Variable meter shunt concept - could this work?

Discussion in 'General Electronics Chat' started by Dyslexicbloke, Sep 11, 2010.

  1. Dyslexicbloke

    Thread Starter Active Member

    Sep 4, 2010
    420
    19
    Hi folks,
    In have 5 deep cycle batteries in paralell
    With my inverter running at 1.5kw the current is obviously quite high on each battery so I would need big low value shunts to measure it.

    However I also want to know what is going on when the inverter is only drawing a couple of wats or when its off all together and I need to see any battery to battery current that could indicate a problem, this would need a much higher value shunt resistor.

    It has occured to me recently that a FET or an IGBT might just do what I want.
    The thought process ... If I were to measure the voltage across the channel and hold it stable by adjusting the gate, would the gate voltage be proportional to the current in the channel?

    If that is even plausable, is there any field effect device that will control DC current in both directions?

    Could I use a diode bridge to effectivly keep the fet the right way up or would that require a high side driver?

    Its a bit off the wall I know but I'd apriciate anything you have to offer :)

    Thanks folks
    Al
     
  2. awright

    Well-Known Member

    Jul 5, 2006
    84
    7
    Doesn't sound very practical to me. Not confident that the relationship between gate voltage and channel resistance is well controlled or temperature stable.

    Why not use LEM current transducers built for the purpose and with well controlled parameters and very good accuracy? Check the LEM website and ebay for some amazing deals on LEM transducers.

    awright
     
  3. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,789
    945
    Big ones like these 8D's or like the smaller group 37's?

    [​IMG]

    Don't worry too much about battery to battery current if they are all the same type and age and have been used together and remain parallelled when the inverter is off then the battery to battery current will stabilize very quickly. It is actually a desirable thing for the batteries to exchange small amounts of current and equalize themselves between uses.
     
  4. timrobbins

    Active Member

    Aug 29, 2009
    318
    16
    Suggest another method is to borrow a DC tong meter (the ones I have are typically with 400A range), or temporarily connect a shunt in series, and measure the current in each battery cable - and at the same time use a mV meter and measure the DC drop along the length of the cable. You can then use the mV meter to measure current levels in the future without having to have the shunt, or tong meter. Accuracy is not too bad.
     
  5. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,789
    945
    Tim!

    That's how I do it. Worked out that my negative leg 2/0 cable is .0091 Ohms. So now I just measure the millivolt drop along the cable from battery to inverter and convert it to amp draw.

    Example:700mV/.0091Ohm=76.9 Amps The 48 volt inverter draws an amazing amount of juice. I hate to think what the draw would be at 12 volts. :eek:
     
  6. Dyslexicbloke

    Thread Starter Active Member

    Sep 4, 2010
    420
    19
    mmmmm .... I hadnt considdered stability with temperature.
    I wasnt too worried about it being linier because I would be able to compensate for that with the control system ... PC to develop, PIC in service but trying to compensate for temp would be going to far I feer.

    The driver here is being able to get relatively accurate readings when the system is only lightly loaded without droping a chunk of voltage, and power, when the current go's up.
    My fridge starting makes the inverter grunt a bit to say the least, particulaly if the PC or TV is already on.

    I will look at the LEM devices but I am a little worried about cost there not to mention the extream range I want to look at.

    Batteries are not too well matched and worse still not in that good a condition either.
    All circa 100Ah but not exact and diferent ages to boot.
    YES I know .... it one of the reasons I want to keep an eye on them.
    (No other choice right now, cant afford new ones)

    I may be attempting to do too much here ... but I am going to try anyway :)

    What is the minimum current you can accuratly measure with your .0091 Ohm cable?

    I would have to add an instrument amp anyway ... FET input opamps spring to mind with some sort of gain control from the micro so the system can auto range.

    Do you think that aproach is any more practical?
    I supose any noise, inverter feedback and the like, could be dealt with at the amp stage.

    Is there a comertial solution to this problem?
    All the meters I have ever seen that handle a huge current range require that the leeds be moved arround so that at least two diferent shunts are involved.

    Thanks for the input thus far guys

    Al
     
  7. timrobbins

    Active Member

    Aug 29, 2009
    318
    16
    Tong / clamp meter has a jaw that opens to go around the cable - some LEMs also have the feature.

    I suggest you don't discard the mV cable drop method too soon. It is eminently acceptable for 'relative' measurements. You can also characterise it at the nominal operating currents and temperatures you experience, in comparison with a known shunt - it will be then quite repeatably accurate. You could use a single channel opamp, or IA (but I doubt you need to get too exotic), and use a mux for the different batteries. This method requires a common point for all measurements.

    You will have twice mains frequency ripple current, so absolute accuracy will require rms measurement, but again not an issue for relative measurements and control.
     
  8. Dyslexicbloke

    Thread Starter Active Member

    Sep 4, 2010
    420
    19
    Yes that sounds like a plan .... Its not too hard and based on the comments earlier almost certainly better than trying to muck about doing somyhing off the wall with a FET.

    I will build an instrument amp, have several opamps kicking about ... I might even have an LM13700 transconductance amp left so I could use that as the final buffer and adjust the gain with a reference voltage.

    could I activly cancle the ripple (noise) by seperating the AC componat and summing it back inverted?
    (just for the measurment obviously)
    Would that help if it worked?

    Thanks All
    Al
     
  9. Dyslexicbloke

    Thread Starter Active Member

    Sep 4, 2010
    420
    19
    Just had a thougt ....
    Wouldnt the voltage drop on a fuse of a given size and type be fairly consistant?

    Al
     
  10. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,789
    945
    The meter I use is very accurate down to around 20-25 mV.

    Which will get me down around 5 amps draw on the negative cable. For a 2500W inverter rated @ 48 volts, 80 amps is the largest current draw I've seen in use, and that was transient. It usually settles down to about 45-50 amps under a heave steady load. The spikes are when the AC window unit kicks on. I don't use the setup on a regular basis, but exercise it once a month or so. It got me through Hurricane Ike and three days without power. Kept me spiked with lots of coffee, internet and AC cooling. I was thinking I might have to skip a day and let the solar panels play catch up. 25 amp charging ability and 50 amp spikes in usage was slowly wearing them big 8D's down. :)

    load you unit down to the limit you know it can take, then based on the wattage of the inverter, and the voltage of the battery line exactly at that load calculate what current would be used. Compare that figure with your millivolt drop on a feed between battery and inverter and you'll get the ohm value within a fairly close range. You can make up a chart for current vs mV and then just monitor the lines from each battery.(unless your cables are to short to monitor individual batteries) Do you have them paralleled at the battery terminals? or are you running cables to a central tie point near the inverter? Changing to an individual cable run per battery will give you the ability to monitor each individual battery for a fall off in current.
     
  11. Dyslexicbloke

    Thread Starter Active Member

    Sep 4, 2010
    420
    19
    The batteries all run, via cables to a large pair of copper busbars .. OK flattned 15mm tube with 8mm bolts through it.

    Anyway it means I can disconnect or add any DC elliment at any time.
    Fusing needs work but that another story, I want to add DC breakers when I can afford it.

    My clamp is only cheep and will not do DC but one of my meters will go to 25A I think so I can extrapolate from there.

    I might replace all the battery leeds with ones of the same length, they differ at the moment because the bats are on two levels with the bus above.

    I am actually more interested in the small loads, I want to see what sort of overhead the inverter is taking to work out weather it worth getting a smaller one to use unless I needc the bigger unit.

    We are off grid here, a relativly new situiation, so even small savings translate into significant propane savings at the genny. My biggest issue is probably batteries ... performance, charging strategy, just understanding them in situe realy.

    Only have 300W of solar and just puting together a wind turbine (Small)

    Thanks for the advice

    Al
     
  12. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,789
    945
    If you are still interested I'll make some pictures of my method for checking current draw next time I exercise my stack of cells.

    Nice to know someone who is 'off grid' completely. I'm totally addicted to the grid, but can do without for a short time. Looking forward to the day when I can move further out of town and set up more power generation schemes.

    I'm reading about pumped hydro storage for wind power right now. Lots of good articles on it. The idea of using wind turbines and also wave power to pump water uphill to provide 24/7 power from hydro generators is appealing to me. Seems to make a lot of sense to use waves to pump water. :)
     
  13. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    If you've got plenty of wind day and night pumping water up into a tower is a pretty neat method of storing potential energy.

    I'd never dump the grid, if you get good enough at it you can sell power back to them.

    I live in a state where you can't really rely on anything and our biggest costs seem to come from heating and cooling. When the day comes that I finally get to build out in the country I'm hoping to utilize a decent geothermal system as a key component for moderating what I need to add or subtract to my ambient indoor temperature.
     
  14. Dyslexicbloke

    Thread Starter Active Member

    Sep 4, 2010
    420
    19
    Ooooh for a grid I have to make a decision about ....

    I live in a large static caravan in a beutiful spot neer Leek in staffordshire (UK).

    No grid of any sort, water is from a spring some 15m/20m below the van, power is what I can collect or generate and my phone and internet are mobile.

    The pumped hydro sounds interesting, particulaly as it could be augmented with rain which we get quite alot of here. The site has a huge slope, unfortunatly the water source is near the bottom.

    I would be very interested in your method of testing and maintaining batteries, I will leave my subscription to this thread active so you have a convieniant way to inform me later, Thanks.
    Al
     
  15. timrobbins

    Active Member

    Aug 29, 2009
    318
    16
    What was your plan to digitise the current level for importing to uP? Battery current is bi-directional, and if you want a reasonable look at low current levels then you probably have to aim for bipolar ADC, and keep to a common 0V point in the analog sensing (such as your busbar), and sense current on the neg leads. If you need to get simpler - then you could split the pos and neg current signals in to two analog levels for digital sensing - use an inverting opamp for the neg current to get to pos signal.

    The twice mains ripple from the inverter can be a pain - analog LPF should suffice - or if your keen you can try to include digital filtering of some kind.

    Once you have battery current, and if your into programming then you can have a field day with Ah accumulation, and SOC adjustment with current and temp and OCV and periodic equalising.
     
  16. Dyslexicbloke

    Thread Starter Active Member

    Sep 4, 2010
    420
    19
    I had thought of going for a single analouge input, always positive, so I'm not halving the effective resolution of the PIC and utilising an additional input to give polarity indication. Input to the instrument amp would be two voltage buffers with unity gain, one inverted, so I always have a posative level.
    No design yet ... Still just a concept in my head, should be relativly easy though with opamps.

    I'l probably do it in stages starting by getting a usable reading. It will need multyplexing as well unless I go with a much biger PIC than I have right now ... Perhaps later.
    AL
     
Loading...