Why, in a lot of applications (e.g for bioengineering where very small devices are developed for the recording of biosignals), after an amplifier there is a comparator (with two inputs, one for the analog signal and the other for a triangular waveform) to modulate an analog signal to a PWM signal?

Why should we convert the analog signal to a PWM signal? What is the purpose?

So far I have understood that we want a PWM signal because it is the first step to convert an analog signal to a digital signal through an ADC. Is that true?

enter image description here

Paper: "Using Pulse Width Modulation for Wireless Transmission of Neural Signals in Multichannel Neural Recording Systems", Ming Yin, Student Member, IEEE, and Maysam Ghovanloo, Member, IEEE


  • \$\begingroup\$ Examples? Because electronic engineers are usually pretty reasonable people. In most cases they don't do what is not needed. \$\endgroup\$ – Eugene Sh. Sep 29 '17 at 14:14
  • \$\begingroup\$ Not really.. By examples I meant some concrete design (schematic or a block diagram) which is puzzling you. \$\endgroup\$ – Eugene Sh. Sep 29 '17 at 14:21
  • \$\begingroup\$ Well, this presented example is taken from a paper whose abstract is claiming this method has a reduced power consumption. I don't have an access to the full paper, but I would expect an analysis proving this claim to be presented in the paper. \$\endgroup\$ – Eugene Sh. Sep 29 '17 at 14:35
  • \$\begingroup\$ @Eugune Sh. I haven't got something specific in my mind.My problem is that generally, I can't understand the reason for someone to produce PWM signals. \$\endgroup\$ – vl123 Sep 29 '17 at 14:35
  • 4
    \$\begingroup\$ PWM is usually produced to generate an "analog-like" signal using two voltage levels only - such as digital output pins of microelectronics. I.e. it is a convenient method of controlling analog components with digital controllers. \$\endgroup\$ – Eugene Sh. Sep 29 '17 at 14:37

It appears they are trying to avoid noise. I found an open (legal) copy of their paper on the NCBI website

However, analog samples are susceptible to noise, and the transitions from one channel to another in short sampling periods can result in significant crosstalk among adjacent channels on the receiver side

This is for implantable devices with the signals going directly to RF. Adding an ADC as suggested in the comments would be counter productive since then a microcontroller would have to read said ADC and then modulate on the signal.

doing a direct AM or FM would likely be difficult since the values are so small and they are multiplexing 15 channels. by going to a PWM they get a digital representation that they can send using Phase Shift Keying (PSK) which is 'digital-like' in that the value are very distinct 1/0 (0 deg/ 180 deg) over the transmission medium. Since they need low power the more distinct the signal out the better.

In short they get the value of a digital representation without the added cost of digital components

  • \$\begingroup\$ Thank you so much @Bageletas.Very good explanation! I appreciate so much your help! \$\endgroup\$ – vl123 Sep 29 '17 at 15:46
  • 4
    \$\begingroup\$ Those keywords very small devices and biosignals suggest that a power budget is going to be very important. PWM is extremely power-efficient compared to analog. As Bageletas has answered, PSK (or FM) is compatible with PWM modulating signals. The RF power amplifier likely uses PWM techniques as well to maximize power efficiency. \$\endgroup\$ – glen_geek Sep 29 '17 at 15:51

Another reason to convert analog signal to digital PWM is that it can easily passed though an opto-coupler when isolation is needed between power domains.

You can pass an analog level through an opto-coupler but the response curve is NON linear and temperature dependent.


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.