ƑƠɭįƠ is a simplistic, yet automatic outer foil tester for capacitors

[Elvee]

ƑƠɭįƠ is a simplistic, yet automatic outer foil tester for capacitors

Following some discussions about the subject, Outer foil cap measurement with Hantek 6022BE
Cap Foil Side Tester,
here is a small project based on the principles outlined here: Outer foil cap measurement with Hantek 6022BE

The process is rendered completely automatic: one just has to insert the cap to identify into the instrument, and a LED instantly indicates the foil side:

No more need for manipulations, comparisons or measurements.

The principle of operation is to send two antiphase waveforms to the CUT, and detect the polarity of the signal electrostatically leaked to the outside with a synchronous detector.

A single IC performs all the signal processing, and two transistor-based buffers handle the higher currents needed for large value caps (the instrument is designed for up to 4.7µF, but still gives usable indications with 10µF).
For fun and challenge, I built the whole thing into a small box of ear-plugs ("boules Quiès"):

The box is tiny, made of thin and soft PP (thus non-conducting, providing no shielding), and barely contains the circuit and 9V battery.
The power push-button is located underside: you just need to push on the box to operate the tester:

For large capacitors, the tester is equipped with an aluminum bracket and a phosphor-bronze strap, helped by an elastic band:

The smaller caps are handled thanks to a phosphor-bronze spring:

The construction is not especially tidy, but it works well:

The usefulness of such a gadget might seem questionable, but in fact a surprising proportion of capacitors are mismarked: more than 10%.
An example is the small yellow-block shown above: it is from the Philips/MBLE/RTC etc. group, not some dodgy Asian manufacturer

 

[Osvaldo de Banfield]

Interesting project, but it lacks utility. As the outside foil of the cap will have a small capacitance to chassis and this capacity is negligible in comparison with the intrinsic capacity of the capacitor, then the cap is a short to those frequencies in which parasitic capacitances enter into account. Then, being this fact, it is careless which armature of the cap has more capacitance to chassis as both will be equally important or not.

 

[Elvee]

but it lacks utility. As the outside foil of the cap will have a small capacitance to chassis and this capacity is negligible in comparison with the intrinsic capacity of the capacitor, then the cap is a short

I agree -partly-: to be honest, I rarely cared very much about the "polarity" of capacitors, and when I did, I relied on the marking where it was present, which in retrospect was a mistake: out of the 6 examples I randomly picked for the pics, 4 had a clear outer foil indication, and one of the 4 was wrongly marked, thus 33.3% ambiguous and 16.6% plain wrong.
Informal statistics on a much larger sample size showed something like 12% incorrectly marked for those that were clearly marked.

One can argue about the usefulness of such information, but the marking convention has been established for sound technical reasons: in the audio field for example, it is important to have the input HF-squashing cap of a sensitive preamp properly oriented, because it will be a few hundreds of pF, be PS or PP and will ruin all your efforts of shielding and ground-plane if the cm² of "hot" metal lays bare, exposed to any 50/60Hz stray field.

It is not the only instance where the orientation matters: I often work with peculiar topologies of A/D converters, and the charge injection into a 47nF integration cap operating at +/-200mV from a 0/5V step in the vicinity can very easily ruin the accuracy, even if the stray capacitance is a few hundred femtofarads.
I now realize that some of the problems I encountered could have been caused by an incorrect marking: that's something I never really questioned, but now that I have more data available, I am certainly going to check the actual orientation for the few cases where it really matters

 

[Elvee]

Something else to consider: if an effect is relatively easy to detect with such a crude and rudimentary test set-up, it will necessarily have some impact on real circuits and situations, where you try to chase the last 1/10th of dB.

After all, an XOR operator used as an amplifier with an unfiltered/unregulated supply is not exactly an ultra-sensitive, top-notch LNA, yet it easily detects the effect for a capacitance range of 0 to 4.7µF

 

[thumb]

It is great I like your device, thanks for sharing. Also I guess this outer/inner folie thing is well knowledge in HF electronics were every small coil/caps counts

 

[Zung]

Differences, however small, do add up in the end. I've been using the thumb and index method so far, but sometimes it's not so easy. This will go into the queue of my to do's.
Thanks.

 

[Elvee]

It is great I like your device, thanks for sharing. Also I guess this outer/inner folie thing is well knowledge in HF electronics were every small coil/caps counts

Foil-based capacitors are rarely used in HF (sometimes small PP or PS types), and ceramic, mica, etc do generally not have the equivalent of an outer foil: they tend to be more or less symetrical (except special types like feed-thru).


A health warning: this project looks easy, and it is in a way, but if you try to replicate my flaky construction methods, a cramped layout, flying leads and isolated case, you will run into problems.

I am an old fox, and I know what is important or really off-limits, so I managed to have the circuit working despite the obvious construction slopiness, but a naive builder trying to replicate my construction will be confronted to intractable problems.

So, use a reasonably aerated construction, both for the PCB and the case and use a metal case.
The PCB should ideally have a ground-plane, like the perfboard I used (it is the only redeeming feature of my build).

As indicated on the schematic, the critical section is in the vicinity of U3: the components in this zone must be very close to the IC, with minimal track lengths.
It is very important to use the operator indicated, and to respect the pin numbers: the input structure of the XOR is asymetric, and only one input will work properly in linear mode.
In addition, the operator chosen is the quietest, and the input pin 9 sits between V+ (an AC ground) and the output (providing feedback).

This seems finicky, but if you know what to do, there is no problem: the circuit has first been tested on a breadboard, meaning no ground-plane and severe parasitic capacitances, yet it worked.

It is important to keep the test wires well clear of the input zone, because the stray capacitance will skew the result.
In my build, I had to use a shielded cable for the right-hand cord, because it had to pass close to the input.

I have indicated optional capacitors across the LEDs: their purpose is to average the pulsed waveform.
Even when a good, in-phase or out of phase signal is detected, small phase errors remain, meaning the OFF LED is not perfectly dark: with our logarithmic perception, a 1 to 100 brightness ratio looks much smaller.

With the caps, even 10 or 20% out of phase signals will not light the LED, leading to a better contrast.
I didn't use them in the final build, because I found the contrast was sufficient, but you are free to implement them.

There is a small omission on the original schematic; here is the correction:

Note that any CD4070 or 4030 will work: the circuit has been tested with Motorola, Harris, Philips, ST

 

[Zung]

Any volunteers to draft a PCB and setup a group buy?

 

[Elvee]

Here are some additional informations about the design, and the measurement amplifier in particular.

I opted for an operation frequency of ~140Hz as a tradeoff.

To maximize the capacitive pickup efficiency, the frequency has to be as high as possible, to transfer more energy, but this conflicts with the necessity to drive the CUT at a high level (the current becomes too large).
140Hz is sufficient for detection, allows up to 4.7µF of CUT and is not multiple of the mains frequency.

The high gain amplifier (U3) is a bit peculiar: it is coupled via an input capacitance of typically <1pF.
With a classical, resistive-input amplifier, the resistor would need to be as high as several gigaohm to pass comfortably the 140Hz, which is not practical.
If the resistance is much lower, like 22 Meg, the behavior will be that of a differentiator, which is not a problem in itself, but also requires a phase-shifted reference signal for the synchronous detector, which would add some complication.
Here, the stray capacitance supplemented by C2 imparts an integrator behavior to the amplifier, straightening the phase.
R7 is still needed for DC stabilization, but AC feedback is prevented by the filter R6 C6.
The structure of R7, R6, C2, C6 results in fact in BP filter, having its center frequency between 100~150Hz approximately.
The exact frequency is not easy to compute, simulate, or even measure in realistic conditions, because too many parameters are indeterminate, but the frequency is in the right ball-park, even if it is not perfectly exact, which is sufficient to be selective and keep a ~0 phase of the amplified signal.

If a key parameter, like operation frequency is altered, none of this will hold anymore, compromising the operation of the tester.
If you want to tweak, or improve Folio in any way, it is certainly possible, but you need to keep all this in mind, and correct what needs to be

 

[Rmcoelho]

Did someone else build this little wonder?

 

[Elvee]

I don't know: people have expressed their interest by PM, but I didn't have a PCB to offer, and I don' knowt whether they went on to build their own version or not

 

[Hearinspace]

I would love to have this as well as your HumBoy and other cool test accessories but my experience has been that in the majority of cases , there's more to a circuit than the schematic.
The circuit and it's trouble-free best physical layout are part of a single package that is the designer's own personal way of thinking and doing things. Just having the bones of the circuit in schematic form gets less and less promising as the complexity increases. I have had as much as I need of diagnosing problems in a circuit I didn't create.

You come up with some of the coolest things that can be found on any forum - not only this one. If your circuits were available in a tried and true form I could be assured I wouldn't have to spend precious time troubleshooting, I'd have them all !

 

[Rmcoelho]

I'm from Brazil, and my English is not good. It was a little difficult to understand how the device really works. Yes, I understand that the LED lights up on the external foil side, but I had some doubts.
If someone who built it could make a video, it would be a great help.
Before reading this thread, I was thinking of using a vu meter, but this one looks better and more accurate!

 

[Elvee]

I would love to have this as well as your HumBoy and other cool test accessories but my experience has been that in the majority of cases , there's more to a circuit than the schematic.
The circuit and it's trouble-free best physical layout are part of a single package that is the designer's own personal way of thinking and doing things. Just having the bones of the circuit in schematic form gets less and less promising as the complexity increases.

You are perfectly right, of course, and all of it is particularly relevant for Folio, which has a small, but highly critical part of circuit.
The circuit itself is only 50% of the potential problems: the mechanical part needs to be crafted carefully: the machining of the enclosure, the fabrication of custom-made parts, like the metal bracket or the flexible metal tongue.
The circuit works by detecting a tiny 140Hz signal through a capacitance of 1pF or less, and this demands serious precautions and attention.

There are many nuances and styles in DIY: at the top end, you have "modern" HAMs buying something like a SDR receiver "kit": they need to tighten 4 screws, et voilà! they have built their equipment all by themselves.

At the other end of the spectrum, you have people floating interesting ideas, in a purely theoretical form, without quantitative or practical details, leaving the would-be builder with the task of caring for practically everything.

I personally sit somewhere between those two extreme: my projects are dimensioned, actually built and tested, and they come with recommendations and guidelines, but they don't include a PCB or detailed BOM: just recommendations about critical components when present.
Some, like Mark Johnson have a more caring approach towards their would-be builders: they are pampered, have detailed instructions, a complete BOM, etc.
My projects are mostly designed for my personal use; when I estimate that one could be of value to the community and other people, I share it, but in the way I built it for myself: as a prototype.
I could develop a PCB based on the prototype, but for me it makes no sense, as I already have a working one, and in general I only need a single piece.
I sometimes develop projects for others, when it is practical and can be tested on a breadboard, like RickTH's 38K oscillator, or when it simply involves the modification of an already tested project, like a high-current regulator for a large tube heater, but normally everything is for my personal consumption.

This project has a critical section, it is even indicated on the schematic, and even I, if I designed a PCB for it, would probably need one or two iterations to make it work perfectly.
The best way to ruin it completely would be to enter it into some CAD software, and let the auto-router loose: the result would not just perform poorly, it wouldn't perform at all.
If you build it on some cm² of perfboard, respecting the recommendations, it may not work perfectly when finished, but it will work, and you will easily be able to make the alterations required to make it work perfectly.
On a PCB, it will be a butcher's work, and once it is done, you will need V2.

Some of my projects are easier/more tolerant than others: take the Circlophone for example. It has specifically been designed to be absolutely bullet-proof, and anybody anywhere in the world can reproduce it using local, or scavenged components. If you are attentive and make no mistake, you do not need test instruments, not even a multimeter: it will play as soon as the last solder is made.
It would be difficult to make other projects, including Folio in the same spirit: it would complicate them, and a delicate part would inevitably remain.

To make it completely safe from the builder's POV, you would need a complete kit, with not only the PCB and the components, but also the mechanical parts, like the machined case and the capacitor test receptacle.
Let's hope a Chinese vendor gets interested and decides to sell the complete kit....

 

[Hearinspace]

Elvee,

Well, your post certainly deserves a reply but what can I say. Of course you aren’t going to learn anything new from me but I still want to offer something in return and I'm afraid all I have at the moment is a pic of a circuit I’ve just finished in a format I'm trying for the first time. A little more than four screws though far less on board than you could fit in.

Not sure how it happened but in addition you’ve got me thinking about the Folio circuit and how I’d lay it out. . . . . the slippery slope . . .. . . .

 

[Elvee]

Congratulations: it is nice, clean and orderly, something highly advisable when dealing with tubes and the high voltages they need.

 

[Elvee]

Not sure how it happened but in addition you’ve got me thinking about the Folio circuit and how I’d lay it out. . . . . the slippery slope . . .. . . .

Think twice if you want to experiment with my projects: Folio is precisely one that is 95% plain, ordinary electronics, but the tiny part around the input and its connection to the outside world is absolutely critical, and if you want to avoid diagnosing and fixing, you might want to try a safer project first.

If you go on anyway, I recommend you make a breadboard test first: it will make matters somewhat finickier, but it is doable (my first tests were breadboard-based), and thanks to its simplicity it will take less than an hour to complete, and it will give you a practical feel of the circuit and its behaviour.

 

[TB]

This is interesting information on topic:
Are Your Capacitors Installed Backwards? Build this and find out - YouTube

 

[Elvee]

This tester is not automatic, not standalone as it requires an oscilloscope and not fully deterministic as it depends on ambient hum to operate: every pitfall Folio was designed to avoid.
It is also slightly more complex, with two chips instead of one.

Contrast that with Folio: place the capacitor into the test-pod, connect the two alligator clips, push the button and the lit LED indicates the foil side: simple, quick and unambiguous.

 

[ZUM911]

If I'm not mistaken Carlson's latest version is stand alone and indicates the outer foil side with LED's. I think you need to be a Patreon to get access to this latest version.