Testing Windings: Tools & Gear for Appliance Repair

  • Thread starter Thread starter Guineafowl
  • Start date Start date
  • Tags Tags
    Testing Windings
AI Thread Summary
The discussion focuses on testing electrical appliances, particularly brush motors and transformers, for faults like coil shorts. Participants share their experiences and suggest tools such as a Megger insulation tester, Variac, clamp-on ammeter, and LCR meters to diagnose issues effectively. They explore methods for testing field coils and inductance, emphasizing the importance of comparing measurements to known good units. The conversation also touches on creative testing methods, including using oscilloscopes and Hall sensors to detect magnetic flux and phase shifts in inductors. Overall, the thread highlights the need for a variety of testing equipment and techniques to ensure thorough appliance repair.
Guineafowl
Messages
865
Reaction score
404
I get given a lot of brush-motor tools that are "burning', as well as assorted other damaged electrical items. I'm not an electrical engineer, just a fiddler with some background in physics.

Now, Jim Hardy and Tom.G have just walked me through testing a 12V alternator, which was found, as far as I can see, to have shorted/low insulation stator windings. I bought an old Megger from eBay for the purpose.

If I see a "burning' brush motor tool again, what tests can I do on it? The Megger found shorts between the phases on the alternator, but how do you test a single phase motor for a coil short that is not to ground - a ringer?

The same question applies to transformers - if the short is not to ground, what test gear will find it?

In other words, I have a multimeter, scope, bench power supply, ESR meter and a Megger insulation tester. What other gear do I need to fully test appliances found in the home and workshop?
 
Engineering news on Phys.org
Guineafowl said:
What other gear do I need to FULLY [/color] test appliances found in the home and workshop?
The manufacturers Quality Control Department.:rolleyes:

A "growler" https://www.google.com/?gws_rd=ssl#q=growler+tester Also lists instructions and make-your-own.

A way to measure Inductance (perhaps an LCR meter. Decent battery powered ones start around $150USD.)

A variable voltage transformer, commonly called a Variac. Be sure to get one that can be wired to provide output above line voltage. A 10Amp rating is plenty for run-of-the-mill consumer stuff. (Again around $100 to $150) 5Amp if you must keep costs down. The ones with a built-in voltmeter and outlets are handy but you can always add an outlet and voltmeter (and maybe an ammeter.)

A clamp-on ammeter is handy to have too. If you get one that also reads DC current you can use it on electronic equipment. Since the stuff I do is more often electronic, I have an AC/DC clamp-on without a built-in meter. It has output jacks for a meter or 'scope connection. Works for power-line stuff, electronic equipment and automotive electrical systems troubleshooting.
 
  • Like
Likes Guineafowl
Tom.G said:
The manufacturers Quality Control Department.:rolleyes:

A "growler" https://www.google.com/?gws_rd=ssl#q=growler+tester Also lists instructions and make-your-own.

A way to measure Inductance (perhaps an LCR meter. Decent battery powered ones start around $150USD.)

A variable voltage transformer, commonly called a Variac. Be sure to get one that can be wired to provide output above line voltage. A 10Amp rating is plenty for run-of-the-mill consumer stuff. (Again around $100 to $150) 5Amp if you must keep costs down. The ones with a built-in voltmeter and outlets are handy but you can always add an outlet and voltmeter (and maybe an ammeter.)

A clamp-on ammeter is handy to have too. If you get one that also reads DC current you can use it on electronic equipment. Since the stuff I do is more often electronic, I have an AC/DC clamp-on without a built-in meter. It has output jacks for a meter or 'scope connection. Works for power-line stuff, electronic equipment and automotive electrical systems troubleshooting.
I vaguely remember growlers from when I was younger (insert joke here). Between your link and Jim Hardy's in my other thread what I conclude is that I should build a time machine and go back to a time when simple articles in popular magazines would show you how to construct mains-connected devices. Our voltage is 240V, but still, we seem to have lost that understanding that once you're an adult, you can be assumed to be capable and responsible.

My ESR meter does do inductance, but I find the values meaningless without a known good unit to compare.

Variac - yes. On my list, as is an isolation transformer.

I do have an AC/DC clamp meter, a cheap UnitT one but it does seem accurate.

What about testing the field of a brush motor?
 
Guineafowl said:
What about testing the field of a brush motor?
Sorry, don't know of a good way; I've always used the smoke test for that! With practice, you could probably learn ballpark values for different size motors. @jim hardy, with his power background, may well have an answer though.
 
Tom.G said:
@jim hardy, with his power background, may well have an answer though.

I wish i had a handy field coil tester. It's so seldom i test such motors i never got around to building one. But I've thought about it.

Since it's the job of a field coil to make flux, it'd be logical to measure how much flux it makes for given current.
Which takes you right back to basic physics definition of inductance which is # of turns X flux per amp.
L = NΦ/I
i suppose an inductance meter would be a natural tool to try.

But what value to look for? That's the sort of knowledge you'd pick up with practice.
At first you'd just compare coils to one another.
In a simple drill or circular saw motor with one could pass a known AC current through both field coils and measure voltage across each.
If they read the same voltage it'd mean either they are both good or are equally bad.

An old car battery charger has inside it a current limited transformer that makes a suitable amount of AC voltage .

I've never done it. But that's the direction i would begin experimenting.
with precautions of course for working with line voltage.

If one is already fixing such power tools can we presume him capable ?

old jim
 
  • Like
Likes Guineafowl
jim hardy said:
Since it's the job of a field coil to make flux, it'd be logical to measure how much flux it makes for given current.
Could that be done using a clamp-on ammeter as a field detector while the coil is energized with a low voltage?

I know the clamp-on I use detects the Earth's magnetic field. For very low DC currents I have to either take two readings turning it around and averaging or orienting it in East-West direction.
 
Tom.G said:
Could that be done using a clamp-on ammeter as a field detector while the coil is energized with a low voltage?
Worth a try.
I was thinking of a small search coil, perhaps the coil from a little solenoid, connected to an AC millivoltmeter.
 
Tom.G said:
I know the clamp-on I use detects the Earth's magnetic field. For very low DC currents I have to either take two readings turning it around and averaging or orienting it in East-West direction.
Going to tool cabinet to retrieve my Klein Tools CL200 clamp-on,

cl200.jpg


to see if it'll measure the Earth's magnetic field. :oldcool:

Edit: Indeed it does. Well, Earth's field will induce a few mV when you move it around.
 
Last edited:
jim hardy said:
I wish i had a handy field coil tester. It's so seldom i test such motors i never got around to building one. But I've thought about it.

Since it's the job of a field coil to make flux, it'd be logical to measure how much flux it makes for given current.
Which takes you right back to basic physics definition of inductance which is # of turns X flux per amp.
L = NΦ/I
i suppose an inductance meter would be a natural tool to try.

But what value to look for? That's the sort of knowledge you'd pick up with practice.
At first you'd just compare coils to one another.
In a simple drill or circular saw motor with one could pass a known AC current through both field coils and measure voltage across each.
If they read the same voltage it'd mean either they are both good or are equally bad.

An old car battery charger has inside it a current limited transformer that makes a suitable amount of AC voltage .

I've never done it. But that's the direction i would begin experimenting.
with precautions of course for working with line voltage.

If one is already fixing such power tools can we presume him capable ?

old jim
As you say, without a manufacturer's schematic that includes expected inductance values, or a known-good unit, all we can hope for is a relative value that builds our experience.

What about this: http://www.ebay.co.uk/itm/AnaTek-Bl...315165?hash=item2a2ea6e1dd:g:plkAAOSwBLlVDzSA

Looks exactly what I need.

Or... maybe I could feed pulsed DC into the coil and count the ringing on my scope?
 
  • #10
Guineafowl said:
Looks exactly what I need.

Or... maybe I could feed pulsed DC into the coil and count the ringing on my scope?
Now THAT's interesting.

Please pardon the following ramble. Old guys just do that.I once devised an inductor test that seemed to work
but it gave results comparable to a much easier to use computerized inductance meter we already had in the shop.
I mention it because it's sort of similar to your "Ringing" test.
I established about an amp DC through the coil
then interrupted it and captured the 'inductive kick" in a capacitor trapping it there with a rectifier..
Connected to the capacitor i had a Fluke DMM set for peak capture.
I sized the capacitor for about a hundred volts on the coils i was measuring.
It would find coils with shorted turns but so did the little computerized inductance tester .

I do hope you build yourself some sort of tester. Just brainstorm about the properties of inductance and experiment away.Another thing i tried is Lissajous patterns on a 'scope.
Remember that sinewaves are a special math function - their derivatives have the same shape as the undifferentiated one.. That's why for a sinewave voltage you get a sine-shaped current, really cosine because either one is the derivative or integral of the other.
I built an integrator and produced the time integral of the induced voltage in the inductor. That should be a measure of the flux in the inductor because e=Nturns X dΦ/dt..
I applied that to the Y axis of my 'scope.
I measured the current through the inductor with a low ohm resistor and applied that to the X axis of my 'scope. That too should be a measure of the flux in the inductor because Φ = μNIA/Length.<---- observe absence of derivative
A perfect inductor should give a straight line on the 'scope because flux and current are in phase
any phase difference between the signals makes the trace more of an open circle.
I had the luxury of a second winding on my inductors which i used for a zero current "Flux detector" to measure induced voltage. You'd need either a search coil(as i had in that second winding) or a Hall sensor to get your flux signal.

A perfect inductor will give a straight line for a Lissajous pattern of ( Φ by integrating induced voltage or direct measurement ) versus (current by direct measurement).
Its open-nes approaching a circle indicates there are other currents flowing besides what you're putting through the windings yourself.
On my inductors that showed me just what lousy cores they had.
With core out i got straight lines as expected . With the core partway inserted i got ellipses. as i inserted more iron core the trace got more circular.
My cores were not laminated and had so much eddy current flowing in them they were suitable for only less than 10hz operation.
I got about 30 degree phase shift because of eddy currents. (Incidentally that explained their odd temperature coefficient - as they warmed their inductance went up because the hot core's resistivity went up, reducing eddy currents)

So I'm thinking- since you have a 'scope - and we can now buy Hall flux sensors reasonably
One of these glued to a popsicle stick should detect flux:
http://www.ti.com/lit/ds/symlink/drv5053.pdf
And the most striking thing i noticed about my inductors was their phase shift when extra current flowed...
And since power tools have a laminated core, eddy currents should be low,
meaning more than a teeny bit of phase shift between measured current and measured flux should be a decent indicator of shorted turns.
.......................
I envision a source of line frequency AC current an amp or less, a current measuring resistor, and a Hall flux detector probe.
Readout could be either a 'scope, or an analog meter indicating phase shift in degrees.
Just squaring up and AND'ing two sines gives a PWM proportional to phase difference... a DC meter reporting average value of that PWM would work.
...
You'd soon either become skilled at its use or figure out what's wrong with the idea.
And if you said it worked i'd buy one.

old jim
 
  • Like
Likes Asymptotic and Guineafowl
  • #11
jim hardy said:
Now THAT's interesting.

Please pardon the following ramble. Old guys just do that.I once devised an inductor test that seemed to work
but it gave results comparable to a much easier to use computerized inductance meter we already had in the shop.
I mention it because it's sort of similar to your "Ringing" test.
I established about an amp DC through the coil
then interrupted it and captured the 'inductive kick" in a capacitor trapping it there with a rectifier..
Connected to the capacitor i had a Fluke DMM set for peak capture.
I sized the capacitor for about a hundred volts on the coils i was measuring.
It would find coils with shorted turns but so did the little computerized inductance tester .

I do hope you build yourself some sort of tester. Just brainstorm about the properties of inductance and experiment away.Another thing i tried is Lissajous patterns on a 'scope.
Remember that sinewaves are a special math function - their derivatives have the same shape as the undifferentiated one.. That's why for a sinewave voltage you get a sine-shaped current, really cosine because either one is the derivative or integral of the other.
I built an integrator and produced the time integral of the induced voltage in the inductor. That should be a measure of the flux in the inductor because e=Nturns X dΦ/dt..
I applied that to the Y axis of my 'scope.
I measured the current through the inductor with a low ohm resistor and applied that to the X axis of my 'scope. That too should be a measure of the flux in the inductor because Φ = μNIA/Length.<---- observe absence of derivative
A perfect inductor should give a straight line on the 'scope because flux and current are in phase
any phase difference between the signals makes the trace more of an open circle.
I had the luxury of a second winding on my inductors which i used for a zero current "Flux detector" to measure induced voltage. You'd need either a search coil(as i had in that second winding) or a Hall sensor to get your flux signal.

A perfect inductor will give a straight line for a Lissajous pattern of ( Φ by integrating induced voltage or direct measurement ) versus (current by direct measurement).
Its open-nes approaching a circle indicates there are other currents flowing besides what you're putting through the windings yourself.
On my inductors that showed me just what lousy cores they had.
With core out i got straight lines as expected . With the core partway inserted i got ellipses. as i inserted more iron core the trace got more circular.
My cores were not laminated and had so much eddy current flowing in them they were suitable for only less than 10hz operation.
I got about 30 degree phase shift because of eddy currents. (Incidentally that explained their odd temperature coefficient - as they warmed their inductance went up because the hot core's resistivity went up, reducing eddy currents)

So I'm thinking- since you have a 'scope - and we can now buy Hall flux sensors reasonably
One of these glued to a popsicle stick should detect flux:
http://www.ti.com/lit/ds/symlink/drv5053.pdf
And the most striking thing i noticed about my inductors was their phase shift when extra current flowed...
And since power tools have a laminated core, eddy currents should be low,
meaning more than a teeny bit of phase shift between measured current and measured flux should be a decent indicator of shorted turns.
.......................
I envision a source of line frequency AC current an amp or less, a current measuring resistor, and a Hall flux detector probe.
Readout could be either a 'scope, or an analog meter indicating phase shift in degrees.
Just squaring up and AND'ing two sines gives a PWM proportional to phase difference... a DC meter reporting average value of that PWM would work.
...
You'd soon either become skilled at its use or figure out what's wrong with the idea.
And if you said it worked i'd buy one.

old jim
Apologies for the late reply - you've given me lots to think about!

I was thinking along these lines, based on the properties of inductors as follows:

1: Flyback, as in the device you made. I made an electric fence energiser based on a 555 timer that could use this effect...
2: Inductance, as measured by a commercial meter.
3. Flux, as per your Hall sensor idea. This seems the most 'life-like' test, as magnetic flux is the goal with transformers and stators.
4. Resonance - a tunable LC oscillator, L being the device under test? Needs a variable cap, though.
5. Ringing - I found this:

IMG_0364.PNG


It appears the timer pulses the coil, and the ringing is used to clock a shift register.

The ideal properties of a coil tester would be:
Battery power, for portability
Standalone use, ditto
Quick go/no go results

Off to the bench!

Cheers,
Andrew
 
  • Like
Likes Asymptotic and jim hardy
  • #12
Right, I've breadboarded (several times) the circuit above and can't get it to work. The 556 stays high and pulses low for about 2ms, which I think is right, but the first op-amp output resultant from this only dips a volt or so (does not go fully low). The collector of the transistor stays high and thus no pulse is transmitted to the test leads. They stay low. The 4017 outputs a zero regardless of what the test leads are connected to.

I've subbed in a 2907A general purpose PNP for Q1 and 1N4004 diodes - are these OK?

I'm sure it's something silly, but what?

EDIT: I'm 'away from bench' at the moment, but have I confused the transistor connections? I would call the one at 1 o'clock Emitter, then collector at 5 o'clock, then of course base at 9 o'clock.
 
  • #13
Guineafowl said:
I'm sure it's something silly, but what?

555's take a huge gulp of current when the output switches.
Try a 10 uf tantalum immediately adjacent its power supply pin, or a big aluminum electrolytic and 0.1 ceramic?

A small one across R6 might help too.
 
  • #14
jim hardy said:
555's take a huge gulp of current when the output switches.
Try a 10 uf tantalum immediately adjacent its power supply pin, or a big aluminum electrolytic and 0.1 ceramic?

A small one across R6 might help too.
Thanks, Jim. It was actually a loose connection from IC2 to ground - a case of "et tu, breadboard?"

The circuit works well on the primaries of switch mode transformers, showing a 9. It doesn't like the secondaries of same or my old car ignition coil, although the latter scored 2 on the secondary. It looks to be suitable only for higher impedance/Q windings without bulky iron cores.

Do you think the design could be optimised for lower impedance/inductance applications, ie can it be made more sensitive? I noticed the ringing could be scoped on the secondary if the primary was being rung. Also, a good winding will ring (according to the scope) a good 20-30 times, suggesting the 4017 could be chained to another, giving an output count from 0-19, to allow range to measure lower inductance coils as well as higher.
 
  • #15
Guineafowl said:
It was actually a loose connection from IC2 to ground - a case of "et tu, breadboard?"

Good Find .

Next question - why doesn't it propagate the rings you see on your 'scope ?

Guineafowl said:
Do you think the design could be optimised for lower impedance/inductance applications, ie can it be made more sensitive? I noticed the ringing could be scoped on the secondary if the primary was being rung. Also, a good winding will ring (according to the scope) a good 20-30 times, suggesting the 4017 could be chained to another, giving an output count from 0-19, to allow range to measure lower inductance coils as well as higher.

That's a good one.

Here's my first thought. Usually my first thought turns out wrong, but here it is.

TL082 is operating outside its recommended range of input common mode voltage
snips from your schematic and TL082 datasheet

CoilRinger1a.jpg


It's very confusing because later in datasheet (table 6.5) they show typical common mode range equal to V+ .
Note with just a 9 volt supply, recommended input range of (V- plus 4) to (V+ minus 4) is only between 4 volts and 5 volts. That circuit has it at 8.4 volts.

So i take that to mean some amps might work and others might not.

I'd try about 4K in series with that diode to bias the inputs close to mid supply, and reduce R9 tenfold to bring them closer together.
Then see if it counts more rings .

That'll show whether we understand that part of the circuit.

What do you think ?

old jim
 
  • Like
Likes Asymptotic
  • #16
PS
Probably a good idea to add diode clamps from node R8-R9-C4 to V+ and V- , to protect TL082's inputs from the inductors

old jim
 
  • #17
image.jpg


Ignition coil secondary (good, I think). Reading: 1 ring
image.jpg


Field coil of vacuum cleaner motor (good). Reading: 4 rings.

image.jpg
Small wall adapter SMPS transformer primary (good). Reading: 9 Rings.I also held the test leads like brushes across the rotor commutator, and turned. Each rotor coil gave a reading of 4 or 5.

So as you see, it's more useful than I thought. What I really need is some known good and known shorted coils of the same type to calibrate it. I played around as you suggested with the TL082 voltages, managed to reduce them but couldn't get any decent ringing off the ign coil.

I suspect the failure of the ign coil to ring is something to do with either the sheer size or the fact that the negative side is common to primary and secondary.

I did try connecting a neon and 33k resistor in parallel with the coil secondary, then touching 6V to the windings. As I disconnected, the neon flashed briefly. Crude, but effective? No good for the primary, though.

My inductance meter, however, gives a reliable reading of -30.6 uH for the ign primary and 30.0 uH for the SMPS secondary, both of which didn't read well on the ringer. Not sure why the minus sign but it may serves as a benchmark for others. Perhaps, amongst the eBay devices, breadboard builds and my Heath-Robinson approach I have all I need to test any coil?!
 
  • Like
Likes Asymptotic
  • #18
image.jpg


Last one: Current-voltage characteristic curve for the ign coil secondary, using the scope's component tester. Horizontal line is open circuit, vertical is short. The very slight upward angle suggests high resistance, the looping a phase shift (?).
 
  • #19
I wonder if your ignition coil has shorted turns in secondary. How does its primary ringdown look ?

Can you perhaps get a junk one from an old lawnmower or weedeater ?
Q of that one is mighty low. Great scope shots !

Ignition coil primary lissajous will be interesting. The more circular the higher the ratio X/R
Those secondaries are tiny wire because the sparkplug current is small. That might explain low Q. Primary on other hand has to handle a few amps.Interesting work you're doing here !

old jim
 
  • #20
jim hardy said:
I wonder if your ignition coil has shorted turns in secondary. How does its primary ringdown look ?

Can you perhaps get a junk one from an old lawnmower or weedeater ?
Q of that one is mighty low.Great scope shots !

Ignition coil primary lissajous will be interesting. The more circular the higher the ratio X/R
Those secondaries are tiny wire because the sparkplug current is small. That might explain low Q. Primary on other hand has to handle a few amps.Interesting work you're doing here !

old jim

image.jpg


Ign coil primary - a score of 2 on the ringer.

Scope Lissajous is a near-vertical line, as for a short, but shorting the leads 'twitches' the line slightly thinner, so there is a slight circularity.

The coil was a cheap one I bought online for use with my homemade electric fence energiser. It's done lengthy service sitting in the chicken house, pulsing away open-circuit (very seldom did it get a 'fox' path to ground). It could well have some damage.
 
  • Like
Likes jim hardy
  • #22
image.jpg

Vacuum cleaner motor field coil ringing pattern again. Scope timebase 0.5 ms/div.
image.jpg

Test rig - note ringing score of four (First LED is zero).
image.jpg

Ringing now damped to a score of 1. Scope timebase advanced to 0.1 ms/div.
image.jpg
The culprit!It's fun as a toy AND it can test stuff. What more could you want?
 
  • #23
Well now ,,,, There's an epiphany !

Guineafowl said:
Ringing now damped to a score of 1. Scope timebase advanced to 0.1 ms/div.

Shorted turn by driving a screwdriver into the coil ?
upload_2017-6-6_18-9-3.png

If i read your 'scope right
Made it decay in just a couple cycles, tightening up the decay severalfold
and frequency went from 9 peaks in 1 msec = 9khz to 5 peaks in 0.2 msec = 25khz?

Your gizmo looks like quite a handy addition to a small motor guy's toolbox. One would quickly learn what to expect.

Congratulations, and thanks for the lesson !

old jim
 
Last edited:
  • Like
Likes Guineafowl
  • #24
Good Grief i forgot to link the TL082 datasheet

http://www.ti.com/lit/ds/symlink/tl082.pdf

Guineafowl said:
I played around as you suggested with the TL082 voltages, managed to reduce them but couldn't get any decent ringing off the ign coil.
I think it has shorted secondary turns.

I'm sure you will use your 'scope to perfect your gizmo. Watch TL082 output clock and ringdown on dual trace (chopped mode if it's fast enough) to see how effective is your detection ?

Above all keep having fun.
Thanks for a fun thread !old jim
 
  • #25
jim hardy said:
Good Grief i forgot to link the TL082 datasheet

http://www.ti.com/lit/ds/symlink/tl082.pdfI think it has shorted secondary turns.

I'm sure you will use your 'scope to perfect your gizmo. Watch TL082 output clock and ringdown on dual trace (chopped mode if it's fast enough) to see how effective is your detection ?

Above all keep having fun.
Thanks for a fun thread !old jim
That's a really good datasheet - gives lots of info and example configurations. Also a good tutorial on layout - bypass caps, etc.

I'll certainly play around a bit more, and let you know. Many thanks for your help and encouragement so far.
 
  • Like
Likes Asymptotic and jim hardy
  • #26
Back onto this ringer circuit now, after a brief foray into inverter repair...

To recap, I'd like to increase the range and sensitivity. A good SMPS transformer primary will show 9 rings, whereas the secondary will show 0. However, the scope shows the secondary is ringing.

According to the datasheet, the 4017 can be daisy-chained using the 'carry out' pin. I assume the reset input is commoned, and the clock input is from the first 4017's CO pin. This should increase the range from 0-9 to 0-19

If only I could amplify the clock input signal of the first 4017, I could surely get a decent 15+ ring count for the SMPS primary, and a more modest count for the secondary. I tried splicing in an LM358 in an inverting amplifier configuration, gain 10, between the second output of the TL082 and the clock input pin of the 4017. It didn't work - would it help to establish a DC bias on the LM358, then AC couple the output to the 4017? I'm not sure how to do this - any help would be much appreciated.
 
  • #27
Tough to make suggestions without knowing circuit, pin voltages, and scope traces on input and output. (We just aren't that psychic.:frown:) When you take the dual-trace measurements of signal input and amplifier stage output, please document the scope vertical scale, taking account of probe multiplier (x1, x10, ?). (for both existing TL082 and any add-ons)

Instead of adding an LM358, maybe use some more TL08x, much higher Gain BandWidth (GBW) product. Using a TL084 gives you two more OpAmps to play with in one package.
 
  • Like
Likes jim hardy
  • #28
Here is the circuit again, for reference:

IMG_0364.PNG


Measuring at the clock input pin, here is the trace from the SMPS primary, which measure 10 rings:
0.2 V/div x10
0.2 ms/div

Ringing frequency is 6.25 kHz. Pulse height 6.2V + 1.6V offset. Pulse width 30 us.

image.jpg
Again, measuring at the clock input pin, here is the trace from the SMPS secondary, which measures 1 ring:
0.2 V/div x10
20 us/div

Ringing frequency 83.3 kHz. Pulse height 7V + 1.6V offset.

image.jpg



Now, 3-4 good rings are clearly visible for the secondary, But the 4017 is not counting them. The frequency (83.3 kHz) is well within the bandwidth of the 4017 (10 MHz) and the pulse width (5 us) is too (min 45 ns). The pulse height is comparable to the primary result (in fact, they are both clipped at the top of the screen). This does not seem to be a signal amplification problem - why is the 4017 not counting the secondary rings?
 
  • #29
What's pin marked "EN" ?
I don't see such a pin in the datasheet.
upload_2017-7-14_18-0-39.png

http://www.ti.com/lit/ds/symlink/cd4017b.pdf
If it's "Clock Inhibit" it can block counting.
i'd give that LED its own pulldown resistor. The LED's should each have their own pulldown anyway.
 
  • Like
Likes Guineafowl
  • #30
jim hardy said:
What's pin marked "EN" ?
I don't see such a pin in the datasheet.
View attachment 207178
http://www.ti.com/lit/ds/symlink/cd4017b.pdf
If it's "Clock Inhibit" it can block counting.
i'd give that LED its own pulldown resistor. The LED's should each have their own pulldown anyway.
I think it's the clock inhibit, as you say. If pulled low, it enables clocking. I've only provided one common 1k resistor for all the LEDs, since they're only lit one at a time. I'll have a scope of that pin if I get time tomorrow.
 
  • Like
Likes jim hardy
  • #31
Guineafowl said:
I've only provided one common 1k resistor for all the LEDs, since they're only lit one at a time.
Oooops my bad - i assumed this was an old fashioned binary output...
You're right you only need one..

Thanks !
 
  • #32
Guineafowl said:
Now, 3-4 good rings are clearly visible for the secondary, But the 4017 is not counting them. The frequency (83.3 kHz) is well within the bandwidth of the 4017 (10 MHz) and the pulse width (5 us) is too (min 45 ns). The pulse height is comparable to the primary result (in fact, they are both clipped at the top of the screen). This does not seem to be a signal amplification problem - why is the 4017 not counting the secondary rings?

Check your RESET pin with 'scope ?
C3 and R13 values...
10,000 pf(103?) and 4.7 K ohms is 47 microsecond time constant, by which time your ringdown is finished.

So maybe it misses the show while waiting for end of RESET ?
 
  • #33
The 4017 datasheet shows the VIH Min. as 70% of VCC. With 9V Vcc the Clk must exceed 6.3V to be recognized. The scope trace shows the second pulse to about at that limit.

Conclusion: You're going to need more gain to capture that signal.

See pg. 3 of: http://www.ti.com/lit/ds/symlink/cd4017b.pdf
 
  • Like
Likes Guineafowl and jim hardy
  • #34
Tom.G said:
The 4017 datasheet shows the VIH Min. as 70% of VCC. With 9V Vcc the Clk must exceed 6.3V to be recognized.

And VIL below about 30% (to be recognized low. )
upload_2017-7-14_20-26-52.png


Dont know if that applies to Reset which doesn't have the Schmitt trigger like CLOCK does
upload_2017-7-14_20-32-3.png
but just looking at datasheet and your circuit, i'd say it's plausible
that exponential decaying voltage on reset pin holds counter in RESET for as much as 50 microseconds .
Use second channel on scope to see how long before it starts counting ?
ch 1 on clock, trigger from there,
ch 2 on "1" digit to show first count ?

old jim
 
Last edited:
  • Like
Likes dlgoff
  • #35
Tom.G said:
The 4017 datasheet shows the VIH Min. as 70% of VCC. With 9V Vcc the Clk must exceed 6.3V to be recognized. The scope trace shows the second pulse to about at that limit.

Conclusion: You're going to need more gain to capture that signal.

See pg. 3 of: http://www.ti.com/lit/ds/symlink/cd4017b.pdf
If the right-hand op-amp is in open-loop, do I just need a higher OLG op amp?

Or is the problem that the TL082 doesn't swing close enough to the +ve rail? If so, which characteristic on the datasheet details better rail-to-rail figures?
 
  • #36
Guineafowl said:
Or is the problem that the TL082 doesn't swing close enough to the +ve rail? If so, which characteristic on the datasheet details better rail-to-rail figures?

this is with 15 volt supply
upload_2017-7-15_8-20-3.png


i don't see any load on it, so looks like it should go reliably within 3V of rail , maybe 1.5
 
  • #37
Guineafowl said:
If the right-hand op-amp is in open-loop, do I just need a higher OLG op amp?

Or is the problem that the TL082 doesn't swing close enough to the +ve rail? If so, which characteristic on the datasheet details better rail-to-rail figures?

@jim hardy in post #15 had a good insight about the Common Mode range of IC2. Try replacing the Diode between IC2-8 and IC2-5 with 4.7K resistor and see if that helps. It could cause an extra count in the LED display though.

Other than that, no direct answer yet. Need more data. How about some measurements and a couple scope traces?

For all the 'scope measurements, the inductor chosen should have the highest ringing frequency available that counts at least two rings.
  • To start, please measure the "+9V Supply."
  • With the 'Test Leads' shorted could you measure the voltage difference between IC2-5 and IC2-6 for verification? I expect 128mV. That is the minimum ringing voltage that will switch the OpAmp output.
For the 'scope measurements, please document both Vert and Horiz scale factors including probe attentuation (x1, x10, ?) and, if not obvious, where the Zero level is on the traces.

Use External trigger from Q1 Collector. If the Chop frequency of the 'scope is not high enough for good photos, set 'scope for Alternate trace. In addition to levels, I'm trying to get relative timings here.
  • A dual trace scope shot of IC2-6 and IC2-7 with the 'Test Leads' open. Also document which, if any, LEDs light.
  • A dual trace scope shot of IC2-6 and IC2-7 with a good inductor connected. Also document which, if any, LEDs light.
  • A dual trace scope shot of IC2-7 and IC3-15 with a good inductor connected. Also document which, if any, LEDs light.
  • A dual trace scope shot of IC2-7 and IC3-3 with a good inductor connected.
The purpose of scope measurement with 'Test Leads' open is to see what the maximum No Ringing voltage is at the OpAmp input. I expect this to be between 700mV and 1.5V. This would guide us in possibly changing the 128mV input offset of the OpAmp.
The purpose of scope measurements on IC3 is to check if the RST pulse is interfering with counting.
 
  • #38
Guineafowl said:
Or is the problem that......

Clarification of post 34:

When Q1 turns ON it yanks RESET pin high through C3. That sets counter to zero.
It won't start counting again until that pin goes LOW.
That won't happen until C3-R13 bleeds down to VIL threshold and we're not sure what that threshold is.

IF
RESET takes 47 usec to decay down to VIL (10,000 picofarads X 4.7 K ohms = 47 usec time constant)
THEN
your ringdown is over by the time RESET goes away.
and counter stays at zero.

That's the only reason i can see for it to count slow ringdowns but miss fast ones.

47usec.jpg


Worth a look ? Ch 1 on RESET, trigger on + transition and watch the decay
ch2 on clock ?Or try a 102 in C3 ?old jim
 
Last edited:
  • Like
Likes dlgoff
  • #39
I may be confused as to when counting occurs - Q1's ON or OFF transition...
 
  • #40
jim hardy said:
I may be confused as to when counting occurs - Q1's ON or OFF transition...
OFF.
According to the data sheet: http://www.ti.com/lit/ds/symlink/cd4017b.pdf
The RST input is positive logic, active high = RESET.
Input pins are diode protected to 10mA for anyone input at a time.
 
  • #41
Tom.G said:
The RST input is positive logic, active high = RESET.

And it's capacitively coupled...

When Q1 turns ON it yanks RST high through C3,
That resets the counter to zero and inhibits counting clock pulses
RST decays back to LOW with 47 usec time constant.
Q1's ON transition also raises voltage at junction R10 R11 C5 by one diode drop (0.6V?) which drives CLOCK low through R10, C5 C4 and opamp's inverting input...
After RST bleeds back down to LOW, CLOCK will follow whatever the coil does, pesumably it'll ring .After 47 microseconds RST gets back down LOW and enables counting. If ringing persists that long it should count.

When Q1 turns OFF, it drives RST below 0 volts and the 4017's internal protective diodes give C3 a discharge path. So no reset on OFF transition.
It also drops voltage at junction R10 R11 C5 back to zero driving CLOCK high. It should count rings again ?

I'm awaiting 'scope traces to understand better.
 
Last edited:
  • #42
Sorry, lads, I'm 'away from bench' at the moment. Your posts are nonetheless much appreciated - I'll get to it soon!
 
  • #43
OK, back at bench.

+9V supply is 9.12V from bench source.
IC2-5 to IC2-6 is 114 mV with test leads shorted.

2ms/div, CH1 is top trace IC2-6 0.5V/div x10. CH2 is IC2-7 1V/div x10. Zero volts for each channel is two divs either side of from centre line. Test leads open.

image.jpg
If I do the same measurement, but AC couple CH1 and alter to 50mV/div x10, you can see the signal better:
image.jpg


As usual, with leads open the 0 LED is lit and the 1 LED is pulsing.
 
  • #44
IC2 6 and 7 with good inductor. 9 LED is lit. Both channels AC coupled or no sensible signal can be seen:

image.jpg


1ms/div, CH1 IC2-6 50mV/div x10, CH2 IC2-7 10mV/div x10.
 
  • #45
Good inductor connected. 9 LED lit.

1ms/div
CH1 top DC coupled, 0.1V/div x10. IC3-15.
CH2 bottom AC coupled, 10mV/div x10. IC2-7.
image.jpg
 
  • #46
Good inductor connected. 9 LED lit.

1ms/div.
CH1 top DC coupled, 5mV/div x10. IC3-3.
CH2 bottom AC coupled, 10mV/div x10. IC2-7.

image.jpg
 
  • #47
jim hardy said:
Worth a look ? Ch 1 on RESET, trigger on + transition and watch the decay
ch2 on clock ?

No-ring inductor connected, LED 0 lit.

0.5ms/div.
CH1 top DC coupled, 0.2V/div x10. RESET.
CH2 bottom DC coupled, 0.5V/div x10. CLOCK.
Ext trigger on Q1 collector as before.

image.jpg


Same settings, but good inductor. LED 9 lit.
image.jpg
 
  • #48
jim hardy said:
Or try a 102 in C3 ?

Done. With open leads, 0 lit and 1 pulsing as usual.

With shorted leads, 0 lit.
Good inductor reads 9.
Bad inductor (smps secondary) reads 2.
image.jpg


1ms/div
CH1 top DC coupled. 5mV/div x10. RESET.
CH2 bottom AC coupled. 5mV/div x10. CLOCK.

I think the 102 has done it.
 
  • #49
A quick test of the vacuum motor field coil, which previously read 4, now reads 5 with a 102 in C3. It appears the faster RESET decay is allowing a ring or two more to be counted.

A lot of scope shots posted here - I hope you don't feel too swamped.
 
  • #50
Post 45
IC2 7 is clock?
IC3-15 is RESET ?
So in post #45, RESET is yanked HIGH by Q1 turned ON and returns low well before ringdown is completed. Looks like that's when it starts counting clocks?
When Q1 turns OFF , RESET (IC3-15) 15 is clamped to -1 volt by IC3's internal diodes.

Post 46
Guineafowl said:
CH1 top DC coupled, 5mV/div x10. IC3-3.
CH2 bottom AC coupled, 10mV/div x10. IC2-7.
So CH 1 is the "0" LED lit briefly? I can't quite tell whether it's aligned over first clock pulse on ch2, but RESET should light it about same time clock pulses start and it looks like that happens... .
CH2 is clock again ?

Post 44 IC2 pin6 swings only ± a diode drop, as expected. Pin 7 swings full range

Thanks, i feel a lot better now about the circuit. It makes sense . And it's clever.

I thought of you and this thread the other day. I was checking polarity of windings in an AC motor that a friend is rebuilding.
Put small amount of DC through it and verified its four poles were N-S-N-S with a Scout compass.
Found out you can overdo it and re-magnetize the compass backward so it points South...old jim
 

Similar threads

Back
Top