Inductor ring tester

Discussion in 'The Projects Forum' started by wayneh, Jun 14, 2012.

  1. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    I've breadboarded a circuit that is basically working but could use a tune-up. It isn't as fast as it could be, and more importantly I'd like opinions on whether it will work as intended.

    I need to test a TV flyback transformer. There's a commercial ($50) tool for this, but after reading how it worked, I realized my breadboard from a previous project already had the ICs on it I could use to create a similar circuit. The schematics and explanation for the commercial device are given in these pdfs.
    First gen View attachment 43636
    Second gen View attachment 43635

    My circuit:
    [​IMG]
    Note: I haven't drawn in all of the LEDs, but they're there.

    One advantage of my device is that by using the 4017 counter, it lights only one LED at a time, saving power and reducing the parts count, since all LEDs can share a single resistor.

    Here are the issues:
    1) The Blue device runs at 10Hz, producing a 2ms low pulse every 100ms. My circuit runs well up to 10 Hz (set by timer R1, R2 and C2), but fails to give the same results when the low time, the discharge, is reduced below about 15ms. This may not really matter to actually using the tester, but it annoys me that it can't seem to go as fast as the Blue device.

    2) The RC (C3 and R5) of the reset pulse sent to the 4017 counter is 47µS versus just 4.7µS for the Blue device. I've been able to get a proper reset of the 4017 with a 18µS pulse but never less than that. This could be due to the LM358 being a bit slow to hit the reset threshold of the 4017, but using a faster op-amp (MC33178) didn't change things. The 4017 counter should be able to reset more quickly, shouldn't it? As with (1), I don't think a slower reset is fatal to the useful operation of the tester, but I don't see why my circuit should be slower than the Blue device.

    3) My biggest problem is I don't have any "good" flyback transformers to test versus "bad" ones. I've tried my tester with some transformers from the junk box and can get up to 8 "rings" indicated. (Yay!!) But plenty of other inductors I have give little to no response. The Blue documentation says you can get some result just by wrapping the test leads around each other. Not me. I knew going in this would be the biggest problem: I'm concerned I'll use my tester and not know how to interpret the results.

    Any ideas?
     
    Last edited: Jun 15, 2012
    absf likes this.
  2. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    I had an error in my drawing, now fixed. This affects my concern (2). I was not including R10, R11 and the diode in calculating the reset pulse duration through C3. I think it's still longer than the 5µs that the commercial device achieves, but less than half what I calculated previously.

    [update] Nuts, another error. The values of R1 and R2 are reversed even in the "corrected" drawing now in post #1.

    Regarding (1), setting R1 to 100K instead of 470K speeds up pinging to 10Hz and I still get 8 rings from my transformer under test. But I cannot reduce the value of R2 from 22K to 10K without reducing the number of rings I see. R2 sets the discharge time at about 15ms. Replacing C2 with a 0.47µF instead of the 1µF likewise changes the result because of the discharge time becoming too short. The Blue device supposedly uses 2ms, but I need a much longer time to count rings. Not sure why.

    Shouldn't a sufficiently large inductor have a slow enough ring frequency to require a long watching period to rack up 8 rings? I mean, maybe I'm only detecting rings with big inductors which also then require a long window?
     
    Last edited: Jun 14, 2012
    absf likes this.
  3. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    The drawing in post #1 is now error-free, as far as I know.
     
  4. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    Tonight I rigged up a sound-card-based oscilloscope on an ancient Mac I had in the basement. Wow, incredibly useful to see what's going on with my ring tester. Many mysteries solved.

    First of all, the inductor under test rings when the injected current pulse turns on, and again when it turns off. That little detail was not mentioned in the documentation but obviously affects the count. I was wondering if it rang on the up or down edge of the current pulse - and of course it's both. Mirror image.

    Secondly, if the current pulse ends before the ringing fades away from the positive going edge, a few rings at the tail end don't get counted that would have if the current pulse had been longer. The negative-going edge restarts the ringing. At 10Hz, the time between pulses is plenty to catch everything. The commercial device counts for only 2ms, which I don't understand. My inductor rings for 10ms or more.

    Third, the 4017 doesn't latch, and so a "13" count looks just like a "3". When I optimized timing to catch all the rings of my particular inductor, the count went well past 10. Cool!

    I'm still concerned that I'm only getting counts on a couple of transformers I have in the junk box.

    My oscilloscope can't see into the µs range to shed light on my counter reset timing.
     
  5. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    Well thanks very much, but I don't even have one sample let alone two, and I could buy a replacement for about $20-30, perhaps less than the shipping costs back and forth to you.

    I'm a bit puzzled at the moment. The tester I built is working but only for a big inductor. Since a big inductor rings slower than a small one, I need to count ring peaks over a much longer period than the commercial device I mimicked. Smaller inductors that might ring as fast a flyback transformer don't seem to give me enough amplitude to trigger counts, even though my op-amp arrangement should give as much or more gain than the commercial device. Maybe my op-amp is too slow? I just haven't had a chance yet to sort this all out. And unfortunately the transformer I need to test is 30 minutes away from my Mac-based oscilloscope. Can't get the two together without serious effort.
     
  6. szhighstar

    New Member

    Jun 26, 2012
    27
    5
    the transformer is applied in flyback circuit, can you supply power and input voltage and current, output voltage voltage and current and schematic? I design the transformer, then you can compare with your existing transformer.
    Can you test transformer? turns and inductance and DCR and Q and leakage inductance Cw.
     
  7. R!f@@

    AAC Fanatic!

    Apr 2, 2009
    8,754
    760
    Hey....Are u trying to make a better one off the shelf. May be I could build tht and check.
    U know, I have to deal with a lot of flyback due to my work.
     
  8. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    I've just been working on a major upgrade to this circuit, current version below. I confirmed that the LM358 was too slow. It would give a nice result for big (slow ringing) inductors but nothing for smaller ones. Now that I've replaced it with a TL082 (chosen because it's faster and available at Radio Shack), I get over 10 ring counts on just about any coil or transformer I try. It might be too generous, and make even a bad flyback look good. I may never know.
    [​IMG]
     
  9. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    I should mention: The 556 and op-amp are there because that's what was already on my breadboard when I started this project. The original circuit used a LM393 for the oscillation and ring detection functions. The 4017 counter was on my board as well, and I think that's an improvement over the commercial device.

    The positive pulse from the 556 timer, although it is the right sense, is not big enough to trigger the 4017 timer reset and to ring the inductor. So the transistor and one of the two op-amps is used to boost it. The transistor shouldn't really be necessary but I simply could not get the TL082 op-amp to behave as a non-inverting amp. So I use it as an inverting amp and let the transistor invert it again. This gives a nice solid reset and ringing.

    The other op-amp in the TL082 does a fine job of counting the peaks, possibly better than the LM393 in the original circuit. But I may need to add an offset hurdle to keep it from counting peaks smaller than, say, 150mV. i believe that's what the commercial device does.
     
    Last edited: Jul 7, 2012
  10. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    Here's yet another update, with a small offset on the op-amp #1 to shave off a few smaller ring counts, as practiced in the original published circuit. See R8 and R9. I've biased the op-amp to the high side because the TL082 can sense to the positive rail but not to within 3V of the negative rail.

    I tried to eliminate the transistor, to use only the non-inverted clock out from the op-amp, but it did not do as good a job as the circuit shown. It was harder to get a reset of the 4017 and I don't think the ring was as solid. This circuit works nicely.

    [​IMG]
     
    Last edited: Sep 5, 2012
  11. Redcatimaging

    New Member

    May 8, 2014
    2
    0
    Thank you very much wayneh :), this is great. Your circuit works very nice, was a really cool project. I've build this up with a little modification for me. Used a 555 because I had one :) and for R10 i used a poti to adjust the number of rings in addition with a Resistor (4,7K) in series with the diode near the testleads (to get higher ring-peaks --> for those ones who want even some of the smaller ring peaks count). This way it can be adjusted better and can show more or less of the rings. And I have populated 9 LEDs (0-8 and 10).
    Nice greetings and Merry Christmas :).
     
  12. ScottWang

    Moderator

    Aug 23, 2012
    4,854
    767
    I can't see the attached file.
     
  13. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    Do you mean the PNG file in #10? Can you see other png files?
    If your problem continues, I can send it in any other format.

    Oh wait, I can't see whatever was in #8 either. Hmmm....
     
    Last edited: Dec 21, 2014
  14. wayneh

    Thread Starter Expert

    Sep 9, 2010
    12,120
    3,046
    Full disclosure first; it wasn't my circuit. OK, I altered it, but I didn't conceive the idea.

    Anyway, I'm very glad to hear it worked for you. I just happened to complete my build, moving the circuit from breadboard to a permanent PCB. The other half of the 556 drives a capacitor ESR tester on the same board. I'm very close to submitting it as a completed project.

    Over time I've made the same changes you did, to alter the threshold cutoff for counting rings. The circuit behaved a little differently on the PCB than it did on breadboard. To make the changes, I took every big inductor I could get my hands on and looked at how many rings I got with each. Then I tweaked the resistors to get 10 LEDs to light on my "best" inductors.

    I found some LEDs at DigiKey that are perfect for the build, wide in one dimension but narrow in the other - perfect for 0.1" spacing. I used red for LEDs 1-5 and green for 6-10.
     
  15. ScottWang

    Moderator

    Aug 23, 2012
    4,854
    767
    I can't see it in #8, only can see a block, the others are ok.
     
Loading...