I am finally getting some time to work on the radios I have collected over the last 12 months. Deciding on which ones first is always hard.

In preparation I bought a range Caps from two different sources about six months ago.

I wanted to check the new Electrolytic ones before I started to replace old ones, and I recently acquired a Megger, for checking transformers etc.

My questions is could not the Megger be used set at the Votlage Rating of the Caps to check them?
I put a few through a test on an anolgue first to check if they behaved as expect, that is the meter would defect some resistance then go to open circuit. I did this both for the new Electros and the yellow poly caps.
When I tested them with the Megger for 250 and 500V which was below the 600 V rating, the yellow caps went open as I would expect but the new Electros that I tested went short.

Have I missed something ( most likely) or are these new Electros failing ?

Please enlighten me

George
Ps I do not currently own any other Capacitance Test capability.

Oh dear...….um...………...I think the vital step you missed was FORMING the caps before "testing"
The electro is a tube of mush until forming then it turns it into some sort of 2 plate insulated device.
Electros need to have a DC current applied and chemically changed, that's what the forming is.
Most people just wire them into a set and turn the juice on and they get "formed" in a real hurry, most survive and work normally.
Some don't and just heat up and go poof. Putting a stress tester like a megger could form them or poke a hole in them!

Experts like Ian Rob and Marcc go on endlessly about being patient in forming new and old electors...listen to the experts.


Fred.

I don't own a megger or a capacitance tester. All I use is a DVM and sometimes a scope. And a series light bulb!

If I suspect a capacitor of leaking at a high voltage I will test it in circuit by measuring voltages. Look at the circuit, observe where there should be voltage and where there shouldn't. Then test.

If I suspect a mica capacitor of intermittent leakage I will disable the active part of the circuit (pull valves), hook the scope up to the cold end of the mica cap and set it to capture an event. Go away and do something else for a couple of hours. Come back and I will know. The same technique is useful for many other intermittents. This technique with the scope can find that tiny little "twitch" in the picture that most people will ignore but can drive you nuts trying to find.

I do just about all component testing in circuit, I never clip parts to test them without a very good reason. It risks inserting a fault that wasn't there previously!

I guess this approach came from the many years I spent in the workshop and the field fixing TVs the other service techs couldn't. Particularly intermittent faults. And working to a deadline, and getting service callback reviews!

Thanks Fred and Ian,
In my situation, I have purchased some new capacitors and thought it would be a good idea to test them before I go to the trouble of soldering them into the circuit. As I have read somewhere that the quality of some of the new 600V Electros are bad before they have ever been used.

Did I suspect mine to be BAD no, but I like to be sure as I am not aware of how long they may have been sitting on someone's shelf when they were sold to me or if that even makes a difference. The results of my initial test with the Megga left me unsure.

Fred did raise an interesting point, I had not thought of, FORMING the Caps before testing them.

I have read about Forming old Caps but not ones that are supposedly new. I have seen some devices that you can build to do this.

Is it really required that you should use such a device on NEW Caps, before installing them or using a Megger to test them??

I have been questioning Silicon Chip re old capacitors: No satisfactory answer to date?

One of the things in respect of capacitors that is not commonly written is the amount of allowable leakage: It is important. There have been considerable improvements in them since the days of the early radio's. Up until the introduction of polymers, Mica was pretty much the most reliable dielectric.

As a generalisation any cap from day one until the early 70's and the introduction of polymers should be treated with suspicion even some Tantalum's & others like the red topped disk ones, in transistor radios, had their moments. many oil filled types were not much better than waxed paper types.

Testing caps does require an understanding of what they are and one does not test an Electrolytic as you would a Non-polarised type like Waxed paper. The waxed paper types are one I rarely test, this is because I rarely find a good one. Do note, there is one standard here, I get payed to fix them and I only ever want to see it once.

Because of what it is the paper deteriorates & wax goes hard, then there is adsorption of moisture. This all conspires to create a bad cap. With the non polarised, aside from breakages in the hard jacket types, they fail two ways. There can be internal breakages & they short, or loose capacity. More common with wax paper they develop leakage. This is what one uses the Megger for Not Electrolytics. The only info I have re-leakage is from a Lafayette tester. It states that for a coupling cap (Plate to grid 1) 200Meg or less is a fail and 50Meg or less is a fail for a screen bypass cap.

This is one of my beefs: How can you say a cap is ,when there is no evidence to suggest it was tested? An ohmmeter is only a short circuit test. My method is a Megger (That is a trade name) / Insulation tester at rated voltage, working, or the test voltage where specified. My opinion is, that a new cap will not leak,so any NP cap that leaks is a goner. If it leaks I do not bother using a capacitor tester on it to check if its value is correct.

We must remember that a valve grid as applied in most radios, draws insignificant current, it is controlled by voltage: Therefore any leakage from a highly positive plate will drive it more positive and that can impinge detrimentally on bias and the valve.

The electrolytic is different, it is chemical. Its insulator is an oxide coating on its plate and it has to draw a current in order to sustain it. With no use, that oxide layer will be lost & it will present as a short circuit & they are known to explode (as do small oil filled types). This is why one does not power a set that has been sitting to see if it goes. That's more likely to present a big bang! from the short, or fry something (Transformers are not cheap). One does not use a Megger on them.

There have been circuits for reformers in Silicon Chip & I built one on the end of my bench PSU & the ancient VCT tester had a version in it. 225V and 25V. I am not a great believer in "Reforming" very old Electrolytic caps. The reformer often gets used to power a rail with issues, or test & ping those "New" caps that decide to get to a voltage below their rating and suddenly start breaking down & conducting. It also bays to bring up NOS caps that have been sitting a while. The Lafayette book quotes 10mA as a fail point but I would toss any that tried that. The only formula I have for them in Micro amps is 0.01CV where "C" is in mfd & "V" the voltage applied. Reforming is not instantaneous.

The reformer in the PSU (Rodney Champness) is a stepper from 25V to 400V. This uses an LR8 Regulator which has a maximum of 450VDC. One starts low & builds up the voltage to rating. NB as before: Duds can explode... Cage them & protect the eyes minimum.

Note that the NP cap leakage was known about early as the VCT has a leakage tester. It uses the 225VDC in series with a Neon tube (With appropriate series resistance & bighty warning) a "go no go", where the intensity of the Neon glow shows how bad it is. An NP cap should not pass DC.

Do study your multi-meter, some do have a capacitor tester built in. Beware, many modern meters will not handle Valve radio voltages & RF. Price is not your friend.

OK, re testing all new caps before you use them:

Our assembly factory at Caringbah uses many, many thousands of capacitors and resistors in our products, mostly these days installed by robotics.
No-one these days does incoming testing of all components. For critical locations they will be batch tested. All serial numbers and batches are recorded so in the very rare occasion if a part fails in testing or in the field we can track down the source and catch it before it goes too far.

Having said that, the factory did have occasion a while back to wish they had tested a reel of 2500 capacitors before using them. We had a product that stopped working in Melbourne as winter approached. It was due to some X5 spec 0603 SMD ceramics that change value dramatically below 20 degrees C. We would never have known unless we had put the product in a fridge as part of the testing!
Tightened the spec to X7, all good.

You do need to know a bit about the chemistry and physics of your components to avoid some problems!

For service work FORGET ABOUT testing new parts! It's a waste of time! Unless you are using a NOS can electro that's maybe 20 or 30 years old, it won't need to be reformed before you use it!

It's not so hard folks!! Not with valve radios anyway. Provided the voltage rating is above what will be applied and you install electros the right way around all will generally be well.

Marc,
Thanks for your very thorough explanation, As always your extensive experience comes through and helps people like me who are just learning as we go along.

The clear message I am getting is forget trying to test my Electros with my new Megger, I do not have a Cap tester on the either of the two Digital Multimeters I have or the old Analogue meter.

I an still not clear if it safe to put a "New" replacement Electro in a Valve Radio I am repairing without first FORMING it?
From what I guess your saying it is a good idea to do so.

Also in doing so, how can you be sure it is suitable and working to spec?

Thanks Ian, it appears our messages might have crossed as I was writing it. So sorry if it appears I was ignoring it in my last post did not see it when I posted.

I am glad to hear it is not really required, and in fact up to now I had not been testing any new components.

But you know how it is, you get a new testing device like my Megger and you go around looking for different things to do with it.

So thank you all I have learnt something new and hope anyone else who might be crazy like me and reads these post get something from this discussion.

George

So from all that probably the best way to approach valve radios and putting in NEW UNFORMED electros is to power them up gently for the first time. When I make a new piece of gear now, I either power at 1/2 voltage or use a series lamp on the first turn on and MONITOR the caps for any temperature rise with a non contact temperature gauge. That also sorts out any stupid wiring mistakes shorting any thing out!
Once the gear has been powered up and the initial charging current from forming caps has died away then I just switch on full bore and get on with it.
There is nothing wrong with being kind to new parts and "bed them in".
Fred.

Totally agree with Fred and Mark.

Even though my electros are made to order ( only two weeks old when I get them ) , I will always push the voltage up slowly. Just yesterday I switched my Mullard on and made a few adjustments like replacing crunchy insulation and one of the old ( 56 years ) electrolytics started smoking . I didn't think it was the electro at first as I thought I might have buggered up the wiring. But yes even though the set was performing well , the electro decided to go bye bye.

No issues with my new electrolytics with their 650 volt surge capacity though.

There are two things with the Electrolytics.

One is that if its new stock like Carl is saying, that is not an issue. It becomes a possible issue if you order in a heap & don't use them quickly: That's NOS (New Old Stock) I have seen quotes that two years is about it. That also applies to those who buy a spare board for something & it sits for ages: That will impinge on its reliability.

Virtually no modern Electrolytic cap has a "surge", or "peak" voltage written on it, as did the old Ducons (often 525SV or 525PV): That is a trap.

Filament rectifiers & semiconductor diodes, are two things that produce voltage rapidly. When feeding heater valves that will produce a surge EMF approaching in many cases twice that of the voltage that the valves will run on. That surge will be pulled down a bit by voltage dividers (where fitted) and to working volts by the valves, as the heaters cause them to conduct.

It is therefore important to note what is on the original caps. Everything on the "B" rail has to be able to stand that initial surge. With #80 / 5Y3 and a rail of 250VDC on load, I will not use caps below 500V (too many fails, & a lot of 450V ones are crap). As I have noted many times if the rail is at 250V the voltage on the first cap of a choke input filter may be around 300VDC. Which is rather why it is the filter cap with the highest probability of failure.

With modern materials, we have a better chance of eliminating variables & ensure that the material is suitable before it is even let out of the factory, this is what specifications are all about & that has reduced the fail rate substantially. With things like resistors it is actually possible to set up to measure every one individually, when it is on a "rail", & that's all computerised.

On meters & caps even the old AVO 7x can measure some of them, but what is required to do so, is not actually intrinsically safe. I have a couple of meters with the ability to measure caps but there are very few of anything that can handle pF caps accurately, if at all.

I recently acquired a Chinese gadget that can measure transistors inductors caps & such for under $15. It also struggles with pF caps & you will kill it if the cap has charge: Others have proven this.....Oops!

I recently had a decades old Micronta meter die, its had a hard life & plenty of love & Araldite: No complaints. I replaced it with a Digitech QM1549 which has a voltage rating more suited to Valve radio. It has a frequency counter, that I have not tried, but that tells me that it may handle RF which has been known to cause flash overs in some Digital meters: So do be wary. Also cap tester etc. (see their site)

This is probably another meter designed by the art department as, unlike the Micronta, they incorporated the "ON" switch in the range switch. That is extremely inconvenient, as you cannot leave it on a range. I consider that a major oversight. Its one thing to make it look pretty; practicality & functionality is a very different kettle of fish.

One thing I will note is when looking for specific specs on a electrolytic which is a well known brand had a 500 volt electro with only a 500 volt surge capacity. In other words I would rate it as a 450 volt with 500 volt surge capacity . I believe it was MIEC. This was very strange considering most of their values had better surge capacity.

A possible reason why I had 500V one get to 470V & toss in the towel.