• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 64 subscribers
  • Views 2383 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM1086: Linear regulator suitable for fan motor power

Josiah Moore
Prodigy 50 points
Part Number: LM1086
Other Parts Discussed in Thread: LM2674, LM2574

I am seeking a linear voltage regulator to provide 12V power to computer cooling fans. I will be using a switch mode power supply to provide an input voltage that could be varied between 15 and 27 volts (absolute maximum, but could be reduced if needed). I would like to power (2) fan motors that each operate at 12 volts and draw 0.24 amps during normal operation. The LM1086 regulator seems to be a good option to provide power to both fans, but I would like to know if there are issues (e.g., inrush current potentially causing a problem for the regulator?) that I am overlooking. I need a very robust solution that will allow continuous fan operation for the duration of my test (potentially 24 hours or more). A loss of power to the fans would cause my thermoelectric elements to catastrophically fail (unless I provide thermal overload protection).

  • 0
    TI__Expert 7015 points
    Hi Josiah,

    The biggest problems that I see here are potentially the LDOs running into thermal shutdown when they are running from 27V and that there may be some power supply overshoot when switching between 15 and 27V. If there is any overshoot on the power supply it can cause the LDOs to fail. Also, if the LDOs enter thermal shutdown the fans will turn off, which can cause the thermoelectric elements to fail.

    Have you considered adding a heat sink to the LDO's, or adding a pre regulator before the LDO?
    Also, do you know what will be the maximum voltage during your transients?

    Best Regards,
    Mark
  • 0
    Guru 140200 points

    Hi Josiah,

    have you read this app note?

    www.ti.com/lit/an/slva683a/slva683a.pdf

    The supply current of your fan is much lower, though. :-)

    If you take the linear voltage regulator LM1086 I would keep the input output differential voltage at about 3V. Then you are close to the maximum of short circuit current (see figure 2 of datasheet). This in order to be able to deliver inrush currents. You could even spend each of the two fans an own LM1086, but I think this isn't necessary.

    Take the LM1086 in TO-220 package and provide a suited heat sink. With two fans powered by one LM1086 (the two fans in parallel) you will have to dissipate 2 x 0.24A x 3V =1.44W. With a 10K/W heat sink you get a junction to ambient overtemperature of under 20K, which seems to me to be more than comfortable.

    Kai

  • 0 in reply to Mark Sellers
    Prodigy 50 points
    Hi Mark,

    Thanks for your response. I should have clarified what I meant by "varied between 15 and 27 volts." The voltage will not actually be varied during operation of the fan. Rather, the voltage will be set on the power supply prior to applying power to either the thermoelectric element or the cooling fans, and the supply voltage will remain constant for the duration of the test. The voltage will be selected based on the required cooling capacity of the thermoelectric element, which could vary depending on the designated test characteristics. But, the fans will always need to operate at 12V.

    Yes, I was planning to add heat sinks to the LDOs. What would constitute a "pre regulator"? Is the 10 µF capacitor shown in the typical application circuit diagram considered a "pre regulator"? Or, is a pre regulator still necessary in light of the clarification I provided above?

    Thanks for your help!

    Josiah
  • 0 in reply to kai klaas69
    Prodigy 50 points
    Hi Kai,

    Thanks for your reply and the application report reference.

    I'm not sure that I completely understand Figure 2 in the datasheet. Does the graph illustrate the maximum current that can flow through the LM1086 at various input-to-output differentials before the LM1086 shorts out, drops out, or fails?

    Thanks!

    Josiah
  • 0 in reply to Josiah Moore
    Guru 140200 points

    Hi Josiah,

    the LM1086 has some circuitry inside to protect the chip against internal overheating caused by too high output currents and too high heat dissipation. The chip manages this by limiting the output current. Figure 2 shows, that the maximum short circuit current can be delivered when the input voltage is about 5...7V higher than the output voltage. When furtherly increasing the input voltage the maximum deliverable output current will decrease. This is the performance which is usually expected from a SOA-protection.

    This limiting of output current is normal operation and does not destroy the regulator. When the short circuit current limitation invokes, the output voltage goes down in order to reduce the output current. That's all.

    So, If you have an application with high inrush currents, it's wise to choose an input voltage which is not too high. That's why I suggested you to have an input voltage of 15V for 12V output voltage.

    It's a good idea to produce the 15V by the help of a switcher, or by other words to use a switcher as pre-regulator. This is what Mark asked for. The idea behind is to have the very high efficiency of the switcher and at the same time the low noise of the linear regulator, while having only low heat dissipation in the linear regulator because of the low input voltage.

    One last word about inrush current handling capability of regulators. It should be noted, that the switcher often is the better solution when high inrush currents have to be delivered. This has to do with the SOA-protection scheme of linear regulator explained above. Look again at figure 2. If an input voltage of 15V is applied to the regulator and the output is forced to deliver a very high current, the output voltage can momentarily drop down to 0V. Think of the charging of a big cap, for instance. Figure 2 tells, that then only 1.2A can be delivered. All of the input voltage drops across the regulator then, which has to dissipate 1.2A x 15V = 18W in this situation. This can make the chip tremendeously heat up. If you would choose an input voltage of 20V, then the LM1086 can only deliver a very very small output current, according to figure 2 it's only about 100mA! So, if too much output current is requested from a linear regulator, it's possible that the output falls down to 0V and stays there.

    A switcher works totally different. It has to dissipate only the voltage drop across the switching element, which is very small. The input output differential voltage does not play the same role as with the linear regulator. Because of this, switchers can usually better handle inrush currents.

    Kai

  • 0 in reply to kai klaas69
    Prodigy 50 points
    Hi Kai,

    Thanks for your help. I did a little bit of research on switchers, and I believe I understand what you're saying. So, the question in my mind is whether or not noise/EMI generated from the switcher would be detrimental to the cooling fans and if I really actually need the LDO to reduce that noise/EMI.

    Josiah
  • 0 in reply to Josiah Moore
    Guru 140200 points
    Hi Josiah,

    in one of our products we have a SUNON fan (5V/0.18A) powered by a LM2674.

    Kai
  • 0 in reply to kai klaas69
    Prodigy 50 points
    Hi Kai,

    Is an LDO also used in the product though too? I read in the LM2674 datasheet that it is often used as an efficient preregulator for linear regulators.

    In short, what would be your recommendation for my application in which I need to step down to 12 volts from an input voltage that could vary from 15 to 27 volts (though it won't be varied DURING operation; instead, only before power is turned on) to power 1 or 2 fans that each pull 0.24 amps during normal operation?

    Thanks for your time and help.

    Josiah
  • 0 in reply to Josiah Moore
    Guru 140200 points
    Hi Josiah,

    the LM2674 is directly driving the fan. Not as a pre-regulator. We had to add the fan afterwards, after the project was already finished. We had a 18V supply and decided to add a 5V Fan to it, powered by a LM2674 switcher. The LM2674 has a 33µH choke and a 100µF cap.

    Because the cabling to the fan is rather long we decided to add an output ripple filter to the LM2674, this in order to eliminate any radiation from the switching. The output ripple filter consists of a 0R47 resistor in series with a 68µH choke and another 100µF cap to GND. The fan is connected to the output of the output ripple filter.

    Have a look at figure 18 of datasheet of LM2574 to see how an output ripple filter looks like.

    I think there is nothing that speaks against the use of a simple switcher in your application.

    Kai