Synchonous demodulation / phase detector for light sensor

Log in or register to post comments
Go To Last Post
7 posts / 0 new
Author
Message
krazatchu
krazatchu's picture
Level: Resident
Joined: Sat. Dec 29, 2007
Posts: 558 View posts
Location: Ontario, Canada
#1
Posted by krazatchu: Wed. Mar 12, 2014 - 03:54 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi freaks,

I have a IR reflected light sensor, an IR LED and a photo-transistor to which I'd like to add some modulation for ambient noise rejection.

My first thought is to just use an IR remote demodulator, like the TSOP58038 or the TSOP4038 which don't have AGC.
But I'd prefer to have lower power consumption (under 0.7mA) and a faster carrier, perhaps 50kHz or more, to allow less phase delay for a faster MCU wake up.

As the MCU is driving the LED from PWM, I have the option to perform synchronous demodulation on the receiver.
Which seems like an XOR/XNOR function, followed by a lowpass.

But I suspect making this low power will be a challenge, especially if I run the the LED at non-symmetrical low duty cycle.

Any suggestions or examples?
I might be searching with the wrong terminology.

Thanks!

Michael
--
http://krazatchu.ca/

Tags:

Learning and Information, General Electronics
Top
krazatchu
krazatchu's picture
Level: Resident
Joined: Sat. Dec 29, 2007
Posts: 558 View posts
Location: Ontario, Canada
Posted by krazatchu: Wed. Mar 12, 2014 - 07:26 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I should add, the target is oscillating at up to 1kHz.
Which is another reason for the higher frequency carrier.

The 38kHz continuous demodulator (TSOP4038) needs about 7 carrier cycles before reporting on the output.
Which gives a nysquist ratio of ~5.4:1 and a resulting phase delay.

My intention is to wake the MCU when the target is present and measure the frequency of the target by reflection.

Michael
--
http://krazatchu.ca/

Top
krazatchu
krazatchu's picture
Level: Resident
Joined: Sat. Dec 29, 2007
Posts: 558 View posts
Location: Ontario, Canada
Posted by krazatchu: Thu. Mar 13, 2014 - 04:31 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Step 1:
Replace the slow photo-transistor with a fast photo-diode and use a trans-impedance amplifier.

Michael
--
http://krazatchu.ca/

Last Edited: Thu. Mar 13, 2014 - 07:40 PM
Top
John_A_Brown
John_A_Brown's picture
Level: Raving Lunatic
Joined: Sat. Sep 20, 2003
Posts: 6054 View posts
Location: Surrey, England
Posted by John_A_Brown: Thu. Mar 13, 2014 - 06:24 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Good luck!

Four legs good, two legs bad, three legs stable.

Top
Kas
Kas's picture
Level: Resident
Joined: Sun. Nov 6, 2005
Posts: 710 View posts
Location: Vilnius, Lithuania, Continental EU
Posted by Kas: Thu. Mar 13, 2014 - 09:30 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
As the MCU is driving the LED from PWM, I have the option to perform synchronous demodulation on the receiver.
Which seems like an XOR/XNOR function, followed by a lowpass.

Rather an analogue switch - one have to deal with analogue signals, imho.

Quote:
non-symmetrical low duty cycle

This should not have impact since detection is clocked by PWM frequency.
MCU pulsing IR LED > IR optical filter > IR photodiode > synchronous detector > LPF > MCU. Low power op-amps for analogue part shall make it achievable. A smart code for locking frequency readout in software PLL could be a nice add-on.

Top
krazatchu
krazatchu's picture
Level: Resident
Joined: Sat. Dec 29, 2007
Posts: 558 View posts
Location: Ontario, Canada
Posted by krazatchu: Fri. Mar 14, 2014 - 03:26 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Instead of XOR, I think analog multiplication maybe more ideal. Where a low is treated as a negative, a high as a positive. Two negatives make a positive as does two positives...

Michael
--
http://krazatchu.ca/

Top
John_A_Brown
John_A_Brown's picture
Level: Raving Lunatic
Joined: Sat. Sep 20, 2003
Posts: 6054 View posts
Location: Surrey, England
Posted by John_A_Brown: Fri. Mar 14, 2014 - 09:36 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

krazatchu wrote:
Instead of XOR, I think analog multiplication maybe more ideal. Where a low is treated as a negative, a high as a positive. Two negatives make a positive as does two positives...

Which, at the digital level is XNOR, as you've already hinted at.

Four legs good, two legs bad, three legs stable.

Top