H bridge circuit evaluation

Thread Starter

JAdams5

Joined Mar 9, 2018
12
Hi. I've received a project for a design course whereby we have to build a variable frequency drive that will run a shaded pole motor for as little as possible. The frequency of the drive should be set using a voltage of 0V-5V DC. I have come up with the attached design. This is my first design project and i need to know if this is okay or will make the motor burn up.

As far as the design is concerned i have a 555 timer generating a 50Hz square wave on top. This is fed to a pair of IGBT's and an inverted signal is fed to another pair of IGBT's. The voltage at the bottom left is the reference voltage. The bottom 555 timer produces a 120Hz wave that is 180 degrees out of phase with the other. This basically fakes deadtime of a pwm signal. Any suggestions? The output from the load is attached.
 

Attachments

MrAl

Joined Jun 17, 2014
11,342
Hi,

Not everyone gets on here at the same time you do so it may take a little while to get replies.

To start, you dont drive MOSFET's with 4.7k gate resistors. That is way too high and could cause high power dissipation in the transistors once a load is applied.
It's a little hard to tell what you are doing here though with the 555's. Do you intend to vary the phase using two of them? Or the pulse width?
Your output wave looks right though. That is what to expect with a pulsed output that has to vary in width.
It has to be tested with the intended motor though in any case.

I just noticed you have two gates tided together for each of two transistors. You cant do that because the upper transistor needs a higher gate drive tahn the lower.
Are those N channel MOSFET's ?
 
Last edited:

Thread Starter

JAdams5

Joined Mar 9, 2018
12
Not everyone gets on here at the same time you do so it may take a little while to get replies.
Apologies for that.

Are MOSFET's and IGBT's the same thing? I don't believe they are. Also an amendment to the top, this H-Bridge is going to be driving a mains powered load(the motor) with rectified AC, i just made it 10V for simulation and forgot to change it back, it's supposed to be 310V. I intend to use STGW20NC60VD IGBT's for this.

I'm not a proper electronics student so forgive my methods. What i intend to do is to change the frequency of the wave that is coming from the 555 timer so that the switching is faster or slower depending on the control voltage at the left (V2). As far as simulation goes, it works, but i know that simulation seldom agrees with experimental results.

Would i need to change the pulse width as well to vary the frequency of the wave?

5555.PNG
The blue wave is the output of the top 555 timer and the green wave is the inverted blue wave.


5556.PNG
This red wave is the output of the bottom 555 timer.


5557.PNG
Here the waves are superimposed together and different scales are used for clarity. Basically when the red wave(from the bottom 555) is above either the green or blue wave, it pulls those waves down to ground. It basically "cuts" both signals so that they have off time to simulate a modified square wave.
 

MrAl

Joined Jun 17, 2014
11,342
Hi,

Oh so you are using IGBT's then?

The pulse width is varied to control average or RMS output amplitude. If your input is stable then you may not need that, but it is usually part of converters since they are already handling the power anyway. I suppose it is not mandatory though.
 

jpanhalt

Joined Jan 18, 2008
11,087
Using the control pin of a 555 to get a VCO is a standard circuit.

What is unclear is how you keep a 50 Hz signal and 120 Hz signal 180° out of phase. Your simulation seems to show they are the same frequency. Have you tried the circuit with real components?
 

Thread Starter

JAdams5

Joined Mar 9, 2018
12
I think i made a mistake with my simulation. It should have been a 50Hz signal and a 100Hz signal. While simulating it shows up as 60Hz and 120Hz respectively. Haven't tried it with real components yet, just wanted some input for my methodology, whether it would be viable for something like this or not work at all.

Well i suppose they aren't technically 180 degrees out of phase. The on time of the 100Hz signal is supposed to be half the width of the 50Hz signal. So it cuts out the last quarter of the positive pulse and the first quarter of the inverted pulse, thereafter the last quarter of the inverted pulse and the first quarter of the positive pulse for every wave. Giving the waveform in my first post.
 
Last edited:

Thread Starter

JAdams5

Joined Mar 9, 2018
12
If your input is stable then you may not need that, but it is usually part of converters since they are already handling the power anyway. I suppose it is not mandatory though.
I'm getting a constant 5V, 500mA max supply from elsewhere to control the driver circuit.
 

MrAl

Joined Jun 17, 2014
11,342
Hello,

Well no i meant the Vcc bus like 200v or whatever you are going to use, not the low volts power supply voltage.
The pulse width is often changed to make up for any input line variations. For example, if you had 220vac coming in to power the DC bus, it may go down to 200 or up to 240 and so the output voltage of your convert er would follow that if it was not 'regulated', and regulated means vary the pulse width.

If you want 100Hz and 50Hz you should probably use a flip flop as divider. The problem with two 555's is that the phase of the second is not locked to the phase of the first.
But also, you probably dont want to knock off 1/4 and 1/4 you probably want to look into the average value or RMS value and go from that, unless you've already done that and it came out to 1/4 and 1/4. But then again it also depends on your DC bus voltage.
 

Thread Starter

JAdams5

Joined Mar 9, 2018
12
I see :rolleyes:. The voltage driving the motor is a rectified mains supply, so ~310VDC. I don't think that fluctuates much. As far as my main goal i need to be able to vary the frequency (and or pwm) using only that reference voltage on the left. The frequency works fine, how can i do that for pwm such that they are independent of each other?
The point of this system is to be automated from that reference voltage so i won't have to use pots or MCU's for this.
With the Flip flop divider, i haven't seen that approach before. I shall look into it. However, do you know of a method where i can alter the phase angle of one of the waves if i were to use the flip flop approach? The 50Hz signal is the main one with the 100Hz chopping it up. so i would need to vary the pwm of the second one independently of the first (which i can do now from the second 555 timer) but i don't think its possible with the flip flop method :confused:. I may be wrong though :p.
 

ebp

Joined Feb 8, 2018
2,332
SG3525 is an ancient part that is suitable. There are many more available.
The high side IGBTs will need either a high side driver with high voltage level shifter or transformer drive.

"... for as little as possible"

In my opinion that is an absolutely idiotic demand to make of inexperienced designers.
 

Thread Starter

JAdams5

Joined Mar 9, 2018
12
SG3525 is an ancient part that is suitable
You know where i can get a simulation model of that from? Preferably in multisim. I'm very reluctant to experiment with stuff due to my lack of knowledge as well as budget that im on (R500 btw) which is why i rely so heavily on simulation.

In my opinion that is an absolutely idiotic demand to make of inexperienced designers.
I agree. It's pretty shite seeing that this is a first time design project. Also i'm in computer engineering and for this they want a pure electronic design, so no MCU's. I have no clue why we are under such restrictions. My skill-set is practically useless because of this.
 

ebp

Joined Feb 8, 2018
2,332
Have a look at this thread:
https://forum.allaboutcircuits.com/threads/full-bridge-pwm-controller.146133/
There is discussion of some models there. There have been other discussions of the 3525 here at AAC.

I also suggest going to the TI website and looking in the Power Management section. TI had a few PWM controllers of their own and then acquired Unitrode which had many and National Semi which had some more. You will probably find simulation models for a number of their parts. Some will be more expensive than the 3525, but it may be worth considering a slightly more expensive part if you can't find a good model for the 3525.
 

MrAl

Joined Jun 17, 2014
11,342
I see :rolleyes:. The voltage driving the motor is a rectified mains supply, so ~310VDC. I don't think that fluctuates much. As far as my main goal i need to be able to vary the frequency (and or pwm) using only that reference voltage on the left. The frequency works fine, how can i do that for pwm such that they are independent of each other?
The point of this system is to be automated from that reference voltage so i won't have to use pots or MCU's for this.
With the Flip flop divider, i haven't seen that approach before. I shall look into it. However, do you know of a method where i can alter the phase angle of one of the waves if i were to use the flip flop approach? The 50Hz signal is the main one with the 100Hz chopping it up. so i would need to vary the pwm of the second one independently of the first (which i can do now from the second 555 timer) but i don't think its possible with the flip flop method :confused:. I may be wrong though :p.
Hello again,

Well actually if you dont care to regulate the output then this is a simple task, assuming your 1/4 and 1/4 cut method is correct. I think you may want to verify that, and perhaps mention where you got that method from.

If your 1/4 and 1/4 cut is OK, then the 1/4 cut on start and 1/4 cut on end just results in a square wave all above zero. So in this case all we need is an oscillator running at 4x the frequency and two flip flops. One flip flop output Q will be at 1/4 frequency, so that brings it back to 50Hz if the input is 200Hz. To get 100Hz output you just increase the oscillator frequency to 400Hz, and so on for other frequencies.
The first flip flop generates the signal, all above zero, the second flip flop generates the polarity, plus or minus./
So when the first flip flop output is zero there is zero bridge output, and when the second flip flop is 1 the output is positive and when 0 it is negative.

Im sure you can figure out the drive for the H bridge switch elements right?
You do need to verify that a 50 percent duty cycle is what you want though too.
What is the normal peak (of the normal sine) voltage for the motor?
 

Thread Starter

JAdams5

Joined Mar 9, 2018
12
Okay.

My design mentor is being cryptic about the stuff and says he doesn't know if it'll work. Which is lies because he most likely does.

Anyway the goal is to control the speed of a shaded pole AC motor from a reference voltage "they" are going to supply to us (0-5V i believe). The motor is a Mellor Electric motor rated at 230V AC 50Hz 24W. What i'm afraid of is burning up the motor. If it blows up, so do my marks and i get suspended for the year. Do you know of any tried and trusted methods that i could use to control it? I have read that a VFD is used sometimes, which is kind of what i tried to simulate with the above design, but i'm sure that's not the ideal output. Any help would be appreciated.

Thanks
 

MrAl

Joined Jun 17, 2014
11,342
Okay.

My design mentor is being cryptic about the stuff and says he doesn't know if it'll work. Which is lies because he most likely does.

Anyway the goal is to control the speed of a shaded pole AC motor from a reference voltage "they" are going to supply to us (0-5V i believe). The motor is a Mellor Electric motor rated at 230V AC 50Hz 24W. What i'm afraid of is burning up the motor. If it blows up, so do my marks and i get suspended for the year. Do you know of any tried and trusted methods that i could use to control it? I have read that a VFD is used sometimes, which is kind of what i tried to simulate with the above design, but i'm sure that's not the ideal output. Any help would be appreciated.

Thanks
Hello again,

The only method i know for sure that will work is a synthesized sine wave converter. That has a pulse pattern that emulates a sine wave.
You still use an H bridge, but instead of a single pulse for each half cycle, you have multiply pulses of various widths that make up a synthetic sine wave.

Also, for AC motors there is usually a voltage vs frequency relationship. You'd have to look that up.
In other words, when you lower the frequency you also lower the voltage, and when you raise it of course you raise the voltage too.
I dont think you can just lower the frequency because the motor will start to draw too much current.

Did you mention what frequencies you want to use with this?
I assume it is a 50Hz motor? But what frequencies do you want to drive it at?
 

Thread Starter

JAdams5

Joined Mar 9, 2018
12
Here's a link to the motor we are using: https://za.rs-online.com/web/p/ac-motors/2483637/

I'll have to play around with the frequency once i get the motor. It's going to be used as a fan, so 3 speed setting of high, med and low (and then off once the frequency/voltage gets too low). So not 100% sure about that.

I have seen that method of SPWM being used to drive the IGBT's or Mosfet's of an H-Bridge. I know that the signals are a difference between a triangle wave and a sine wave and a triangle wave and an inverted sine wave for the negative cycle, but i have no clue how to generate those signals with analog only. Also simulation is also a restriction so i can't be 100% sure of what i need to get. If you know of a circuit that can generate an SPWM and be able to use it to drive the gates, please send it to me, or show me where to go. I'll probably buy the components and test myself when i get a chance.

Thanks.
 

MrAl

Joined Jun 17, 2014
11,342
Hi again,

Well there are sine wave oscillators and triangle oscillators. A sine wave oscillator coupled with a comparator and op amp would generate a triangle, but the triangle as you know has to be at a multiple of the sine frequency so you would need a PLL design or something to keep the triangle sync'd with the sine. But i would have to question the practicality of doing it that way anyway. The more normal way is to use a sine table where you look up the values and use that as a pulse width reference.
 
Top