Interesting low voltage capacitor leakage tester

[PsychMan]

The guy (Mr Carlson) claims this performs better than traditional high voltage units, quite remarkable

https://www.youtube.com/watch?v=LhovRIM5xAo

Adam

[PsychMan]

I see no detail on how it works until you pay via Patreon for the build info, which I will likely do. Fair enough, can't fault the guy for asking for a few quid for his design. The testing looks very promising indeed.

[MrBungle]

Can't really lose for $2.

I imagine it's an exponential current source/sink of some kind looking at the simple BOM and behavior.

[ukcol]

I was very sceptical about an effective leakage test at low voltage. The reason being that I thought that the leakage resistance doesn't obey Ohm's law and that you need a high voltage to see leakage. I use an insulation tester that can be set to 250, 500 or 1000v DC to test paper and poly types. I will often see leakage resistances below 500k ohms using this tester and the same component measured with a DMM on the 20 meg ohm range will show Over-Range [that is a leakage resistance greater than 20 meg ohms].

Watching the video has challenged my view on testing at low voltage and it will be interesting to find out how the tester works.

[GrimJosef]

I'd need some convincing I'm afraid. A high voltage test does what it says on the tin. If DC current flows then the capacitor fails i.e. it leaks at high voltage. But if a capacitor passes then it doesn't leak at high voltage. So what's wrong with it ? Well, Mr Carlson's box doesn't like it. Hmmm.

I confess I haven't watched the whole video, so maybe some explanation will be forthcoming. But I did get as far as seeing the HV tester fail to spot what Mr Carlson said was a 'known bad' capacitor. I'm afraid my first question would be "Is that a 1kV capacitor which, in circuit, misbehaves at 900V ?". If it is then a test at 600V might well not detect the fault and the reason could be that at 600V there really isn't a fault. That capacitor might outlive me at 600V (the capacitor at 600V, not me at 600V ). As Colin says, leakage very definitely isn't Ohmic. But if we want to do an HV test to find out if it's OK at 900V then the test has to be at 900V.

A jolly good case can be made for not using high voltages if we're not competent to do so. But we'd only use high voltages to test high voltage capacitors. And surely we'd only be testing high voltage capacitors if we planned to build something which used them. At high voltage. I think you can see where I'm going with this one can't you .

Cheers,

GJ

[Diabolical Artificer]

Was going to post this myself. I used be a patron on said channel but though he did cover some good stuff, I learnt nowt. I will say this for him, he does seem to know his stuff but also likes to big up his "invention's" like they're the best thing since sliced bread. Nowt wrong with that, but it makes me suspicious.

I'll be sticking to the HV method; if a cap in circuit sees 450v I want to know it can take 475v whilst smiling nonchalantly and whistling oob la di oob la da : )

Andy.

[Radio Wrangler]

Capacitor leakage can look like a breakdown process.

Something has changed in a bad capacitor to reduce the voltage at which breakdown occurs below where it should be in a good part.

So does this change in a capacitor produce other electrical parameter changes which can be detected by low voltage measurements? With its high dielectric constant, water getting into paper dielectric will certainly affect the C value. There may be other fault mechanisms which can be deduced from low voltage tests.

But will the low voltage tests find ALL the mechanisms of high voltage failure? We aren't looking for certainty in "Yep, that's definitely a bad capacitor" we are looking for certainty in "Yep, that's definitely a good capacitor".

A real test at the real voltage has the advantage of comprehensively testing all the possible mechanisms at once. There is no risk of missing any mechanism which doesn't give clues at low voltage.

The dangers of high voltages are real, but if you are working on things which need these high voltage parts, then you already need to have developed safe methods of working.

David

[PsychMan]

I think what attracts me to this is the low cost of building it above the safety concerns. If I didn't have a slight clue in that regard I'd probably be dead by now

A lot of original test gear is very expensive (as he states) and takes up more space in my tiny workshop. I did build the reformer / leakage tester circuit listed on this site and it's good, but I need to connect a meter to it and it doesn't go higher than 375V. If I can't be bothered to apply high voltage I might use a cheap component tester and go by the capacitance reading, if it's out of tolerance I may not care why.

Assuming Mr Carlson's gadget is what he claims it is, then it's nice to have the versatility in such a small package.

[GrimJosef]

The beauty of leakage testers and reformers is that they hardly need to deliver any current at all. In fact they're safer if the current is limited, say, to a milliamp or less (of course the risk of a serious belt from a charged capacitor still applies). The low current requirement means that the high voltage can be derived from a voltage multiplier circuit based on almost any old mains transformer's HT winding. So a DIY tester can be dirt cheap to build. The most expensive item can easily end up being the box. Or the meter if you use one.

Cheers,

GJ

[MotorBikeLes]

30 odd years ago when I did a lot with Grundig CTVs, I did a service exchange repair on many of the modules. The Tuner/IF modules on the CUC 120 through to CUC3*** series had quite a few 0.1UFd capacitors, some on supply rails, but many in high impedance sections. Earlier versions had black coloured ceramics, but later they had what looked like the same ceramic, but dipped in a blue (epoxy?) resin. Many of the black ones failed, but never on the supply rails, but the blue dipped ones were always OK. I pulled the black ones on sight, and checked them with my Selectest on the high ohms range, which, like the Avos, have a 15v battery for that range. Most showed significant leakage, but went in the bin irrespective. Testing on the standard range (1.5v battery supply) NEVER indicated the slightest leakage. I never investigated the voltage at which it showed, but the two tests were "chalk and cheese". I am in the high voltage for high voltage caps team.
I believe the fault mechanism was moisture adsorption on the ceramic surface, almost certainly prevented by slight local "drying" on those on the supply rail.
Les.

[Diabolical Artificer]

If you have a variac PM, you can build a reformer tester that fits into a matchbox sized enclosure. Two caps a few resistors and a neon, bish bash bosh job's a good un. Yes you do need to connect a meter or two but there on your bench anyroad.

Cap testers have gone up in price cos there are several channels on Utube banging on about them. One releases a "How to test a capacitor" video, the others follow. In the last few years I've seen a big upsurge of interest from beginners in toob amplifier's and vintage test gear on Youtube due in part to Youtube videos; it's a self fulfilling prophecy.

Lest folk get excited that there's a wave of spotty boys all interested in old kit and therefore our hobby will live on, it doesn't appear to be the case. Most of the new interest is by folk in their 40's and older according to a look at his stats on one such channel. He was rather worried as he's been stashing valves away as an investment for his kids.

Apologies for the digression. Andy.

[ukcol]

Going back to Mr Carlson's video, whatever the merits of his tester may or may not be, I think the film probably tells us more about the shortcomings of the Heathkit and Jackson testers than it does about the merits of his tester.

Both instruments, when dealing with capacitance, are measuring instruments but when it comes to leakage they are merely indicators. They simply ask the question “is this component leaky, YES/NO”.

In order for the indication to be of any use at all it is desirable to know the threshold for the YES/NO decision.

I managed to find some information on the net about the Heathkit IT-11 but not what I wanted to know. The leakage calibration process in the manual mentions using a 1.5 meg-ohm resistor connected to the test terminals and adjusting internal pre-set(s) until the eye tube display just closes. This suggests to me that a capacitor will have to have quite a lot of leakage before the instrument condemns it. The specification of the IT-11 given in the manual offers no useful information in this regard (see below).

The Jackson 591 manual is also available on the web but only if you are prepared to pay for it.

The insulation tester I mentioned in post #4 above is the Vici VC60B+ and is available from various sources for about £30. I offered a review of the VC60B+ in post #10 of this thread.

The VC60B+ is not suitable for testing electrolytic capacitors.
.

[ms660]

I watched the video but unless I missed it I could not see or hear a technical explanation of how his tester works, anyone know?

Lawrence.

[Stevie342000]

Yes I've seen the full version. It's simple it treats the capacitor like a resistor under test. If you're testing paper caps that test good on his kit or a heathkit tester you can change the type to mica on his tester and on his tester it can still test faulty but passes on the heathkit. Essentially he is testing down to 1uA on the mica setting.

Days after I had bought an Avo Type 1 bridge tester, he came out with this and from what I can see the Avo performs a similar function down to 1uA. I was not happy at first but have to say the AVO is a stunning piece of kit internally and externally.

Either or all of the testers are yes or no function testers in my opinion the AVO is the better tester and if I did not have one already I would build his piece of kit.

[ms660]

Thanks, must admit I've never used a capacitor leakage tester much except for a brief period with a homebrew neon one back in the 1960's.

Lawrence.

[Stevie342000]

Me either it is a useful tool for us who rebuilding, repairing or making new equipment using valves.

Mr Carlsons attitude on reforming is not worth the effort, they are still going to fail at some point and cause damage. Capacitors in his opinion are cheap enough these days to just replace them, I have to say I tend to agree.

For me the importance of a capacitor leakage tester is to repair all the kit I have and to bring it back into working order. That is my main aim, I am loathe to fire up equipment which has been stored for a long time, yes I have done it.

The Williamson amplifiers worked fine but they did have paper in oil reservoir capacitors which are military grade and rarely go out of kilter. But it may have had issues none the less. If I tried 5881 in there they used to red plate which could have been one of two things, bias was wrong or there was DC on the grids from the previous stage, probably the former. They were not really rated for the circuit it is at their upper limits.

If you do have leaky grid capacitors at worst it can take out kit like output transformers which will not be so easy or cheap to replace, so to me it is worth the expense to buy one, build or borrow one.

[Argus25]

I watched the video too. There are some things I could point out, positive and negative.

Firstly there is a bit of hyperbole over the voltages involved with the high voltage/eye testers "500V 1" and "600v" etc, a bit dramatic. It is not the voltage that risks electrocution, it is a potential field. It is the current. You can charge yourself to over 1000v walking on carpet and discharge that though your hand later. It is the current and energy that is important. Off hand I think it takes at least 70mA or more to be at risk of cardiac arrhythmia. Most household RCD's trip at 30mA and medical ones at 15mA. So I would think Heathkit would have limited the current to less than 10mA on the high voltage terminals. So it likely "safe" but perhaps unpleasant. Though it could be that a large uF value high voltage cap could acquire enough total charge and energy to be a bit more risky.

Secondly, in the past I have tested capacitor's where the leakage only shows up at higher applied voltages >50V. Also, for a low voltage leakage test, all you need is a standard digital meter, and if it doesn't go off scale to >20 meg after a while you know the cap is leaky.

I do agree that different types of capacitor's require different types of tests, for example an electrolytic as it dries out can simply lose capacity and have normal or even lower leakage and raised ESR. So the ultimate in capacitor tester would measure leakage + ESR + capacity and perform a rated voltage check too and a waveform check (see below) which a low voltage leakage tester cannot do.

Some capacitors have an interesting version of leakage and behave as though they have a much larger capacitor value with a large series resistance placed in parallel with the main value , without significant parallel resistance. Because they can have little parallel leakage they get missed on a leakage test. Also depending on the uF testing method and the impedance of that circuit they can read a normal value. The way to pick up this defect involves applying charging waveform with a discharge into a high resistance >1Meg and looking on the scope for the presence of two exponential waveforms of different time constants.

Also, clearly the Leakage sensitivity on the Heathkit tester is a bit low, meaning the leakage current from those defective caps tested in the video wasn't displayed in the magic eye because that current, which generated a voltage across some resistive load in the unit, was not amplified enough to see it, but it was there. Which is just a sensitivity issue.

Having said all that it is nice that the inventor of this more sensitive capacitance leakage tester has made the design available at a reasonable price. But there is clearly a motive to demonstrate superiority of the device for financial reward albeit small.

I'm of the opinion though that every capacitor he identified as faulty I could with the aid of my digital meter on Ohms range (which is a low voltage test), and an ESR tester and uF meter for the electrolytics. But as noted there is still the high rated voltage withstand test and the discharge waveform test to do, to be 100% sure about the capacitor. Also to go the whole hog the capacitor can be charged up, briefly shorted and if the voltage climbs after that is a measure of the dielectric absorption.

I agree with his remarks about reforming capacitors, which surprised me because seldom have people agreed with me on this issue.

[Diabolical Artificer]

If you think about it someone working on gear that uses HV and has HV rated caps inside should be able to work with HV safely and shouldn't be freaked out by 600v. If your scared, your not thinking straight, best leave alone. Therefore a cap tester that has 600v on it's OP terminal's is no biggy.

Where I'd disagree about reforming is where a scope for instance has about 20 big smoothing can's or a bit of gear you don't use much. Good 500v rated can's ain't cheap (in my book), blowing £200 for caps on a scope you don't use often isn't a go'er.

Hadn't heard about the waveform test, will have to look it up.

Andy.

[GMB]

Since leakage is non-linear I can't see how a low voltage test is going to work unless you made measurements at different voltages and then tried to extrapolate - which might be possible but I suspect it could be misleading. The apparent resistance of a duff capacitor measured at low voltage is often huge and hard to measure in my experience.

AC bridges don't measure leakage either as they cannot distinguish between series and parallel resistance. Do be aware that some bridges present the measurement as capacitance and parallel resistance - but this does not mean it is like that.

[G8HQP Dave]

If you want to know the leakage of a capacitor with 180V across it then the only way to find out is to put 180V across it. It is reasonable to suppose that if it doesn't leak with 500V then it won't leak with 180V. It is unreasonable to suppose that if it doesn't leak with 5V then it won't leak with 180V, whatever waveform or fancy calculations you use.