Loading ...
Sorry, an error occurred while loading the content.

Re: Current Sensing --- WOW - That's a neat chip!

Expand Messages
  • lcdpublishing
    Crap, you think you were confused - HELP! Reality check for me at this point. I was looking at the adapters from Sparfun then it dawned on me - I am running
    Message 1 of 28 , Jun 1, 2008
    • 0 Attachment
      Crap, you think you were confused - HELP!

      Reality check for me at this point. I was looking at the adapters from
      Sparfun then it dawned on me - I am running 6.5 amps through that thing
      and up to as much as 30 amps (if used to capability). How on earth is
      that little device (along with tiny traces) going to handle 30 amps?

      After a moment of panic, I then realized I must be wrong with how to
      wire this thing. I am thinking that the ACS712 is wired in series with
      the positive leg to the motor. What I mean by that is my +24 wire goes
      into pins 1&2 comes out at pins 3&4, then continues onto the motor.

      I am now thinking that I put this in a parallel circuit next to the
      positive lead of my motor so that not ALL the current passes through
      this thing.

      HELP :-(

      Chris
    • rtstofer
      ... Well, you certainly need wide traces on both the top and bottom layers. ... That s the way I see it. The current conduction path is INSIDE the package.
      Message 2 of 28 , Jun 1, 2008
      • 0 Attachment
        --- In Electronics_101@yahoogroups.com, "lcdpublishing"
        <lcdpublishing@...> wrote:
        >
        > Crap, you think you were confused - HELP!
        >
        > Reality check for me at this point. I was looking at the adapters from
        > Sparfun then it dawned on me - I am running 6.5 amps through that thing
        > and up to as much as 30 amps (if used to capability). How on earth is
        > that little device (along with tiny traces) going to handle 30 amps?


        Well, you certainly need wide traces on both the top and bottom layers.


        >
        > After a moment of panic, I then realized I must be wrong with how to
        > wire this thing. I am thinking that the ACS712 is wired in series with
        > the positive leg to the motor. What I mean by that is my +24 wire goes
        > into pins 1&2 comes out at pins 3&4, then continues onto the motor.


        That's the way I see it. The current conduction path is INSIDE the
        package. Look at the "Functional Block Diagram" where it shows the
        internal path in bold.


        >
        > I am now thinking that I put this in a parallel circuit next to the
        > positive lead of my motor so that not ALL the current passes through
        > this thing.


        I don't think so...

        Look at it this way: not only do you have to conduct as much as 30A
        through this device, you also have to have similar traces going to the
        MOSFET and whatever terminals you use. The trick will be to keep the
        traces thick, wide and VERY SHORT.

        There is a table (Figure 4) that relates trace width/copper thickness
        to trace width and temperature rise. Basically, you will want 2 oz
        cladding and 1/4" traces to get 12.3A ampacity with 10 deg C rise.

        See:
        http://www.egr.msu.edu/classes/ece480/goodman/fall/group03/Documents/app_note_ben_levesque.pdf

        Since you can have traces on both sides they can be smaller but I
        would go as large as possible. It wouldn't bother me if the traces
        were twice too large.

        In my 30A application, I was using a commercial golf cart controller
        (GCC) and measuring with an external 50 mV shunt. None of the motor
        current was anywhere near my controller. I just used optically
        isolated PWM as the input to the GCC. Now, I am certain that the GCC
        had high current traces on its' PCB but I didn't have to design them.

        Richard
      • Shawn Upton
        It s hard to believe, but yes, it s meant to take 30A continously. Well, perhaps by using the PCB traces as heatsinking too, so don t let the traces do all
        Message 3 of 28 , Jun 1, 2008
        • 0 Attachment
          It's hard to believe, but yes, it's meant to take 30A continously. Well, perhaps by using the PCB traces as heatsinking too, so don't let the traces do all the heating--the sensor is a 1.2milliohm resistor afterall. Just make sure other high current paths don't come too close, or you may have crosstalk. There is an app note on it, and I want to say keep all other high current wires a good inch away.

          If you think 30A is high, check out the ACS760. 20A continous on a 28 pin TSSOP. Granted, it uses 6 pins per side; but it seems just as bad! Worse, it's spec'd to take some serious fault currents too, way in excess of that 20A. [But this sensor isn't what you're looking for, just bringing it up for reference.]

          Shawn Upton, KB1CKT



          ----- Original Message ----
          From: lcdpublishing <lcdpublishing@...>
          To: Electronics_101@yahoogroups.com
          Sent: Sunday, June 1, 2008 6:37:34 AM
          Subject: [Electronics_101] Re: Current Sensing --- WOW - That's a neat chip!


          Crap, you think you were confused - HELP!

          Reality check for me at this point. I was looking at the adapters from
          Sparfun then it dawned on me - I am running 6.5 amps through that thing
          and up to as much as 30 amps (if used to capability). How on earth is
          that little device (along with tiny traces) going to handle 30 amps?

          After a moment of panic, I then realized I must be wrong with how to
          wire this thing. I am thinking that the ACS712 is wired in series with
          the positive leg to the motor. What I mean by that is my +24 wire goes
          into pins 1&2 comes out at pins 3&4, then continues onto the motor.

          I am now thinking that I put this in a parallel circuit next to the
          positive lead of my motor so that not ALL the current passes through
          this thing.

          HELP :-(

          Chris




          [Non-text portions of this message have been removed]
        • lcdpublishing
          Oh boy, this is going to be a fun one to let the smoke out of :-) I do have some heavy PCB stock here. I will try to pick off a bit of copper to measure its
          Message 4 of 28 , Jun 1, 2008
          • 0 Attachment
            Oh boy, this is going to be a fun one to let the smoke out of :-) I
            do have some heavy PCB stock here. I will try to pick off a bit of
            copper to measure its thickness but I suspect it is 2 oz. and is
            double sided.

            Thanks Richard.

            Chris


            --- In Electronics_101@yahoogroups.com, "rtstofer" <rstofer@...>
            wrote:
            >
            > --- In Electronics_101@yahoogroups.com, "lcdpublishing"
            > <lcdpublishing@> wrote:
            > >
            > > Crap, you think you were confused - HELP!
            > >
            > > Reality check for me at this point. I was looking at the
            adapters from
            > > Sparfun then it dawned on me - I am running 6.5 amps through that
            thing
            > > and up to as much as 30 amps (if used to capability). How on
            earth is
            > > that little device (along with tiny traces) going to handle 30
            amps?
            >
            >
            > Well, you certainly need wide traces on both the top and bottom
            layers.
            >
            >
            > >
            > > After a moment of panic, I then realized I must be wrong with how
            to
            > > wire this thing. I am thinking that the ACS712 is wired in
            series with
            > > the positive leg to the motor. What I mean by that is my +24
            wire goes
            > > into pins 1&2 comes out at pins 3&4, then continues onto the
            motor.
            >
            >
            > That's the way I see it. The current conduction path is INSIDE the
            > package. Look at the "Functional Block Diagram" where it shows the
            > internal path in bold.
            >
            >
            > >
            > > I am now thinking that I put this in a parallel circuit next to
            the
            > > positive lead of my motor so that not ALL the current passes
            through
            > > this thing.
            >
            >
            > I don't think so...
            >
            > Look at it this way: not only do you have to conduct as much as 30A
            > through this device, you also have to have similar traces going to
            the
            > MOSFET and whatever terminals you use. The trick will be to keep
            the
            > traces thick, wide and VERY SHORT.
            >
            > There is a table (Figure 4) that relates trace width/copper
            thickness
            > to trace width and temperature rise. Basically, you will want 2 oz
            > cladding and 1/4" traces to get 12.3A ampacity with 10 deg C rise.
            >
            > See:
            >
            http://www.egr.msu.edu/classes/ece480/goodman/fall/group03/Documents/a
            pp_note_ben_levesque.pdf
            >
            > Since you can have traces on both sides they can be smaller but I
            > would go as large as possible. It wouldn't bother me if the traces
            > were twice too large.
            >
            > In my 30A application, I was using a commercial golf cart controller
            > (GCC) and measuring with an external 50 mV shunt. None of the motor
            > current was anywhere near my controller. I just used optically
            > isolated PWM as the input to the GCC. Now, I am certain that the
            GCC
            > had high current traces on its' PCB but I didn't have to design
            them.
            >
            > Richard
            >
          • lcdpublishing
            It sure is amazing to think that all this current can pass not only through the chip, but also the traces! I can t help but try to compare it all to house
            Message 5 of 28 , Jun 1, 2008
            • 0 Attachment
              It sure is amazing to think that all this current can pass not only
              through the chip, but also the traces! I can't help but try to compare
              it all to house wiring and the heavy gauge conductors used there -
              while at less amps!

              This should be fun and exciting to watch if it all goes up in smoke :-)

              Chris


              --- In Electronics_101@yahoogroups.com, Shawn Upton <kb1ckt@...> wrote:
              >
              > It's hard to believe, but yes, it's meant to take 30A continously.
              Well, perhaps by using the PCB traces as heatsinking too, so don't let
              the traces do all the heating--the sensor is a 1.2milliohm resistor
              afterall. Just make sure other high current paths don't come too
              close, or you may have crosstalk. There is an app note on it, and I
              want to say keep all other high current wires a good inch away.
              >
              > If you think 30A is high, check out the ACS760. 20A continous on a
              28 pin TSSOP. Granted, it uses 6 pins per side; but it seems just as
              bad! Worse, it's spec'd to take some serious fault currents too, way
              in excess of that 20A. [But this sensor isn't what you're looking for,
              just bringing it up for reference.]
              >
              > Shawn Upton, KB1CKT
              >
              >
            • Shawn Upton
              Don t forget: household wiring goes for many many feet, while most PCB traces are considered long when they hit like 2 inches. Also, what is a 2V drop on a
              Message 6 of 28 , Jun 1, 2008
              • 0 Attachment
                Don't forget: household wiring goes for many many feet, while most PCB traces are considered "long" when they hit like 2 inches. Also, what is a 2V drop on a 100' run of 14g, running your TV or lights or ___? Nothing, really. But what is a 2V drop in your 24V circuit? In a 5V circuit?

                Sometime ago, I was concerned about trace heating. I found a graph that indicated that I could actually run an amp through a 1oz 10mil wide trace. Ridiculously lossy, but supposedly it won't fuse. Now, I would not go about design 10mil traces to carry that, but IMO for most (low power) electronics the voltage drop becomes an issue sooner than heat and fusing problems. Just my opinion, though. If you make a board, I'd spring for 2oz copper if you can. Stitching both sides of the PCB together also helps with resistance issues.

                Shawn Upton, KB1CKT



                ----- Original Message ----
                From: lcdpublishing <lcdpublishing@...>
                To: Electronics_101@yahoogroups.com
                Sent: Sunday, June 1, 2008 4:05:41 PM
                Subject: [Electronics_101] Re: Current Sensing --- WOW - That's a neat chip!


                It sure is amazing to think that all this current can pass not only
                through the chip, but also the traces! I can't help but try to compare
                it all to house wiring and the heavy gauge conductors used there -
                while at less amps!

                This should be fun and exciting to watch if it all goes up in smoke :-)

                Chris

                _._,___




                [Non-text portions of this message have been removed]
              • cresswellavenue@talktalk.net
                Sadly 2 pieces of wire arent always the same. I did some experiments on wire a good few years ago and some wire is much more lossy than others. This means you
                Message 7 of 28 , Jun 1, 2008
                • 0 Attachment
                  Sadly 2 pieces of wire arent always the same.

                  I did some experiments on wire a good few years ago and some wire is much
                  more lossy than others.
                  This means you will get serious voltage drops if much current is taken.

                  We bought various makes of wire and found the best and used that.



                  -------Original Message-------

                  From: Shawn Upton
                  Date: 06/01/08 23:57:06
                  To: Electronics_101@yahoogroups.com
                  Subject: Re: [Electronics_101] Re: Current Sensing --- WOW - That's a neat
                  chip!

                  Don't forget: household wiring goes for many many feet, while most PCB
                  traces are considered "long" when they hit like 2 inches. Also, what is a 2V
                  drop on a 100' run of 14g, running your TV or lights or ___? Nothing, really
                  But what is a 2V drop in your 24V circuit? In a 5V circuit?

                  Sometime ago, I was concerned about trace heating. I found a graph that
                  indicated that I could actually run an amp through a 1oz 10mil wide trace.
                  Ridiculously lossy, but supposedly it won't fuse. Now, I would not go about
                  design 10mil traces to carry that, but IMO for most (low power) electronics
                  the voltage drop becomes an issue sooner than heat and fusing problems. Just
                  my opinion, though. If you make a board, I'd spring for 2oz copper if you
                  can. Stitching both sides of the PCB together also helps with resistance
                  issues.

                  Shawn Upton, KB1CKT

                  ----- Original Message ----
                  From: lcdpublishing <lcdpublishing@...>
                  To: Electronics_101@yahoogroups.com
                  Sent: Sunday, June 1, 2008 4:05:41 PM
                  Subject: [Electronics_101] Re: Current Sensing --- WOW - That's a neat chip!

                  It sure is amazing to think that all this current can pass not only
                  through the chip, but also the traces! I can't help but try to compare
                  it all to house wiring and the heavy gauge conductors used there -
                  while at less amps!

                  This should be fun and exciting to watch if it all goes up in smoke :-)

                  Chris

                  _._,___

                  [Non-text portions of this message have been removed]





                  [Non-text portions of this message have been removed]
                • azhar aslam
                  friend can someone send me some conceptual data or notes about bandwidth.
                  Message 8 of 28 , Jun 4, 2008
                  • 0 Attachment
                    friend can someone send me some conceptual data or notes about
                    bandwidth.
                  • azhar aslam
                    friend i find it difficulty in coupling capacitor.please help me in this matter
                    Message 9 of 28 , Jun 4, 2008
                    • 0 Attachment
                      friend i find it difficulty in coupling capacitor.please help me in
                      this matter
                    • Leon
                      ... From: azhar aslam To: Sent: Thursday, June 05, 2008 7:53 AM Subject: [Electronics_101] about
                      Message 10 of 28 , Jun 5, 2008
                      • 0 Attachment
                        ----- Original Message -----
                        From: "azhar aslam" <azhar_eaggle1@...>
                        To: <Electronics_101@yahoogroups.com>
                        Sent: Thursday, June 05, 2008 7:53 AM
                        Subject: [Electronics_101] about bandwidth


                        > friend can someone send me some conceptual data or notes about
                        > bandwidth.

                        <http://www.google.co.uk/search?num=100&hl=en&sa=X&oi=spell&resnum=0&ct=result&cd=1&q=bandwidth&spell=1>

                        Leon
                      • John Popelish
                        ... (snip) In general, a coupling capacitor is intended to pass an AC signal from one DC bias voltage to another. The value of capacitance required to perform
                        Message 11 of 28 , Jun 5, 2008
                        • 0 Attachment
                          azhar aslam wrote:
                          > friend i find it difficulty in coupling capacitor.
                          (snip)

                          In general, a coupling capacitor is intended to pass an AC
                          signal from one DC bias voltage to another. The value of
                          capacitance required to perform this task is related to the
                          AC impedance of booth the source and load sides of this
                          circuit. At the lowest AC frequency that must be coupled,
                          the AC voltage lost across the capacitor should be a small
                          compared to the signal that is passes to the load side. For
                          instance, if an amplifier stage with a 5000 ohm output
                          impedance is being coupled to a second amplifier stage with
                          a 20,000 ohm input impedance, and the lowest frequency to be
                          coupled is 20 Hz, half of the signal energy will be blocked
                          by the coupling capacitor if its impedance at 20 Hz is equal
                          to the sum of the output impedance of the source amplifier
                          and the input impedance of the load amplifier (25,000 ohms,
                          total).

                          The formula for the magnitude of capacitive impedance is
                          Xc=1/(2*pi*f*C). Setting Xc to 25k ohms and f to 20 Hz and
                          solving for C produces a result of 0.32uF. Any capacitance
                          smaller than that would drop a more significant part of the
                          total signal and capacitance larger than that would drop a
                          less significant part of the total signal.

                          If you want the coupling capacitor to act as something of a
                          filter that couples the signal while blocking AC frequencies
                          of lower frequency, you might choose a value near the "half
                          energy lost" size for the lowest "intended to be coupled"
                          frequency. But if you want the coupling capacitor to be
                          essentially invisible at the lowest "intended to be
                          coupled" frequency, you might choose a capacitor 10 times
                          larger than the above formula suggests, with the
                          understanding that it will not act as an effective filter to
                          block somewhat lower frequencies from passing through, if
                          they are also present in the source signal.

                          If you choose an electrolytic capacitor as the coupling
                          component, then you must make sure it is oriented so that
                          the DC voltage across it has the correct polarity, so that
                          its oxide insulation is not damaged by reverse DC.

                          --
                          Regards,

                          John Popelish
                        • Shawn Upton
                          One note on what John stated: make sure the capacitor is rated at the frequency you want to use it at, in particular for RF applications. The capacitive
                          Message 12 of 28 , Jun 6, 2008
                          • 0 Attachment
                            One note on what John stated: make sure the capacitor is rated at the frequency you want to use it at, in particular for RF applications. The capacitive reactance needs to be low enough, but often it does not need to be micro-ohms. It threw me for a while, looking at HF circuits, that 0.01uF coupling capacitors would be used instead of 0.1uF--why would they do that, I wondered, when it would be more tightly coupled with a 0.1uF? Took a long time before I found out that a leaded 0.1uF might have an SRF of maybe 10MHz--and much less with the typical experimenters lead lengths. As you continue up into UHF, expect coupling capacitors to go to 1nf and then 100pF, finally down to 10pF up in the GHz region.

                            Shawn Upton, KB1CKT



                            ----- Original Message ----
                            From: azhar aslam <azhar_eaggle1@...>
                            To: Electronics_101@yahoogroups.com
                            Sent: Thursday, June 5, 2008 2:57:43 AM
                            Subject: [Electronics_101] coupling capacitor


                            friend i find it difficulty in coupling capacitor.please help me in
                            this matter

                            __._,___




                            [Non-text portions of this message have been removed]
                          • Roy J. Tellason
                            ... Good point. Now if one were buying caps from some large mfr. it would probably be easy enough to find out what the SRF was, but what if one does a lot of
                            Message 13 of 28 , Jun 7, 2008
                            • 0 Attachment
                              On Friday 06 June 2008 20:31, Shawn Upton wrote:
                              > One note on what John stated: make sure the capacitor is rated at the
                              > frequency you want to use it at, in particular for RF applications. The
                              > capacitive reactance needs to be low enough, but often it does not need to
                              > be micro-ohms. It threw me for a while, looking at HF circuits, that
                              > 0.01uF coupling capacitors would be used instead of 0.1uF--why would they
                              > do that, I wondered, when it would be more tightly coupled with a 0.1uF?
                              > Took a long time before I found out that a leaded 0.1uF might have an SRF
                              > of maybe 10MHz--and much less with the typical experimenters lead lengths.
                              > As you continue up into UHF, expect coupling capacitors to go to 1nf and
                              > then 100pF, finally down to 10pF up in the GHz region.

                              Good point. Now if one were buying caps from some large mfr. it would
                              probably be easy enough to find out what the SRF was, but what if one does a
                              lot of salvage, as I do? Are particular types better at higher frequencies?
                              (I see much more use of ceramics at higher frequency circuits.) Is there an
                              easy way to measure it?

                              --
                              Member of the toughest, meanest, deadliest, most unrelenting -- and
                              ablest -- form of life in this section of space,  a critter that can
                              be killed but can't be tamed.  --Robert A. Heinlein, "The Puppet Masters"
                              -
                              Information is more dangerous than cannon to a society ruled by lies. --James
                              M Dakin
                            • Shawn Upton
                              Well, not all datasheets are created equal. :) A few years ago, I wanted to find the thermal impedance of the quarter-Watt 1% resistors we used at work.
                              Message 14 of 28 , Jun 7, 2008
                              • 0 Attachment
                                Well, not all datasheets are created equal. :) A few years ago, I wanted to find the thermal impedance of the quarter-Watt 1% resistors we used at work. Could not find it, nor anything for any quarter-Watt resistors. To this day, I'm still not sure just what kind of drift these 100ppm/C resistors will have, per mW of heating. I have read that the Pd of surface mount resistors are strongly affected by the traces used on them--but sometimes a conservative estimate is all that is needed.

                                Similarly, you might not find ESR, ESL, SRF, and other factors in all types of cap datasheets. But looking at a few should give some idea as to what a particular type will give for results. I have an AVX SMT cap datasheet at work that shows about 15MHz SRF for 1206 caps. Oddly enough, SRF does not move with respect to case size (0805 does not resonate higher) but IIRC, the 1206 definately showed more of a null at the SRF--meaning, less ESR--which is *not* specified either. Q was specified but at a pretty low frequency.

                                Anyhow, I figure, short leads, a 0.1uF cap should be up around 10MHz; the SMT's are at 15MHz. Assuming the ESL is "fixed" for a given capacitor package, going down by a factor of 10 should move the SRF up by a factor of 3. Mono's (and ceramic disk) are good at high frequencies; electrolytics are ok at audio and perhaps a bit higher. Tantalums I've read are upwards of a few MHz. Poly's should be inbetween. Lastly, you can use a cap on the other side of the SRF number--but the impedance will be rising instead of falling. Which may or may not matter, depending upon the application.

                                I've thought about making an ESR tester, but in the end I figure I'd be better of just getting a wide band RF generator. Why measure at 1kHz or 10kHz or whatever, when I could just try to measure at the desired frequency? My first thought would be to pass RF through the capacitor and into a 50ohm dummy load, with an SWR meter off the generator. When the SWR rises to (3:1? 2:1? I'd have to do the math to find the right value), then the capacitor is on the other side of the SRF. With enough measurements one might be able to figure out where the SRF is, but this test would probably only be good to quickly find the upper limit of a capacitor, not where the SRF actually is.

                                So my second thought would be to to run the generator into an attenuator (to keep the generator "happy"), then pass the RF through a resistor and then to the cap, to form a LP. As the frequency is swept up, the voltage out of the filter should go down--until the SR is hit, after which it will start rising. The resistor size has to be proper in order to avoid trying to measure microvolts out of the network; and stray capacitance across the capacitor under test will kill the testing at very high frequencies.

                                Shawn Upton, KB1CKT



                                ----- Original Message ----
                                From: Roy J. Tellason <rtellason@...>
                                To: Electronics_101@yahoogroups.com
                                Sent: Saturday, June 7, 2008 2:03:09 PM
                                Subject: Re: [Electronics_101] coupling capacitor




                                Good point. Now if one were buying caps from some large mfr. it would
                                probably be easy enough to find out what the SRF was, but what if one does a
                                lot of salvage, as I do? Are particular types better at higher frequencies?
                                (I see much more use of ceramics at higher frequency circuits.) Is there an
                                easy way to measure it?

                                ,___




                                [Non-text portions of this message have been removed]
                              Your message has been successfully submitted and would be delivered to recipients shortly.