Forum challenge: Ultra low voltage oscillator

Discussion in 'General Electronics Chat' started by tom66, Oct 20, 2010.

  1. tom66

    Thread Starter Senior Member

    May 9, 2009
    2,613
    214
    Here's a cool challenge.

    Design an oscillator, capable of any waveform and any frequency that operates on supply voltages below 1V (it is allowed to have a higher voltage to start up but it must continue oscillating at the target voltage), using commodity parts, which include:

    • resistors
    • capacitors
    • inductors
    • diodes incl. LEDs
    • transistors
    • logic gates
    • op-amps
    • 555 timers (including low voltage variants)
    The output swing does not matter, but bonus points for one which spans the entire supply voltage more or less (i.e. 0V to V+.) You cannot use a voltage booster to obtain 5V then run a 555 timer off it; it must not have any voltage boosting circuitry. The frequency is allowed to vary with voltage.

    Here's my submission, it operates down to ~362mV but requires ~395mV to start up:

    [​IMG]

    The next person should have an oscillator working below 362mV or so.
     
  2. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    tom66,

    Do your rules disallow the use of a joule thief?

    hgmjr
     
  3. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    Can I assume that your oscillator is using germanium transistors?

    hgmjr
     
  4. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    Last edited: Oct 20, 2010
  5. retched

    AAC Fanatic!

    Dec 5, 2009
    5,201
    312
    I guess you would have to.

    Diode junction.
     
  6. tom66

    Thread Starter Senior Member

    May 9, 2009
    2,613
    214
    If it were used to boost voltage then yes, but if it were only used as an osillator I think it would be okay.

    It is using silicon BJT's which is why it does not work very well (the waveform is distorted.)

    Mark your submissions are interesting. 0.16V is pretty impressive.
     
  7. retched

    AAC Fanatic!

    Dec 5, 2009
    5,201
    312
    Without .7v how are your BJTs working?
     
  8. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,536
    I'm in the same boat. I don't see it working in real life. A TLC555 will go down to 1.5VDC, but it is out of the spec range of the datasheet.

    I'll be interested in this thread, but I have enough trouble getting some of my weird concoctions working with 5V.
     
  9. tom66

    Thread Starter Senior Member

    May 9, 2009
    2,613
    214
    Here's the secret: transistors don't necessarily need 0.7V to switch on. 0.4V will make them conduct a small amount of current, enough for the oscillator to work. The oscillator requires 395mV to start up, because the capacitors are used to drive the bases of the transistors. When the caps switch off, about 500mV is produced at the base of each transistor, enough to make it conduct the small current for the 47k resistor. The oscillator doesn't stop at 360mV, but it is no longer able to maintain the voltage at the caps required to drive the transistors so within about 10 cycles the oscillations stop.

    This isn't in violation of the voltage boosting rule because the oscillator runs off 0.36V and the caps happen to drive the bases as required to keep the oscillations going.

    Here's another version which works as low as 330mV. It's difficult to get it to go lower, as base voltages below ~410mV don't work, but it may be possible to get it to go even lower.

    [​IMG]

    Code ( (Unknown Language)):
    1.  
    2. $ 1 4.9999999999999996E-6 19.867427341514983 50 0.5 50
    3. t 288 288 256 288 0 1 0.01764707741450966 0.27535693973683845 100.0
    4. t 384 288 416 288 0 1 0.28389833612909143 0.3971094858264473 100.0
    5. r 256 240 256 176 0 330000.0
    6. c 256 256 304 256 0 1.0E-8 -0.1393996235041185
    7. r 304 240 304 176 0 330000.0
    8. w 304 240 304 256 0
    9. w 256 256 256 240 0
    10. w 256 256 256 272 0
    11. w 368 256 368 240 0
    12. w 416 240 416 256 0
    13. r 416 240 416 176 0 330000.0
    14. c 368 256 416 256 0 1.0E-8 0.1621457900394826
    15. r 368 240 368 176 0 470000.0
    16. w 304 256 384 288 0
    17. w 368 256 288 288 0
    18. w 256 304 256 320 0
    19. g 256 320 256 336 0
    20. w 416 304 416 320 0
    21. g 416 320 416 336 0
    22. w 416 256 416 272 0
    23. w 416 176 368 176 0
    24. w 368 176 304 176 0
    25. w 304 176 256 176 0
    26. w 416 256 464 256 0
    27. O 416 176 464 176 1
    28. 172 256 176 224 176 0 6 0.3304 0.395 0.3 0.0 0.5 Voltage
    29. o 23 32 0 35 0.9568131466127622 9.765625000000001E-205 0 -1
    30. o 14 64 0 34 1.093625362391506 4.2719740718418204E-5 1 -1
    31. o 13 64 0 34 1.1692013098647223 4.5671926166590717E-5 2 -1
    32. o 25 64 0 33 0.29230032746618057 9.134385233318143E-5 3 -1
    An added advantage of this ultra low voltage oscillator is it also works on 5V and in the low voltage mode consumes <1.8μA (when operating on 5V <58μA.)

    I'll be back to see if it's possible to get below 200mV using BJTs.
     
  10. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
  11. Tesla23

    Active Member

    May 10, 2009
    318
    67
    The ones from Dick's Website intrigued me, I had to try one.

    [​IMG]

    With a J310 FET and a 70turn:1turn transformer wound on an old Amidon toroid (so only 2 components), it reliably oscillated at 3.2MHz from 43mV drawing 1.6mA. It needed 43mV to start and dropped out below 41mV.

    At 43mV it happily allowed a 1m 50 ohm coax connected to the drain to a 1Meg input impedance on a scope to observe a reaasonable looking sinusoid of about 15mV p-p. Connecting the coax only changed the frequency about 3kHz.

    Placing a 10:1 scope probe on the gate reduced the frequency to 2.4MHz and increased the supply voltage needed to 66mV, but then showed a 1.13V p-p reasonable sinusoidal waveform.
     
    Last edited: Oct 22, 2010
  12. Tesla23

    Active Member

    May 10, 2009
    318
    67
    Schematic added for earlier post.[​IMG]
     
    • osc.png
      osc.png
      File size:
      7.9 KB
      Views:
      296
  13. eblc1388

    Senior Member

    Nov 28, 2008
    1,542
    102
    I played around with this circuit a bit using LTSpice and can get it to oscillate at 10mV supply voltage.

    I don't know if it would oscillate in real life.

    It might be possible to get it to oscillate at lower voltage still. Who knows?

    [​IMG]
     
  14. tom66

    Thread Starter Senior Member

    May 9, 2009
    2,613
    214
    Very impressive!
     
Loading...