Forum Update - Announcement about May 10th, 2018 update and your password.

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

1246

Comments

  • Mark_T
    Sounds like a request for a high resolution bench multimeter with a serial interface?

    I am still leaning towards the ICL7135, which is capable of communicating with a uC.


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • jmgjmg Posts: 11,273
    edited May 6 Vote Up0Vote Down
    idbruce wrote: »
    However, I would have to agree that a single probe would be the cat's meow.

    You still need to have some differential signal, to sense the drop.


    That leaves the problem of 'calibrate' of the fuse (+ and short wire) voltage drop, to extract the milliohms.

    You could (with enough meter battery grunt :) ) inject a brief test current across the Fuse+Cable stub under test, to then extract the milli-ohms, and then that can calibrate the mV you measure, back into mA

    You would need the highest precision ADC you can find for this, with a reasonable sample rate, but this would auto-calibrate for fuse resistance/cable resistance, and could allow a common connect point (+12V) at some fuse-entry, and a single flying probe for the individual fuses.

    If we inject 10mA into 10m Ohms (your example above was 51m ohms) that's 100uV, needs a 1uV LSB to calibrate to 1%
    Or, you inject 50mA into 10m Ohms, and now the 500uV drop, needs 5uV LSB to calibrate to 1%.

    A SMPS part like the TSOT26 AL8860WT, can give 600mV at 2%, up to 2A, for calibrate current injection.
    idbruce wrote: »
    I am still leaning towards the ICL7135, which is capable of communicating with a uC.

    Depends how many you plan to make, I guess ?
    The ICL7135 needs split Vcc and is very much trailing technology, with a high price and large package.

    The NAU7802 looks hard to beat here, it has 24b engine, and comes under $1/100 in SO16N, and the internal LDO could be 'good enough' if you want to scrimp on skipping an external VREF.
    For VRefs the
    AZ431AN gets you to 2.5V, 0.4%, 20ppm/°C Typical at 15c/100, and
    TLVH431BIL5T gets 1.24V, 0.25%, 30ppm/°C Typical at 20c/100
    or
    MCP1501T-18E/CHY gets choice of Vo, 0.1%, 10ppm/°C Typical at 60c/100

  • idbruceidbruce Posts: 5,648
    edited May 6 Vote Up0Vote Down
    jmg

    In many ways, my idea for a tester, is electrically way beyond my comprehension. After reading all the input of this thread and after looking at numerous datasheets, it appears to me, that maxim/intersil is the leading technology in this arena, but I could be wrong.

    In fact, I might go as far to say, that even the gurus have been having trouble with this one. At first, I thought it would be a simple fix, with an adc, but the more I investigated a solution, the more I realized how complicated a solution really was.

    The biggest problem that I foresee is a blown fuse, at which point the tester and driving IC would see 12V. With that in mind, whatever circuitry I utilize would certainly have to require some type of auto ranging feature, which is not commonly available with standard ADCs.

    EDIT: Otherwise, the entire fuse block would have to be tested for a blown fuse, before actual voltage drop tests could commence.


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • @jmg - you haven't read my post. It achieves a differential using a single probe (+ ground), a calibration sample, and switching in constant current loads and measuring the amplified difference from a sample and hold. This has nothing to do with using the fuse as a shunt but does require some resistance in the wire and can't be damaged by over-voltage since it is in fact measuring the actual voltages.

    @idbruce - good luck with voltage only :) There are much better DMM chips around, modern ones that handle 60,000 counts and have serial built-in etc. Personally though, I think I will build my differential sampler just for the fun of it.
    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    --->CLICK THE LOGO for more links<---
    Tachyon Forth News Blog
    TAQOZ documentation
    Brisbane, Australia
  • jmgjmg Posts: 11,273
    idbruce wrote: »
    The biggest problem that I foresee is a blown fuse, at which point the tester and driving IC would see 12V. With that in mind, whatever circuitry I utilize would certainly have to require some type of auto ranging feature, which is not commonly available with standard ADCs.
    I don't know if it needs auto-ranging, but it does need to tolerate a 12V Fault condition. An over-range output is enough to indicate a blown fuse.
    It would be generally useful to measure 12V VBAT anyway, and that becomes a question of do you want 2 probes, or more.

    idbruce wrote: »
    After reading all the input of this thread and after looking at numerous datasheets, it appears to me, that maxim/intersil is the leading technology in this arena, but I could be wrong.
    Intersil is now part of Renesas, and they do have some nice 24b ADC and also 60V V-I Monitors, very similar to the TI INA233 mentioned earlier.
    Lowest cost 24b ADC that is easy to handle, looks to come from Nuvoton.
    To complicate things even more, there are cheap Microcontrollers like the EFM8LB1 that have 14b ADCs included, for sub $1/100
    idbruce wrote: »
    At first, I thought it would be a simple fix, with an adc, but the more I investigated a solution, the more I realized how complicated a solution really was.
    Yes, if you want to make this easy to use, it rapidly gets complicated to design :)
    eg Autosense of fuse resistance milliohms is certainly nice to have.

    The next complication is around how do you power all of this ?

    - A battery like a multimeter is easy, but also bulky and it seems there is always 12V about anyway.
    If you then look at 12V powering, that also gets murky, as you are measuring current up at the +ve PWR pin.
    A -ve LDO can give simple enough 'high side' raw power, but then the better references and ADCs have dropouts below +ve PWR pin.
    Another approach is a flyback converter that regulates above +ve power pin, technically not complicated, but that's a less common topology.
    Isolated DC-DC modules are certainly available, but those start at ~$7.66 in wide VIN.

  • jmgjmg Posts: 11,273
    @jmg - you haven't read my post. It achieves a differential using a single probe (+ ground), a calibration sample, and switching in constant current loads and measuring the amplified difference from a sample and hold. This has nothing to do with using the fuse as a shunt but does require some resistance in the wire and can't be damaged by over-voltage since it is in fact measuring the actual voltages.
    Yes, but that includes the battery impedance, and you are trying to sense millivolts on a volts-offset, not to mention trying to infer some DC current path, amongst all those offsets.

  • jmg wrote: »
    @jmg - you haven't read my post. It achieves a differential using a single probe (+ ground), a calibration sample, and switching in constant current loads and measuring the amplified difference from a sample and hold. This has nothing to do with using the fuse as a shunt but does require some resistance in the wire and can't be damaged by over-voltage since it is in fact measuring the actual voltages.
    Yes, but that includes the battery impedance, and you are trying to sense millivolts on a volts-offset, not to mention trying to infer some DC current path, amongst all those offsets.

    You don't have to try and be an expert on every aspect of technology, known or unknown. Besides, a car battery's ESR is so minor compared to the wiring and also although you may have read my post, you still haven't grokked it obviously. Tesla's wise professor once said "Tesla will achieve great things, but an AC motor, this he will never do" :)

    The "impossible" has been achieved on this forum more than once so there is no point in trying to denounce thought out ideas as such because none of us can know everything, but if we did know something then we'd be on the other side of the table as the proponent instead of the antagonist. Which side of the table are you on? :)

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    --->CLICK THE LOGO for more links<---
    Tachyon Forth News Blog
    TAQOZ documentation
    Brisbane, Australia
  • Bruce
    I didn't know that the plastic fuses came with the metal buttons now. And I see in the vidieo that the guy is using just an average low tech maybe $10-$20 multimeter. So now I'm baffled just use a $10 meter to check the current draw across your fuses.
  • jmgjmg Posts: 11,273
    edited May 7 Vote Up0Vote Down
    You don't have to try and be an expert on every aspect of technology, known or unknown. Besides, a car battery's ESR is so minor compared to the wiring and also although you may have read my post, you still haven't grokked it obviously....

    The "impossible" has been achieved on this forum more than once so there is no point in trying to denounce thought out ideas as such because none of us can know everything, but if we did know something then we'd be on the other side of the table as the proponent instead of the antagonist. Which side of the table are you on? :)
    I'm not following what you are trying to say here ?

    Google says a lead acid impedance varies with load, and that means as other circuits time-out as Bruce has mentioned, your loop impedance varies.
    It is usually some milliohms, at heavy loads, but climbs at lighter loads, so is not an insignificant factor, given all the fuse pathway currents flow thru this shared resistance.
    One link even shows a dynamic impedance change, early in the discharge curve. Minor to a user, but massive to any system hoping to extract millivolts ....

    That still leaves the fundamental problem of trying to infer the current thru your selected fuse, whilst having all the loom, other loads, and battery variation in the mix.
    A measurement across a fuse, has none of that crosstalk or common mode effects.
  • frank freedmanfrank freedman Posts: 1,329
    edited May 7 Vote Up0Vote Down
    @digitalbob, this would go back to the table of fuse resistances you would have made before (or after) measuring all the voltages. Can't determine current through the fuse with only a voltmeter unless the fuse resistance is known.
    Ordnung ist das halbe Leben
    I gave up on that half long ago.........
  • jmg
    I don't know if it needs auto-ranging, but it does need to tolerate a 12V Fault condition. An over-range output is enough to indicate a blown fuse.

    You are quite correct. Auto-ranging would be unnecessary and it is over-range that I would actually need.
    It would be generally useful to measure 12V VBAT anyway, and that becomes a question of do you want 2 probes, or more.

    My concept is a 2 wire device, designed specifically for measuring the voltage drop across the fuse. Keep in mind, the goal of this tool is to quickly and easily locate current draws at the fuse block, with no other function. There is always a right tool for any given task. For instance, let's look at a similar example, of an under the dash fuse block, but with this example, I am looking for blown fuses. In this case, I would use a classic bulb style circuit tester such as this one: https://store.snapon.com/6-and-12-Volt-Circuit-Testers-Classic-Bulb-Style-Circuit-Tester-P744276.aspx. Just clamp it to the nearest chassis ground and you are ready to go. On the other hand, if I needed to know the specific voltage at any given point in the vehicle, I would just use the multi-meter.
    The next complication is around how do you power all of this ?

    - A battery like a multimeter is easy, but also bulky and it seems there is always 12V about anyway.

    I was thinking of using a 9V alkaline battery. One way or another, I will need to use an enclosure to house the electronics. I suppose the battery and enclosure issues will have to be determined by the requirements of the electronics.
    To complicate things even more, there are cheap Microcontrollers like the EFM8LB1 that have 14b ADCs included, for sub $1/100

    I have not looked at the specs for this uC, but I will need some capability such as perhaps an SD reader, perhaps a display, quite a few GPIO pins, and quite a bit of memory, as well as being capable of being programmed in C.


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • idbruce wrote: »

    I have not looked at the specs for this uC, but I will need some capability such as perhaps an SD reader, perhaps a display, quite a few GPIO pins, and quite a bit of memory, as well as being capable of being programmed in C.

    Nah, it's all about mobile devices today. Give it Bluetooth comms and create a couple of apps, one for Android and one for iOS.

    PropBASIC ROCKS!
  • Mickster
    Nah, it's all about mobile devices today. Give it Bluetooth comms and create a couple of apps, one for Android and one for iOS.

    At first I was considering wireless, but then I was thinking that it should all be hardwired. I am still uncertain which way to go. Either way, I keep running various designs through my head. While answering this post, I have a new idea for the interface that I will have to think about.

    I believe the K.I.S.S. policy should apply here, or at least as much as possible. A complicated device, yet simple to use.

    Ahhhh..... I think I have it now!


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • Heater.Heater. Posts: 20,698
    Mickster,
    Nah, it's all about mobile devices today. Give it Bluetooth comms and create a couple of apps, one for Android and one for iOS.
    Which brings us back to my suggestion to use the Mooshimeter a few pages back:

    https://moosh.im/mooshimeter/



  • Heater

    I did not look at it before, but I must admit the Mooshimeter appears to be a nice tool, however it would be incompatible for what I have in mind.


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • Heater. wrote: »
    Mickster,
    Nah, it's all about mobile devices today. Give it Bluetooth comms and create a couple of apps, one for Android and one for iOS.
    Which brings us back to my suggestion to use the Mooshimeter a few pages back:

    https://moosh.im/mooshimeter/



    OMG....how did I miss this and how many times could I have used this product!!!!

    I am ordering right now...many thanks.
    PropBASIC ROCKS!
  • Tracy AllenTracy Allen Posts: 6,221
    edited May 7 Vote Up0Vote Down
    Bruce, the shunt circuit shown on page 32 of the data sheet for the ADS1015 is a good reference. It measures current in both directions. The classic differential amplifier at its input is quite easy to protect from over-voltages due to the resistors at both inputs.

    The way I see it, your invention might have a couple of spring-loaded pogo pins that are spaced just right to come down on top of the fuse nibs, then give you an instant reading on your display. When powered by your 9V battery, it remains a two terminal device. The app-note refers to "low-side current shunt monitoring", but if you power it with 9V independent of the car battery, it doesn't care, low, high or middle, all the same.

    Since there will be a microcontroller (Prop) in this, it is open for rampant feature creep. I really like the idea of auto-test of the fuse resistance by means of a pulse of current, so your display could show amps instead of just mV, and without having to look up the typical fuse resistance in a table. A pin could trigger the current pulse in a two step measurement. How to produce and measure a short current pulse that won't wear down your 9V battery is an interesting circuit problem, keeping it simple but sufficient.

  • kwinnkwinn Posts: 7,820
    With a Propeller in the circuit it might be possible to measure the discharge time of a small (1 - 10uf) capacitor to calculate the resistance, or provide a constant current for a short period of time and measure the change in mV across the fuse.
    In science there is no authority. There is only experiment.
    Life is unpredictable. Eat dessert first.
  • jmgjmg Posts: 11,273
    ...
    Since there will be a microcontroller (Prop) in this, it is open for rampant feature creep. I really like the idea of auto-test of the fuse resistance by means of a pulse of current, so your display could show amps instead of just mV, and without having to look up the typical fuse resistance in a table. A pin could trigger the current pulse in a two step measurement. How to produce and measure a short current pulse that won't wear down your 9V battery is an interesting circuit problem, keeping it simple but sufficient.
    kwinn wrote: »
    With a Propeller in the circuit it might be possible to measure the discharge time of a small (1 - 10uf) capacitor to calculate the resistance, or provide a constant current for a short period of time and measure the change in mV across the fuse.
    A variant of that would be a deliberate skew of charge:discharge of a impulse Cap, to ease peak battery loading.
    eg for a simple example, constant current charge at 2mA and constant current discharge at 100mA is balanced with a 50:1 time ratio.

  • I was thinking the capacitor discharge too, measuring the peak voltage across the fuse=shunt. It would be a fast discharge phase indeed, thinking maybe 50 microseconds to discharge 1000µF through 50 milliohms. It's 100V/second slew at even 100mA constant current. Peak current could be slowed to 100mA by the current limiter, or maybe even by a simple series inductor if the initial capacitor energy is fixed.

    Rather than a faster ADC, I'd be inclined to put a peak-detector in front of a slower ADC. Another prop pin would serve for reset and for a straight-through connection when measuring the normal automotive amps. Recharge the capacitor to 9V could be at leisure.
  • jmgjmg Posts: 11,273
    ... Peak current could be slowed to 100mA by the current limiter...
    Yes, once you do that, the impulse becomes flat, and a peak capture is no longer needed.
    Some simple discharge ball parks
    100u*6/60m = 0.01 ie 100uF and 60mA is 10mS to discharge 6V, or 10uF gives a 1ms flat top current capture window.



  • I see there has been activity during my absence. In the meantime, I have been tossing dirt with a shovel. Why am I tossing dirt with a shovel, you ask? Because I cannot be trusted with a Bobcat. $175.00 towing bill. It took the driver longer to write up the ticket, than it did to tow me out :) LOL That was a scary ride :)

    I am sure by now, that many have learned a lot about this particular subject, and many have danced around the subject, without divulging too much. Anyhow I am envisioning a handy and useful tool for fuse block testing, and I can see it all clearly in my minds eye for the perfect tool, except for the electronics to make it work reliably.

    As it appears now, several gurus are in agreement. Is it possible for us to work on a collaborative effort?
    1512 x 2016 - 2M


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • pjvpjv Posts: 1,898
    Tracey;

    How would you get an accurate fuse resistance with just two pogo pins onto the fuse nibs? A Kelvin connection will be needed as you can't be sure of the pogo pin-to-nib resistance.

    Cheers,
    Peter (pjv)
  • idbruceidbruce Posts: 5,648
    edited May 7 Vote Up0Vote Down
    Tracy
    How would you get an accurate fuse resistance with just two pogo pins onto the fuse nibs? A Kelvin connection will be needed as you can't be sure of the pogo pin-to-nib resistance.

    I agree. Besides, I already have a plan. At this point, the circuitry is the most important aspect of this project. Whether you all trust my minds eye is different subject :)

    Keep in mind, this is always in my mind and at my disposal: http://www.porta-melt.com

    Let the tool casting commence :)


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • Tracy AllenTracy Allen Posts: 6,221
    edited May 7 Vote Up0Vote Down
    Bruce, that's good you have a plan, you always do!

    The test points on the only one I have here are recessed below the surface of the plastic (rather than nubs that stick out on the top), and they take the form of short forks, formed not surprisingly out of the same material as the blade. It does look like a suitable spring-loaded pogo could lock in there. But variation? On the Bussman ATC data sheet it says you can order them either with or without the test points. The DC resistance is not among the specifications, only the holding current and the graph of break time vs fault current. That is the bottom line after all.

    Testing a 4 amp ATC fuse here, I see a resistance of 0.03 ohms. How consistent that is that, and across manufacturers? Higher amp values will have less resistance--Is it inversely to amperage rating? 100mA thru 0.03 ohm is only 3 millivolts. What say to a heftier pulse of test current? Peter, good point indeed about contact resistance, especially in a product that is to be used casually in a potentially grimy environment. Keep it simple and just read the millivolts and the color of the fuse.


  • jmgjmg Posts: 11,273
    pjv wrote: »
    How would you get an accurate fuse resistance with just two pogo pins onto the fuse nibs? A Kelvin connection will be needed as you can't be sure of the pogo pin-to-nib resistance.
    True, if you were chasing micro-ohms of precision, you would inject the current on another connection.
    The most accurate reading will be the voltage, and the current injection allows some indication of current, via the indicated resistance, and it would be good for sanity checking any fuse table.


  • I still think leaving the shunt resistor onboard makes the most sense. It will be in parallel to the resistance of the fuse, so the fuse resistance can be determined from the parallel resistance formula. The big advantage is that you don't have to worry about the over-voltage consequences of a blown fuse.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • jmgjmg Posts: 11,273
    I still think leaving the shunt resistor onboard makes the most sense. It will be in parallel to the resistance of the fuse, so the fuse resistance can be determined from the parallel resistance formula. The big advantage is that you don't have to worry about the over-voltage consequences of a blown fuse.
    ? You mean on the breakout board ? - if the fault that caused the fuse to blow, is still there, you have just placed 12V across the shunt resistor. Instant smoke.
    Even if not fuse-blow levels, some automotive (eg) lamp loads will stress a SMD 0.1R

  • idbruceidbruce Posts: 5,648
    edited May 8 Vote Up0Vote Down
    Tracy

    Yes, they would be recessed, because if any metal object was to contact the two protrusions, you would have instant problems. They are recessed to prevent shorts.

    I really did not want to share this information, although it is readily available, but just so everyone has a better understanding. https://www.powerprobe.com/fuse-voltage-drop-charts/

    EDIT: one protrusion and Ground = short
    EDIT: however, as per
    if any metal object was to contact the two protrusions
    you would eliminate the over-current protection device


    Novel Solutions - http://www.novelsolutionsonline.com/ - Machinery Design • - • Product Development
    "Necessity is the mother of invention." - Author unknown.

  • Tracy AllenTracy Allen Posts: 6,221
    edited May 8 Vote Up0Vote Down
    Ahh, that chart is sort of what I was looking for. I looked up my 4 amp ATC fuse, and the chart says 5mV across it would mean 219 mA flowing through it. That translates to 5/0.219 = 22.8 milliohms. Checking other mV readings for that fuse in the chart--what a surprise!--They obey Ohm's Law.

    I had measured 30 milliohms using the ohmmeter (Agilent 34401 6.5 digits) earlier on my specimen. To check, I hooked it up to a constant current power supply (Agilent 6612) set at 0.219A. The voltage across the fuse was 6.6mV, which does translates to 30 milliohms. So there is a start on an error bar, 22.8 mΩ vs 30mΩ.

    In the ballpark anyway.

    About the Adafruit breakout board. Assuming you do remove the shunt to try it out, do add two resistors ~1kΩ between the (+) and (-) inputs and your fuse under test, and on the breakout side a pair of back to back parallel silicon diodes to clamp the differential voltage. Run the ADC on its own battery power, not from the car's power.
Sign In or Register to comment.