This post is from the Garage Journal, and was created by "stick", one of the members here.  It's written in the first person, something we don't normally do, but it is a really good discussion of how parasitic draws happen, and how they can be found.

For the InfinityWire users, things are both easier and harder.  Easier in that by just looking at the relevant LED on the powercell that corresponds to the output in question, you can see if it is "on" or not, or if there is another problem.  It's harder in the sense that the entire InfinityWire system represents a draw all the time, as it is actively running microprocessors, and managing data on a CAN bus network, and even more draw if you have the InTouchNet option, which is running a wifi network (though InfinityWire recommends running a switch to turn the transmitter off when not needed, saving power).  With that as context, enjoy the article!

Standard Disclaimer applies - As with everything automotive, "it depends" is the only 100% correct answer for a given situation. This is only how I look at things, and how I approach a problem. You may have a better routine that works for you, and I would encourage you to find a routine or system that works for you.
A couple of quick notes before we get started.
1. I work on a pretty limited ammount of "normal" problems in my day to day routine, and what I consider as "normal" some of you may find interesting as a discussion topic. So, if you have any suggestions for our monthly thread, feel free to suggest them in the Any interest in a monthly how-to thread? discussion.
2. Along the same lines, if you've worked on a project, or are planning a project that may be interesting, take some pictures and we can make it a how-to thread for the month. I primarily do driveability, and brakes/chassis/PM services on our fleet, while most "normal" problems don't really surface, or are taken care of well ahead of time, so as a result the stuff I write about is the stuff I run into everyday.
Sorry about missing December's thread, but I had all sorts of things going on with kids hockey, work, and the holidays. I've had a couple of PM'd questions on this topic, and had a vehicle in recently where I needed to check for a parasitic draw, so I took some pictures for this month's how-to thread.

Let's go ahead and define what parasitic draw is... All modern vehicles will draw some power from the battery when the vehicle is shut off. The reason is pretty straightforward, the vehicle needs a small amount of power to maintain stored values in the PCM and to keep settings in things such as clocks, radios, and alarm systems. The low power draw used to keep those settings is normal and expected, and usually the amp draw is low enough that the vehicle can sit for long periods of time without problems. When other items are using more power than expected, your battery will discharge faster than normal and in a worst case sceneario, leave you with a dead vehicle.

So how do we test it? First we need to provide a way to access the interior of the vehicle during our testing, so we can check fuses and connections. Opening a door will usually turn on dome lights (and possibly blow the fuse in our meter), but if we make the door think that it's actually closed, the dome lights should stay off and we'll still have access for testing.

I've got one of these handy dandy "door buzzer sanity tools", but you can just stick a screwdriver in the bearpaw latch and gently close it. The only reason I use the tool is that I've already got it, and it also keeps me from slamming the door closed with the latch already shut.

To figure out how much power the vehicle is consuming when shut off, and determine if it's a normal amount vs. an excessive amount of current draw, we first start the vehicle, then shut it off to bring everything to a "normal" state. We then hook up our test leads with the meter set to measure Amps. One test lead goes to the battery post, and the other to the battery cable, and being *very careful* we unhook the battery cable from the battery leaving our test leads hooked up in series with the battery cable and the battery terminal

Because hooking up test leads on some vehicles can be a real PITA, I went ahead and made up this adapter. It's nothing more than a $10 battery disconnect switch with a couple of #6 machine screws placed on either side of the disconnect. It took me about five minutes to put the screws in place, and also make up an adapter for testing side post batteries (which is just a short piece of 6g battery cable with some ends crimped on it.

Hook up your leads like so. I use some alligator clip style test lead extensions clipped to the two screws, and once the test leads are hooked up, start the vehicle/shut it off (to bring everything back to "normal"), and then give the green knob a twist to unhook the battery and put the meter in series with the rest of the system.

I usually use a graphing meter for something like this, but really any meter will do (you'll just need to check on it a bit more often). The timebase on this graph is 1 hour, and you can see the baseline amp draw of 280mA while the communications buss is still active. The amp draw then steps down to ~150mA after 45 minutes or so, and then again 10 minutes later to the final value of 37mA. What's considered excessive? That depends on the vehicle. Most GM products seem to run 25-35mA, where Ford and Chrysler vehicles run closer to 50mA or so. If you look at some of our fully upfitted police vehicles, they run around 75mA because of all the additional electronics. Generally I like to see under 50mA after all of the modules go to sleep. This vehicle appears to be ok, but occasionally you'll find a vehicle that has a module "wake up" after several hours, so if you continue to have a problem you might want to test overnight to see what's going on (min/max values on the meter work well).

So whats next if you do have too much parasitic draw? You need to figure out where it's coming from... The best way to start is to figure out what circuit the draw is on.

I've got more than one meter, but if you've only got one, you'll need to hook the battery cables back up and go from there.

I prefer to avoid pulling fuses whenever possible, and instead opt to do a voltage drop test on them instead. Some modules will wake up when fuses are pulled/re-installed, and doing things this way keeps that possibility to a minimum. If you've never done a test this way it's pretty simple. Set your meter to the millivolt scale and (pressing the "range" button on an autorange meter will allow you to manually set voltage range). Put one probe on each divot on the top of the fuse and see what it reads. Make sure to check *every fuse* in the fuse box to see what they read.

What you don't want to see is the numbers rapidly fluctuating like this (ghost voltage). Any time you see voltage moving on the meter and it isn't supposed to, it means you have an open. In this case the two most common possibilities are that either the fuse is blown or you don't have the probes firmly on the fuse test points. 

This is what you want to see, true zero voltage. When you see true zero voltage, it means two things: that you have a good connection (try it with just your test leads), and that no current is flowing in that circuit. No voltage drop = no current flowing, it's that simple.


Now what happens when we find something like this? Obviously it's not what we want to see (true zero voltage), but it's not ghost voltage either... That's where this chart comes in handy. (pdf)Assuming that we're checking a 10A mini fuse, a reading of 2mV (.002V) means that there is .280A (280mA) of current flowing through that fuse. In that same pdf file it says that anything over 40mA is worth investigating, which seems to be a pretty reasonable guideline to go by.

Seems pretty accurate considering that the 194 bulb that's in this glovebox draws .27 amps at normal battery voltages...
Now not every parasitic draw is going to be as simple as a glovebox bulb that's staying on (I did that just for the writeup), but it's not that much more work to track down the source of the draw. If you followed along with October's how-to thread, you should be comfortable with reading wiring diagrams. It's just a small step from there to finding the assorted components in that circuit and unhooking them one by one until your draw goes away. Occasionally you'll need to go as far as clipping wires in a harness, but for the most part it's just a matter of accessing the right connectors and unplugging them.