0
\$\begingroup\$

In my application, a microcontroller with 5V logic needs to drive two 12V DC fans in voltage. The microcontroller will use a PWM drive signal with a carrier frequency of 32 kHZ. For several reasons, I want to drive the fans with constant voltage sources, ie no unfiltered PWM should reach the fans. The fans draw a maximum of 0.3A per fan, which gives 3.6W at maximum power. I need to evaluate two kinds of designs: The first one, which is a complete mosfet based push-pull configuration with an LC filter to linearize the PWM. The second one, which is a simpler high side switch, directly driven from the microcontroller, with an LC filter at the output. I am aware that the first option is more efficient, but due to cost/complexity I would prefer the second option if viable. What are the benefits/drawbacks of one configuration over the other?

First Option uses an LM27222 driver and a dual mosfet IC: enter image description here

Second option uses an ISP452 (I should go for a faster-response IC): enter image description here

\$\endgroup\$
2
\$\begingroup\$

I need to evaluate two kinds of designs: The first one, which is a complete mosfet based push-pull configuration with an LC filter to linearize the PWM.

This isn't actually a push-pull design. The lower FET never pulls. Current always flows through it from source to drain, not drain to source.

This configuration is usually called an "active switch" configuration. Because the lower FET is replacing the diode of your second design.

What are the benefits/drawbacks of one configuration over the other?

The active switch design can be more efficient. The diode based design is likely to be less expensive. Exactly as you already determined.

Deciding which is best for you will depend on the exact balance of the importance of cost and efficiency in your application. And it how well you can optimize it will depend on all the detailed choices you make in designing either of the two choices, and your ability (or your organization's ability) to find good prices for the different types of components.

Unfortunately (or fortunately) there are thousands of different controller chips, MOSFETs, and diodes out there to choose from, and pricing changes dynamically so it's unlikely you'll ever find the absolute best choice.

Edit: As another answer points out, either one of these is probably overkill (and excess cost) to solve your actual problem. The fan itself can be used as an inductive element to smooth a PWM signal. Or, if the reason you want "no unfiltered PWM" is to reduce audible artifacts, finding a controller that can provide a higher PWM frequency may be the most cost-effective solution.

\$\endgroup\$
  • \$\begingroup\$ The two transistor version can conduct current through the lower transistor in the normal way when on very light loads i.e. what would be described as discontinuous operation if it were a non-synchronous design. \$\endgroup\$ – Andy aka Sep 20 '17 at 17:55
  • \$\begingroup\$ Exactly. I want to filter out PWM for several reasons: in order to reduce audible noise, to be able to read the tachometer signal without interference, and finally to reduce EMI. I have the feeling that the LM27222 configuration is overkill, due to the low power required. But I would stay with the LC filter circuit at the output. \$\endgroup\$ – Francesco Sep 20 '17 at 17:55
  • \$\begingroup\$ @Francesco, it's also possible to find fans with a fourth wire is used to do PWM control inside the fan. Sanyo SanAce family offers this, for example. \$\endgroup\$ – The Photon Sep 20 '17 at 18:03
  • \$\begingroup\$ Yes. I am aware about the PWM fans. But unfortunately I need to use one specific model which does not support this feature. \$\endgroup\$ – Francesco Sep 20 '17 at 18:04
2
\$\begingroup\$

If you want low cost/complexity you could use an integrated motor driver IC.

For example this one should be able to drive both fans...

The one with the diode also gets much simpler if you can put the switch in the low side (with the fan between +12V and the switch). 300mA can be handled by a little FET with very low gate charge, so you can also remove the driver and just use one of the micro's PWM output pins.

\$\endgroup\$
  • \$\begingroup\$ How can I filter out the PWM signal if I use a low-side FET? Should I keep LC on the +12V line while switching the low-side? \$\endgroup\$ – Francesco Sep 20 '17 at 17:47
  • \$\begingroup\$ +1: Better to answer the "X" question than the "Y" one. \$\endgroup\$ – The Photon Sep 20 '17 at 17:49
  • \$\begingroup\$ Yes, you can put the filter on the positive side... you'll need a decoupling cap very close to tie the +12 to ground, it should work just fine. Also, don't switch too fast if you're concerned about noise. \$\endgroup\$ – peufeu Sep 20 '17 at 18:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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