Me and my partner are making an inverter (12VDC to 220VAC) which oscillates at 10Hz<f<20Hz. However, the device can not be used for a long time. The device can only run for 1 minute. The problem is the mosfets I use are heating-up quickly (have been attached to heatsink). But it goes normal when the inverter is used at normal frequencies (50/60 Hz). Can anyone help me to solve this problem?

Here I post the mosfet circuit I use.

Thank you.

 

Capacitive coupling is not a good idea here.

At low frequencies the capacitors are charging up, reducing the gate voltage to where they start to come out of conduction.
Re-design it for DC gate drive.
Since the MOSFETs source is grounded, just drive the gates directly from the flip-flop.

 

Try to use a 100Ω/0.5W resistor to replace C1, R1, D2.

 

Tell us about the transformer. Low frequency transformers are usually lossy at high frequencies.

Edit: The time constant no the drive coupling is only 4.7 milliseconds!

 

Hello,

The mosfets are driven by a capacitor. That will give problems with the switching of the mosfets.
Most circuits are direct driving the mosfets via a resistor.
Here is a circuit for a modified sine wave inverter, wich also pervents that the mosfets are on at the same time:



The description for the circuit can be found here:
http://krelllab.blogspot.com/2014/11/homemade-modified-sine-wave-power.html

Bertus

 

Please clarify: Your title says that the MOSFETs get hot at low frequencies and your post says that the MOSFETs are fine at 50/60 Hz but they get hot at high frequencies. Which is it?

 

Hello,

Why do you want to run the inverter at 10Hz<f<20Hz?
The transformer will very likely giving the problems at that low frequency.

Bertus

 

Capacitive coupling is not a good idea here.

At low frequencies the capacitors are charging up, reducing the gate voltage to where they start to come out of conduction.
Re-design it for DC gate drive.
Since the MOSFETs source is grounded, just drive the gates directly from the flip-flop.

Thank you for your advice. I will discuss it with my partner

Try to use a 100Ω/0.5W resistor to replace C1, R1, D2.

Thank you for your advice. But, can you explain why these components should be replaced with 100Ω resistors? I'm sorry if I ask too much

Tell us about the transformer. Low frequency transformers are usually lossy at high frequencies.

Edit: The time constant no the drive coupling is only 4.7 milliseconds!

I still use a regular transformer (Center tap).

Hello,

The mosfets are driven by a capacitor. That will give problems with the switching of the mosfets.
Most circuits are direct driving the mosfets via a resistor.
Here is a circuit for a modified sine wave inverter, wich also pervents that the mosfets are on at the same time:

View attachment 156074

The description for the circuit can be found here:
http://krelllab.blogspot.com/2014/11/homemade-modified-sine-wave-power.html

Bertus

Thank you for your advice and for schematic. I will discuss it with my partner

Please clarify: Your title says that the MOSFETs get hot at low frequencies and your post says that the MOSFETs are fine at 50/60 Hz but they get hot at high frequencies. Which is it?

I'm sorry about that, maybe I'm a bit wrong in writing the title. What I want to say is that the mosfets experience an abnormal temperature rise when they are switching below the normal electrical frequency.

Hello,

Why do you want to run the inverter at 10Hz<f<20Hz?
The transformer will very likely giving the problems at that low frequency.

Bertus

I want to run a device that can only accept frequencies in that range

 

To summarize, I think two major problems have been identified. First, the gate drive circuitry does not provide sufficient low frequency response. You can use much larger capacitors, or just direct DC coupled from the CD4047. Your output transformer is likely designed fro 50/60 Hz. As such, the transformer will quickly saturate when trying to pass a 10 - 20 Hz signal. You will need a transformer designed for these low frequencies (probably a custom designed unit) and it will have **much** more iron.

 

Tha

To summarize, I think two major problems have been identified. First, the gate drive circuitry does not provide sufficient low frequency response. You can use much larger capacitors, or just direct DC coupled from the CD4047. Your output transformer is likely designed fro 50/60 Hz. As such, the transformer will quickly saturate when trying to pass a 10 - 20 Hz signal. You will need a transformer designed for these low frequencies (probably a custom designed unit) and it will have **much** more iron.

Thank you for your advice. Do you mean I should choose larger capacitor for C1 and C2 or not using those capacitors at all? Sorry if I ask too much

 

Thank you for your advice. But, can you explain why these components should be replaced with 100Ω resistors? I'm sorry if I ask too much

I'm afraid of the big problem is that the frequency response of Transformer, when the working frequency is too low and exceed its range of working frequency then the current transformer will be become more like resistance and it doesn't like the coil, and the second is the driving waveform, when you used the CR circuit that it is a Differential circuit then the driving waveform will be as straight up and slower discharge and straight down and slower charge, it will be affecting the transfer efficiency, the most of driving waveform are using the sine wave or square wave or approaching both, on the other hand is that the Rds(400mΩ) of MOSFET is too high, you can choose some others that the Rds less than 20mΩ or more less.

The transfer efficiency of traditional transformer about 60~80%, and that is working at the 50/60Hz, if the frequency is more lower then the transfer efficiency will be more less, today, the working frequency of switching power supply already exceed 200Khz and the transfer efficiency is exceed 98%, and it is already proved that the frequency will be affecting the transfer efficiency then why you want to use the lower frequency?

 

Me and my partner are making an inverter (12VDC to 220VAC) which oscillates at 10Hz<f<20Hz. However, the device can not be used for a long time. The device can only run for 1 minute. The problem is the mosfets I use are heating-up quickly (have been attached to heatsink). But it goes normal when the inverter is used at normal frequencies (50/60 Hz). Can anyone help me to solve this problem?

Here I post the mosfet circuit I use.

Thank you.

View attachment 156065

Use an IRF540 mosfet, and change the frequency to 50Hz.

 

Capacitively coupled drive is sometimes used where there is a possibility that something will go wrong and the load will have power applied for an excessive time, causing damage. When the load is a transformer, particularly at high current, it becomes difficult to be sure the FET remains fully turned on for the required time but would turn off before the transformer core entered saturation (that is, you reach a point where the magnetic flux will no longer increase with increasing magnetizing force). Transformers just aren't designed with the extra "margin" to allow them to have unipolar voltage applied for more than a small amount of time longer than normal. You can't achieve keeping the gate voltage high enough for a normal cycle while blocking it quickly enough if something goes wrong. Directly driving the FET through a resistor will do the job just as well. Since there is no need to protect the gate against negative gate to source voltage, the diodes from the gates to common are not necessary (actually, the gate of a FET can normally be driven to the same magnitude of voltage below the source as it can above e.g. if the gate can withstand plus 20 volts relative to the source it can withstand minus 20 volts relative to the source).

By far the best way to protect everything is to use a proper control IC that provides good gate drive and supports cycle-by-cycle current limiting. This type of current limiting does add some complexity but it will protect both the transformer and the FETs not only against input-side problems but against overload on the output of the transformer. There are lots of ICs available for this purpose.

About the only way you will be able to use a standard transformer at a frequency much lower than it is designed for is to use a transformer rated for higher voltage. For example, if you use one designed for 480 V to 24 V at 50 Hz, you could use it at 240/12 V at 25 Hz - it is the "volt-second product" that determines the magnetizing force. But you still need the same current rating, so the transformer will be much larger and more expensive.