500 VDC Capacitor Leakage Tester 

Hahaha! I think you're a little biased, being my son! But thanks!!! ❤️

Walt Wimer Hopefully you can teach me more tutorials about these capacitor leakage testers and then I’ll be able to get the idea. I hope to learn some new computer skills to benefit myself so that I become smarter at these skills. Perhaps I’m not yet at that level but I hope to get there soon!

In the Vintage Test Equipment page He always contributes with very useful comments aswelll..

Thanks, Steve!

@@waltwimer2551 I hope I don't sound critical! I have a tendency to not quite finish my own test equipment. I love doing the whole thing with rub-on lettering, etching, custom labels and meter scales... but these days, I get it working (for work) and use it. I built a few things to help automate testing at work, all are currently still half on protoboards. At home, I take my time.

@@AlienRelics Not critical at all. There is always room for improvement... I'm not trying to sell this as a real product. If I were, I would redo it and make several changes myself. As it is, my tester represents a compromise between trying a few construction ideas vs building something for the very narrow purpose I needed -- checking DC leakage current for small-valued capacitors. My next big project is a (passive) switch box for switching among 12 different open reel and cassette tape decks. It will have 12 ON-OFF-ON toggle switches, two 12-position rotary switches, and 56 RCA jacks... I want to do a nice job labeling it, so that it looks like a professional audio component. I'm still researching the best approach for accomplishing that. 🤔

Good point, Alan. I really only designed my tester to test film capacitors with capacitance values of about 1 μF and less. In the case of electrolytics, I usually just test for ESR using my LCR meter (which uses very low voltage AC, less than a PN-junction drop so that the meter can be used in-circuit). And/or I just replace electrolytics based on age or ripple measurements in-circuit. And my tester is for use at about 50 VDC and higher, since I use a variac to set the voltage (not very precise below around 50 VDC at the final output). For lower voltages, I just rig up my bench DC power supply (max 32 VDC on one channel, or 64 VDC with both channels in series) and a DMM in ammeter mode. (I'm currently working on restoring a 40-year-old Revox open reel tape deck. It has a muting circuit to suppress power-on/off pops in the audio chain. The muting circuit was stuck in "mute" mode, preventing the deck from playing any music at all. I traced the problem to an electrically-leaky 22 μF 16V tantalum capacitor, verified by using my low-voltage DC bench power supply technique. In the steady state, it was passing 5 mA of leakage current. This robbed a BJT of base current, thus keeping a relay de-energized, which kept the audio muted. I temporarily replaced the tantalum cap with an aluminum electrolytic I happened to have on hand, and voilà, the muting circuit now works correctly.) 🙂

@@waltwimer2551 Got, thanks! I'm currently assembling a leakage tester offered by Antique Wireless Association that I'll review on my channel. It is a direct measurement instrument to - can test from about 10V up to near 500V, with sensing/reading of current across a 1K resistor (to give 1V/ma response). It is designed to use external metering for capacitor voltage as leakage current, but I'm putting a vintage voltmeter in the housing to read the test voltage (as you may have seen on my Twitter feed). Interesting failure mode on the muted audio. Reminds me of a tricky problem I had on a old Kenwood TS-830S. CW sidetone was stuck "on" even when not keyed up. But, if I put my fingers across the key, it would stop. Turned out to be a leaky switching diode in the sidetone oscillator - when presented with the full open-key voltage (about 60V) it would leak enough to go low-Z and cause the oscillator to run. Putting my fingers across the key lowered the open circuit voltage far enough to lower the leakage in the diode to the point where the oscillator wouldn't start up. Replaced the diode, all good. Interesting that the prior owner had the problem, couldn't figure it out, and cut a wire off of the mode switch to keep the sidetone oscillator off. Found and fixed that first, then discovered the original/real problem.

I may eventually get around to making a video about my design, with internal views of my tester. However, I highly recommend a safer, low-voltage design by Paul Carlson, of Mr. Carlson's Lab. Check out his design (and all of his other excellent videos): ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-LhovRIM5xAo.html

Thanks! It's straightforward. Nothing innovative or anything. But it does the job. I went overboard by purchasing everything new and selecting high-quality components. So it certainly wasn't a cost-effective build. But I enjoyed assembling it and I'm generally pleased with the final results. I guess one reason that I went a bit overboard is that when I was a teen, I did something stupid with a lawnmower. My family had a bare-bones inexpensive walk-behind lawnmower from Sears. It had a recoil-start engine and no choke or throttle or anything. To stop the engine, there was this little metal tab that you moved to ground out the exposed terminal on the spark plug. (A design like this wouldn't pass safety regulations today...) The metal tab eventually rusted and broke off. For a few years, we used a big screwdriver to ground out the spark plug. One day I tried to be clever and wire up a garden-variety 120VAC momentary toggle switch with a metal bat handle that I could use in place of the screwdriver. I mounted the switch up high on the lawnmower handle so I didn't have to bend down. (I used some wire with thick rubber insulation, figuring that it would handle the high voltage.) Well, I started the lawnmower and let it run for 30 seconds or so. Then I confidently grabbed the toggle switch bat handle (avoiding the contacts) to switch off the engine... ZAP! I got one hell of a shock! I got out the big screwdriver, stopped the engine, and disassembled my abominable contraption... Eventually I came to understand the concept of withstanding voltage... There was no way the internal insulation inside that toggle switch could insulate against several thousand volts or more. It basically acted as a capacitor and effectively just passed the ignition pulses through almost like a wire! With that painful memory in mind, I wanted to be sure that all the components in my leakage tester could effectively insulate the human operator from the 500 VDC. The DISCHARGE/TEST switch was the most challenging. Most of the switches that I could find at Digi-Key, etc. maxed out at 250V. Everything with higher ratings were humongous and prohibitively expensive. I eventually found a switch on Amazon that was rated for 660 V that was suitable and didn't break the bank. That's why my tester has an "AVOID" position... I didn't want that position, but that's how the switch works...

Nothing bad will happen with electrolytics (assuming you observe proper polarity), but with the schematic as-is, the charging current is (intentionally) limited to 50 μA. Electrolytics will take a LONG time to charge (hours or more...). For use with electrolytics, you will probably want to modify the circuit by adding a rotary switch with various precision resistors to shunt across the ammeter. With the correct value shunt resistors, you can scale the 50 μA ammeter to represent 500 μA, 5 mA, 50 mA, 500 mA, etc. You will also need to switch out the 10M ohm series current-limiting resistor. One of the reasons I didn't design these features into my tester is that I didn't feel very comfortable using switches rated for only 125 or 250 volts. When I looked for switches with higher voltage ratings, they were either impossible to find, or they were huge, expensive things that made no sense for my tester. I guess the only option is to use a 250 volt switch and hope for the best.... And use an all-plastic knob. If you do this, it's at your own risk. (I apologize that the schematic is currently unavailable. The link points to my home web server, which is currently offline. I'm not sure when I will be able to get it going again. I wish RU-vid would let me post a .png file.)

As long as the secondary is 380 volts and you're feeding the primary with the correct rated voltage (220 in your case), there shouldn't be any changes needed to resistors, etc. Have you tested multiple capacitors? If you've only tested one, then that particular capacitor may be shorted. Also, you're using a variac, right? Also, my tester is mostly intended for testing low-capacitance capacitors (say 1uF or less). If you try my tester on big electrolytic capacitors, the charging time will be very long, so the ammeter will be pegged for a long time. In that case, you *could* reduce the 10M ohm resistor, but you will be making more current available at the test terminals, thus increasing the dangerousness of your tester. And if you decrease the 10M ohm resistor, you will want to place an appropriate shunt resistor in parallel with your ammeter, in order to protect the ammeter. This will change the calibration, such that a reading of say, 20 microamps might be (depending on your choice of resistors) 20 milliamps or whatever.

@@waltwimer2551 first thanks for the response.... i am using an isolation transformer then a variac then the tester..i found that the 10 mg resistor was defective and was only doing .998m ohms...fixed that and now the meter works...didi not catch the part about low value caps in the vid..i need this to test audio amp filter caps...mostly 400uf and up..sometime way up...lol so i am going to have to play with the resistor or mr c's tester...i like this one more..it looks good on the bench,,,,,,,so keep the vids coming...Gator...Brasilia,, Brazil...

@@robertskolnick8162 The nice thing about the low current is that you can use it to reform electrolytics slowly. Personally I'd add a switch to allow the the 2nd voltmeter to read directly off the test terminals so you could watch the charge build. If you leave it across the test all the time it'll effect the leakage reading.

Let's see if this hyperlink works: www.hypervolt.net/misc/leakage_tester_schematic.png

Try the link again, or a little later. It goes to my home Linux server, so connectivity may be spotty sometimes. I just tried it via the cellular network from my cellphone and it worked for me.

Thank you!

Once again, I highly recommend Paul Carlson's much safer low-voltage capacitor leakage tester: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-LhovRIM5xAo.html. My design is mostly a "brute force" linear power supply design, similar to what you would typically find in a classic piece of vacuum tube equipment, such as an audio amplifier. I have a step-up transformer, a bridge rectifier, filter capacitors with balancing / bleeder resistors, a 10 Mohm series current-limiting resistor, and a 50 microampere DC meter. Plus voltmeters and a funky switch to flip between "discharge" and "test" mode (from the perspective of the external Capacitor Under Test). I was very careful to select capacitors, switches, safety banana jacks, etc. that could handle 600 VDC or higher. If you are trying to reach 1500 VDC, first I would say, "Don't do it." Try Paul Carlson's low voltage design first. If that doesn't make you happy, then try designing and building something around 500 VDC or lower. Whatever target voltage you choose, you must carefully select your components (transformer(s), capacitors, rectifiers, switches, jacks, etc.) to handle the expected voltage. I found it somewhat challenging to find modern components that could handle my goal of 500 VDC. To find components that can handle 1500 VDC (actually I would design with a safety margin, so I would select components rated for 2000 VDC or better), will be quite challenging. Designing and building a 1500 VDC tester is probably completely unnecessary, and is probably *not* worth the safety risk. (Also, to go that high, you probably want to use a voltage multiplier circuit that uses a combination of multiple diodes and capacitors to achieve higher output voltages. But again, just say no... Try Paul Carlson's low-voltage tester!) :-)

Good eye! Yes it is. It's one of these. I forget which one exactly. Obviously one of the larger ones. Instead of the intended orientation, I turned it 90° and added my own feet to the new bottom. Everything just worked better that way for this particular project. www.hammfg.com/electronics/small-case/general-purpose/500-515-519

Please see the description for a link to the schematic. Cheers! 😎👍

That is one frustrating aspect. Ideally, it would indeed be nice to see the same reading on the "capacitor voltage" meter as on the "supply voltage" meter when the test is in progress. Unfortunately, if the "capacitor voltmeter" were connected across the capacitor during the test, the current drawn by the voltmeter itself would affect the current displayed by the ammeter, changing the results. So I elected to keep the "capacitor voltmeter" in parallel with the discharge resistor. Thus, the operation is: - In the TEST position, the "capacitor voltmeter" is inoperative (dead). - In the DISCHARGE position, the "capacitor voltmeter" shows the present voltage across the capacitor. Since I'm testing small-valued capacitors (e.g. 0.1 μF), the discharge resistor very quickly discharges them, thus the voltmeter is not fast enough to respond. If I were to test a much larger capacitor (say 1,000 μF), then the voltmeter would be more interesting. Another option that I considered was to wire the voltmeter to the control switch such that in the TEST position, the voltmeter would show the voltage "upstream" of the ammeter, rather than the voltage directly on the capacitor, and then in the DISCHARGE position, the voltmeter would show the voltage across the discharge resistor. I don't recall why I ultimately chose not to do that. It may have been a limitation on the number of switch contacts. Or, since the purpose of the "capacitor voltmeter" is for safety, I may have decided that it should be permanently wired across the discharge resistor. 🤔

No. The cap may "cook" but the power transformer will fry to a crisp sooner. The transformer is harder to replace and much more expensive. BTDT. Old caps are bad. Period.

I chose a power transformer from Hammond Manufacturing: Model 269EX. It has a 380 VAC center-tapped secondary. I left the center-tap disconnected and used a bridge rectifier to convert the AC output to DC. If you multiply 380 VAC RMS by 1.414 to get the peak voltage (the voltage to which the filter capacitors are charged up), the result is around 535 VDC. I use a variac on the primary side of the power transformer to adjust the final output voltage. So if I was careless, I could likely output 600 VDC or perhaps a little more. The voltmeters I chose hit full scale at 500 VDC, so I use care to avoid exceeding that. Here is my schematic: www.hypervolt.net/misc/leakage_tester_schematic.png

Thanks for the answer, all explained! 73 from Germany

Well, I kind of went "top shelf" with everything... The metal case and the power transformer were the most expensive components. I included nice touches like safety banana jacks and quality test leads, and a perfboard inside mounted on nylon stand-offs for wiring the bridge rectifier, filter capacitors, and resistors. So, I sunk about $200 to $250 into this project. If instead you dig up a used power transformer and different case solution (re-use something else, or make something out of scrap wood, whatever), skip the fancy jacks and test leads, etc., I'm sure you can get the cost down to around $50 or less. The meters were remarkably cheap Chinese meters from Amazon -- $11 each! Good Luck! Walt

@@waltwimer2551 That's a professional clean build. Respect. I dig the $11 meters! Thanks again!

He is actually a big reason that I made this first video. I mentioned to him that I have a ~1952 Heathkit S-2 "Electronic Switch" (similar in concept to the Sylvania one I first saw on his channel) that I am planning to restore. He immediately encouraged me to make a video of my efforts. I haven't found time for that project yet, but one of these days. I will try to make a few more videos in the future. They'll never be up to Paul Carlson's standards! He has set the bar! But I'll try to make them half-decent. :-)

@@waltwimer2551 Mr C isn't from here. He arrived here using a VTWD (Vacuum Tube Warp Drive) spacecraft. But don't tell anyone. Him and Uncle Doug are both irriely awsome.

I can already tell that my dad’s about to become the next Mr. C. Exciting! 💯

You guys need to watch some of Joe Smith’s videos and the test gear he built for testing multimeters for high voltage transient spikes. He and Paul along with Alan(w2aew) have the best engineering chops on RU-vid. Joe found serious flaws in a Gossen Ultra that a simple magnet could switch a relay so the meter would show low voltage when actually hooked to a high energy source. He wrote them about it and showed them the video, they never responded. An $800+ meter that could get you killed. He also called out Dave Jones with many issues of his 121 GW. Dave didn’t have much of a sense of humor about it. Joe has a great relationship with Brymen meters, I have purchased 4 after watching his tests. They are very robust and took 12kv hits before minor failures like clamping diodes. Brymen and Fluke were the clear winners in high voltage transient tests. UniT meters failed from a single grill starter hit. A static charge in the winter could produce as much energy and fry one. Anyways, Joe’s videos are great to watch. His sense of humor makes you watch the whole video. ru-vid.com/group/PLZSS2ajxhiQBTCU8Mq_i9jidT024A0dV6

@@waltwimer2551 You might find these interesting. Joe Smith’s tests of a Gossen meter. Joe’s skills rival Paul Carlson in electronics. Some of the test equipment he builds for doing meter robustness tests are amazing. ru-vid.com/group/PLZSS2ajxhiQBTCU8Mq_i9jidT024A0dV6

I always plug my tester into a variac.

@@waltwimer2551 I don't have a vriac existing another way to change voltage thanks for your reply

@@radiotvrepair1059 Well, first I must state to EVERYONE that you are proceeding entirely at your OWN RISK here. Any advice I give is strictly for qualified personnel. If you have *any doubts* about your skills, knowledge, or abilities, DO NOT proceed. Stay safe! That said, there are a number of possibilities... 1. You could use an inexpensive solid-state dimmer control that you normally install in the wall to dim lights. You could wire such a dimmer on the primary side of the tester's transformer. This is kind of a "nasty" approach since the output waveform from a solid state dimmer is "spikey" -- it's full of harmonics that aren't particularly kind to the transformer. It could make the transformer audibly buzz, and could heat up the transformer. But if you are only testing small capacitors with the tester, then the current will be very low, and the abuse from the solid-state dimmer will probably be very minimal. On the other hand, if you have modified your tester to provide higher output current to the Capacitor Under Test for reforming electrolytic capacitors and such, then the current through your power transformer will be higher, and the undesirable effects of a solid-state dimmer will be more pronounced. 2. If your transformer has a secondary winding with a center tap (or multiple taps), you could wire a switch to select between the center tap and the "top" outer leg in my schematic. This would give you a coarse voltage control: half vs full. You could use this "coarse" control in conjunction with the other approaches to provide a wider total range of control. 3. You could create a variable voltage divider using a rotary switch to select different resistors, thus giving different voltages. You could include a potentiometer to serve as a fine adjustment control. This resistive voltage divider could be wired on either the primary or secondary side of the transformer, based on different trade-offs. On the primary side, the resistors don't have to handle the full 500 volts of output voltage, so that reduces the chance of a shock to the human operator indirectly through the switch(es), and also reduces switch contact arcing. But on the primary side, the resistors and switches will have to carry more current, so that represents more heat, the need for bigger resistors, switches that can handle the current, etc. On the secondary side, the trade-offs are reversed. Engineering is all about choosing among various compromises... Study and analyze well before choosing a particular approach. And then once you have chosen a particular approach and built it, check things like the current draw, etc. using a multimeter. Test and evaluate your project to ensure that it is _actually_ working as you expect.

@@waltwimer2551 thanks Walt

I've made a schematic available: www.hypervolt.net/misc/leakage_tester_schematic.png

If you live in an area where the standard mains voltage is 240 VAC, then definitely use a transformer with a 240 V primary. If the secondary is 300 V, then your final DC output from the tester may be a little lower than 500 VDC, but it should still be quite high. By my quick calculation, you should end up with a final DC output of around 425 to 430 VDC. Still quite effective for this kind of testing. 👍

Good question. For signal coupling capacitors (which should block large DC bias voltages while passing small AC signals, like audio signals in an amplifier), basically any measurable leakage is bad. So if I don't see the ammeter needle go *all* the way back to zero, I would consider the capacitor bad. New, high-quality caps go completely to zero. So anything else is suspect. :-)

Like the author said, good new (non-electrolytic) caps will not register leakage current. From a Cap Checker instrument's manual: Max milliamps acceptable leakage is uF x K + .03, where K = .01 for 3-100 volt rating, .02 for 101-250 v, .025 for 251-350 v, .04 for 351-450 v. So a .01 at 400v rated cap should have less than .01 x .04 + .03 = .0304 milliamps, or 30.4 microamps. This works for electrolytics too. For non-electrolytics, you can measure in-circuit if there is no voltage at one end and high volts on the other: With the amp on, measure DCV at the zero-volt end. It might be up to .04 volt and be fine for most circuits, if it fluctuates between + and - then you KNOW it's fine. If there is DC volts at both ends (like a cathodyne or Williamson PI) lift the low (or 'no') voltage end of the cap and test volts at the lifted end. It will have a static charge at first which will gradually discharge through the voltmeter, and you should end up with very close to zero, again, fluctuating between + and - very low volts like .04 is fine. The fluctuating + to - lets you KNOW it's not leaking.

It's really a judgement call. I find that new capacitors return all the way to zero current flow on my 50 μA ammeter, so I generally prefer zero leakage. But it is possible to use Ohm's Law and work out the equivalent series resistance of a leaky capacitor, and then decide for yourself what the effect would be in the original application (such as an amplifer). So, say the ammeter showed 10 μA of current at 500 VDC. My tester has an internal 10 MΩ resistor in series with the ammeter and the external capacitor under test. 500 V divided by 10 μA equals 50 MΩ. Then 50 MΩ minus 10 MΩ equals 40 MΩ. So the example external capacitor acts like a 40 MΩ resistor in parallel with whatever its capacitance is. Draw or imagine the schematic for your application (e.g. an amplifer) and add a 40 MΩ resistor in parallel with the capacitor. What happens to the circuit? If the circuit's DC resistance at that point is something like 4 MΩ (10%) or less, then the 40 MΩ leakage probably won't matter. But if the circuit's normal resistance is above 4 MΩ at that point, then you could begin to have problems. So, to be safe, I look for zero current on the ammeter. It's what a good quality new capacitor will give you. If you're willing to accept some leakage, then do the math and decide what the impact would be on the particular circuit involved.

@@waltwimer2551 This is in ceramic capacitors and for chemical capacitors they have a high leakage current how I will judge that a capacitor is leaking.👍thanks

I built my tester primarily for testing foil capacitors with a small capacitance value (typically 1 μF or less). My target is wax-sealed paper-and-foil capacitors, polyester-and-foil capacitors, metalized polypropylene capacitors, etc. These are the types of capacitors typically used as inter-stage coupling capacitors in vacuum tube amplifiers and similar vacuum tube equipment. For other types of capacitors, you will have to do your own research. My tester design may not be appropriate, especially since my tester maxes out at 50 μA of current. You could alter my design to add shunt resistors around the ammeter in order to scale the current value by 10, 100, 1000, etc. Ceramic capacitors are very reliable and rarely require replacement. For electrolytic capacitors, I use an ESR meter to measure ESR rather than leakage. I have a DER EE DE-5000. Good luck!!! 😎👍

@@waltwimer2551👍👍 very good explanation thank you very much. I have a esr tester in circuit: mesr-100 made in china

@@waltwimer2551 ok I will add shunt resistors on ammeter but I do not know the voltage in mv for a maximum ammeter deviation.thanks

One of these of the appropriate size. I turned it 90° on its side and added feet on the new "bottom". After shooting the video, I also added a handle to the top. I chose a smooth simple metal handle from Home Depot or Lowe's that is normally sold for use on kitchen drawers and cabinets. www.hammfg.com/electronics/small-case/general-purpose/500-515-519

Than you. I will send a photo when I get it put together. I am waiting on a transformer. @@waltwimer2551

Ah. Sorry about that. It is a temporary problem with my home server. I will fix it within the next few days and let you know. Thanks!

@@waltwimer2551 thanks sir for your following .

I've fixed the issue with my server. It should work now. If it doesn't work right away, try waiting an hour and then try again. By then, any cached DNS records should have expired and your browser should see my server's new IP address.

@@waltwimer2551 thanks sir the link work good

I see on the Vedio that you varied the voltage from 200 v to 400 v while the power supply is not variable on the schema diagram how did you do sir and thank you very much.

Sadly, if the right-hand voltmeter were connected across the capacitor during a test, the current drawn by the voltmeter itself would be seen by the ammeter, thus causing the ammeter to show "leakage" current that isn't the fault of the external capacitor under test. So, I wired my "test" switch so that the right-hand voltmeter only shows the capacitor voltage when the "test" switch is in the DISCHARGE position. The charge on small capacitors is so small that the right-hand voltmeter barely moves when the test switch is moved from TEST to DISCHARGE. Thus, the basic idea is to *always* move the test switch to the DISCHARGE position before touching/moving the test leads. In the DISCHARGE position, the right-hand meter shows the voltage across the external capacitor under test. If the right-hand meter shows 0 volts, then it's safe to touch the test leads. If we were to connect an electrolytic capacitor to the test leads, move the switch to TEST followed by DISCHARGE, then the discharge process would take longer, and the right-hand voltmeter would show the voltage across the external capacitor gradually decaying. But I didn't really design my tester to test electrolytic capacitors, only small film capacitors. 🤷‍♂️

Must have been a RU-vid or network problem. It looks fine to me and many other people. 🤷‍♂️😞

@@waltwimer2551 Thank you. It doesn't matter. I have a Hickok 203 I just restored and a Triplett 630. Both will tell me the same thing. Old caps have to be replaced. Duh... Regards, KN4SMF

@@totallysmooth1203You must be fun at parties.