Cool Cap Engineer

Engineering by an anime nerd

1 Comment

Electronics: SN754410

Today’s article focuses on a popular chip that many robotic hobbyists use. Of course I’m talking about the SN754410, or the Quadruple Half H Driver.

Figure 1: Standard H-bridge Schematic

First things first; what’s an H-bridge? As shown in figure 1, an H-bridge is an electronic circuit that allows current to flow in different directions depending on the state of the switches. In fact, the 7 level cascaded inverter project I’m working on is nothing more than levels of H-bridges.  When S1 and S4 turn on, the motor’s shaft rotates clockwise, but when S3 and S2 turn on, the motor’s shaft rotates counter clockwise.  If you want to build your own H-bridge, then you’ll most likely implement it using transistors (my recommendation is to use MOSFETs).  But, if you’re lazy and wish to avoid building the circuit, then you can buy ICs that have one or more H-bridges in one chip.  The SN754410 chip is among the plethora of H-bridge chips you can buy with H-bridges already inside.

SN754410 Pins

Figure 2: SN754410 Pinout

So let’s take a look at the pins of the SN754410, which can be seen in figure 2. I will give a detailed explanation on the purpose of each pin.

1,2 EN( Pin 1) and 3.4 EN(Pin 9) : For the SN754410, you can connect up to two DC motors to the chip, which I’ll explain where you connect the motors to on the chip later on. However, to turn on either motor, 4.5-5.5V must be applied to the EN pins. For example, applying 5V to 1,2EN will turn on the motor on the 1, 2 EN side and applying 0 volts to the 1,2EN will turn the motor on the 1,2EN side off.  It’s possible to apply a PWM (pulse width modulated) signal to the EN pins to control the speed of the motor. Although it will dissipate a lot of power for your motor control project, you can place a pull up/down resistor at the EN pins to prevent the pins from ‘floating’. In digital logic, floating pins will cause some unexpected behaviors for your device. The pull up/down resistor will make sure the EN pins have a default state.

1A/2A (Pin 2/Pin 7) and 4A/3A(Pin 15/Pin 10):  Any pin that has an A in the name is where you put your digital logic into the chip. The 1A/2A pins controls the motor connected to left side of the chip while 3A and 4A controls the motor connected to the right side of the chip. The digital logic you put into the chip will determine whether you’re the shaft of the motor rotates clockwise, counter clock wise, or stays in position. For the logic pins, it sees a high state when 4.5-5.5V is applied, while it sees low state when 0V is applied.

1Y/2Y (Pin 3/Pin 6) and 4Y/3Y(Pin 14/Pin 11): These pins are where you connect your motors to.  1Y and 2Y are responsible for controlling the motor on the left side of the chip, while 4Y and 3Y controls the motor on the right side of the chip. It’s recommended to add a .1uF capacitor to both motors as it will cut down on the motor’s noise, which could harm the main controller of your project.

VCC1 (Pin 16)/VCC2( Pin 8): These pins are used to power the chip. Unlike other chips, you cannot power both pins using the same power source. VCC1 should receive 4.5-5.5V since this power the logic gates inside the chip. VCC2 is used to power the motors and can receive up to 36V.

Ground Pins (Pins 4,5,13, and 12): Simply connect these pins to ground (0V) of your power supply.

Example Circuit

Figure 3: SN754410 example circuit

Figure 3 shows a simple example circuit of the SN754410 running a small DC motor. Please keep this in mind; 9V is a poor power source. Sure, it has a high voltage, but suffers from a small current rating (around 500-750mah). By having a small mah rating, the battery will discharge at a fast rate.The only reason I include a 9V battery in this picture is to show the different power sources necessary to power the chip.

SN754410 Recommendations

Despite the fact this chip can accept motors that require up to 36V, the current it can handle is very limited. This chip can handle up to 1A per side, which can be really limiting. Sure it can handle most of the Pololu’s Metal Gearmotors and all of the Tamiya motors without any problems, but you’ll run into problems for beefier motors that requires a higher current.  If you need an H-bridge that can run high amperage motors, you can either build one yourself (trust me, it’s pretty straight forward) or look into the L298.

Thank you guys for reading this article! If you have any compliments, comments, or concerns, feel free to post a comment!

Leave a comment

Projects: Single Level Inverter Implementation #2

Hey guys! I have an update for you guys on my 7-level inverter project.  As I said in my previous post, I was deciding on what parts I wanted to use for the project. Well, I finally finished selecting the parts that I want to use.

Before I talk about the parts, let me tell you guys the design parameters that I wanted to work with. The first parameter is the peak voltage of the AC waveform. I don’t want to work with dangerously high voltages, so I chose the peak voltage to be at least 12 volts.  As for the operating current, 1A seems like a reasonable number. Since each level is connected in series with each other, the current will not increase (KCL equations).  Finally, I don’t want the inverter to operate at a high speed, so I’m going with a 60 Hz operating frequency, which is the “ideal” frequency of most grid tied applications. With the parameters listed, let me tell you the major parts I will use for the project.

Figure 1: FQP27P06 and FQP30N06L representation

As shown in figure 1, I will use the FQP27P06 P Channel Mosfet, and the FQP30N06L N Channel Mosfet.  Both of these transistors are rated to handle a maximum of 60V between the drain and source.  This is much more than the expected peak voltage I will be working with. Although FQP30N06L is rated to handle 32A while the FQP27P06 is able to handle -27A, this is much more than the 1A requirement. The only thing I’m kind of concern about for these transistors are the ratings for the rise and fall times. The FQP30N06L max rise time is 430ns (.43 us) while the FQP27P06 max rise time is 380ns (.38us).  However, since the operating frequency of the inverter will be 60 Hz, I doubt this will hurt the output waveform. In fact, the gate driver will reduce the rise and fall times of the Mosfets.

Figure 2:  HCPL-3120 representation

For the gate driver, I will go with the HCPL-3120, which can be seen in figure 2. There are two reasons why I’m going with this gate driver. One of the reasons why I’m using this chip is because I’ve use this chip before. Back at my power electronics research, I built the application circuit listed on the HCPL-3120 datasheet due to sheer boredorm. Another reason is the 2A output current, which is needed to reduce the rise/fall time of the Mosfets.  Since the gate of a Mosfet functions like a capacitor, the more current you put into the gate, the shorter the time it takes to charge and discharge the capacitor.

Figure 3: HCPL-3120 example circuit

The rest of the parts I’ll go into brief detail. I’ll just following the example circuit shown on the datasheet (figure 3).  The parts I need to get from the datasheet are the 270 ohm resistor and a .1uF capacitor. The final part I wanted to include in the project is an 8 pin dip, which will make it easier to swap defective gate drivers. Another thing to keep in mind is that I need to get every part I listed in multiplies of three.

What about the PCB? I’ll most likely go with dorkbot for the PCB manufacturer. I’ve been using Dorkbot for two years now, and I trust them with my project requiring a PCB. Although professional engineers will cringe at this service, this is a really good service for small projects like mine.

Now, when I am ordering all of these goodies? Ideally,I want to say as soon as possible, but realistically I’m going to have to say two-three weeks from now.  The reason why is due to Dorkbot. The last time I ordered from Dorkbot, it took 4 weeks for my PCB to arrive home, even though it usually takes 2 weeks for PCB orders to ship out. Since I’m going back to college two weeks from now, I thought it would be better to order the parts when I go back.  Not to mention, I can use the lab equipment at my college to properly test the single level of the inverter.

Assuming $10 of shipping and buying the parts in multiples of three, the cost for the project’s parts should come to approximately $64. The worst part is that it does not include the cost of the PCBs which should be between $15-$25.

Well, that’s it for my update. If you guys have any compliments, or criticism for this project, please post a comment down below. Thanks guys!

1 Comment

Projects: Single Level Inverter Implementation #1

So a couple months ago, I did an REU power electronics experience at University of Tennessee at Knoxville. Despite the initial frustration, the experience was worthwhile and made me realize the importance of power electronics in today’s industry. I spent most of the research at the REU program running ideal simulations of a seven level cascaded that’s powered by a PV module for each level.  While participating in the research, I learned various parameters that must be taken in account when implementing a real multilevel inverter.  After participating in the program, I went to work on designing the circuitry, PCB, and selecting the parts for my own multi-level inverter.

So how does a multilevel cascaded inverter work? First, let me explain how a single inverter level works. A single level of an inverter consists of 4 switches and a DC source connected between them.  A single inverter level must be able to generate 3 different switching states: 0 volts( when only the top two switches turn on), VDC (when the top left and bottom right switches turn on), and -VDC (when just top right and bottom left switches turn on).

Now, let say we connect another inverter level to the original level, but the switching sequence is out of phase of the original. By connecting out of phase inverter levels, you’ll get a stair case waveform, where the number of stairs is just two times the number of inverter levels added by one. By adding an infinite amount of out of phase inverter levels in series with each other, you’ll get a perfect sinusoidal waveform. The easiest way to understand this is to think of the multilevel inverter as those baby stacking toys you use to play with. Figure 1 shows an example of a S level inverter and figure 2 shows the waveform associated with the multilevel inverter.

Figure 1: Multilevel Cascaded Inverter example

Figure 2: Multilevel Inverter Waveforms (n level)

The goal is to successfully implement a single level of the inverter. Not only am I doing this for sheer curiosity, but it can be use for my microcontroller class I will be taking this fall. By implementing the single level correctly, the multi-level inverter can be implemented later. So, there were a couple things I had to take in account when selecting the parts for a single level of a inverter.

Transistor Blocking Voltage:  The blocking voltage has different terminologies for different transistors. For example, the VDS rating is the blocking voltage of the mosfet, while the VCE rating of a BJT is also the blocking voltage. The blocking voltage must be 150% higher than the peak voltage of the inverter. The 150% may seem excessive, but I want to make sure your transistor will not burn out due to the peak voltage of the inverter.

Transistor Current Rating: Since my project will use mosfets, we want to make sure the drain current rating of the mosfet is much higher than the possible peak output current of the inverter. Like the blocking voltage rating, you want to choose a current rating 150% higher than the expected output current of the inverter.

Gate Driver: Let’s get one this straight; transistors are not ideal switches. Transistors have a lot of imperfections and limitations that makes everything I learned from my ideal seven level cascaded inverter simulations absolutely useless.  Among them, transistors cannot instantaneously turn on or off due to the limited rise and fall times associated with them.  The rise and fall times for a mosfet can be further reduced by adding a gate driver to the gate of the mosfet. The gate driver will allow a bigger current to go through the gate of the mosfet, allowing a faster charge time for the mosfet’s gate capacitance.  By increasing the charge time of the mosfet’s gate capacitance, you achieve a faster rise time for the mosfet. Of course, the gate driver has its own rise/fall time as well.

Although I did not select the parts yet, I managed to design the circuitry and the PCB associated with the inverter with Eagle Cad since the parts will keep the same format for the most part. I’ll update you guys on any further progress I make with this project.If you worked on multilevel inverters, and you have a complement, or compliant, let me know by posting a comment! Thank you and have a great day!

Figure 3: Single Level Inverter Schematic

Figure 4: Single Level Inverter PCB Design