Capacitor leakage tester

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

diydidi

Active member
Joined
Mar 17, 2014
Messages
39
Location
South Africa
Hi Guys!
I want to build a capacitor leakage tester/ checker instead of buying something expensive.
I was thinking of using my variac, followed by a voltage doubler to reach about 700V DC.
I'm not sure how to do the metering. I could either use a 50uA moving meter in series with a 10k resistor followed by the cap under test, or I could measure the voltage across the cap instead of the uA meter keeping the series 10K resistor.
I'm open for ideas and if any one has a better layout, please share it with me.
Thank you!
 
diydidi said:
Hi Guys!
I want to build a capacitor leakage tester/ checker instead of buying something expensive.
I was thinking of using my variac, followed by a voltage doubler to reach about 700V DC.

I'm open for ideas and if any one has a better layout, please share it with me.

What you have in mind is very similar to what I use to reform capacitors, should suit your needs with your voltage doubler instead of the 1:4 step up transformer needed here, jpg attached.

We are 120 mains, so this plugs into variac, a 1 amp wacky circuit breaker with a long red button from an old '60's TV set or something, transformer is 120V --> 480V step up, with center tap.

Bridge rectifier from the 480V, and a pair of 10uF/450V  filter caps in series, center point of caps to center tap of transformer, the same voltage range output as what you have in mind.

The implementation is drawn out on the jpg: A precise 1K resistor from positive for monitoring current, 1V = 1 mA. 1mV = 1 uA. Those test sockets take meter probes snugly, just plug in the meter leads and leave it there to monitor current.

After that, is a 47K / 10 watt green ceramic, shown peeking out at the bottom of the photo, it is ballast for reforming caps so you can leave them reform overnight, unattended. Last, test jacks to monitor the cap voltage with a second meter, if you want. The buss wires sticking out is where you actually connect the cap with alligator test leads.

Just a simple "I need something in 20 minutes" project, that I still use.

Gene
 

Attachments

  • Reform school for caps.jpg
    Reform school for caps.jpg
    55.7 KB · Views: 47
Of course be careful messing with several hundred volts.

I rolled my own crude 480VAC supply  by connecting 4  standard transformers with all the secondaries connected together in parallel, but the primaries in series forming a bastard auto former...

I was able to crudely measure insulation resistance for the item I was testing by connecting a standard VOM in series... as long as the current leakage is low the VOM is not exposed to dangerous voltages.

I feel compelled to repeat, be very careful, and if messing with caps don't ignore that they can hold a charge and still zap you after they are removed from the test rig.  :eek:

JR
 
Can you get any one of these old tube-based testers where you are? The Eico 950A/B and the Heathkit models appear to be most common, at least they are ones I've used or seen at hamfests in the USA.

http://www.ohio.edu/people/postr/bapix/capchkr2.htm

These "only" go up to 450V. 700V is a lot (well, so it 450V when stored in an electrolytic cap), do you really have capacitors rated to 700v?

I second John's concerns. My Eico works fine (works for both reforming and testing leakage), the case is insulated from everything (though I really should put on a 3-wire cord with the case going to ground for better safety), but I've been very careful (one hand in pocket before reaching for a know, other precautions) when going over 50 to 100 volts or so.
 
A cheap digital insulation resistance meter can provide a variety of useful functions, including as an excellent capacitor leakage tester.  My unit tests at 100, 250, 500 and 1kVDC, with the benefit that it is designed to be safe, convenient, and has a current limited supply.  Even if testing below a capacitors rated voltage, any leakage is likely to show up.

In the proposed variac tester, using a multimeter, rather than a dedicated sensitive meter, has the advantage of easily starting at a high voltage range and then stepping down (manually or automatically depending on what meter you use).

An alternative variable high voltage DC supply can be made using the cheap ebay dc/dc supply that provides a pot settable output from 45V to 390V.  That supply has a current limited output.  Two of those supplies can have their outputs connected in series for up to 780VDC testing - with the 'upper' dc/dc powered from a small isolated 12V battery (eg. 7Ah VRLA monobloc that are very common thanks to ubiquitous use in domestic/commercial UPS) and using a preset voltage such as 390V.
 
benb said:
Can you get any one of these old tube-based testers where you are? The Eico 950A/B and the Heathkit models appear to be most common, at least they are ones I've used or seen at hamfests in the USA.

These are a good way to go, I paid 10 bucks at a hamfest for a working Heathkit IT-28 about 20 years ago. But after reforming a gabillion caps over the years, the magic eye tube phosphor was barely lighting up.

After putting a NOS 6E5  tube in the Heathkit, I built the transformer gismo pictured above for reforming, to not wear out the new tube from days of unattended operation, while no one is having the joy of looking at it. I really love those magic eye tubes, newfangled digital consoles should come with them for VU meters. Scribe the dB scale right on the phosphor target, put red phosphor above the 0dB point.

The IT-28 has voltage settings in 50 volt increments up to 600V, open circuit voltage is shown on the meter.

And yeah, resist the urge to lick the binding posts.

Gene

 
Hell yeah!

Just bought an IT-28 waaay to expensive but I'm looking forward to get this beauty back in business. Just need to get those paper/wax caps and electrolytes out of there so I dare switch it on to check if the magic eye still glow bright (I keep my fingers crossed because I do not need that extra cost :)...)
 
synthiaks said:
Hell yeah!

Just bought an IT-28 waaay to expensive ...

Congratulations, they are cool.

I recapped the two main PS caps with modern ones and added dividing resistors. Went through the cal procedures that are all over the net, and with modern line voltages, increased the value of the resistors that feed the plate/target of the 6E5 magic eye tube, it was about 215V, should last longer at the recommended 200V. 6E5's aren't all that rare, I got a NOS one from a friend that is in the NOS tube business, $10.

In case you haven't found it already, the schematic is attached. With a little searching, the complete service manual is out there, also.

Gene
 
Hi Gene!

Thanks a lot, the unit came with a copy of the manual all tidy in a plastic cover.

Interesting point with the line voltage and the plate voltage. I'm going to check mine to if it gets to much.

I love the simplicity of the circuit but hate point to point...

Here in Germany the 6E5s are available but they are pricey. You have to pay 20$ for one at least. And I already have paid to much for the IT,

The IT-28 are quite rare here and when one go up for sale on eBay it usually sell fast but I figured they won't get any cheaper as they are quite useful still and with caps from 50s,60s,70s,80s and90s now all seem to fail, a cool leakage checker with Bias up to 600V DC seem like a smart thing to own :)
 
trobbins said:
An alternative variable high voltage DC supply can be made using the cheap ebay dc/dc supply that provides a pot settable output from 45V to 390V.  That supply has a current limited output.  Two of those supplies can have their outputs connected in series for up to 780VDC testing

Hi can you let us know the models names or the ebay source for those dc/dc supplies?

thanks
 
trobbins said:
Easiest to search eBay for 45 and 390  and dc.

It's this one?
http://www.ebay.com/itm/DC-DC-Boost-Converter-8-32V-to-45-390V-High-Voltage-ZVS-Capacitor-Charging-Board-/291625317020?hash=item43e639329c:g:rtcAAOSwZG9Wi4W~

Thanks
 
Yes.  I reverse traced the schematic, and put in another forum to see if anyone new the control ICs used, but no feedback as yet.
 
Whoops, this was the link to more info on that dc/dc module:
http://www.diyaudio.com/forums/power-supplies/297112-yh11068a-10-32v-45-390v-dc-dc-guess-control-ics.html
 
trobbins said:
Whoops, this was the link to more info on that dc/dc module:
http://www.diyaudio.com/forums/power-supplies/297112-yh11068a-10-32v-45-390v-dc-dc-guess-control-ics.html

Nice I will a buy few of those.

Seems like a really good circuit to do the Tech Bench  Variable PSU that I want to do for a long time.

 
One of the reasons I went on that forum was to try and identify the control IC's in that psu.  The short amount of observation I did is that the output, although regulated, does include varying levels of ripple and transient response due to to load changes - whether that is in any way noticeable in your application I'm not sure.
 
trobbins said:
One of the reasons I went on that forum was to try and identify the control IC's in that psu.  The short amount of observation I did is that the output, although regulated, does include varying levels of ripple and transient response due to to load changes - whether that is in any way noticeable in your application I'm not sure.

Could that ripple be smoothed out?
 
The PSU output response could be 'isolated' by using a series diode in to say an RC filter that is your application supply - that extra C would need a separate bleed resistor.  Application currents would then mainly flow through the extra C.  The PSU has a common 0V, so grounding needs to be looked at.
 
diydidi said:
Hi Guys!
I want to build a capacitor leakage tester/ checker instead of buying something expensive.
I was thinking of using my variac, followed by a voltage doubler to reach about 700V DC.
Building a high-voltage unregulated PSU is a rather trivial matter; the most difficult part is finding the xfmr. JR's solution of connecting several xfmrs in series is probably the easiset and most flexible. It is important to be able to set the voltage as desired.

I'm not sure how to do the metering. I could either use a 50uA moving meter in series with a 10k resistor followed by the cap under test,
The 10k resistor in series will not be sufficient because, in case your capacitor is shorted, the current will be many times the lethal dose for the galvanometer. You need to add protection diodes across the galvanometer.


or I could measure the voltage across the cap instead of the uA meter keeping the series 10K resistor.
That would be very risky; if for any reason the voltmeter goes short-circuit, you galvanometer is instantly vaporized.
 
Back
Top