Shop OBEX P1 Docs P2 Docs Learn Events
Electrical Gurus Needed :) ---- Measuring The Millivolt With High Accuracy - Page 3 — Parallax Forums

Electrical Gurus Needed :) ---- Measuring The Millivolt With High Accuracy

1356

Comments

  • Was thinking on another project, and down a garden path I went. It may be worth looking at the LM331 V/F converter. There may be objects in the exchange for using this device. Its voltage input max is well above your 24v requirement.
  • jmg
    Depends on what you call Parasitic.
    If a large fuse drops 10mV at rated current (eg 25A) , that 0.1mV LSB is 250mA - rather higher than most consider 'Parasitic'

    HeHe.... Personally, I would call any unwanted and unintended current draw parasitic, because it is just like a parasite feeding off the life blood of the battery :)

    My figures tell me something different...

    Different fuse types will have different mA results.

    However, for the sake of discussion, let's assume a Mini ATM fuse, with a 25A rating. With a 9.9 mV reading across the in circuit fuse, according to my resources, the current draw on that circuit would be 4195 mA and 10.0 mV reading would be a 4237 mA current draw. Thus for that particular type fuse, the 0.1 mV difference between 9.9 mV and 10.0 mV would be equivalent to 42 mA current draw.
  • idbruceidbruce Posts: 6,197
    edited 2018-05-05 03:50
    Phil
    The INA219 breakout board includes a sense resistor. To take a reading across a fuse, you will want to remove the fuse first to get an accurate reading.

    Then I guess the INA219 breakout board will not work, because the fuse must be measured in circuit. Otherwise, you might as well just remove all the fuses and take an amperage reading between all the fuse block terminals.

    Measuring the voltage drop across in circuit fuses is common practice and all I need is a way to get accurate readings with these fuses in place and without removal.
  • INA219 is a great starting point but if you're serious about supporting 24v automotive you'll want more than the 26v abs max the ina219 offers. Keep in mind there are spikes/surges above the nominal voltage, and when charging nominal might be as high as 14v in a 12v car, so 28v in a 24v system

    My suggestion in this case would be the max9611/9612, they're rated to 60v and have a few ranges so you can 'zoom in' on smaller leakage currents. There's a maxim breakout/eval board around $20ea.

  • Tracy AllenTracy Allen Posts: 6,656
    edited 2018-05-05 17:59
    Bruce, you can easily remove the sense resistor from the breakout board and use the fuse itself as the resistor, but that will measure only to the accuracy with which you know the fuse and contact resistance. Is that your intent? I've solved that kind of problem in some cases involving batteries by using a two-sided shim that slides in between the battery contacts, so that the current is diverted through the meter, but I don't see how that gimmick could work with plug fuses.

    About this, you said:
    Please allow me to correct myself. I would need two channels to enable differential readings. This would eliminate the need to have the probes in a specific orientation.

    Be careful with the specs on devices like the INA219. The input itself is differential, sufficient, you don't need two channels. You might read in the data sheet that the differential input range can be +/-26 volts, however, both inputs have to stay within the common mode input range, which extends down to only -0.3 volts. These TI switched capacitor ADCs are quite happy to convert voltages down to -0.3 volts for the current measurement, but if the fuse happens to be blown and -12 volts appears at one of those inputs, ¡magic smoke!.

  • idbruce,

    Does the car manufacturer have a procedure for finding these faults?
  • Heater.Heater. Posts: 21,230
    I don't follow Tracy. Where is that -12 volts going to come from?

    Normally both inputs are either at 0v, when the power is off, or about 12 volts when it's on. Give or take the drop across the fuse.

    If the fuse blows one of the inputs drops to 0V.

    Which all seems OK.
  • Do not forget the measure contact shunt, and resistance of you measuring.


    By the way
    You do remember me a lovely car battery problem with dady's car.

    After a weekend, most of time the car don't want to start. Low battery

    Change battery, alternator, all the wires,...... take months, go ten times to the garage big money.

    One day, coming home after a late night party, I saw light coming out of the trunk of my dad's car.
    I found the light switch for truck light was bended, guess after putting some luggage in the car trunk.
    So after a couple of day not using the car the battery was discharged due to the car truck illumination.
    It has to be very dark to notice this when the car truck was closed

  • jmgjmg Posts: 15,140
    Heater. wrote: »
    I don't follow Tracy. Where is that -12 volts going to come from?

    Normally both inputs are either at 0v, when the power is off, or about 12 volts when it's on. Give or take the drop across the fuse.

    If the fuse blows one of the inputs drops to 0V.

    Which all seems OK.

    Tracy has a point, in that the INA2xx series are not really 2 terminal devices.

    INA233 Data says
    VVS Supply voltage 6 V
    Analog Inputs, IN+, IN– Differential (VIN+ – VIN–)(*2) –40 40V
    Common-mode –0.3 40
    VVBUS VBUS pin voltage –0.3 40 V
    (*2) IN+ and IN– can have a differential voltage between –40 V and 40 V. However, the voltage at these pins must not exceed the range of –0.3 V to 40 V.
    IB Input bias current (IIN+, IIN– pins) 8 μA ( Note Graph shows ~17 μA ?!)
    VBUS input impedance 830 kΩ


    So the parts ideally need to be GND connected in the vehicle, to have 12V,12V on a normal fuse, and 12V 0V on a blown fuse. ie 3 wire measurements.

    The Bias current is not super low on these parts, as they expect battery type connections, but you could still use the usual series R and clamp diodes as smoke protection.
    1k series with 8uA will add 8mV of offset, 17mV if that 17 μA is correct, 100 ohms is 1.7mV, but gives higher clamp currents.
    Data sheet mentions 10 ohms filter/protection.
  • idbruceidbruce Posts: 6,197
    edited 2018-05-05 20:39
    I have decided to toy around a bit and do some testing.

    Power Supply: Linear Power Supply Outputs 14.94VDC, 2.5A

    To this power supply, I have wired in series, a 2A Mini ATM fuse, a 5W 27 Ohm 5% resistor, and another 5W 27 Ohm 5% resistor.

    The current draw across this series is 281mA. When probing the two fuse nibs with the Fluke, the reading is 15 mV. Although my resources do not make an indication for 15.0 mV, my calculations have determined that the voltage drop is right in the ballpark.

    Touching on Tracy's subject
    Please correct me if I am wrong... Unless one of my probes is touching ground, my reading device should never exceed, let's say 20 mV.

    Anyhow, I also have on hand an Adafruit ADS1015 ADC ( https://www.adafruit.com/product/1083 ) and a Propeller Board of Education, with the support code for setting it all up. My goal is to try to attain that same Fluke reading of 15 mV from my test rig, with the ADS1015 and the Propeller Board of Education. Is this a fool's endeavor or can it be achieved with these two items?
  • Heater.Heater. Posts: 21,230
    The way I'm interpreting the data sheet just now this is not a two terminal device.

    The chip is powered from 0V and 6V with respect to ground. As per any normal circuitry.

    The inputs can zoom up and down between -0.3 and 40V with respect to ground. They can be 40V different from each other, either way around.

    In tha scenario if the fuse blows then one input drops to ground whilst the other stays up at 12v

    I see no problem with this. Where would that -12V come from?

    Why would one ever imagine to have the ground of the chip supply up at 12V?

  • idbruceidbruce Posts: 6,197
    edited 2018-05-05 20:55
    In tha scenario if the fuse blows then one input drops to ground whilst the other stays up at 12v

    I see no problem with this. Where would that -12V come from?

    Why would one ever imagine to have the ground of the chip supply up at 12V?

    Heater

    I am just saying that if I have one probe on the 12V side of a blown fuse, and as long as the other probe never touches vehicle ground, the meter itself should never see more than approximately 20 mV. I certainly would never want the other probe to touch vehicle ground, if the chip could not handle 12V.
  • jmgjmg Posts: 15,140
    Heater. wrote: »
    The way I'm interpreting the data sheet just now this is not a two terminal device.

    The chip is powered from 0V and 6V with respect to ground. As per any normal circuitry.

    The inputs can zoom up and down between -0.3 and 40V with respect to ground. They can be 40V different from each other, either way around.

    In tha scenario if the fuse blows then one input drops to ground whilst the other stays up at 12v

    I see no problem with this. Where would that -12V come from?

    Why would one ever imagine to have the ground of the chip supply up at 12V?
    See my post above. Yes, it is not a 2 terminal device, and provided you connect gnd, you are fine.

    I'd expect that clamp diodes from IN-, IN+ to GND, could protect for the case of that GND clip falls off, when you have a blown fuse case, or just during any-order attach of the 3 leads.

  • idbruceidbruce Posts: 6,197
    edited 2018-05-05 21:00
    AHHHH.... but if the fuse is blown, then the mV becomes Volts
  • idbruceidbruce Posts: 6,197
    edited 2018-05-05 21:06
    Doesn't setting up the ADS1015 for differential input make it a two terminal device?
  • ... but if the fuse happens to be blown and -12 volts appears at one of those inputs, ¡magic smoke!.
    That's one reason I suggested leaving the shunt resistor onboard and pulling the fuse for whichever circuit you're testing. Of course, the resistor has to be rated for whatever it's called upon to dissipate, if the circuit happens to be "on." The other reason is that you always know what your shunt resistance is without the fuse in place.

    -Phil
  • jmgjmg Posts: 15,140
    edited 2018-05-05 21:23
    idbruce wrote: »
    Please correct me if I am wrong... Unless one of my probes is touching ground, my reading device should never exceed, let's say 20 mV.

    Anyhow, I also have on hand an Adafruit ADS1015 ADC ( https://www.adafruit.com/product/1083 ) and a Propeller Board of Education, with the support code for setting it all up. My goal is to try to attain that same Fluke reading of 15 mV from my test rig, with the ADS1015 and the Propeller Board of Education. Is this a fool's endeavor or can it be achieved with these two items?

    You do need to tolerate a 12V across the probes case.

    Data on that says :

    If it is possible that the voltages on the input pins may violate these conditions, external Schottky clamp and/or series resistors may be required to limit the input current to safe values (see the Absolute Maximum Ratings table).


    Since you chase millivolts, you could just use STKY diodes negative and silicon diode or LED positive.
    It also says for highest gain :
    ±0.256V Differential Input Impedance = 710kΩ
    If you use 1k series R in each probe, you have ~ 6mA of 'oops' fault current (thru the diodes), and a 0.3% reduction in gain.

    To help limit common mode variations, 1k from each probe to GND would keep everything within millivolts of GND, and be not significant to a fuse. (53mΩ in your example above)
    You might like to add a 99k to one probe, to allow 12V as 120mV reading, for simple sanity checking.

  • Since you want to check the current through the fuses by measuring the voltage across them, you would have to measure the resistance of each fuse to be tested so that you can calculate the current through them. If it was me (realizing I am about suggest a brute force solution), I would simply measure each fuse and write the values in a table. Then given the conditions you wish to measure the voltages, measure each of the voltages across the fuses and determine the current. No special test device required, no development time spent, if the fuse is blown probably measure +12V across it so no damaged devices; the meter should handle 12V. Unless as I gather you are building a commercial device for this purpose. But as long as the fuses have to be checked sequentially a meter and chart of fuse resistance would actually be fastest way to test. Alternatively you could just fabricate a fuse holder that would allow you to put an ammeter in line with the fuse for direct readout. But that is back to sequential testing.
  • Tracy AllenTracy Allen Posts: 6,656
    edited 2018-05-06 01:25
    Doesn't setting up the ADS1015 for differential input make it a two terminal device?

    It approaches a two terminal device if you power the measurement circuit from its own battery and connect only the differential (+) and (-) to the two sides of the fuse (=shunt?). No connection to the car battery as power or ground. In that situation it is acting like a hand-held voltmeter where you are touching the red and the black leads to the two sides of the fuse. Caveat: there may need to be a couple of resistors (~MΩ) added within your "voltmeter" to provide necessary bias currents. I'll also echo the caveat to add some kind of protection to those inputs in the way of resistors and clamp diodes.

    Once you make a connection to the a third node in the car's electrical circuit, it is no longer a two terminal connection and you have to consider a triad of possible conditions.
    Anyhow, I also have on hand an Adafruit ADS1015 ADC ( https://www.adafruit.com/product/1083 ) and a Propeller Board of Education, with the support code for setting it all up. My goal is to try to attain that same Fluke reading of 15 mV from my test rig, with the ADS1015 and the Propeller Board of Education. Is this a fool's endeavor or can it be achieved with these two items?

    I think that is feasible. Although, I'd say, go for the ADS1115 at 16 bits. The best resolution on the ADS1015 is 0.125mV, whereas on the ADS1115 it is about 7 microvolts. Smaller shunt, higher accuracy, and the Prop code is about the same. Don't forget the input protection!

  • Heater, I grant you that the bad situation I pointed out (-12 volts common mode) would not occur if everything is connected up correctly with the fuse in the positive lead.

    I was thinking in my own experience of using the ADS1115 to monitor both charge and discharge current of a battery, with the charge current coming from a solar panel, using low-side shunts. Bad things could happen (if there isn't additional protection) if the shunts were not set in properly. The circuit made good use of the capacity of the ADS1115 to measure voltages down to 0.3V below its own Vss power supply, on the 256mV full scale range, resolution 7.8µV.

    My point was really to bring up the need to consult the data sheet and to stay within the common mode range. Poking around on a fuse block is an invitation to automotive gremlins!
  • This is getting complicated.

    What is the battery drain time? Over night? Over nights? Week? Month?

    That should tell you the rough size of the drain.

    If it's a short drain period, pull fuses one by one and evaluate battery. It is also a high current, could be found by pulling fuses with current measure in line with the battery.





  • idbruceidbruce Posts: 6,197
    edited 2018-05-06 03:34
    potatohead
    This is getting complicated.

    What is the battery drain time? Over night? Over nights? Week? Month?

    That should tell you the rough size of the drain.

    If it's a short drain period, pull fuses one by one and evaluate battery. It is also a high current, could be found by pulling fuses with current measure in line with the battery.

    You are missing the point!

    I believe a new battery has already solved my problem, but I am working on a commercial Propeller solution.

    As Frank hinted a few posts back, by knowing the resistance of each fuse, tables can be created, for quickly performing voltage drop tests across the fuse, without removing the fuses. This is my goal.

    Testing voltage drops across the fuses, is certainly a lot less labor intensive than pulling all the fuses and testing for current draw. Fast and simple, providing you have the proper circuitry.
  • idbruceidbruce Posts: 6,197
    edited 2018-05-06 04:03
    I am certain that there are probably many ways to solve this circuitry, but as DigitalBob indicated several posts back, there are ICs available for this specific task:
    DigitalBob

    Another good choice is the ICL7135 4-1/2 digit DMM IC. It already has a built in integrator and it interfaces easily with a PIC. Just add a resistor network like a Caddock.

    AN046: Building a Battery Operated Auto Ranging DVM with the ICL7106
    https://www.intersil.com/content/dam/intersil/documents/an04/an046.pdf

    AN028: Building an Auto-Ranging DMM with the ICL7103A/ICL8052A
    https://www.intersil.com/content/dam/Intersil/documents/an02/an028.pdf

    ICL706, ICL7107, ICL7107S Datasheet
    https://www.intersil.com/content/dam/Intersil/documents/icl7/icl7106-07-07s.pdf

    ICL7135 Datasheet
    https://www.intersil.com/content/dam/Intersil/documents/icl7/icl7135.pdf
  • It would be easier to find the source of the current drain on your car. This can be done with any cheap multimeter. Just disconnect the battery terminal and put the meter in series with the battery post and terminal. The meter needs to be set on 10 amp D.C. At first it should draw about 1 amp then settle down to 25 Ma. If it doesn't drop then pull the fuses one by one until you find the drain.
    Even if you had a device that you are talking about you still need to pull the fuse to attach leads.
    Or insert some fuse bus extender either way the fuse still has to come out. Unless your talking about old cars with glass buss fuses. Any later vehicles have plug in plastic fuses I don't see how you would get test leads in there.
  • idbruceidbruce Posts: 6,197
    edited 2018-05-06 04:07
    Actually the ICL7135 that DigitalBob suggested has Six Auxiliary Inputs/Outputs Available for Interfacing
  • DigitalBob

    First I credit you with a wonderful IC suggestion, but then you come back with that post :(

    If you read the thread, then you will know that I know how to test for total current draw and I also know how to pull the fuses and test for current draw across the terminals for the individual circuits.
    Even if you had a device that you are talking about you still need to pull the fuse to attach leads.
    Or insert some fuse bus extender either way the fuse still has to come out. Unless your talking about old cars with glass buss fuses. Any later vehicles have plug in plastic fuses I don't see how you would get test leads in there.

    Not true. Read the thread or watch this video
  • idbruce,

    The 7106 has been used in 3-1/2 digit multimeters from decade such as the Harbor Freight freebies.

    My Korean made Actron CP7676 Automotive Digital Analyzer uses a JRC (Japan Radio Co.) version of the 7106 inside.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2018-05-06 07:21
    If you use a single probe that injects a programmable constant load current and then amplify the differential of the previous sample and hold under probe load with no probe load into your A/D you should be able to calculate the ESR and thereby the fuse load current all without having to connect across the fuse. In fact you can probe before the fuse if you like but closer to the load will yield greater ESR and more differential.

    You could probably deduce the source voltage from zero current loads (maximum hold) or calibrate by touching the battery, and use that for one end of the differential in addition to the extra loading differential to establish the ESR as that offers better resolution. I may have missed some fine point but a single probe that can be used anywhere would be more useful. You could even stick a small LCD on the end of it too.
  • Peter

    Much of this is already beyond my understanding, but after what you just said... all I could do was smile :)

    However, I would have to agree that a single probe would be the cat's meow.
  • Sounds like a request for a high resolution bench multimeter with a serial interface?
Sign In or Register to comment.