help amplifying a triangle waveform

Discussion in 'The Projects Forum' started by badbaud, Aug 21, 2012.

  1. badbaud

    Thread Starter Member

    Mar 5, 2009
    26
    0
    I have an adjustable triangle waveform generator circuit that puts out 4V pp and can be adjusted from .1 Hz to 22Hz.

    It uses a CD40106 with a pot and a cap, the output waveform taken off of the cap.

    My power supply is 12VDC.

    I need to maintain the same waveform symmetry but amplify the waveform to travel from rail to rail, 0VDC to 12VDC and back to 0VDC.

    Tried connecting the signal to the base of a darlington transistor, cap coupled with a .1uF, 1N4148 to ground to diminish the below ground swing.

    All I got was a square wave... that went rail to rail.

    Tried a 2N3904 but the signal was disjointed and floating above ground.

    What kind of amplifier am I trying to make to amplify this signal (servo, DC..?)?

    The amplifier should have a high input impedance but be able to track the DC input voltage and duplicate it as a higher output voltage.

    Where can I get some kind of schematic to try out?
     
  2. Austin Clark

    Member

    Dec 28, 2011
    409
    44
    This is going to be terrible, but I'm gonna give it a shot anyway.


    You can't connect the output directly to the base of a BJT, for one you'll essentially be shorting it out if you don't use resistors at the base and/or emitter, and you'll quickly bring the transistor into saturation, which is why you're getting a square-wave out (it's saturating really quickly after it gets past the forward voltage of the base-emitter junction, making it look like a short or a piece of wire).

    You need to forward-bias the base using a resistor voltage divider, and use a capacitor and resistor in series with your output signal to the base. This will ensure that no DC is allowed in and that the lower portion of the wave isn't cut off.
    Next, the output should have a capacitor in series as well (probably need a large value cap, if you didn't want to add another stage), this will block the DC voltage coming out.

    In the end, the voltage coming in will dictate how much current is going through the base resistor (and the forward voltage will be taken care of), which will in turn dictate how much current will want to flow through the transistor Collector-Emitter junction. If you have a resistor in series with that, and you know that the current through it is, you can calculate the correct resistor value needed to get the 12V at peak current across it.
     
  3. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    An opamp can amplify the signal. A non-inverting opamp circuit has an extremely high input resistance.
    Some opamps have an output that goes rail-to-rail if the load resistance is high.
     
    Austin Clark likes this.
  4. Austin Clark

    Member

    Dec 28, 2011
    409
    44
    Yeah, or you could do that...

    MY WAY WAS COOLER!:p
     
  5. badbaud

    Thread Starter Member

    Mar 5, 2009
    26
    0
    I tried the transistor approach but by the time I got a high enough base resistance where the sawtooth waveform would not be distorted the transistor would not fire up.

    Tried a Darlington transistor and learned how easy it was to turn a triangle waveform into a square wave.

    Currently have a op-amp hooked up (uA1458). The input triangle waveform looks symmetrical but I get a positive going spike (1/2 triangle waveform). The other half of the triangle waveform was flat-lined on the op-amp output. The spike amplitude was 10 Volts, well within the voltage I need to control the LM317 adj. pin.

    How do I get a symmetrical output using the op-amp approach.

    100K input resistor, 1M gain resistor.

    Should I try a audio amp?
     
  6. wayneh

    Expert

    Sep 9, 2010
    12,087
    3,027
    I'm guessing the problem is that the op-amp cannot sense rail to rail and is confused during the spike portion. One or both of the inputs are out of its common mode range. You can apply bias resistors to fix this, or choose a different op-amp. Some can sense only within a narrow range between the two rails, some to one rail or the other, and some to both rails.
     
  7. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    A uA1458 has opamps that are extremely old. They are noisy, have a poor bandwidth and have a poor slew rate.
    Their inputs cannot go anywhere near rail-to-rail and their outputs too.

    I have never used rail-to-rail opamps so I can't recommend a good one.
     
  8. badbaud

    Thread Starter Member

    Mar 5, 2009
    26
    0
    I ordered some TLC2272 op amps, TI says they are rail to rail.
     
  9. DickCappels

    Moderator

    Aug 21, 2008
    2,647
    632
    Those are excellent opamps.

    Be aware that the input common mode range does not work near the positive power supply. Also, "rail-to-rail" means "really close to each rail as long as their is no significant load".
     
Loading...