Tonight I am experimenting with an LTC 6992 Voltage Controlled PWM IC. I had an application where I wanted to control a motor's speed using a variable voltage and the LTC6992 came up in my search. Here is the chip as listed on the Newark site:
And here is the link to the Data Sheet from Linear Technology.
http://www.farnell.com/datasheets/1642034.pdf?_ga=1.70056336.1418853850.1451543312
My first challenge after receiving the chip from the vendor was putting it onto a workable board. I should be starting to get used to the size of these new surface mount ICs but they always seem to surprise me. The device that I received is a TSOT-23 package and way to small to deal with unless it is mounted on a board. I had some converter boards on the shelf but none of them were perfectly suited to this device. I finally ended up using an SOIC8 adapter board and although the legs didn't match the spacing on the board exactly I was able to solder bridge the difference. Here is the finished device now the size of a regular 8 pin DIP.
Now it doesn't take much to impress me but I found the features of the LTC6992 impressive and easy to use. The device comes in 4 variants designated by a -1, -2, -3, or -4 after the number. Page 17 of the data sheet explains that the differences between them have to do with the duty cycle ranges that they cover. The device I am testing is a LTC6992-3 which means that it will have a duty cycle range of 0% to 95% as the Modulation voltage varies from approximately 0.1 volt to 0.9 volt.
The device can run on a voltage supply of 2.25 Volts to 5.5 Volts. I ran my experiments at 5 volts. It is important to put a 0.1 uF decoupling capacitor between Pin 5 V+ and Pin 2 Ground for proper operation. Pin 1 is the Modulation and expects an input from 0.1 volt to 0.9 volt which will cause the chip to vary its PWM output from 0% to 95%. Voltages outside this range have no further effect but if the voltage is more than 0.3 volts above V+ or more that 0.3 volts below ground damage can result. For the purposes of this experiment I built a voltage divider with a 1K resistor and a 10K resistor and took my Mod input voltage across the 1K resistor. I did this since my power supplies are inaccurate and have poor resolution in the 0 to 1 volt range and the voltage divider allows me to use 0 to 10 volts on my power supply divided by approximately 10. I put a bench meter on the Mod pin and placed it near the Oscilloscope so that I could take simultaneous pictures to relate the voltage on the Mod Pin 1 to the Duty Cycle of the output on Pin 6.
This chip has two circuits that allow a designer to choose the frequency of the output. The first method uses a single resistor with a value between 50K and 800K connected between the #3 Set Pin and ground. A 50K resistor will generate an output of 1 MHz and the 800K will produce an output of 62.5 kHz. The second method uses the #4 Divide pin to introduce a divide by factor to further control the output. Internally the chip has a 4 bit ADC which takes the voltage on Pin 4 and converts it to a digital divider of the frequency. From 0 to 1/2 of V+ the ADC will move through 8 division ranges that will bring the output frequency to as low as 3.815 Hz. Once the voltage on Pin 4 is greater than 1/2 V+ the Polarity of the duty cycle reverses and 0 volts on the Mod Pin will produce 95% PWM and 0.9 volts on Pin 1 Mod will produce 0% PWM. This is very handy as it can serve as a negative feedback loop to stabilize a dynamic system.
My first experiments with the LTC6992 utilized an LED with a current limiting resistor on the #6 Output Pin. The device can supply or sink 20 mA but no more. Later in the experiment I will expand the output to an N Channel MOSFET and use it to drive a 24 volt motor.
Here are pictures of the Mod voltage and the Oscilloscope output for voltages of 0.5 volts, 0.714 volts, and 0.105 volts. I am using a 470K resistor on the #3 Set pin which produces a frequency of 106.4 KHz on the output. The Data Sheet provides a simple formula for calculating the frequency. I found that the actual frequency very closely match the calculated frequency. For this part of the experiment I had the Divide Pin 4 tied to ground so that my divide by factor was 1. This also put the Mod to Output in a positive correspondence. PWM increases 0% to 95% as Pin 1 voltage increases 0.1 volt to 0.9 volt.
The next step in the experiment was to add a resistor divider to the Div Pin 4. I chose to use 1K from V+ to Pin 4 and 100 Ohm from Pin 4 to Ground. This choice should produce a divide by factor of 4. I calculated that this should produce an output frequency of 26.6 KHz. Here is a picture of the Oscilloscope and you can see that the frequency is indeed 26.6 KHz.
My next step was to switch the configuration of the resistors on Pin 4 so that the correspondence of the change of the input to the output would become negative. In this case as the Voltage on Pin 1 changes from 0.1 volt to 0.9 volt the output would change from 95% PWM to 0% PWM.
Here is a close up of the bread board at this point and a shot of the Oscilloscope showing the negative correspondence with the 0.772 volts corresponding to a very low % PWM.
The last part of my experiment was to add a 24 volt motor and a N CH MOSFET to drive it. Here is the breadboard set up.
I had a lot of fun tonight exploring the functionality of the LTC6992. I believe it will have some good applications in some of my projects in the future.
Top Comments