simple audible continuity tester - any interest?

[duck_arse]

people,

I found a circuit today for a "sonic continuity tester" using a single 4046 chip. a very rudimetry search here and at googoo showed no signs of same. PRR mentioned one recently, didn't search return, though. I don't have any 4046's to test it with, but I know some here might want to try it.

so, anyone interested? I'll redraw and post  if yes.

[therecordingart]

Interested. I have a continuity buzzer on my meter, but would be game for building something smaller. I use the continuity buzzer more than anything else on my meter.

[R.G.]

It would be good if that thing used a test voltage below a junction potential - that is, below 0.1V or so, to account for germanium.

If your continuity tester uses such a low voltage, it will not give false positives because it turns on semiconductor junctions. This was a feature of separate continuity testers back in the 60s. Most multimeters then and now did not have this feature. Multimeters tend to use a volt or two for powering the resistance test, and that's also what they use for continuity, only adding a buzzer output for "hey - I smell low ohms!"

[duck_arse]

I didn't redraw, I just hacked it all up a bit. I'm afraid I know 0 about the use and abuse of the pll, so whether it can be modded to lower test V's, I don't know.

[Shoeman]

That'd be a handy rig.  Years and years ago I had what I can best describe as an audio signal injector that I got from my grandfather the Heathkit Builder.  It had a very skinny probe and a ground clip and an internal speaker.  You could switch it between testing continuity tests using the internal speaker, or switch the speaker off and use the probe to inject a tone in a circuit listening to or for the tone in the circuits own output. i assume it was made for TV and stereo/radio diagnostics.  Wish I still had it, but I was a young teen in the '70's then and stuff just got lost over the years.   Simple tools like that and this are so handy at times.

[R.G.]

I would be sorely tempted to redesign to the concept.

I might use an LM324 or its x2 friend the LM2904 to make a 0.1V or smaller square wave signal and to detect it and amplify it back up with another section. An LM386 to drive a micro-speaker-beeper-etc would round this out.

Probably the LM324 now that I think about it.  A couple of sections to send/receive signal, some comparators to tell when it's shorted (i.e., resistance less than X threshold) and to clean up the signal. The LM324 works happily with its inputs at 0.000V, so you could generate 0.1V through some resistors, and use this voltage and real ground. The signal would be "shorted" to ground by a resistance below X, and make the LM324 in comparator mode flip state compared to a 0.05V reference also generated by those resistors. This gives you an off/on signal for driving the beeper. The signal gates two sections of the LM324 into oscillation. It's possible that the fourth section maybe plus a transistor could drive a small beeper directly. Maybe. It might be as useful to drive an LED on one of the probes so you also get a visual "beep". I'm pretty sure the LM324 would do that.

[R.G.]

Mr. Simulator says the LM324 will do it, albeit in a different way.

A 5.6V zener (for freedom from 9V battery voltage variations) provides a reference to two resistor strings. One string is 560R and 10R to ground. The junction of these two is the "hot" probe. Ground is the ... well, ground probe. The other string is a 5.6K and two 100R's in series. The hot probe and the junction of the two 100Rs are applied to the two inputs of the LM324, in such a manner that the output goes high when the "hot" probe is below the voltage of the two 100Rs. It works out that this is the equivalent of about 7.5 ohms. So resistances under 7.5Ohms from probe to probe will make a high output, higher than that up to infinity will leave the 324 output low.

A second section of the LM324 is set up as an oscillator, much like the LFO in the MXR Phase 90, but with its frequency adjusted to a few kHz by judicious choice of resistors and timing cap. A series diode/resistor from the output of the first LM324 to an input pin forces this to NOT work if the output is low. First section output goes high, second section oscillator starts.

Mouser sells an audio "speaker" that's 9mm square and 4mm thick. Its resistance is about 40 ohms, and it outputs 80+db when fed from a 5Vpeak to peak. This is LOUD and 40 ohms is more than an LM324 can drive directly. A 1k-2K resistor from the oscillator output to the "speaker" lets it drive the speaker 40 ohms OK.  So the beeper beeps when the probes are connected by less than 7.5 ohms. The open circuit voltage is 100mV, too little to turn on even a germanium or Schottky junctions. So you see actual copper/wire resistance, not semiconductors turned on. It's DC, so capacitors may delay it turning on, but they will charge up to the 100mV open circuit fairly quickly, and will not interfere with you actually seeing "shorted or not".

This would be enough for a basic audio beeper. An LM2904 in an 8-pin dip would work. Driving an LED is a good mod, too. A MOSFET or transistor from the output of the first LM324 section could turn on an LED, or you could use one of the two remaining sections of the LM324, and have yet another left for other mods.

I should explain - when I look for continuity, I don't want junctions and other parts connected to the wires to give me a false "shorted" indication. It's a PITA and I always wonder ... is that really shorted?

This doesn't tell you resistance by varying the tone, but for me, that's both a crude indicator, and a distraction. Just tell me where the wires connect! 

[newperson]

Want to post a quick drawing of what you think a good tester could do?  i would be interested for your layout. 

[PRR]

> when I look for continuity, I don't want junctions and other parts connected to the wires to give me a false "shorted" indication.

As I understand the Everyday Electronics plan, junctions will give a significantly different peeep than a good wire/trace. ~~45hz for solid metal, 1KHz for silly-state junctions.

I'd think even tone-deaf low-pay workers could learn the difference. Certainly everybody here could.

It's not clear to me that it will tell you <<100 Ohms, which is often what we *need* to know. Assume pitch-shift of 10 cents (0.6%) off the 120 ohm resistor.... looks like we all could hear less than 1 Ohm, but not the 0.1 Ohm to 0.05 Ohm as you slide down the track towards the un-wanted short.

We could lower the 1K+120r pair but they are already a heavy load on the 9V battery. (We can't use a One-Wort unless we are SURE there's no other path from general ground to the PCB, which may not be obvious to novices.)

I'm not too concerned about regulation. Telling 50 mOhm from 60 mOhm. You get a reference pitch with probes shorted (whatever the battery feels like giving you) and listen for pitch changes up from that.

There should be zero problem driving a speaker with nearly any chip. First proof: the 4046 is a pretty lame thing yet they felt it was publishable. Using logic: your ear will be close. (With my PD 105, less than a foot.) The tone is probably a square-wave into a teeny speaker, so it's all high harmonics of the reference pitch, where the ear is sensitive. The workroom "should" be quiet for low distraction (yes production work is different). 10mA-30mA is often ample for peepers. And if it must be used in loud places (or in the library), put a 1/8" jack on and use iBuds.

[R.G.]

Yeah, I got that.  For me,  I just want to know metal connection or not. My own quirk, I guess. The interpretation mode might be useful at times in random troubleshooting, but for pedals I think the distinctions between junctions and shorts are a distraction. Now, if one were debugging a wire-wrap nest of several hundred logic chips, yeah - tell me all about the connection you can. In fact, a TDR would not be amiss. 

[R.G.]

OK, still attracted by this Shiny Thing. 

I sometimes want to find short circuits on PCBs, and have just found them other ways. You can find these easier if you force a current through the shorted path, and follow that path by the changing voltage.

One way to do that is to feed an AC current through it, and then follow the AC signal with a tiny magnetic tape head. Works great, but is a bit clumsy. A Hall Effect device might be better, but setup is tricky.

Another way is to feed DC into the shorted path, and follow the voltage. Every wire is really a low-ohms resistor. If you feed current into it, it obeys Ohm's law and generates a voltage of its resistance times the current. If you feed a current and sense the voltage along the wire, the voltage goes from maximum to zero as the probe goes from the current injection point to the the current exit point.

Here's the good part. If there are wires attached in the middle of this string leading off somewhere else, there is no current flowing in them. So if you follow a decreasing voltage from one point to another point down a wire or PCB trace, come to a "T" and follow a path where the voltage no longer decreases, then the voltage stops changing because with no current down that wire, the voltage remains constant down the non-conducting path. You can literally track the current down the correct PCB trace. Note that this only works if the branch paths have much higher resistances than the main path, and do not, for instance, have any turned-on semiconductor junctions.

I sketched out what I think is a good sensor. It's a voltage limited constant current source. It will supply up to 100ma, but that voltage limits at 0.200V. So you can apply 100ma into a dead short, or any resistance up to 2 ohms, and you get a voltage proportional to the resistance. The 0.2V is chosen because it's below a typical junction turn-on (although some germaniums will  ) and it can be read with the very common 200mV range of a digital multimeter. It takes one dual opamp (but it's a picky one, a TLC2272, about $0.75) and two TL431 voltage references, about $0.32 each, plus some 1% resistors and such.

Each millivolt is 10 milliohms, and you can resolve that with a common meter. This is usually enough resolution to measure the resistance of a trace with a careful setup with probes.  For reference, a typical 25mil-wide trace on 1-oz copper PCB material is 19milliohms per inch. A 12-mil wide trace is about 40mOhm per inch. So with 10milliohm resolution, you can tell when you're within a half inch of the probes being truly shorted (* If you set it up carefully - this is tiny stuff we're measuring here.)

That's OK, but it could get tricky with only 10millohm resolution. So I designed a 10:1 and 100:1 after amplifier. This increases the resolution, but decreases the accuracy, as any errors in the setup as well as offsets and gain errors in the amplifier get magnified too.

But that's probably OK. Tracing out short circuits as opposed to measuring wire resistance is a relative thing. So minor offsets and inaccuracies won't keep you from the goal.

Having thought this out, the only objection I have to the 4046 is the high voltage it lets get onto the circuit. You'd get false positives from semiconductor junctions. However, if you could use an up-front low voltage sensor and some kind of oscillator to tell you the scoop, you'd be in good shape. I think the voltage-limited current source would be a great front end to this. Another dual opamp gives the 4046 a high-resolution incoming signal to work with.

I've also thought about wanting a go/no-go indicator instead of a relative "you're getting warmmmmer..." signal. I think one could gate the thing on and off below some threshold where it would have no indication at all until a resistance below X was encountered, then trigger on an LED and the audio squawk for homing in.

So that's where I got in the thought design. The pieces seem to be OK. I don't know how interesting this is to anyone.

[armdnrdy]

OK, still attracted by this Shiny Thing. 

I don't know how interesting this is to anyone.

You have my interest.
I've built 90% of my own test gear. Some I've built from projects and some I've designed.
There is nothing like having the proper tools to work with even if you only use them now and again.

[R.G.]

I'll post some things here when I get further along then.

Still messing with analog accuracy.

Are you scared of PICs? I only say that because it's a cheap, simple way to wind up the results in one 8-pin chip. Looks like the analog frequency out and trigger on/off could take a couple of 8-pins plus related RCs.

[DIMstompboxes]

I'll post some things here when I get further along then.

I'm also interested...What I want also is a simple capacitor tester....can anyone link me to that. Thanks

[armdnrdy]

Are you scared of PICs?

I've never messed with them, know very little about them, but what I do know is when I learn to use them I'll be thinking....why didn't I make this leap sooner!

[tubegeek]

I'm also interested...What I want also is a simple capacitor tester....can anyone link me to that. Thanks

https://www.sparkfun.com/products/9485

[duck_arse]

I'm still reading. not keen on pics.