trust your measuring equipment not your device under test.

[Colt45ws]

Re: Teardown and investigation of "LP" 150W sinewave inverter

I found this page a bit ago about reengineering a "3.5kw" inverter.
http://ludens.cl/Electron/chinverter/chinverter.html

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

I know it. That one is a fair bit more complex than what i have, and it actually had sinewave output. Goes to show just how much people designing these things actually know about them.

[Colt45ws]

Re: Teardown and investigation of "LP" 150W sinewave inverter

He makes a comment about the battery charger circuit, about it being for a power supply and not a battery charger. Makes me imagine they took various circuit designs, or parts of them, and put them together like puzzle pieces to make what they wanted.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

There's nothing wrong with reusing an existing design - if it fits your needs that is.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

Okay, i'm back at it. Be prepared to see some really lame electronics. The design looks very rudimentary, i would have been better off just buying an inverter case and cooking my own.

The 12v side is done with a pair of IRFZ44N. Average. They are driven directly off the 3525. There's room for an extra pair. I haven't bothered to check the topology but it's likely push-pull.

The output is indeed full bridge, using four IRF830s. There are only two TO92 transistors beside them, it is probably using the internal transistors of the 494 to form two totem poles. There are no SMD transistors on the bottom side, only resistors and diodes.

The soldering is really really lame. I found little solder blobs all over the place. Any one of them could have caused a short. At least the heatsinks are screwed on rather than soldered. Thermal contact of devices to heatsink is good. There is a temperature sensor fitted to the 12v heatsink. Unfortunately it wasn't glued into place so it could have easily slid out by itself. So much for protection.

I'll see what i can do about that hiccuping 55Hz squarewave output. I have a 300W passive PFC coil - that oughta make the output a tad more sine-ish. I'll be back with more measurements like operating frequency and such. I'll try taking some waveform shots too but my camera and my scope don't get along.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

Um, yeah. kaboom was right. The output freq is decided by a timing capacitor. A crapo lytic at that. A 1uF cap and a 1k5 SMD resistor. No trimpot to tweak the frequency no nothing. It's going to get a high quality polyester cap and a 2k multiturn trimpot, that way it can output 50Hz and STAY there. I also noticed that the TL494 locks up and outputs DC if i attempt to make a reading on pin 5 of the 494 (the timing cap). I still have to find the reason for the hiccups. The step-up switcher runs at 66.6kHz. Evil! Since it's a push-pull, the effective frequency on the transformer is 133.2kHz.

Edit: Center tapped primary gave it away. It's a push-pull. The secondary ie HV side is NOT center tapped however, because the center tap should be ground and in this case it's simply taken from the primary side because there is no need for isolation from the battery. Input and output grounds aren't separated, the earth wire of the socket is connected to the negative side of the battery. There's no inductor, only the 100uF 400v capacitor. That explains why there is no feedback from the DC bus - whoever designed this didn't feel competent enough to make a feedback loop that would prevent flux walking of the transformer and hence switching transistor destruction, so it just runs at a fixed duty cycle.

The HVDC bus has no feedback as far as i can see. The fullbridge (ie the output) however does, but it doesn't seem to be of much use. I'll try to wrap my head around as to what it does but it's likely simple peak sensing using a diode and feeding that to an error amp of the 494.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

Substituting a "proper" 1uF film cap made things worse. Along with the 494 going dumb when i attempted to check pin 5 (timing capacitor, ramp), something clicked in my mind. Larger cap... longer leads... more trouble, means it's sensitive to interference. Wait - since when is the TL494 spec'd down to 50Hz??? And the million dollar answer? IT'S NOT, IT NEVER WAS AND IT WILL NEVER BE. Using a cheap lytic in what should be a sensitive timing circuit doesn't help either.

To quote the Chilean guy with the "3500W" inverter - this isn't being cheap, it's being incompetent. It should say on the label - "We decided to make this an universal 50/60Hz inverter so you'll get 55Hz... if you're lucky. At this point it seems like a much better idea to crank up the frequency and make some sort of 50Hz oscillator to feed the 494 for true sine wave output.

[Krankshaft]

Re: Teardown and investigation of "LP" 150W sinewave inverter

At this point it seems like a much better idea to crank up the frequency and make some sort of 50Hz oscillator to feed the 494 for true sine wave output.

It seems at this point it's time to hang this thing up in the yard and use it for target practice.

The old saying holds true you can't polish a turd. The fact that they lied about it being a sine wave inverter is the peak of hilarity.

Hey why are my transformers buzzing? Why does my CRT have interference in the raster ?

Back in 03 I could get a 150 watt stepped sine inverter better designed for 25 bucks. The price must be in the toilet at that wattage (with modified sine) today.

Don't put too much cash in something not worth much. You can only go so far before you realize you have reached the limits of the circuit design.

You probably should just design your own instead of working with this garbage design.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

I never said i will put any money into it. But agreed, time is money.

Btw it looks like i was barking at the wrong tree with that oscillator. The 1uF cap was for the soft start. It has a 100nF film cap (hey at least it's a film) and a 122kOhm resistor. I think the high value resistor is the reason for the oscillation (1uF film caps aren't cheap). But the resistor on its own shouldn't cause this. The real issue is piss poor layout of the board.

If i put something powered by a 50Hz transformer on it it won't even run properly btw.

[Th3_uN1Qu3]

Re: Teardown and investigation of "LP" 150W sinewave inverter

I was using the inverter to test up my modded ATX supply. The series lightbulb was too much current limiting, so i figured out the inverter would still be better than connecting it straight to the mains. Plus, i can also scope the primary this way.

At a small load things were going fine, so i upped it. I was delivering 180W to the load, and the inverter went into protection mode. It started cutting off the output voltage, flickering the fault LED, and sounding the buzzer (well, no more sounding the buzzer 'coz i cut that off ). Normal behavior for this inverter. Then strangely, the fault led went solid on, but the inverter started delivering full output! I turned it off at that point.

When i turned it back on... i heard popcorn. Along with some smoke and that familiar smell. So much for overload protection i guess. We'll find out what popped in a few minutes.

Edit: Casualties of war: Two of the IRF830s in the full bridge failed short, and the two IRFZ44N on the 12v side cracked. That's where the popcorn came from. Why? Well... the push-pull converter runs open-loop. And as i just found, it has no protection of any kind either. Apparently the protection circuit only shuts down the TL494, if the full bridge fails short, it's bye bye 12v side parts. Lovely... You'd think they'd at least put in a fuse on the DC bus or something... Nope.

There's even a current sense resistor on the DC bus, but look, it's turning off the wrong supply. It's wired to the 494. They didn't even bother to make a latch circuit, as in "fault" mode the output just keeps switching off and on.