Boost converter and H-Bridge

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
Hi all,

I a currently reaching the deadline for my Senior project and I am trying to get this to work. I am working on a DC/AC converter (Pure sine wave) and I am currently attempting to couple a boost converter with the H bridge. The DC/DC converter has a PI feedback control with a microcontroller (PSoC 5LP) performing both PI control and SPWM generation.

My problem is: Whenever I couple both the boost converter and the H-Bridge, the output of the boost converter drops by a significant amount even though the controller compensates for load changes.

Below is a a picture of the LtSPICE schematic of both the boost converter and the inverter. Note I did not include the PI controller for this part but below the schematic, there is a picture of the simulink model for the converter control. For the feedback system, I used a voltage divider network.


Any recommendations or thoughts as of what could be causing this?
upload_2019-4-13_14-32-57.png
upload_2019-4-13_14-35-35.png
 

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
Your PSOC PWMs configed with adequate dead time set ?

Did you try this with no load on H Bridge ?


Regards, Dana.
The output was tested with not load, just the filter.

Regarding the dead time set, what exactly do you mean?
Here is a picture of the scheme I am using to generate the SPWM:
WaveDAC8_1 generates the 60 Hz reference sine wave, WaveDAC8_2 Generates a 10 kHz triangle wave. both are being compared and the output is inverted by the not gate and then fed into digital output pins 2[1] and 2[0]. In order to avoid the same sides of the H-bridge to be on at the same time, I generated a 50% duty cycle square wave to which SPWM waveform will be ON. Whenever the non-inverterted square wave is on, the no inverted SPWM waveform will be on, and viceversa. This way, both the in-phase square wave and SPWM waveform will feed opposite gates of the H-bridge generating an alternating current at the load.


upload_2019-4-13_18-58-20.png

Even w hen using this method, the boost converter seems to act weird.

P.S The boost converter works perfectly when powering everything else but the H-bridge.
 

danadak

Joined Mar 10, 2018
4,057
OK, I understand the Sine approach.

If you have a scope look at overlap if there is any on a side,
leg, has both hi and lo side fets on, needless to say
thats a short of power to ground. Dead time is a concept
where one MOSFET in a leg turning on and off, there is
always a delay before off and on of the other fet to eliminate
this problem.


http://data:image/png;base64,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

Regards, Dana.
 
Last edited:

MisterBill2

Joined Jan 23, 2018
18,169
Of course the output of the boost converter will drop a bit when it goes from a no-load condition to a loaded condition. How much is too much is a separate consideration. Probably there is some portion of the physical construction of the boost converter that has more resistance or inductive reactance than the design. You can prove or disprove any interaction by substituting a resistive load that draws the same current as the H-bridge portion and see if the voltage drop is the same. And then you can also try adding an additional fairly large capacitance to the DC connection between the boost converter and the H-bridge. Of course, if all of the systems are only simulations at this point, getting actual results of real world operation will be a challenge. But adding the large capacitance will be simpler and cheaper.
 

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
Of course the output of the boost converter will drop a bit when it goes from a no-load condition to a loaded condition. How much is too much is a separate consideration. Probably there is some portion of the physical construction of the boost converter that has more resistance or inductive reactance than the design. You can prove or disprove any interaction by substituting a resistive load that draws the same current as the H-bridge portion and see if the voltage drop is the same. And then you can also try adding an additional fairly large capacitance to the DC connection between the boost converter and the H-bridge. Of course, if all of the systems are only simulations at this point, getting actual results of real world operation will be a challenge. But adding the large capacitance will be simpler and cheaper.
I tested the boost converter with the dead time and it now works without affecting the boost converter output.
 

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
Dead time -View attachment 174897

Regards, Dana.
I implemented the deadtime with the following circuit:
upload_2019-4-15_12-48-24.png

The top VDAC is set for 400 mV and the bottom one is set for 600 mV so that this way, they will compare with the sine wave with a little time difference which works out.

Now my problem is that whenever I feed the output waveform into the driver, the driver output does not modulate the way it used to before without the square wave ON/OFF states.

I am using IR2110 MOSFET drivers. Any recommendations?
 

danadak

Joined Mar 10, 2018
4,057
I am unfamiliar with this approach of modulation.

Is this like what you are doing https://www.cypress.com/file/45956/download

Have you confirmed you are getting dead time with scope on each leg of
H bridge ?

I want to be careful here as I am not a DC/DC design guy, so my help may
be limited. I would post your project and description over on Cypress Website
right away. Schematic also. To get it looked at by Cypress team.


Regards, Dana.
 
Last edited:

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
I am unfamiliar with this approach of modulation.

Is this like what you are doing https://www.cypress.com/file/45956/download

Have you confirmed you are getting dead time with scope on each leg of
H bridge ?

I want to be careful here as I am not a DC/DC design guy, so my help may
be limited. I would post your project and description over on Cypress Website
right away. Schematic also. To get it looked at by Cypress team.


Regards, Dana.
It works and I tested it. I am currently generating the following waveforms:



This waveforms are then being fed into the IR2110 driver. The drivers are being used to switch both legs of the H bridge using the following schematic:


In this schematic, the yellow modulated wave feeds B1 (Left) and the Yellow square wave feeds B1(Right). The other ref waves feed the other two Pins.

My problem is, the output of the high for the SPWM wave looks like this:


As you can see, it never reaches zero which means it fails to modulate meaning I cannot get a Proper SPWM output.
My VDD is 15 V and all the capacitor and resistor values are the same as in the schematic.

Would you happen to know what the issue would be?
 

danadak

Joined Mar 10, 2018
4,057
The output wave, where are you probing to see that ? What is probe ground reference ?
Whats the time base setting on scope when you looked at output wave ?

Sure looks like the low side fet not getting turned on, except 1 spike towards beginning
of your posted output waveform. Is the control loop telling us something ?

Trying to educate myself I found this useful - http://users.ece.utexas.edu/~kwasinski/10_EE462L_H_Bridge_Inverter_Basics.ppt

Looks like plenty of dead time, how much is it ?

One thing that confuses me is the cap in the H Bridge load. Its purpose is ?


Regards, Dana.
 
Last edited:

MisterBill2

Joined Jan 23, 2018
18,169
My guess is that the L and C in the load position on the H bridge are supposed to be a series resonant circuit. And the part about modulating something is quite confusing, unless this is supposed to be a transmitter of some kind. I must have missed, or not understood, the description of what this system is intended to accomplish. I know that I sometimes do miss that part. Sorry about that.
 

MisterBill2

Joined Jan 23, 2018
18,169
IT does indeed look like some portion of the system is not being driven the way it is intended to be driven, So there may be a failed component or a broken or missing connection. And with a system like this, really, the only place that you would see a sine wave is across the load. And it would look a lot more like a sine wave if the circuit were resonant. And we do need to know the reference point for any voltage or waveform measurements. And one more thing, in that circuit shown in post #11 the negative connection for Vbus is not visible. That may have an effect on a simulation if it is actually missing.
 

BobaMosfet

Joined Jul 1, 2009
2,110
I tested the boost converter with the dead time and it now works without affecting the boost converter output.
Normally when you design a boost converter, you design it for a specific voltage and current range. So what voltage and current did you design the converter to output, and how much current is your load drawing? Too much?

Test your boost converter without the H-BRIDGE to start with, and make it work first. Once it's working, then you can test it with the h-bridge. Divide & conquer.
 

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
The output wave, where are you probing to see that ? What is probe ground reference ?
Whats the time base setting on scope when you looked at output wave ?

Sure looks like the low side fet not getting turned on, except 1 spike towards beginning
of your posted output waveform. Is the control loop telling us something ?

Trying to educate myself I found this useful - http://users.ece.utexas.edu/~kwasinski/10_EE462L_H_Bridge_Inverter_Basics.ppt

Looks like plenty of dead time, how much is it ?

One thing that confuses me is the cap in the H Bridge load. Its purpose is ?


Regards, Dana.
The output is been probed at the gate of the transistors. It is the output of the drivers. It, however, shows that it does not go down between on and off and that s my problem. The LOW side only goes down to around 10 V meaning it does not shut off the transistor.
 

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
Normally when you design a boost converter, you design it for a specific voltage and current range. So what voltage and current did you design the converter to output, and how much current is your load drawing? Too much?

Test your boost converter without the H-BRIDGE to start with, and make it work first. Once it's working, then you can test it with the h-bridge. Divide & conquer.
The boost converter works now and I tested it with and w/o the H-bridge. My problem is that the Drivers are now not driving the output signal completely off.
 

Thread Starter

Carloco0306

Joined Nov 13, 2018
46
IT does indeed look like some portion of the system is not being driven the way it is intended to be driven, So there may be a failed component or a broken or missing connection. And with a system like this, really, the only place that you would see a sine wave is across the load. And it would look a lot more like a sine wave if the circuit were resonant. And we do need to know the reference point for any voltage or waveform measurements. And one more thing, in that circuit shown in post #11 the negative connection for Vbus is not visible. That may have an effect on a simulation if it is actually missing.
The schematic I gave you was just so you get an idea of what the circuit looks like. The waveforms are actual results from the physical circuit. As of right now, I am feeding 15 volts as my main supply. The driver however seems to work fine when I do not introduce a dead time which seems weird to me.
 

MisterBill2

Joined Jan 23, 2018
18,169
The schematic I gave you was just so you get an idea of what the circuit looks like. The waveforms are actual results from the physical circuit. As of right now, I am feeding 15 volts as my main supply. The driver however seems to work fine when I do not introduce a dead time which seems weird to me.
If it seems to work OK with no dead time, that means that the turn-off is being supplied by the opposite set of driver connections. So now you need to discover what connection you are missing that is being substituted for by the opposite driver. I hope this explanation makes sense, because that certainly looks like what it is. The intended path is missing for some reason.
 
Top