When is a part “dead”….(length of string question)

[The Philpott]

- Obviously add into the mix how inaccessible a mediocre but functional 'vintage' component is, should it completely fail in the near or middle future (and the cost of the replacement...which isn't usually much)
Dave

[HoverJohn]

So my cheap megger arrived what’s a good value for the caps to come in at I’m getting a range of 8Mohms down to 100K and one seems to die (156kohms)at 300v (350 rated) but it’s reading 2meg down near it’s actual operating voltage??
I have no idea what is acceptable as “ok”

[Ed_Dinning]

Hi John, to some extent it will depend on the circuit it is in, if you can "get away with it"

For example a 5meg leak at working voltage on a cap in a screen grid decoupler fed by a 47K resistor from HT will behave perfectly well, but a similar leak on an AVC decoupler fed from a 10meg resistor will greatly reduce the sensitivity of the set.

That said, any leakage will only worsen, so better replace with a modern part.
I would be suspicious of any cap, when tested at working voltage that gives more than the slightest twitch of the megger needle

Ed

[HoverJohn]

What sort of leakage did they have to start off then??

[Uncle Bulgaria]

As Ed said, it's what the circuit can get away with, and as I mentioned earlier they were quite tolerant of parts as the designers knew the limitations of the technology and rarely specified tightly matched components.

Here's the table I use with my capacitor tester, re-drawn from the Everyday Practical Electronics project front plate. You'll see there's a range of leakage acceptability depending on voltage rating and capacitance value.

As I test more, I am getting more of a feel for what's acceptable.

[Radio Wrangler]

You'll find allowable leakage specs in the datasheets of reputable capacitor makers. I suppose it's one of the ways they became reputable...

David

[HoverJohn]

Ah that chart is magic. I will hook up an ammeter instead of volt meter!!

[Ed_Dinning]

Hi John, you will need both a voltmeter and an ammeter.
The ammeter needs to be quite sensitive and multi range if you want to detect 10's of megohms at 100v or less; you also need to protect it (a fuse would be useless) from a sudden change of leakage or a breakdown as you increase the voltage.

Ed

[Station X]

I measure leakage current by measuring the voltage across a resistor connected in series with the power feed to the capacitor. If a 10k resistor is used each volt measured across it represents 10mA of leakage current.

[ms660]

Originally Posted by HoverJohn

So my cheap megger arrived what’s a good value for the caps to come in at I’m getting a range of 8Mohms down to 100K and one seems to die (156kohms)at 300v (350 rated) but it’s reading 2meg down near it’s actual operating voltage??
I have no idea what is acceptable as “ok”

I've posted this link before somewhere on the forum some years ago, anyways read and digest, from an old time manufacturer in 1935, includes simple formula's, describes effects, shows some simple practical calculations and example circuits including converting leakage resistance in the form of a shunt resistance to an effective series resistance at a particular frequency (Xc squared/Rsh) etc etc:

https://worldradiohistory.com/Archiv...ox-1935-03.pdf

Lawrence.

[Uncle Bulgaria]

Lawrence, that's a brilliant resource. I've saved that to my reference folder, thank you!

[ms660]

No problem.

Lawrence.

[HoverJohn]

Well I think that seeing as my rubbish multimeter is picking up milliamperes some of its fairly bolloxed… (with one terminal of the cap disconnected to avoid it finding another route!!

[HoverJohn]

Also that document is useful as I had no ball park for how many Mohms I need !!

[Station X]

Originally Posted by HoverJohn

Well I think that seeing as my rubbish multimeter is picking up milliamperes some of its fairly bolloxed… (with one terminal of the cap disconnected to avoid it finding another route!!

Depends on how many mA you're measuring. For leakage tests on electrolytics I allow a max leakage current of 1mA per 30uF. Most electrolytics are better than this.

[HoverJohn]

Well the caps are showing resistances in. The 10s of Kohms and about 70ma on a 16uf cap the rest are not much better.
How long would a big cap take to charge to voltage??

[Station X]

10's of kohms is far too low, but you need to wait and see if the reading increases. Initially a capacitor looks like a short circuit and inrush current will be high, equivalent to a low resistance. As the cap charges the resistance should increase into the Mohm region. This effect is very noticeable when measuring the resistance of a cap with a meter. The reading should start low and progressively increase towards infinity. However unless the reading is very low the test is invalid as caps should be tested for leakage at their full working voltage. what you actually want to measure is insulation resistance which requires a high voltage such as that from a Megger.

If a 16uF capacitor is drawing 70mA at the working voltage it won't be long before it overheats and explodes!

How long a cap takes to charge is not something I've ever considered, it depends on its capacitance, the applied voltage and any series resistance. If it's going to charge it'll probably do so within about 10 seconds, often a much shorter time. Of more interest is the time it takes to discharge. If you connect a voltmeter to a good electrolytic it may take several minutes for the reading to drop to zero volts.

If you're endeavouring to reform a capacitor you need to limit the reforming current with a series resistor.

It looks to me like you have a whole load of duff capacitors and your best bet is to replace or restuff them.

[Hartley118]

Originally Posted by HoverJohn

What sort of leakage did they have to start off then??

I recall buying brand new wax capacitors in the 1950s, some of which were already leaky enough to be unsatisfactory as anode-grid coupling capacitors because of leakage as low as a few megohms . By contrast the now-much-derided but then innovative Hunts 'Moldseal' capacitors were a leak-free revelation. At what point in time they became troublesome, I don't know, but I still have well-preserved examples which are OK. I'm sure that their insulation depends on how well the 'Mold' is 'Sealed'. I guess that the working environment may cause micro-cracks to develop in the rather brittle compound: often these are all too visible and may cause the coating to fail. Temperature changes and storage in a damp environment will then draw in water vapour with inevitable, probably irreversible, consequences.

Rather surprisingly, I do have examples of Smiths Radiomobile car radios where the wax capacitors from the 1950s are still perfectly good as coupling capacitors. So some wax capacitors are good enough to survive one of the worst ever temperature cycling environments - automotive!

Martin

[HoverJohn]

Right I’ve ordered everything HV cap wise and the degraded resistors too. I can’t find the right size cans so the 60uf and 16uf Will just be axials inside the chassis and the 100uf will be a 50+50 can doubled up with a packing ring.

I will need to fashion some form of cap mounting under the chassis.
All grid couplers ordered too god knows how it didn’t go bang when I turned it on.

In other news I’ve hooked up the 80s legato linear to the speakers and it knocks the socks off the Technics…

And I’ve even found the 2 halves of the broken selector knob in my grandfathers desk where he must have stashed it in the late 70’s.

Glad I bought the megger - I was going to reuse some right tripe.

Next to look at the gram pickup.

Thanks everyone for help so far it’s great ��

[Hartley118]

Originally Posted by HoverJohn

I will need to fashion some form of cap mounting under the chassis........


Thanks everyone for help so far it’s great ��

I've used ceramic standoffs successfully for mounting replacement capacitors where the originals were chassis mounted. I bought them from an eBay supplier. Fortunately, modern replacement caps are very much smaller than the vintage originals.

The picture shows a in-progress replacement job in my Pye T10 radio which has an impressive array of chassis-mounted electrolytics which I left in place as a support for an earth busbar carrying a row of insulated standoffs which in turn carry the replacement electrolytics and their associated wiring connections.

Martin