[SOLVED] Square Output instead of Sine Wave in High-Frequency Inverter

[Elhassa]

Hello, I am designing a High-Frequency Inverter, 12VDC is Boosted up to ~320 VDC using SG3525 IC on ~35Khz Switching Frequency.
and I am generating SPWM to X2 IR2110 MOSFET driver IC for the H-Bridge Mosfets [IRF840].
H-Bridge is connected to 320VDC, the SPWM signals going to the MOSFET's gate were checked using an oscilloscope, and everything is fine.
the problem is the output wave is square instead of a sine wave?!
but when I connected the glue gun [20W Inductive Load] on the output for testing purposes, the waveform turned into a sine wave [and the voltage dropped immediately, but that's not the main point now]
so any idea on what's the problem? the input signal to the MOSFET's gate is fine [Vgs(ths) = +15VDC] for the High and low side, but the MOSFETs doesn't seem to be switching properly while there is no load, or I missed something or can't measure the actual output..
any help?
*note: this picture is just for demonstration, I Already Built the full circuit on breadboard for testing purposes and tuning
*another note: in the actual circuit, VS of the first Driver is Connected correctly to the Source of Q1 Mosfet and Drain of Q2 Mosfet (before the coil)

 

[KlausST]

Hi,

the "output" is across C1. You can´t have a square wave across a correctly working 2uF capacitor.
So either you did something wrong or the capacitor is damaged.

For better assistance you could upload a scope picture of the waveform. 100kBytes should be sufficient. Mind to clearly show V/div and time/div.

BTW: without (resistive) load the output is expected to wildly ring. So use a suitable load for your measurements.

Klaus

 

[Elhassa]

You can´t have a square wave across a correctly working 2uF capacitor.

Actually, the Capacitor is 3.3UF [335k 400V]

upload a scope picture of the waveform

That's the Output Voltage Across Output Capacitor, I'm working with 12VDC instead of 320VDC to Avoid being chocked [it Really Hurt ]

the Same Output without the capacitor, I Wonder why there is no PWM Shown? it seems that the MOSFET doesn't have enough time to discharge or any pass to the ground I guess? because I wasn't connected any load, but I see in many videos people were just scoping the sine output without any load and it works fine

as shown here, the reference probe is connected to the ground, and I was Scoping the High Side of the MOSFET, Ch1 is the Source Pin, and Ch2 is the Gate
it seems that the SPWM signal coming from IR2110 is Fine, but the MOSFET on channel 2 is not following the SPWM signal

another example of the Gate Signal, the probe Reference is connected to the Source, so that as shown the Vgs(threshold) is more than the rated one for the IRF840 (which is only +4VDC)
I am really confused about how to measure the output AC Voltage? what is missing? everything is supposed to be working fine (i guess)
do I need to pull the output to the ground? but I didn't see anyone doing this.
the main problem is that when the gate signal is turning off, the MOSFET doesn't (or it does but I can't see it on the oscilloscope)

 

[KlausST]

Hi,

There's definitely something wrong.
Show a photo of the circuit, including wiring and the scope connections.

The shown waveform does not fit to the schematic / description.

Klaus

 

[FvM]

You claim to drive the H-bridge with SPWM but you don't explain the modulation scheme. It might be inappropriate.

 

[Easy peasy]

change R5,6,7,8 to 47 ohm, and put 1uF across the driver chips, right on the Vcc / gnd pins

--- Updated Jul 14, 2022 ---

add to C2, C3 1uF right on the pins of the chip, make sure all of the diodes are ultra fast types,

If any of the high side drivers need to be on for say 10mS, then you need to port isolated power to that driver - 10uF will not cut it, if you look on these caps - WITH AN ISOLATED SCOPE INPUT - you may see the voltage falling on this cap due to the long on time and then the UVLO will kick in, stopping that fet.

 

[Elhassa]

There's definitely something wrong.
Show a photo of the circuit, including wiring and the scope connections.

I guess so too, here are some pictures to check if I did something wrong:

*Mosfet Power is Connected to 12VDC For Testing Purposes, it should be connected to 320VDC, I tried that and same results but with higher voltage

another Picture with Another Angle

my Testing Breadboard, the EGS Section is not used now since I am testing with the 12VDC, not 320VDC

You claim to drive the H-bridge with SPWM but you don't explain the modulation scheme. It might be inappropriate.

Good point, I scoped the Output from Arduino

Ch1,Ch2 are for Two High-Side MOSFETs

Ch1, Ch2 are for the Low-Side MOSFETs, with a Dead time of ~60 uSec
as Shown Here:

change R5,6,7,8 to 47 ohm, and put 1uF across the driver chips, right on the Vcc / gnd pins
add to C2, C3 1uF right on the pins of the chip, make sure all of the diodes are ultra fast types,

I tried, but without any different Output, the only difference that I noticed when I Changed the C2, and C3 to 1uF, is that the Capacitor Discharged so Fast that it couldn't generate Vgs Threshold Difference to the Gate of the Capacitor
all Diodes are Ultra Fast Types, D1, and D2 are FR107 SMD [I soldered some Legs to connect them to the Breadboard]
--------------
I will try to Connect my signal Directly to an old EGS002 Board [to the SMD Driver] in order to make sure that the problem is not from the driver itself, since EGS002 is a Ready Commercial Controller

 

[KlausST]

Hi,

a breadboard is not suitable for switching power applications.

Klaus

 

[Elhassa]

Hi,

a breadboard is not suitable for switching power applications.

Klaus

That's Right, but I just wanted to test my controller, do you think that the cause of the problem is the breadboard?
but all high-frequency circuit was just working fine on the breadboard

 

[KlausST]

Hi,

the breadboard is not suitable. But there still is another issue (or more).

Picture 1 of post#3 simply does not fit to a 2mH + 3.3uF series circuit. The rise rate shown is too fast and I expect the signal to ring a lot.

Don´t know what happens here.

Klaus

 

[Easy peasy]

how long do the high side devices need to be on for ? maximum ? this is important, as bootstrap psu will only go for so long before the bootstrap cap discharges ....

 

[Elhassa]

The problem was solved, I found that the problem was from the SPWM Signal from Arduino.
thank you all for helping, really appreciated