Voltage Divider VS Op-Amp Feedback in Push Pull Converters

Thread Starter

Elhassan

Joined Jul 17, 2022
7
Hi, in High Frequency Inverters, Voltage is First Boosted to 311 VDC using push pull converters, then it converter to AC using SPWM and couple MOSFETs or IGBTs.
to maintain the 311VDC, the PWM IC such as SG3525 is getting feedback from the rectified voltage which is the 311VDC.
The point is.. that I found two ways of getting the feedback, the first one is done by just a voltage divider circuit [just couple resistors to reduce the 311VDC to around 4VDC], which is controlling the Duty Cycle of the PWM Circuit Which is Controlling the push pull MOSFETs, this method is found mainly on most of the circuits that I found online.
but when I was checking the Commercial Design of Inverters, I found that they were using op-amp as a comparator [including the voltage divider for the comparator circuit] and isolation optocoupler and the output signal is then goes to the PWM IC, which is just turning the PWM ON with 100% duty Cycle, or OFF, as Shown Here:
1666456269930.png

When I tried the First Method, Which is controlling the Duty Cycle, I Found that the MOSFET's Temperature was Raising Dramatically, Without Loading the Circuit, Actually there is Many Probabilities that may Cause that, but my scope here is just related to the Feedback only...
so, the question is:
1- What is the Difference between the Two Methods? does it just for Isolation?
2- Does the Feedback Method do something with the Temperature/Efficiency of the Circuit?
I mean, Maybe when the MOSFETs is cooling when they are turned OFF [on second method]? while the first Method is just reducing the duty cycle and not turning off the PWM completely, which is doing something with the crossover losses I guess?
some Extra Information:
- Switching Frequency = ~33Khz
- Dead time is Taken in Consideration with the Used MOSFET and Checked with the Oscilloscope
- Soft Start is Taken in Consideration
- The Whole Circuit was Build on the Breadboard, I know that there is may be noise etc..., But I just wanted to check the Feedback Method Before Making the Prototype on the Vero Board [Then I will make the Prototype PCB, but later]
- This is very close to the actual Push Pull Converter Circuit, the only difference that I removed the Feedback here
1666456055777.png

Very Thankful for Reading my Questions. Waiting for your Answer
 

ronsimpson

Joined Oct 7, 2019
2,988
I do not understand how the circuit works. I see the MOSFETS chopping up the 12V, transformer, diodes and cap.
It looks to me that if the 12V drops 10% then the 300V will also drop 10%.
It looks like changing the duty cycle has little effect on the 300V.
Do I understand right?
 

Ian0

Joined Aug 7, 2020
9,668
1- What is the Difference between the Two Methods? does it just for Isolation?
2- Does the Feedback Method do something with the Temperature/Efficiency of the Circuit?
1. Yes. For compliance with safety standards in most countries, the battery terminals may not connect directly to "mains" terminals. Neutral must be connected to earth at source, and earth must be connected to any exposed metalwork. If any terminal of the output is connected to the battery voltage, then the battery voltage would be at about 110V AC, which would be dangerous to touch. In vehicles, one battery terminal is generally connected to the chassis, which would put a direct short on the output of the inverter.
2. No. However, with all the delays due to op-amps and (slow) opto-isolators, without careful design it may go unstable. It is much more likely to be stable with a simple voltage divider.
 

Papabravo

Joined Feb 24, 2006
21,159
There are at least two main reasons for MOSFETT heating:
  1. Sub optimal gate drive circuitry.
  2. Losses due to insufficient dead-time between the phases.
 

Ian0

Joined Aug 7, 2020
9,668
There are at least two main reasons for MOSFETT heating:
  1. Sub optimal gate drive circuitry.
  2. Losses due to insufficient dead-time between the phases.
I'll add to that: "capacitive load". A lot of people of don't know what they are doing with high-frequency circuits forget the output inductor and connect the bridge rectifier straight to the output smoothing capacitor.
 

Papabravo

Joined Feb 24, 2006
21,159
I'll add to that: "capacitive load". A lot of people of don't know what they are doing with high-frequency circuits forget the output inductor and connect the bridge rectifier straight to the output smoothing capacitor.
Design is different from synthesizing from existing schematics. You have to go back and sus out what the original designer was thinking without being able to ask her.
 

Ian0

Joined Aug 7, 2020
9,668
And another two:
1) Avalanching MOSFETs due to badly designed transformers with high-voltage spikes.
2) When choosing the Vds of the MOSFET, forgetting that the voltage on the MOSFET that is OFF is TWICE the supply voltage.
 

ronsimpson

Joined Oct 7, 2019
2,988
in High Frequency Inverters
I am running much faster than that.

Please do some math for me.
Remove the secondary. Pretend you are only driving the primaries.
33uH inductor, 12V and on-time of ? (33uS) = what current?
You have that much current before you start out. (no load current in the FETs)
 
I'll add to that: "capacitive load". A lot of people of don't know what they are doing with high-frequency circuits forget the output inductor and connect the bridge rectifier straight to the output smoothing capacitor.
Makes sense but I have repaired many Chinese inverters and they don't have a buck inductor in the up-converter. The topology is exactly as OP has shown. Push-pull transformer with bridge rectifier on the secondary. Typically 40 KHz or so.

My guess is that the heating problem is inadequate gate drive. The SG3525 only puts out a couple hundred mA.

One reason they might use an op-amp is to support loop compensation. Do you see small capacitors in the signal path? Check out TI app note SLVA662
 
And another two:
1) Avalanching MOSFETs due to badly designed transformers with high-voltage spikes.
2) When choosing the Vds of the MOSFET, forgetting that the voltage on the MOSFET that is OFF is TWICE the supply voltage.
Inductive spikes from circuit layout problems. Show us a photo of the proto-board version.
 

Ian0

Joined Aug 7, 2020
9,668
Makes sense but I have repaired many Chinese inverters and they don't have a buck inductor in the up-converter.
And many of them don't last very long.
It is possible to design out the inductor by designing leakage inductance into the transformer.
If it is just a Chinese copy, then they may not be aware of that.
 

Thread Starter

Elhassan

Joined Jul 17, 2022
7
Thanks for Answering,
It looks like changing the duty cycle has little effect on the 300V.
Do I understand right?
Sure, the Main Idea is to turn the duty cycle on/off to maintain the Selected Voltage

And another two:
1) Avalanching MOSFETs due to badly designed transformers with high-voltage spikes.
2) When choosing the Vds of the MOSFET, forgetting that the voltage on the MOSFET that is OFF is TWICE the supply voltage.
Exactly!, I soldered the whole board into Vero-Board/Testing PCB Board, to Eliminate any noise, the Problem Still There,
The MOSFET is Burning Really Quickly!!!! that was insane!
but then I tried to connect just one MOSFET to see on the scope what will happen, and Just as you said, the main problem was with the Voltage Spikes when the MOSFET is Turning OFF (Hope this to be the problem)..
1667177948944.png
Ch1: Probe Connected to Gain Pin, as Shown the Spikes reaches 80VDC in about 4 us
Ch2: probe Connected to the Gate, to See the Switching Timing

the Problem that I just Copied the Exact Schematic as the Chinese Board, No Snubber or anything is Connected to the Chopper, and it Works Fine, and my Design is Heating up Very Quickly !!
so Any Ideas on How to Dissipate that amount of energy? and Preventing it from Building up on the MOSFET when it Turns off? [the Dead Time Duration], I Guess Something like the Freewheel? but when I Searched online, no one was connecting freewheel.. so I was Wondering..
here are some Specs of the Chopper:
Primary Coil [+12V], L: 4.79 mH, R: 268.2 milli Ohm, C: ~56 nf
Secondary Coil [+300V], L: 7.91 mH, R: 1.885 milli Ohm, C: ~35 nf
*measured with LC Meter, and Low Ohm Meter
 

ronsimpson

Joined Oct 7, 2019
2,988
Sure, the Main Idea is to turn the duty cycle on/off to maintain the Selected Voltage
Please explain how the power supply works.
When the MOSFET is on 12V is across the winding. When the MOSFET is off ......
The output diodes peak detect. The output voltage will remain constant over a large duty cycle range. 12V x turns ratio = output voltage.
Did you copy this circuit from somewhere? Where?

Your scope picture looks like the secondaries are not connected. or The leakage inductance is bad. The primary and secondary are not well coupled.
 
Last edited:

Thread Starter

Elhassan

Joined Jul 17, 2022
7
Please explain how the power supply works.
When the MOSFET is on 12V is across the winding. When the MOSFET is off ......
The output diodes peak detect. The output voltage will remain constant over a large duty cycle range. 12V x turns ratio = output voltage.
Did you copy this circuit from somewhere? Where?

Your scope picture looks like the secondaries are not connected. or The leakage inductance is bad. The primary and secondary are not well coupled.
Thanks,
Yes, I wasn't connecting the secondary coil, does it matter?
Do i need to connect them to try?
what is the difference?
 

Ian0

Joined Aug 7, 2020
9,668
Please explain how the power supply works.
When the MOSFET is on 12V is across the winding. When the MOSFET is off ......
The output diodes peak detect. The output voltage will remain constant over a large duty cycle range. 12V x turns ratio = output voltage.
Did you copy this circuit from somewhere? Where?

Your scope picture looks like the secondaries are not connected. or The leakage inductance is bad. The primary and secondary are not well coupled.
It wouldn't need the secondaries to be connected. It would just need good coupling between the two primaries (and the primaries being wired in the correct phase).
Show the voltage on both drains.
 
Top