Circuit for a simple mains-powered capacitor leakage tester.

[Yorkie]

I thought this simple circuit from January 1961 Practical Wireless might be of interest. It shows a metal rectifier but these days a 1N4007 or similar silicon diode would be fine in that role.

Neon panel lights are plentiful and cheap. EG:

http://www.ebay.co.uk/itm/2-Neon-Panel-L...ltDomain_3&var=630000900010&hash=item53df526401

Hope that's of interest to someone.

[tiv48]

may I suggest that anyone wanting to build this should use 2 back to back mains transformers.
Rob T

[Skywave]

In order to test a capacitor for an acceptable value of leakage - which, ideally, will be zero µA - the voltage applied to that capacitor needs to be approximately the same as the rated working voltage of that capacitor. This important consideration, that unit fails to do. Consequently, a H.V. d.c. supply which has a variable output, albeit over a limited range - or one with pre-set values of voltages - is required, so that capacitors of various working voltages can be reliably tested for leakage. Such a device is not difficult to design: indeed, many have been published over the years; some are available on the 'Net.
Now add to that criticism Rob's highly-relevant comment (as above), and in my opinion, this unit should not be built: it fails to meet its design requirement; it is dangerous.

No disrespect to you David: just my own personal opinion of this design.

Al. / Skywave / Oct. 3, '13 //

[ppppenguin]

You can get the safety of mains isolation and 2 different test voltages by using a shaver transformer.

[Nick]

I completely concur with the sentiments above - this is a dangerous circuit and not fit for purpose.

If you want to check the leakage of a capacitor, charge it with a bench DC supply through a current limiting resistor, disconnect it from the supply and then use a VOM, DVM or other high-input-impedance voltmeter (i.e. NOT and AVO!) to measure and plot its decay.

It'll be pretty obvious very quickly what the leakage is like and whether its acceptable. All caps leak to some extent anyway...

The picture below is an example of the curve expected from a good electrolytic - I plotted this back in 2010 from an 8 y/o extremely low ESR cap used in commercial UPS systems - a day and a half later, the voltage had dropped by 20%... (we swap these caps out every 3 to 5 years as a matter of standard maintenance - they are GBP 90 each and a typical mid-range UPS will have maybe 60 of them...).

[Refugee]

I just use a 100K resistor in series with an isolated HT power supply and measure across the 100K resistor with a DVM and calculate the leakage currant.

[Yorkie]

I take the points made - I think this unit would be safe in our hands, but not perhaps in the hands of a novice, though if the overriding concern is that only equipment isolated from the mains should be entertained, live chassis radios such as the DAC90A - often the first radio that complete beginners work on - would go to landfill! (From my perspective, few restorers - especially novices, use an isolation transformer). The design dates back to an era when it was commonplace in magazines such as PW to feature radios in which the HT was taken directly from the mains and just a heater transformer was used. Very much a case of "ride at your own risk" - I can't imagine any magazines featuring such designs today!

Insulation testers from the Far East are now so cheap that in any event I guess no-one is motivated to design or make a home-brew one. I've mentioned a couple over the last year or so on the Forum - one is digital - the Victor VC60B, (£36.99 post free), (250V, 500V, 1000V, 200M Ohm). The other is a hand-cranked Megger 'lookalike' 'ZC25-4 £32.22 from Amazon post free. (1000M Ohm, 1,000 Volts).

http://www.ebay.co.uk/itm/Digital-Insula...uipment_ET&hash=item4cf50e6dd6

http://www.amazon.co.uk/Insulation-Megoh...15?ie=UTF8&qid=1363123612&sr=8-15

Both are described as 'Megger' which is a trade-mark infringement, given that Megger are still very much in business!

Personally I'm sorted - I built the Radio Bygones cap reformer and tester some years ago, which is excellent. It uses two back to back transformers and a voltage doubler to give a maximum of 500V and two meters - a digital one for voltage, an analogue one for leakage. It also has a neon to give an indication of leakage on non-electrolytic caps. Altogether, a very useful unit.

(I do have a hand cranked Bakelite vintage Megger but don't use it).

[Refugee]

I use a restored real Megger for non electrolytic types and for ones to be used for grid coupling I just shove them between live and earth on a PAT tester with a DMM in series and calculate the resistance by ohms law based on the DMM having a 10 meg terminal resistance. I just turn the voltage control as the capacitor charges up and I can then see the tiny variations in mains voltage as thermostats click on and off on the mains supply on a good one.
I can measure up to gig ohms by this method.

[Radio Fixer]

"I just use a 100K resistor in series with an isolated HT power supply and measure across the 100K resistor with a DVM and calculate the leakage currant. "

Is this good enough for say SM?

Me, I have a Levell Insulation Tester TM14 picked up at an NVCF. Great tester, battery powered with an inverter for voltages from 10 to 1000V and indicates Gig Ohms.

Gary

[pwdrive]

We used to make up similar to what David posted when I was a TV tech, potential live chassis TV's were the norm then, inevitably they would end up in a draw being unused once the novelty had worn off.
To be honest I have always been perplexed why so many folks go for capacitor leakage/ESR meters in the context of valve vintage radio/TV repairs, snip the earthy end of the cap with the set on and stick an AVO on it, job done, if it's an electrolytic that's bulging, oozing or whatever then just replace it, if you suspect it's lost it's value then just bridge it freehand, if normal service is then resumed then replace it.
In the workshops I worked in there was only three kinds of homebrew gizzmo's that were ever used, one was a signal injector (multivib) used when another tech had hoarded the bench signal generator, another gizzmo was a home made tube bopper, the other was an instability probe which was good with the old high gain Murphy TV's.

Lawrence.