1
\$\begingroup\$

I am measuring analog DC voltage at the analog pin of dspic33 series microcontroller. The source of the signal is AC. So, I rectified it with diodes and amplify it with op-amps. In order to reduce the ripple I am using RC filter as first phase filtering and LC filter as second phase filtering.

Single RC filter cannot satisfy my need that it outputs a voltage with 100mv of ripple content in a signal which is going to be measured by ADC. Also, I cannot increase the capacitance since it violates my timing constraints of the requirement due to high RC time constant. So, I decided to go for LC filter as second phase filtering to reduce the ripple content of the DC signal.

My doubt is, whether using LC filter near the analog pin of microcontroller is technically ok or not? Is there any interference come with LC filter when the setup undergoes environmental tests and noise tests?

RC filter has 1k ohm resistance and 1uF capacitance. LC filter has 470 mH inductance and 1 uF capacitance.

Help please!enter image description here

Edit : I have added filter portion of the schematic. As per comments I need to clarify my points.

Input to the circuit is AC (0 - 200V) The output of the circuit is DC (0 - 3.3V)

Timing constraint means, I need to detect the voltage within certain milli seconds to process and do other actions based on this value within a second.

\$\endgroup\$
6
  • \$\begingroup\$ "I am measuring analog DC voltage at the analog pin of dspic33 series microcontroller. The source of the signal is AC" This is contradictory. What is it you are actually interested in measuring? Normally you want to sample the signal that is actually there and you filter in order to properly sample what you are interested in. But you seem to want to filter just to make your samples look a certain way without really caring about what the real signal actually looks like. \$\endgroup\$ – DKNguyen Feb 5 '20 at 2:47
  • \$\begingroup\$ Also, what timing constraints? What are you actually trying to do with all this? \$\endgroup\$ – DKNguyen Feb 5 '20 at 2:54
  • 2
    \$\begingroup\$ Please post a schematic. Realize that an LC filter will amplify at the resonance. I would use a 2nd order (or higher) op-amp filter instead. \$\endgroup\$ – Mattman944 Feb 5 '20 at 2:59
  • \$\begingroup\$ What is the AC frequency (or range)? \$\endgroup\$ – Mattman944 Feb 6 '20 at 2:06
  • \$\begingroup\$ It's frequency is 400 Hz \$\endgroup\$ – CNA Feb 6 '20 at 2:13
0
\$\begingroup\$

The input impedance of the micro is very high .No wonder there is a pronounced resonant peak as Mattman has shown .Damp the LC filter to get low Q and the nasty peak will go .Your coil should be screened like in a can so it wont pick up noise .Otherwise your proposed filter is valid .

\$\endgroup\$
0
\$\begingroup\$

I used CircuitLab to plot the frequency response of the filter. I don't like it very much. Any noise at 10.5 kHz will hardly be filtered at all. It will filter fairly well above 15kHz. So, if that is where your noise is, this filter is OK. If your noise is lower than that, you should use a Sallen-Key butterworth filter.

Or, if you show your entire circuit, there might be better solutions. For example, if your input is always 400 Hz, you could bandpass the AC before you rectify it.

schematic

simulate this circuit – Schematic created using CircuitLab

enter image description here

Edit: An alternative inspired by Autistic's answer. Add a 33 ohm in series with L1 to dampen it. You should subtract any resistance in L1 if it is significant.

enter image description here

\$\endgroup\$
3
  • \$\begingroup\$ Your design consists of 400Hz AC Source directly coupled with filter section. But in my circuit, the AC signal will be rectified and converted to a 3.3V DC signal before it reaches the filter section. I would like to know, there will be any problem due to the LC filter design which is directly connected with the microcontroller's analog pin. I have read somewhere that inductors are susceptible to EMI/EMC interference. \$\endgroup\$ – CNA Feb 7 '20 at 7:24
  • \$\begingroup\$ I understand that the signal is DC at the input to the filter. The DC does not need to be modeled since it passes through the filter. Injecting a sine wave into the filter is used to test the filters response. The response is poor unless modified as I have suggested. \$\endgroup\$ – Mattman944 Feb 7 '20 at 7:36
  • \$\begingroup\$ As for inductors being susceptible to EMI/EMC interference, that depends on the inductor construction. If you are worried about susceptibility, use an inductor with a torroid core. Torroid cores, because they are symmetrical, do not radiate, nor are they susceptible to EMI. How susceptible a non-torriod core inductor would be is extremely complicated. It should not normally be an issue, but could be in extreme EMI environments \$\endgroup\$ – Mattman944 Feb 7 '20 at 8:41

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.