How Do You Deal With IR Drop?

Let's assume you design a power supply to power a piece of equipment a few feet away from the supply. We will stipulate that your design is well done — stable with various loads, a varying input voltage, and temperature. And it's accurate to within a fraction of a percent. You have several feet of wire (the loop length going out and coming back) and the wire gauge is suitably large. And let's assume the current draw is more than just milliamps. What problems do you encounter?

Probably you'll find that the voltage at the piece of equipment (henceforth called the load) is low (maybe even too low). Why? Because of the voltage or IR drop in the wiring. What to do? Well, you could just turn up the supply voltage a little bit to compensate for the IR drop. But if the load current changes, the voltage at the load will change. That's generally not acceptable — after all, that's why you built that high-performance, regulated supply with the excellent specs.

You could increase the wire gauge by a couple sizes. That would help, but not completely solve the problem. Plus, it'll cost more. Another approach that is sometimes used is a load-sensing configuration. That's where you have two extra wires running from the power supply to load. These monitor the voltage at the load and send that back as a differential signal to the power supply's voltage feedback monitoring circuitry. That gives the proper regulation right where you need it. Of course, you need two more wires, so that'll cost more, too.

Here's another approach that splits the difference (or perhaps blends the difference) with the above methods. Linear Technology Corp. (LTC) has a new device that is used as a high-side current monitor. The device (LT6110) measures the load current and produces a proportional signal that can be used to slightly tweak the output voltage of most any standard voltage regulator. It does this by pulling down the voltage feedback point of the regulator just a tiny amount which makes the VOUT rise slightly. This compensates for the IR drop in the wiring. Through selection of a resistor value at the IC, this voltage compensation can be tailored to your specific set of conditions (VOUT and load current).

The resistor labeled RIN provides the necessary calibration. Without voltage compensation, here's what you typically get when the load changes. This is showing a load change going from 1.6A to 1.0A and back.

The voltage regulator pretty quickly compensates for the change as all good regulators do. But at the load, there is a ΔV of 200mV. With the LT6110 in place and functioning on the same circuitry, the load voltage is stable.

Of course, the voltage at the output of the supply steps around, but that is just as expected. Note that there are some transients and ringing as the load current and voltage change. This can be reduced somewhat with proper compensation capacitor (C1) value selection.

Additional bypass capacitors are certainly present in typical designs. These also mitigate excessive voltage transients.

The part is available in an 8-lead plastic DFN package (2mm X 2mm) and an 8-lead plastic TSOT-23 package. It is fully spec'd from -40°C to +125°C.

Have you had problems with excessive IR drop in power connections between source and load? How did you deal with the problem?

— Brad Albing, Editor-in-Chief, Planet Analog and Integration Nation Circle me on Google+

Related posts:

13 comments on “How Do You Deal With IR Drop?

  1. yalanand
    October 31, 2013

    Have you had problems with excessive IR drop in power connections between source and load? How did you deal with the problem?

    @Brad, Thanks for the post. Designers always encounter this problem. We used to increase the size of the wire guage but never used extra devices to compensates for the IR drop.

  2. Scott Elder
    November 1, 2013

    Brad, The first thing that came to mind when reading your post was “The War of the Currents”.  Tesla won.

    Seems like there are a couple of issues to consider when using this approach. 

    One is what happens when the cable is shorted?  Does the regulator charge up large filter capacitors to an overvoltage point such that when the short is removed, the voltage is momentarily out of specification at the load side?  

    And then a second issue would be that multi purposing the source voltage for local loads is limited since the local voltage will be a function of a remote load and will have all of the load noise reflected onto the source supply output.


  3. eafpres
    November 1, 2013

    Hi Brad.  Interesting and cool part.  I looked at the materials found at the link.  They get credit for having lots of documentation, demo boards, even videos to help you apply this.  Linear must have pretty good marketing and product management.

    They note one of the applications is automotive power distribution.  I'm having a hard time understanding that application since you have a plethora of loads and they change, so which ones should the supply respond to?

    In a dedicated application like certain kinds of industrial control this could be really helpful.

  4. Netcrawl
    November 3, 2013

    @Brad that was great! great link! very educational and informative, content still king, with those documentations and demos anayone can pull something on this,I mean you can have everything you need here.

  5. Netcrawl
    November 3, 2013

    Another effective way to avoid IR-drop and temperature rise problems is to use wide planes of thick copper in all of your power connections, copper has the second lowest electrical resistance of all metals.   

  6. Brad_Albing
    November 4, 2013

    @yalanand – Yep – that's how I always dealt with the problem too – sort of the “big hammer” (brute force) approach to fixing the problem. this part provides a far more elegant solution.

  7. Brad_Albing
    November 4, 2013

    @Scott – Well, if there was a dead-short , the [load] filter caps wouldn't charge up at all. Plus the V-regulator would shut down (current limiting that's present on almost all regulators). But let's take a closer look: What happens when there is a high current draw that's within specs of the regulator — and then the load abruptly shuts off? You'll get a voltage transient that rises to some unspecified level; then returns to the nominal value. How long does the transient last? depends on the compensation of the LT6110 — and the compensation for the main regulator. Best bet — run a simulation and see.

  8. Brad_Albing
    November 4, 2013

    @Scott – re second issue, as with any V-sense or other compensation scheme, it is intended for single point loads, not multipoint distribution. In those cases, generally best to use an intermediate bus and the use local regulation schemes at each POL.

  9. Brad_Albing
    November 4, 2013

    @eafpres – quite right on the multiple loads issue — only is practical with single loads. With those automotive loads, the intent is to use this as part of a single load, but used where the V-regulator happens to be a little ways a way from the actual load.

  10. Brad_Albing
    November 4, 2013

    @Netcrawl – glad I could point you at some useful info. And yes, LTC doea provide good and useful info to make your design tasks a bit easier.

  11. samicksha
    November 5, 2013

    @Scott, You are right, number of things to kept in account before approaching but one thing which often see engineers gets confused,  IR is the drop in voltage that appears at the RESISTIVE COMPONENT of any impedance (not it's reactive component).

  12. RedDerek
    November 6, 2013

    @yalanand – I agree that increasing wire size such that one can handle <1% (what ever is tolerable) voltage drop over the distance is one way. But for automotive applications and others that can be cost sensitive, copper is expensive and this is not the way to go.

    The other option is remote sensing, like some good lab supplies have. This is essentially what the LT chip is doing. They basically did remote sensing and add in the offset due to the wire voltage drop and feed into the feedback pin of the PWM controller.

    Oscillations could occur in very high current applications. One would be in welding. Here is where 100's of amps are sent down a long cable and the inductance in the cable can cause oscillations upon start and stop of the weld.

  13. SunitaT
    November 30, 2013

    Another effective way to avoid IR-drop and temperature rise problems is to use wide planes of thick copper in all of your power connections

    @Netcrawl, true thick copper definitely helps us to reduce IR-drop but routing thicker copper wire is not easy because routing space is usually at a premium. So if routing space is not a constraint then we can opt for planes of thick copper in all power connections.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.