voltage regulator oscillation

Discussion in 'The Projects Forum' started by 1-3-2-4, Nov 22, 2013.

  1. 1-3-2-4

    Thread Starter Senior Member

    Dec 26, 2008
    So I have a 7812 Vreg without a 0.1uf cap the voltage regulator would heat up pretty fast within seconds, is it standard to put a 0.1uf cap on it regardless of length?

    I built this SLA battery charger from here


    to add a bit more to this I wanted to add a PWM motor controller which I have here (MIC502) and a voltage monitor to cut power from the battery when it falls under 12.4V (MAX8212)

    being as the MIC502 has a max input voltage of 13.2V I have it on a 12V voltage regulator..

    This whole project is a emergency lighting, as long as power is going to the SLA charger the lights will be off, but as soon as the power drops the lights would come on

    attached is the battery monitor the PWM fan controller I don't think should be too hard to connect, I'm not too sure what size PCB I need.. the fan I will be using is a 12V .60A fan that is temp controlled.

    If this sounds like a mash up of stuff it is, I took a break from the project for a few months to do some other stuff so I was wondering if any other improvements could be made and not to take much space up in the project box I have.
  2. JohnInTX


    Jun 26, 2012
    Rather than cob together a weird bunch of stuff, you might want to take a look at ti.com and search emergency light controller. You'll get some hits on things like the bq24450 which is a SLA charger controller with low voltage cutout. bq is benchmarq, an outfit bought by TI that specialized in battery charging ICs. I don't know about your fan req but see what they say.

    Re: oscillation, the caps are not optional. Put them as close to the LM reg as possible. Take a look at the LM317 datasheet. The circuit you show is clever, the LED on makes the voltage at the ADJ pin such that the LM runs full tilt then when the transistor turns off, Vadj is determined by the voltage divider/pot. The problem is that right at the turnoff point things are unstable. There is no hysteresis. That may be where the oscillation occurs. As it slowly moves over from high current to float current the battery voltage will fall ever so little but enough to fire up the high charge rate again and then Vbatt rises making it float again and then Vbatt drops ... (as I read it.. kind of speculating here..). So it will oscillate as it transitions from full blast to trickle charge.

    What's the fan for in an emergency light? Just curious.

    EDIT: re-reading your post - we strive for accuracy here, :) you indicate a 7812. That won't charge a battery like you have indicated (it wants an LM317 - different altogether). At any rate, a 7812 will be most unhappy if the input voltage ever is less than the output voltage (as it would be if the battery were connected to the output and the input was off). You can protect the regulator by putting a diode across Vin-Vout (cathode-anode) but.. it still won't charge a 12V SLA.

    Personally, I'd consider a PIC and a pass device.. Measures volts/amps and does the logic. A lot like the benchmarq chips.. You might want to visit microchip.com and check out battery management.

    But first.. what regulator are you using and are you starting with the circuit you posted?
    Last edited: Nov 22, 2013
  3. 1-3-2-4

    Thread Starter Senior Member

    Dec 26, 2008
    This is separate from the charger the charger has a LM317, it's already been tested and works very well.. the Fan is to cool the heatsinks off during the charging if I remember right durning testing I was seeing temps up to 56C and that was outside of the project box, this time it's going to be inside a box. The fan is a Delta FFB0412SHN 40mm fan and pushes 24 CFM it's loud but it's going to be in a closet and I doubt it would be at full speed that often or long.