Suoer 2000W Pure Sine Wave inverter teardown

In this page, we will have a look at:
FPC-2000A 2000W pure sine wave inverter
FPC-3000A 3000W pure sine wave inverter
HAA-1000A 1000W modified sine wave inverter

       Up for teardown is a loaner unit from my supplier. It is a Suoer model FPC-2000A 2000W Pure sine wave inverter.

       This unit is a 230V 50Hz unit and I was tasked to modify it for 60Hz operation. I found this a perfect opportunity to do a teardown (not extreme though) to see what parts they used and how well it is built.

       Enough of the talking. Here are the pictures!

As received. Box contains the inverter, manuals, warranty card.. The usual stuff.

Here is where it begins.

The inverter is quite hefty.

Power input connectors and fans.

The power connectors look decent enough to handle the required input current.

The AC output receptacles look ok. Not the best but not bad either.

They do have closing covers to prevent accidental contact of the live terminals.

Nice touch on the sides.

Aluminum inserts add some contrast to the blue case.

Top cover removed.

Nice clamshell case design.

This inverter bites. I did not notice where I got cut.

Sub board mounted to the output side faceplate.

The sub board holds the main power switch, USB jack and 5V DC-DC powersupply.

The 5V supply is based on an XL4001 converter chip.

It's nice that they used connectors on all wires leading to the front panel.

Makes removing it easy.

The output filter components are all populated.

Not much cost cutting going on here.

The main filter inductor (wrapped in shrink tubing) is a little wobbly though. It could use some glue to hold it down.

The main DC-DC converter uses the common SG3525 based driver board as seen in other PSW inverters.

And also the same SPWM driver board that uses the EG8010 driver chipset.

Here is one of the internal heatsinks that hold the output rectifiers and output H-Bridge IGBTs.

A thermistor for over temperature protection.

They could have done a little better on making the thermistor get good thermal contact on the heatsink rather than just slopping some grease on it.

Output rectifiers are RHRP3060

Hyperfast diodes rated at 30A 600V.

Output H-bridge IGBTs are FGW75N60HD

Rated at 75A 600V.

The devices appear to be decently rated for the intended power rating.

Now that's interesting.

The hole for the cap is indeed near the heatsink to make the layout easier.

The cap lead is insulated with heatshrink tubing so I guess it should be fine.

Clamps for the main input switchers are removed to view the part numbers.

There are a total of twelve switches. Four per power transformer.

The one on the extreme left is a TO-220 packaged thermal switch that turns the fan on when the heatsink temp is over 45degC.

Main switches are HY3506



No, it wasn't me. I have no idea why the optocouplers are slanted like they were stepped or pressed onto really hard.

The solder joints looked like they were reworked.

Bottom side is not the cleanest but not bad.

Generous amounts of solder is neatly applied to the high current tracks.

PCB part number.

These jumpers carry high voltage DC and I guess they thought that the solder resist and silkscreen is not enough so they put... Masking tape!

Come on! Use the right type, preferably Kapton tape!

This I didn't see in the other inverters. Not really sure what the blue trimmer is for (have not checked the datasheet) as it is usually unpopulated. The orange trimmer is for trimming the output voltage.

Fuses are soldered directly into the PCB.

And they are arranged neatly.

Output bulk caps are 470uF 450V each Rubycon branded.

Not sure if genuine but they appear to be.

DC input caps are no name "TJ" branded 105degC rated caps.

The QC department missed this.

Solder splatter behind the front panel.

Removed it after taking the pic.

Here is the idle current draw of the inverter. 1.7Amps

       I could not do some power testing as I don't have enough juice in the battery bank let alone enough to run this full tilt at 2000W continuous but judging from the parts used, I think it can meet the specifications.

And of course, the real reason for this being in my hands!

Original configuration: 50Hz output frequency

After modding. 60Hz operation!

Hint: It is in the chipset datasheet *wink wink* but really fiddly to apply in the circuit.

       10 Feb 2015:

       I now have a 500W PV array and it is sunny so I can do the power test. As usual, I used our microwave oven measured at over 1200W and over 1400VA.

Here is the inverter opened up ready for testing.

With no load, the output is pretty clean.

With the microwave oven load, the output becomes a bit noisy but it does not show any clipping or excessive distortion.

The main switching mosfets are warm, heatsink is somewhat hot to touch but the fans did not turn on during the test.

AC side heatsink is also hot to touch but in here, we can see two IGBTs are hotter than the rest.

These are the ones that are modulated with the half sine.

These three hot spots are the primary snubber networks on each of the three power transformers.

This hot spot is the regulator for the SPWM board.

       The test appears to be a success. The frequency modification does not appear to have any adverse effect on the power stages.

Back to top of page

       13 Mar 2015:

       So there is a bunch of inverters now lining up for modification. There are the same 2000W PSW inverters, some 3000W PSW and one 1000W Modified sine for experimentation purpose as I am also going to modify it for 60Hz output.

The 3000W version is basically the same as the 2000W version but with more mosfets, transformers and IGBTs.

I guess they did not take into account of the heatsink getting in the way.

Output IGBTs. They used eight Fairchild SGH40N60 IGBTs.

Devices are rated at 20A 600V.

These are the same ones used in the TBE inverters.

Output rectifiers are MUR3060PT.

These are rated at 30A 600V ultrafast rectifiers.

Wires wouldn't fit in the PCB hole so....

Thermal switch used by the fan.

There are five transformers with four mosfet switches each so there are twenty HY3506


Sub board mounted to the output side faceplate.

The sub board holds the main power switch, USB jack and 5V DC-DC powersupply.

The 5V supply is based on an XL4001 converter chip.

Output EMI filters are also completely populated.

bench setup.

Before modification.

After modifcation.

Back to top of page

Now for the 1000W modified sine model.

It's pink!!!

This uses the typical DC-DC converter and H bridge design.

PCB is a one sided layout and through hole parts in FR1 material.

Output IGBTs

Output modulator board.

Uses the KA7500 chip.

Powersupply controller board

ICs are KA7500 and LM324.

Front panel sub board.

This holds the USB 5V supply and a uC to control the digital display.

H bridge mosfets are 26A 500V devices.

main supply switchers are eight HY3506


Not the cleanest but I've seen worse. Despite the flux mess, solder is laid out neatly.

Idle current with no load.

With 10.5V input, the display shows this.

Output waveform at 10.5V input.

This was measured through a 12V output transformer, notice how the peak and RMS voltages change.

At 12.5V input.

Output waveform at 12.5V input.

At 13.2V input.

Output waveform at 13.2V input.

At 15V input.

Output waveform at 15V input.

       Parts are adequately rated. Voltmeter accuracy is low by 0.2V. Input and output connectors are decent and can handle the currents so I can say this can do its 1000W nameplate rating. Power testing to follow when time permits.


There is a 48V 2000W inverter in the mix and this are what I found to be different.

A small sub board is used to step te supply voltage down to power the control boards.

It uses a TOP104 switch/control IC.

Fuses are now 10A vs 40A in the 12V version.

Input caps are 63V.

Input protection diodes are 6A 1000V devices.

       I was not able to see what mosfets are used since I have to take the board out to see and I did not have the time but it should definitely be a different part that can handle around 100V or so.

Page created and copyright R.Quan © 07 Feb 2015.