"DIY"-ish Oscilloscope current clamps

Discussion in 'General Electronics Chat' started by strantor, Sep 3, 2014.

  1. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    I needed to measure 3 phase current for a job recently and didn't have the clamps to do it. I rushed to eBay and bought the first good deal that popped up. I made an offer and got 3 of the clamps shown below, for $50 ea. I got rush shipping, and as soon as I got my tracking number and a guaranteed delivery date, I promised my client to be on-site the next day.

    current probe 1.png

    What you see there in the screenshot is all I looked at before making the purchase. The negligent eBayer that I am, I did not scroll down and look at the other pictures or descriptions. I am well familiar with the pictured clamps, also sold under Fluke brand, and others. I thought I knew exactly what I was bidding on.

    But as it turns out, these are a different. They are NOT Oscilloscope probes, at least not for any oscilloscope I've ever seen. In the above pic, it looks like they have a shielded BNC connector, but had I scrolled down and looked at the other pictures, I would have seen this:
    current probe 3.jpg
    current probe 2.jpg

    So you can imagine my dismay when I open the package at 7PM the night before I'm to be at the big hot job in the morning.

    I have one of the Fluke brand clamps of the same exterior design, rated for 10mA_AC/A_AC, for use with a multimeter, and I have used it before with my scope by placing a burden resistor across the terminals to provide a mV signal.

    I decided to hack these probes and see what I could see. I did not take the time to trace out the circuit or identify the components on the internal PCB; I just compared it to my Fluke clamp, which was just the CT passed straight out to the leads. When using my Fluke probe with the scope, I placed the burden resistor at the end of the banana jacks, near the scope.

    I desoldered all the components on the AEMC probes PCBs, and bypassed all the traces with jumper wires, straight out to the leads. I used the internal calibration trimpots as burden resistors. I cut off those wacky 3 pin "meter" connectors (what kind of "meter" uses connectors like that anyway? Some kind of power quality/phase angle meter?) and solder/spliced BNC connectors on them. I tested them out with a resistive load up to 12A, and everything looked beautiful.

    After staying up until 3AM with this crap, I went to bed and got up at 5AM to head out to the job. I set up the scope and monitored the current, and it kinda looked like crap. The traces were just really really noisy, like MHz noisy. Granted, this is not a nice resistive load, but the input to a switching power supply. But still, I expected to see pulses in the kHz range, and while they were present, they were very hard to make out and barely stood out above the noise.

    So my questions are:
    1. Is my theory sound? Is it "legit" to do what I've done, and if I had the same probes, designed from the factory to be scope probes, would it look any better?
    2. Is my issue that the burden resistors are inside the clamp, and the whole length of the lead is carrying a minute mV signal unshielded; could I get a better result by placing the burden resistor down at the scope input like I used to with the Fluke probe?
    3. Is my issue that these are strictly AC (as in, sine wave) probes and just can't provide an accurate picture of whatever this fast switching PS is doing (I should be using active AC/DC clamps)? I have verified in the past using the fluke probe with burden resistor that it will display non-sinousidal waves, like DC motor drive phas-angle SCR switching, but that is still on the order of 60Hz; this could be much faster.
    4. Or is what I'm experiencing perfectly normal? Or is there some other issue I haven't thought of?

    Thanks
    Strantor

    (BTW I think I might dig the new site theme, not sure yet)
     
  2. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    Hello there. What do you mean by measure 3 phase current? Do you want the rms of each leg? Do you need to show someone the scope trace of the current? What is the magnitude of the current that you need to measure?
     
  3. #12

    Expert

    Nov 30, 2010
    16,284
    6,797
    Love to help you, but I've only done this by ordering (as a repair part) the inductive pickup for a timing light and using that to pick up spark plug current. Obviously, it's AC only and completely lacks any calibration.
     
  4. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    Ok, so the story goes like this:
    Its an industrial microwave for making diamonds (think diamond cutting tools/abrasives). It keeps shutting down unexplained, without generating any fault output (it has several fault outputs for various faults). It will run maybe an hour, a day, a week, two weeks, then in the middle of the night it simply stops outputting as if someone had removed the run command, yet the run command remains high. So I was asked to set up my scope to monitor the incoming power (voltage) and trigger on a loss of the "running" signal. I put my voltage probes across 2 legs of 480, the running signal, and the run command. 2 days later they call and tell me that my scope triggered and I go out and look at the saved waveform and there was a chunk taken out of the incoming voltage, 2 cycles before the "running" signal went low. So now we have to determine if that's cause or effect; was there a power delivery blip that knocked it offline, or did it short out internally and pull the incoming voltage low. Hence the current clamps. I plan (I did) set them up, 3 clamps on 3 legs, trigger on loss of running signal. And yes, once I obtain a waveform showing a current spike or no spike, I need to be able to show that to my client.
     
    Last edited: Sep 3, 2014
  5. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    See, that wasn't hard. Is the control circuitry that powers the running signal powered from one or more 3 phase legs?
     
  6. BillB3857

    Senior Member

    Feb 28, 2009
    2,400
    348
    Does the equipment have phase loss detection? That wouldn't be too unusual if power stability is required. If the phase loss was temporary, it could command a shutdown, and since the error were only temporary, again sense normal power but require a manual restart. I have a similar setup for my wife's computerized sewing machine and just a blink of the lights will shut it down.
     
  7. GopherT

    AAC Fanatic!

    Nov 23, 2012
    6,045
    3,813
    Is there any chance a disgruntled employee is tripping it manually? Are there any security cameras to check the area?
     
  8. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    Yes it has phase loss detection, but there is a fault output for that, which never comes on. Also, when I remove the run signal, or force/simulate any other fault (except for forcing a phase loss - not keen on yanking live 480v out of the terminal) it takes 10 or more AC cycles to turn off. This fault (whether cause or effect) brings it down in two cycles, making me strongly suspect it is the effect of an internal short. The Event lasts a quarter cycle, hardly enough to constitute a "phase loss" (although we don't have data from the 3rd phase). See attached. IMO Any piece of equipment, especially one designed for an industrial environment, should be able to ride through a glitch like this, if it is an incoming power glitch. A nearby nonlinear load like a motor drive could cause more of a line disturbance than what's seen here. The problem is obviously (to me anyway, speak up if I'm wrong) with the microwave unit. Cause & effect is really irrelevant; either its causing the anomaly because its messed up, or it's messed up because it can't cope with this super minor anomaly. But the client wants all the data they can gather before sending the unit back to the manufacturer, as its already been back twice ans they swear there isn't anything wrong with it. Manufacturer is in germany and each time the gets sent out, it's a month of down time.
     
  9. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    Not a chance. That was my first question too, and client told me not a chance. Just be sure, I programmed a couple of tell-tales in the PLC. If someone were pushing a button or pulling a wire, I would know about it.
     
  10. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    Not that I know of. The manufacturer is super tight on info; it took a series of strings being pulled before they would give their blessing for me to open it up without voiding the warranty. The drawings they provided were very basic, just a step above block diagrams. As far as I know the 480 just goes to a transformer/rectifier, or a some sort of switching boost circuit. It gets stepped up to 5100VDC for the magnetron. The unit has 230Vac single phase labeled "control voltage" going into it, but all the signals are 24vdc.
     
  11. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    Your losing your run signal. You will need more info. Check all conditions know for RUN. Bang on interlocks, etc. Check wire connections and switch actions for erratic contact. It's a shame some companies gouge their clients.
     
  12. #12

    Expert

    Nov 30, 2010
    16,284
    6,797
    I don't expect that this will help you, but...
    A couple of months ago I had a glitch that would reset the brain board in an air conditioner, but it wouldn't blow a 5 amp fuse. The machine kept stopping to go through its safety delays and resets in a heat index of 104. Half a dozen delays like that and the machine wouldn't catch up with the heat load until 3 A.M. I installed a few 3 amp fuses, and finally, after 4 hours, one of them blew. Bingo! The 5KV air cleaner was shorting out intermittently. PIB, but that's what we get paid for.
     
  13. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    Yeah I have a DIN rail I carry around with me that has 30 fused terminal blocks on it and wire leads on one side. For those times when "something keeps blowing the damn control fuse." How do you figure out which of the dozens of input/output devices is blowing it? Fuse each one and see which blows. Yank all the field wires off the output side of existing terminal strip and land the wire leads of the fused terminals in the now-empty spots, then terminate the field wires to the other side of the fused terminals. Apply power and wait (or not).
     
  14. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    No, I'm losing my "RUNNING" signal (feedback from microwave, red scope trace). My RUN signal is just fine (blue scope trace). Note that these signals were scoped INSIDE the case of the microwave, where they terminate to the PCB. There are no interlocks or connections to bang on.


    Does anybody know about the current clamps which are designed for scopes? Do they differ much from the ones designed for meters? Are they able (or should they be able) to provide an accurate picture of kHz switching?
     
  15. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    Pardon me, I did not discern run and running. Do you know the conditions necessary for the running signal? Erratic operation is usually physical(mechanical). Can you stress the apparatus safely? Very feed rate or power levels perhaps?
     
  16. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    It has happened at power levels ranging from 0 to 60%. We have tried bringing it up to 100% for a time to see if it would cause a failure but it did not. It cannot be left at that power level for long. There does not appear to be any correlation to power output. The scope trace shown above captured the event at zero power output. It was left ON (RUN command, blue trace) but no power was commanded; it failed after 2 days sitting in this state. The pic above is a snapshot of a 12 second long recording; if you have FlukeView I can upload it if you're curious. For the 11seconds prior to the trigger event (RUNNING feedback, red trace, going low) the AC waves are absolutely sinusoidal and perfect, and neither dc signal wavers in the slightest. Then the brief hiccup, then RUNNING feedback drops out while RUN command remains high. It happened over the weekend and the plant was shut down. There were no vibrating machinery or movement of any kind. Even the unit's own cooling fan was not running, as it only comes on when outputting a good deal if power and generating internal heat. I can guarantee there was absolutely no mechanical disturbance, even from a passing car with subwoofers, as the location is far back from the road. There were no storms, no roving security guards, nothing. I had the unit open and specifically looked for evidence of rodents or insects and it was clean as a whistle. I shook it upside down and from all angles to dislodge any screws or misc that might be floating around, nothing.

    Conditions necessary for RUNNING signal (supposedly): 1. RUN command present, 2. No faults (faults are latching and require manual reset by RESET button and removing and reapplying RUN command).

    Hence the big mystery, this is an undocumented failure mode. If the RUN command is high, the RUNNING feedback should be high, unless there is a fault, in which case the fault light will be on and an output sent to the PLC to log the fault. There should be no instance where RUNNING goes low like this. It still requires manual reset of the RUN command to start again even though there is no fault (but it does NOT require you to push the actual RESET button, as it would if there were a real fault). If you follow the waveform snap shot forward into the future, that red line will never go high again until the blue RUN command is removed and reapplied after a mandatory restart delay of 15 seconds.
     
  17. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    Well alright then. We will work on your line glitch theory. I propose that the glitch causes a ringing thru your power supply circuits. Causing an un-monitored fault in amp or supporting circuitry. Is there, or can you put line filtering components at input of power supply? Would client consider VFD, just for line conditioning quality?
     
  18. strantor

    Thread Starter AAC Fanatic!

    Oct 3, 2010
    4,302
    1,988
    I appreciate it, but honestly I think I've got it under control. This thread was about amp clamp probes, and it isn't really going the direction I intended.

    If my questions about amp clamps need clarification, please let me know.
     
  19. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    I have never tried to measure current with scope probes, but your reasoning seems correct to me. I use to have to data log motor current in heavy, noisy industrial environments. I always used 4 to 20 milliamp current sensors. They don't cost much and they come at different voltages and current ranges. For instant....a 480 volt with a 20 amp rating. You slip this sensor(which is just a calibrated coil of wire) around a leg. The sensors I used were powered with 12 and 24 VDC. The sensor would output 4 ma at zero amps. The 4 ma. confirms you are connected to sensor. Anything less or no current means you lost your sensor. 20 amps thru the leg will give you 20 ma. out of sensor. With a 500 ohm resistor at logger terminal, gave me a 2 to 10 volt swing to correlate to 0 to 20 amps to log. Some of these sensor were hundreds of feet away from logger. But with a twisted current loop, you don't get induced or interference signals. This was all 60 hertz. Hope that helps.
     
Loading...