AC/DC hall effect sensor circuit

Discussion in 'General Electronics Chat' started by frascati, Jun 1, 2010.

  1. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    I asked this a couple of years back and finally gave up in frustration (cannot even remember my username). I'm a novice electronics tinkerer(very novice), and was trying to add a process to my Tig welder. All I want to do is trigger a common relay to energize a 115vac circuit whenever there is current detected in that cable. I recently found the old attempt in a box in the corner and thought I'd give it another try.

    The welder cable carries from 5 to 300 amps, AC/DC, during operation.
    I got a hall effect sensor along with a simple diagram and a snap-on ferrite choke and constructed the folllowing..http://img201.imageshack.us/img201/7934/imgp0006cr1.jpg

    I drew a model with Google Sketchup of what is inside that...
    http://img234.imageshack.us/img234/8108/ferriteiti0.jpg

    It's been so long that I do not have a copy of the circuit used anymore.
    But I was unsuccessful.

    Would the following circuit work for my purpose...
    http://www.discovercircuits.com/DJ-Circuits/pumpmotormon2.htm

    I discovered that in a search for Hall Effect on this site, although it does not appear to use a hall effect device. I've been under the impression that to detect very low current, ac and dc, that there were no alternatives to hall effect sensors.

    I was hoping that two years later there might actually be a discrete IC of some
    sort that would detect a low amperage magnetic field and have pin
    inputs for low supply voltage and simple switch pin outputs.

    I'm determined, for a number of reasons, to do this with a clip on device. I don't want to hardwire any device to the welder or disassemble or tap into any part of it.

    What the simplest method of accomplishing this?
     
    Last edited: Jun 1, 2010
  2. rjenkins

    AAC Fanatic!

    Nov 6, 2005
    1,015
    69
    I'd try a simple 1000:1 current transformer feeding a bridge rectifier then a chain of 1N4000 series diodes to 'clamp' the voltage; 6 diodes would give about 4V.

    Add eg. a 1K resistor across the diodes and you have 0V when no current and 4V when current flows.

    A current transformer gives out a proportional fraction of the primary current, so a 1000:1 type would give 5 to 300mA out for 5 to 300A through it.

    Using 1A diodes for the bridge and clamp chain would allow up to 1000A on the main cable without problems, and the 1K resistor would cause the voltage to drop down below about 4A.

    You MUST always have a load circuit solidly connected, as the output can give thousands of volts if it does not have a circuit for the current.

    I'd solder the bridge rec and clamp diodes directly at the CT teminals to avoid the chance of any wiring getting disturbed.

    Current transformers are made to be threaded over cables and busbars etc. and are available with various size center holes, quite cheaply - probably $10 to $15, certainly not crazy prices.

    Example:
    http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=582-1017-ND

    And a the makers info, including built current switches:
    http://dkc3.digikey.com/PDF/UK2010/p2418.pdf
     
  3. someonesdad

    Senior Member

    Jul 7, 2009
    1,585
    141
    The current transformer and Johnson's circuit will only work when an AC current is flowing in the cable. If you also want to detect DC, as you stated, you'll have to use the Hall sensor. You could conceivably use the inductive methods with DC, but you'd only be able to detect the transient current on and current off conditions.

    If you use the Hall sensor, you must make sure you have the sensor properly aligned with the direction of the magnetic field. The field around the wire will be tangential (see any elementary physics text for a diagram); thus, for example, in your diagrams, if the sensor detects the field component parallel to the leads, the design is incorrect. The sensor's datasheet will tell you how it should be applied.

    Here's another option to think about if you want a quick solution and you're willing to throw some money at it. There are clamp-on meters on the market that will detect the AC and DC currents. These devices output a voltage, so you'd just need to e.g. add a comparator. Depending on the device you chose, you might have to switch the device between AC and DC detection and perhaps adjust the comparator's input for detecting different current levels. It shouldn't take long to breadboard the comparator circuit and get something working.

    Even though you've indicated you don't want to mess with the welder itself, that's probably your best direction to investigate first. It's possible that the signal you want is already present inside the welder and you just need to get your hands on it. This could probably be decided if you can get your hands on the schematic. And it's not inconceivable that the manufacturer sells an option that does something close to what you want. It wouldn't hurt contacting the manufacturer with what you want to do and see if there's already a solution available.
     
  4. rjenkins

    AAC Fanatic!

    Nov 6, 2005
    1,015
    69
    Another thought for a simple solution.

    Try a tiny magnetic reed switch, as used in burglar alarm door contacts.

    Get a bit of mild steel around half an inch wide and a couple of inches long and form it into a horse shoe 'U' shape around one of the welder output cables, then sit the reed switch across the end of the 'U' so the metal strip is against the glass just in from the ends.

    It won't work down to 5A, but should click in at any reasonable welding current.

    If you want to make it more sensitive, you would have to put two or three turns of cable through the steel core.
     
  5. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    Looping the cable is not an option. It is 12mm water cooled. I want to be able to apply this detection circuit to any welding machine, whereever I happen to be. That eliminates tapping the argon gas solenoid switch, tapping into the pedal/rheo plug, removing accessories, cutting cables, etc.

    What If I were to frame the matter more directly?

    I have a 16ga stranded copper extension cord. Fortunately there is a point where I can separate the hot from the ground wires. I'd like to detect when 5amps of current, ac or dc, is flowing through the hot wire using a clip on sensor. When current flows, a relay lights a 115vac light bulb on my test bench.

    Is there not a relatively simple circuit available that utilizes a common omnipolar hall effect chip that can do this? Seems like it would be a junior high science level experiment. Unfortunately I never even got that far in science. There must be a circuit floating out there somewhere.
     
  6. bertus

    Administrator

    Apr 5, 2008
    15,635
    2,342
  7. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    Bertus thank you very much for that. However, I've accumulated a couple of gigabytes of application notes and primers on Hall effect devices. I've actually read many of them end to end. I know just enough about them to be pretty certain that they (or magnetoresistive devices) are precisely what I need for this little project. But I'm a novice at circuits. I am competent enough only to 'paint by numbers' if I have a diagram at hand. But I cannot seem to locate one for this application. It seems like it would be a very very common application indeed.

    In fact I have a suspicion that because it is such a basic operation, there may even be an IC with pinouts dedicated to just this task. I just can't find it.
     
  8. bertus

    Administrator

    Apr 5, 2008
    15,635
    2,342
  9. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    That was one of the documents I read some time ago. In fact that section, and others, inspired at least this much progress...
    [​IMG]
    [​IMG]
    [​IMG]


    www.gmw.com/magnetic_sensors/asahi/documents/eq-711l_gmw.pdf
    (need to cut and paste in browser to access)
    but the circuit did not prove to be sufficient. Or I erred in constructing it.

    Here was the board..
    [​IMG]

    I'm pretty close I think. I'm just struggling like mad to find a tested proven circuit that can accomplish the comparatively rudimentary task of energizing a relay when current is detected in a cable. This is, along with position sensing, among the most common applications of hall effect sensors. Yet I can't find a simple circuit to unlock the genie.

    Here's a cutaway of the cable that is being measured. It is the braided cover cable that the sensor is snapped onto in the first photo above...
    [​IMG]
    The braided copper is only separated from the hall sensor by the protective jacket, and not by the water jacket.
     
    Last edited: Jun 2, 2010
  10. bertus

    Administrator

    Apr 5, 2008
    15,635
    2,342
    Hello,

    Do you have a schematic of the shown board?

    I also found an other PDF for switching a relays using a hall sensos.
    See figure 20 of the attached PDF.

    The only difference with your shown sensor is , it is a switching sensor and it works on 12 Volts.

    Bertus
     
  11. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    Bertus, here is the diagram I tried
    http://i802.photobucket.com/albums/yy303/svejkovat/SDC11293.jpg

    Another, perhaps even more promising chip is this one...
    http://www.gmw.com/magnetic_sensors/sentron/csa/CSA-1.html
    And here is a wealth of application data on it....
    http://www.sentron.ch/csa_appl_notes.htm

    incorporating it into a snap on clamp would not be hard.


    And here's a page that illustrates just how maddeninging this search can be for a non engineer...
    http://www.dacomwest.de/de/mlx_hall.htm

    And those are just the offerings of one manufacturer. I know exactly what I want to accomplish. I know that it requires a hall effect device. But I don't have quite enough knowledge to identify the right device/circuit.

    In my favor is the fact that I don't require measurement/accuracy/scaling.
    Simply on/off.

    Working against me is the conductor itself...
    [​IMG]
    Which is relatively unique for the purposes of calculating gauss/current effects. Also working against me is the fact that I want to detect current as low as 5amps, but have not been able to determine if a chip capable of detecting that will suffer any damage, or be able to continue activating whatever relay circuit I come up with when the current goes as high as 300amps. I mean, gauss surrounding the cable even at 300amps is still pretty mild! The answer to that, I'm almost certain, is "no problem". But noone has been able to confirm this.
     
  12. bertus

    Administrator

    Apr 5, 2008
    15,635
    2,342
    Hello,

    I did have a look at the CSA_1V sensor and noticed this diagram:

    [​IMG]

    The sensitivity is strongly dependend on the wire diameter and current.
    What is the diameter of your cable that you want to monitor?
    You say the current is AC/DC.
    Can you give more details on the amplitudes and frequencies?

    Greetings,
    Bertus
     
  13. Markd77

    Senior Member

    Sep 7, 2009
    2,803
    594
    I'm pretty certain that you can't damage a hall effect sensor with a high magnetic field. The bit of silicon that senses the magnetic field before amplification wouldn't have more than a couple of millivolts across it in the strongest field.

    From the datasheet of the A1301EAU:

     
    Last edited: Jun 3, 2010
  14. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    Would that need to be remedied in the circuit design? If the circuit is signalling a relay to energize when five amps is detected, will it continue to signal likewise when 300 amps is in the cable?

    [​IMG]
    The ID above is 8mm. The copper stranding is equivalent to about 14 to 16ga copper.

    Calculable amplitude and frequency data is going to be hard to determine. It is an AC/DC tig welding power source. Amps range from 0-300. Because of the variety of processes it is capable of, there is a bewildering range of frequency, amplitude, and even waveform to be concerned with. I wouldn't even want to begin. That's why I'd like to simply start with an ultra sensitive generic ac/dc current sensing circuit and start bench testing.

    Here's a technical manual on the welding power supply...
    http://www.thermadyne.com/IM_Uploads/DocLib_2191_300X4884B.pdf
    It's a large file.

    When I've passed this by anyone that I happen to connect with in industry, they seem fixated on establishing a predetermined measurement threshold involving calculations that are waaaay beyond me. I'm having a hard time getting across that I simply want to energize a relay whenever any measurable current is flowing in the conductor. Anything from .05 amps to about 5 amps is ok by me. I'm not concerned with accuracy, measurement, or scaling.
     
    Last edited: Jun 3, 2010
  15. BillB3857

    Senior Member

    Feb 28, 2009
    2,400
    348
    What about a low ohm/high watt current shunt in the ground lead. Shunts typically supply 50mv/rate amps. Output would be either AC or DC depending upon welding mode. Amplification with proper clipping circuits could then activate your relay.
     
  16. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    Absolutely must be a non contact sensor.
     
  17. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I think you're on the right track.

    Don't worry too much about linearity of the Hall-effect sensor. After all, you're looking for a threshold value.

    Anything above the threshold (let's say >=1A), you want to turn on the relay.
    Anything below the threshold (let's say < 1A) , you want the relay to turn off, right? Just to clarify your problem statement.

    There are lots of different Hall-effect sensors on the market. But to narrow them down a tad:

    1) Some have a relatively linear response over a range of Gauss (strength of the magnetic field).

    2) Some act as Schmitt-trigger switches; if the Gauss exceeds a limit, they will turn on. If the Gauss then falls below a lower limit, they will turn off.

    3) Some act as latches. These are sort of like those in 2) above, but require that the polarity of the magnetic field be reversed before they will turn off.

    While you are experimenting, I suggest that you stick with the linear type.

    You will need to use a comparator with the linear type to detect whether the threshold has been exceeded.

    Most comparators have relatively limited current sink ability. Many don't have any current source ability. You will need to use a transistor or MOSFET output driver to control your relay.

    It will help a lot if you can give more specifics on the relay you wish to use.
     
  18. bertus

    Administrator

    Apr 5, 2008
    15,635
    2,342
    Hello,

    To get back at the circuit you posted in #11.
    What did the circuit do?
    Is there a problem with the sensitivity?
    Did the relais scatter?
    What should the relais switch?
    Is the switched device connected to mains power?

    Here I made some small changes to improve the sensitivity:

    [​IMG]

    Bertus
     
    Last edited: Jun 4, 2010
  19. frascati

    Thread Starter Member

    Jun 1, 2010
    28
    0
    I had the relay energizing a 115vac light bulb.

    What is represented by U3 in the diagram?

    I believe the problem I ran into with the first attempt was simply ham-handedness. I used parts from my workshop parts bins and matched values as closely as possible. Used leds in place of the diodes, etc. I used an ML7805 three terminal regulator in place of the LP-2980-5. I also crowded all of the components onto a small square, perhaps instinctively wanting to make it compact. A mistake for a test board.

    Last night I ordered all fresh (and by-the-numbers) SMD parts for that diagram from Digikey. I doubled up on the comparators and regulators and got 10ct minimums on the resistors. I'm going to salvage SMD caps from the clutter of cannibalized electronic boards in my "cannibalized parts" drawer. It was getting late when I did the digikey order and I didn't feel like messing with caps, and my bill was already close to thirty bucks. I hope I don't regret that. I already regret forgetting to order an SMD project prototyping board. Damn. Maybe I'll go to their online support chat and see if I can modify the order since it's still the weekend.

    Should I swap 250 ohm trim pots for the 200 ohm ones if I'm able? And I ought to replace that 10K with a 2K?

    All the resistors I ordered are thick film 1/4 watt. Sound sufficient for this project? I was told elsewhere that this same circuit will work for testing the more sensitive CSA-1VG chip. If Digikey does not stock this I'll ID the eqivalent and add that to the order.

    Should not be a big deal in any case. The extra parts needed should not exceed 8 oz so maybe I can slip them through the 1.85 U.S.P.S bargain delivery from digikey.

    It's so hard to resist drinking a pot of coffee and succuming to a little mania when it comes to playing with this stuff. And it's not even winter! I sat eyeing my old HP inkjet printer in the corner that did not sell for three dollars at the last garage sale. I wonderd if I might use the rubber clad main roller from another similar printer (three bucks at any goodwill), offset them and put a taut conveyor belt between them, draw an SMD PCB on some freeware diagramming program, and print/etch my own board. DAMN. Of course I'd never get it to work. It's just sort of distracting crap that keeps me from getting any one thing accomplished.
     
    Last edited: Jun 4, 2010
  20. bertus

    Administrator

    Apr 5, 2008
    15,635
    2,342
    Hello,

    The U3 in the schematic denotes the power-connections of the LM393.

    Are you bound to use SMD parts?
    Almost all parts are also available in through-hole.

    An other idea for the lamp driver could be the use of an optocoupler with a triac in stead of the relais.
    For the optocoupler you could take a MOC3023 as example.
    See datasheet for more info.

    Bertus
     
Loading...