Designing the Ultimate Continuity Tester

Thread Starter

Sensacell

Joined Jun 19, 2012
2,582
A continuity tester is a pretty mundane piece of gear, but very useful.

I have several, but I started thinking about designing and building one that meets my obsessive fantasy specifications.

I am often working on some prototype and just want to quickly validate connectivity, with minimum effort.
My fancy Fluke multimeter has a good continuity function, but it's a pain to turn it on and switch the mode, I have to twiddle the dial and push buttons, I need a dedicated tool that suits my workflow - just pick up the leads and test, always ready, nothing to think about.

1) It's just a box with 2 banana jacks and a beeper. Super simple.

2) No ON/OFF switch.
I want it to always be ready, just grab the leads and test.

3) Long battery life.
Standby current of a few uA.

4) Only beeps when it sees less than 5 ohms on the test probes.
Ignores bad connections and most components in the circuit.

5) Ignores diode and transistor junctions- it's a CONTINUITY tester!

6) Outputs a limited, safe current.
Less than 10 mA? Will not damage delicate parts.

7) Robust- withstands my absent-minded probing of live circuits without blowing up.
50 V AC/DC without damage.

8) It should respond really quickly.
This allows me find intermittent connections in connectors and cables, etc.
Some multimeters suck at this! drives me nuts, have the designers not ever used the thing??

Anyone have any circuit concepts?

I will share my design as the project unfolds.
 

profbuxton

Joined Feb 21, 2014
419
Already built one years ago. Just a pvc junction box with battery and alarm beeper and two long leads.Always ready to go loud enough to be able to be heard in noisy area. served well for years , still works, ended up being used by our testers instead of DVM cos they would switch off when idle and were too low volume.
 

#12

Joined Nov 30, 2010
18,210
That's some really tough specs. You have to drive 10 ma into 5 ohms to achieve 0.05 volts for the detector to detect. To avoid triggering silicon junctions, the source voltage must be less than 0.5 volts. That makes the driver impedance <45 ohms, but you want to be able to hit that 45 ohm Zin with 50 volts DC or 70.7 volts peak, in either polarity, with no damage. I did not achieve that in 2 hours of design work.

This ain't gonna be 6 parts and a buzzer.:(
 

Thread Starter

Sensacell

Joined Jun 19, 2012
2,582
Already built one years ago. ended up being used by our testers instead of DVM cos they would switch off when idle and were too low volume.
Exactly why the DVM sucks for continuity testing!!!

Mine always times-out and turns off- just when I am concentrating on tracing out my board.
I end up constantly having to touch the probes together to make sure the stupid thing is still on.
 

#12

Joined Nov 30, 2010
18,210
Seems easy, until you think it through.
Yebbut...how many hours do you want to spend on this?
I tried a 4 ua cmos chip but the input offset was 10 mv. That means you must have 0.1 volt to detect to stay within 10% of spec.
I haven't seen a low offset, micro-power cmos chip yet.
Then the power to the load..I'm thinking a constant current generator with a suppressed j-fet for high impedance current but that won't work because you can't buy a 100V j-fet and the j-fet needs a couple of volts to get to its proper idle current.
Ah ah ah. Can't go above 0.5 volts.

I thought about it.
Just because you can dream it doesn't make it practical for me to design it.:(
This is going to need different chips for different purposes, and bring it all in under the power budget? How much current does a small MPU use?
 

kubeek

Joined Sep 20, 2005
5,730
#12 mybe you could use what you had in mind, but with added protection of say a PTC fuse in series and a TVS or zerner behind that, so that it limits to say 5-10V going to the detector?
 

profbuxton

Joined Feb 21, 2014
419
My goodness, talk about overdesign.What happens if your fancy electronic gizmo fails while you are testing? As I said all you need is a battery , some leads and a sound making device.
 

OBW0549

Joined Mar 2, 2015
3,503
I tried a 4 ua cmos chip but the input offset was 10 mv. That means you must have 0.1 volt to detect to stay within 10% of spec. I haven't seen a low offset, micro-power cmos chip yet.
Linear Tech makes a low-power bipolar op amp, the LT1494, with pretty good specs: 1.5 μA max supply current, 375 μV max Vos. The part is also available as a dual (LT1495) and a quad (LT1496). I doubt it's the key to making the whole thing work, but FWIW it takes care of the Vos problem you cited.

Then the power to the load..I'm thinking a constant current generator with a suppressed j-fet for high impedance current but that won't work because you can't buy a 100V j-fet and the j-fet needs a couple of volts to get to its proper idle current.
Ah ah ah. Can't go above 0.5 volts.
Yup. I think the bigger challenge here is going to be protecting the current source from high external voltage, not protecting the detector portion of the circuit.

This is going to need different chips for different purposes, and bring it all in under the power budget? How much current does a small MPU use?
According to its data sheet, a PIC12F675 draws 8.5 μA (typ) at Vdd = 2.0V and Fclk = 32 kHz. I'm not sure an MPU would be of much use here, though.
 

OBW0549

Joined Mar 2, 2015
3,503
This set of specs is not practical.:(
That's where I'm at with this, too; IMO, the requirements for no ON/OFF switch and ability to withstand high external voltage conspire to make this damned close to a "can't get there from here" situation.
 

#12

Joined Nov 30, 2010
18,210
375 μV max
That cures the offset problem, but the ceiling is still 0.5 volts and the impedance is still below 50 ohms.:(
You can't fire 50 volts into that by mistake without some very clever design work and I've been awake for 22 hours. Got brain fry.:oops:
 

Reloadron

Joined Jan 15, 2015
5,601
Built a box once for checking switches. Used a Calex Model 930 Programmable Current Source (or roll your own) running 10 mA. Measured the voltage drop across a switch, amplified the voltage and ran it into a comparator. Hang whatever you choose at the comparator out be it red and green LEDs (Pass/Fail) or a beeper. Skip the comparator and use a uC. You are alone for the always On. That or add a Push To Test button. What you have here is a very generic descriptor. :)

Ron
 

Thread Starter

Sensacell

Joined Jun 19, 2012
2,582
Ok here is my first go at it:

The 2 transistors form a switch that turns on the opamp and reference section when current starts flowing, Q3 provides hysteresis so it snaps-on cleanly. Now we don't care about the idle current of this whole section.

Once the opamp (comparator) is switched on, it's looking for ~ 50 mV or less on the input to turn on the LED.
A low power, rail-to-rail opamp should work here, an MCP6021 for example.

Idle current is kept low by the 1 meg resistor.

Positive over- voltage sees a 1 Meg impedance- no problem there.

Negative over-voltage will heat up R1 / R2, maybe it needs a PTC fuse in series.
R1 being 10W would be the brute force solution. Maybe R2 could be larger to limit Q1's max base current in this condition.

I need to breadboard it to see what nasty transient mayhem happens when the opamp turns on.
There is probably a solution with a few caps to make it start up in a nice well behaved manner.


contest.png
 
Top