No announcement yet.


  • Filter
  • Time
  • Show
Clear All
new posts

  • Current-Monitor?

    I've been searching here and elsewhere for a decent current monitor setup, and I can't find anything. What the hell!?

    I've got an auxiliary battery that is connected to the main battery over 4-gauge wires, I'd like to be able to monitor the voltage of the aux battery and the current (bi-directionally) going to/from the main battery. This is just to make sure it's charging properly, and to make sure the battery isolator is working.

    Connected to the aux battery I've got a fused distribution block that breaks out into (4) 8-gauge wires that I use to run various circuits for the CarPC. Right now I'm only using 2 - one for the M4 atx power supply and the other that runs the LCD and backup camera, and I want to be able to monitor the current going out to each.

    Why does a black box for this not exist?!?

    Anyway - I guess I'll be rolling my own. The device will consist of a project box that has a few shunts - I'm thinking a 200A and (4) 50A. These will be connected to a custom PCB that contains a shunt monitor op-amp for each one, and a high-precision voltage divider, those inputs will be plugged into an arduino micro, which will control an LCD panel for displaying the voltage, current, and system power, and allow a connected PC to poll that info.

    As far as cost - the arduino micro will run $25, the shunts are under $5 each, shunt monitors are a couple bucks a piece, and the custom PCB will probably be about $25. I'm hoping to build a prototype for about $125.

    Would anyone else be interested in a device like this?

  • #2
    The reason no black box exist is that current is useless for monitoring the system health. It's only use is for experimenters and the exceptionally inquisitive.

    Voltmeters are used to monitor system ad battery health.

    And if you are going to design such a system, don't use shunts. Use Hall-effect sensors. Nobody uses shunts except when the shunt is a normal cable (eg, ground strap or +12V feed).
    Old internal shunt (series) ammeters went out of fashion decades ago, and external shunts not much later.


    • #3
      I was planning on doing the shunts because they are cheaper - about half the price - and I was having problems finding what I was looking for on Mouser... but you're right - it's probably better to pay the extra $$ and go hall-effect. I'll look around and see if I can find something reasonably priced. Even if it's $15 a piece, it's probably worth it.


      • #4
        I suggest you scrap the idea.

        The current to the monitors etc can be done with a once-off reading, though usually their specs are enough. In vehicle systems they are almost irrelevant, it's only off batteries they'd be a concern, and only then for design reason, not battery protection etc.

        How will you use current to determine if a battery is charging properly? If you understand that aspect, you should understand why current is essentially useless - you would certainly have voltage monitoring long before any current monitoring system.
        Same to determine if your isolator is working...

        A big problem with battery ammeter designs is the range of current involved. A main battery shunt may have to handle 100A (excluding the starter motor) yet you may want an accuracy of 1A or better...
        At least if Hall sensors flame out thru over-current it is unlikely to be as hazardous as a flamed shunt. (Ah I miss those old dashboard ammeter fires!)

        But, seriously reconsider. Though maybe expensive and complex "fun" for various loads, for batteries etc it is essentially useless.
        Current won't tell you if the system is operating normally and within specs.
        How will you tell if your battery is undercharging or system is under- or over-voltage (and imminent electronic burnouts), or that the battery is about to explode due to a collapsed cell?


        • #5
          The reason I want current monitoring is so I can measure the power draw of different components and optimize the power usage. In particular, I want to keep a subset of my system running 24/7, and current monitoring will let me know how long I can run on battery power.

          I should be able to monitor the current and see that when the computer & LCD & backup camera are all on that it uses some amount of power, and if I turn off the LCD and the backup camera, or turn off GPS or the sound card or whatever, that it drops in power consumption. I should be able to tell that when I first start the car the isolator keeps everything disconnected, but as soon as the main battery comes up in voltage it kicks over and starts charging the aux battery. My solar power system will eventually be plugged into this, and I can use it to make sure that during night time there is no wasted current going into the panels, and to monitor how many amps the solar system is charging the batteries, and to make sure that if the system is on 24/7 that the solar power is bringing in enough power to keep it all running.

          I would only being using open-loop or closed loop hall sensors, not the ones where you have to wire them in, so they couldn't burn out from overcurrent. The main/aux connection is fused at 60A, and readings down the amp would be fine. If I did +-60A, that would give me 0.1A per tick @ 1024 ticks (arduino ADC resolution). For the other lines I'd be running a max of 20A unidirectional, so about 0.02A per tick (~275mW).

          This would certainly be useless to a lot of people. But might be useful to someone other than me.


          • #6
            Current monitoring has limited ability in determining battery reserve. It assumes a battery that conforms to specs and that you know its initial state (ie, OC voltage).
            I have seen many systems that have let people down because the battery dies quicker than expected. Voltage is definitely required as part of that prediction, and the current determines what the end voltage should be (ie, the drop at the battery terminals due to internal resistance). (Again, the latter current is usually known by one-off tests for the various equipment used, and any actual variation is only significant when used with sophisticated battery condition algorithms which - trust me - have proven to be a very complex issue.)

            If current is wasted, what the heck is that load doing being connected in the first place?
            If you mean selective load shedding, that's cool, but that then requires good data recording and predictive software, plus a specified required target time. Even UPS systems tend merely to initiate a shutdown after a pre-programed delay else when the battery voltage reaches a certain threshold prematurely. Load shedding is done by splitting the power systems.

            If you think you can monitor your currents and hence determine suitable load shedding to achieve a required reserve time, then I hope you have adequate understanding. Even experts tend to adopt a "copious" battery reserve design; the limit being predefined load shedding. UPS battery design and BCMs (Battery Condition Monitors) and their predictive capacity theorems etc are probably a good info source noting that they are usually for defined batteries and highly controlled battery environments (even they have trouble, unless things have changed in the past few years).

            Load currents are usually only used in sizing the original installation - eg, wiring capability and battery reserve. Even then, initial designs usually exceed expectations if the spec'd battery ratings are based on the common 80% capacity aging rule. But then add in temperature impacts, sulfation, etc.

            As to determining charging health, that is primarily a voltage issue. EG - if a battery is charging at 2A (say double its float current), is it charging healthily? And what if it's at 14.6V (Optima batteries specifically excluded) or 13.7V?

            By all means go ahead if you think it will be useful, but ensure you have voltage monitoring in place first.
            Other than that suggestion, I am merely trying to convey that you will probably find it to be of limited value - especially in consideration of cost and complexity.
            And I suggest few will be interested in a commercial product of that type - even the DIY expense with free software is probably unattractive. It might however be a desirable product for audiophiles to match their "stiffening" capacitors, or even the gullible Australian market (dc-dc converters for remote battery charging & "intelligent" battery isolators refer).
            FYI - I have long been intending to obtain some Tamura dc current sensors from Digikey for my varied work, but I have had quite adequate results from my $10 50A shunt - including the testing of 250A starter motor currents (short duration tests obviously!). I'm still waiting to get the +/-15V Tamuras for interfacing direct to cheap voltmeter panels or CPU circuits. But they are for confirming the behavioral characteristics of designs I purport.

            BTW - you said your isolator kicks in as soon as the main battery voltage comes up. That is essentially immediately though voltage-sensing isolators will have an added delay as necessitated by design.
            Do not confuse that with kicking in after the main battery reaches some level of recharge - that's a bit of bullsh marketing used by voltage-sensing sellers to disguise one of their limitations.

            But good luck and best wishes. The learning curve is IMO rewarding even if at times frustratating.


            • #7
              I won't be doing the 'load shedding' as you call it predictively, I'll be doing it by hand, but knowing which components have what power consumption is essential either way. For example, if I find out that one particular system is taking up a lot more power than expected, then I can come up with a different solution or find a different device. In my case I want to keep my video cameras up and running - so I need to be able to see how much current each camera draws, and I can't just use a multimeter and see how much the CAMERA draws, because the computer will be saving the video - which means CPU usage goes up, SSD usage goes up, power supply loses, etc - it's a complex system so the only way to figure out the true power draw of a component is to test it.

              I also need to see how disabling a USB device in software affects power draw. Some devices might correctly power down, other devices might not. It doesn't take much effort from me to implement a USB controlled relay for turning devices on and off, if needed, but... I'm finding it incredibly difficult to get accurate measurements on the bench, let alone in the car under real world conditions, at least without a shunt or hall effect sensor. Every difference between my bench setup and my in-car setup is a source of error, and there is no way I can pull every component out of the car and/or reproduce the setup exactly on my bench.

              btw, I'm not intending to use the current sensing in place of voltage sensing - it will be in addition to voltage sensing. And yes, I am aware that the alternator pretty much instantly brings the voltage up, and it doesn't have much to do with charging (unless your battery is on it's way to battery hell).


              • #8
                Again, that's cool. But that omits some of the intents you originally posted. And I guess you understand why no such "black box" exists other than the plethora of current monitoring attachments available for things like the Arduino and Fusion Brain etc, or other dedicated commercial products. (Capacity prediction is available in the latter, and their reliability keeps improving.)

                I don't know what you mean by the alternator not having much to do with charging. It has everything to do with charging!
                If you mean the system's charging voltage has nothing to do with the state of charge of the batteries, then you too understand the bullsh hype regarding a battery isolator's "priority charging" (except as pertains to a time delay only).