1
\$\begingroup\$

I've been doing some work lately with the DRV103 IC by Texas Instruments (datasheet here, essentially it creates a PWM signal of variable duty cycle and frequency, preceded by an 'ON' signal of variable time), which is a pretty cool chip, but I've hit a snag with it.

For a particular application, I decided to control the duty cycle with voltage rather than resistance. According to the datasheet, this should be straightforward.

enter image description here

I found the right resistors and potentiometer for the values needed, selected somewhat arbitrary delay and frequency values, then set it up as per this schematic:

schematic of application circuit

P1 and P2 are generic wire-to-board connectors, the transistor is a PMF170XP,115 P-type MOSFET which I used to invert the output.

Parts arrived, I soldered it up, hooked up the power, but when I switched it on, rather than the expected PWM, the output went straight to 12V and stayed there. The voltages everywhere seem correct (including the potentiometer range), there are no shorts that I can see (and the benchtop power supply I used shows almost 0 current draw) and I replaced the chip as a sanity check; no luck.

Have I missed something obvious? There's no flyback diode, but the chip is only driving the signal input of another piece of hardware; adding one is the next thing I'll check, along with a pulldown resistor.

\$\endgroup\$

1 Answer 1

1
\$\begingroup\$

I believe I found the problem - one of the example circuits has a pullup resistor between OUT and V+ when working with a transistor. I added a 4.09k and now it seems to be working.

EDIT: The output was still not quite right, so I added a second 4.09k pulldown resistor between the PWM line and ground. This completely fixed it, giving a perfect square wave.enter image description here

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.