0
\$\begingroup\$

I want to sum/mix 2 PWM signals in order to reach a better resolution on a DAC (see here or here).

I need help to understand the limitations of a simple resistors network to sum my two signals as described in the previous links. Indeed, I succesfully implemented this to sum two 3-bits PWM running at 10kHz as a test and I am now wondering what can be the advantages of summing those signals by using an op amp or by using "complex" circuits like the one described in the second part of the first link.

I understand that with the passive resistor summing, I have to be sure my two PWM signals have the same impedance but it should be the case as they come from the same MCU. I also understand that this method will not give the best results in terms of distorsion but as long as I am not using this to carry audio signals or other sensible signals, it should be okay, right ?

Briefly, I'd like to know your opinion on using passive resistors suming to mix let's say two 8-bits PWM signals at around 10 kHz in order to emulate a 16-bits PWM signal and what can be the pros and cons of using another "sophisticated" method. The resulting signal after filtering has low frequency (<1 kHz) but requires good SNR.

\$\endgroup\$
  • \$\begingroup\$ Is your 8-bit PWM source designed with one part in \$2^{16}\$ accuracy (or precision, depending how you're using it)? \$\endgroup\$ – The Photon Apr 17 '18 at 16:05
  • 1
    \$\begingroup\$ Start by defining your signal BW , and THD and SNR desired \$\endgroup\$ – Sunnyskyguy EE75 Apr 17 '18 at 16:15
  • \$\begingroup\$ @ThePhoton Yes, I assume for this example that the source of the PWM is not a problem. \$\endgroup\$ – J. K. Apr 17 '18 at 16:16
  • \$\begingroup\$ openmusiclabs.com/learning/digital/pwm-dac/… PWM has inherent distortion YOu are far better off with an I2C DAC or parallel DAC \$\endgroup\$ – Sunnyskyguy EE75 Apr 17 '18 at 16:17
  • \$\begingroup\$ @TonyStewartEEsince1975 I do not have precise ideas yet of those parameters but the BW might be <1kHz, the SNR has to be quite high, let say over 1000. Once again I only have a rough idea for those parameters right now, are they critical point to consider in order to descriminate one solution or another ? Thank you \$\endgroup\$ – J. K. Apr 17 '18 at 16:30

Your Answer

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

Browse other questions tagged or ask your own question.