[SOLVED] Full Bridge Transformer primary voltage is constant and not varying with gate signal

Hi. I am trying to build a full bridge SMPS DC-DC converter. Attached are my primary gate drive waveform (Blue waveform)
and Primary output Voltage waveform (yellow waveform).



Problem I am facing is that even when the gate voltage is zero, MOSFET is turned on and it turns off only when a -15V is applied across its terminal. I guess that some noise is trigger the MOSFET to be on even when the gate voltage is zero and only when a negative bias is applied we are able to turn off the MOSFET. It would be great if someone can point in a certain on what seems to be going wrong.

Check the source voltage at the moment you turn off the device. Is it possible it goes negative as a result of switching action, etc.? Then it requires for the bias to be negative.

This is in regard to an N-device and its source (more negative) terminal.

Yes I am using a NMOS. Like you said it is going negative because of switching action. Can you suggest me some circuits which can introduce negative bias to the gate drive signal.

Thanks

can you show your gate drive circuit?...is there a low impedance path with which to discharge the fet GS capacitance?

I have attached my h bridge circuit. Input to the Gate pulse transformer is fed from a MOSFET driver in series with Damping resistor of value 10 ohm and 10uF capacitor.


I hoped that the Resistor across gate and source can discharge the capacitance but it is not happening. I checked at the source point of the high side MOSFET Q1 which turns off only when its source is grounded. This happens exactly when the Q2 turns on. That's why I am able to get alternating pulse of 50% duty cycle at the output.

Looking at your scope trace...

During On time, voltage follows a slight downward slope similar to the gate trace. Then as On time ends, voltage jumps higher. Then it stays high and doesn't vary much.

Your schematic has no load. Do you take readings with or without a load? I wonder if the readings change when you put a resistive load across the H-bridge? Or a load to ground? Etc.?

Load as of now is RC snubber circuit. Also I did connect the output Isolation transformer to see if there is any change and checked the secondary of the isolation transformer, it still is the same output as posted before. I would also check by connecting the resistive load to ground and see if makes any change. But if I connect a resistive load to the ground what do I do with the other output pin.

Put some load across the H bridge....!

I hoped that the Resistor across gate and source can discharge the capacitance but it is not happening. I checked at the source point of the high side MOSFET Q1 which turns off only when its source is grounded.

Click to expand...

I think you are misinterpreting the waveforms in post #1. The MOSFET does turn off when Vgs of 0 is applied. But the bridge output voltage is keeping it's previous value by the working of drain-source capacitances. That's normal operation of an unloaded (or light loaded) push-pull switch.

You can e.g. connect a resistive load across both bridge outputs to change the situation.

Here are some full bridge simulations in the free ltspice. You can probe them to try and isolate your problem before going to the lab...as its quicker to probe a simulation than a circuit in hardware.
Sometimes, in fact, IMHO, always, it is best to get a "version" of the smps working on the simulator before actually going into the lab to work on it at all.
IMHO, if you cant get a basic simulation of an smps working on the simulator (ltspice), then its a total and utter waste of time starting on the hardware.
What doesn't work on the simulator, will never ever work in the lab.

Don't get me wrong though, the number of "drongo" engineering managers who wont allow an engineer to even use a simulator is amazing.....they hoik them into the lab without allowing them even 10 minutes on a simulator...totally wrong.

Of course, another golden rule of the simulator, is that something that works on the simulator, may very well not work in the lab…..but in spite of this, you must always work on the simulator before getting your hands on the hardware.

If an SMPS designer tells you that he/she never ever uses a simulator…then he/she is one of the following….
1…A liar, or...
2…Not an SMPS designer, or...
3…He/She is from a distant planet, and has intelligence way above human levels

Thanks a lot guys. It did give me a great insights into the working of full bridge converters. I will now head to the simulators to have a better understanding of full bridge converters.

sorry the simulation was quickly hacked together, you may like to change it to a more suitable turns ratio etc, but at least it shows the general working

- - - Updated - - -

Oh and by the way, now that you have "seen the light" and entered the world of SMPS simulation, be prepared for a massive amount of abuse from experienced engineers who will accuse you of "Playing" on the simulator, rather than getting into the real world of the lab.......no doubt, most of this is from seniors who DO NOT want you to use the simulator, because they fear you will learn much and be a competitor for their job.!
I now formally apologise to yourself profusely for all the abuse that you will now receive...Stand By!

I am hobbyist working my way to make a lab based power supply on my own. I am now learning by doing. I am in a long way from making it as my career.

I tried to simulate them, I am getting an error of Multiple instance of flag error in LTspice. It would be great if you can share any schematics files which I can open in simulator, also helps in visualizing and understand the circuit better.

ill look into it, those sims work ok on my ltspice simulator on my pc.
I wonder if some of the components I have used are now obsolete on the latest download of ltspice.

Hi Treez. Can you please let me know the version of LTspice that you are using. I will also download the version of LTspice that you are using and run the simulation files which you have provided.

I am using version 4.23e
(I hope this doesn't make a difference though)

Hi. I struggled to get the simulator to work for me. But finally in vein, I started to test your suggestions with my real world prototype. I did connect 100k load across the outputs of the H bridge circuit hoping to see if there is any change. I am not able to see any visible change in the output. I thought 100k is not a light load and I hope you guys also think the same ?

here is a 600w full bridge led driver in ltspice.

maybe you have to go into hammer icon and change the solver to "alternate 2" or similar

here is full bridge 600w with voltage output regulation. in ltspice
Change the load as you please

100k is hardly a load, try at least 10W