Power Supply Capacitors

[Neil Purling]

What would be the advantage of using these Paper-In-Oil capacitors in the PSU of a home brew radio rather than electrolytic types?

[turretslug]

Fifty or more years ago, one might have said better temperature range, longevity at higher temperatures, possibly ESR (though not often a consideration in their likely application).

Nowadays, electrolytics are much better than they were, a significant development motive being SMPSU use where compactness, low HF impedance and good life at raised temperature are important.

With those capacitors, disadvantages compared to current electrolytics could be size, low leakage could mean long charge retention, depending on surrounding circuit, to surprise the unwary possibly many days later! Also, they are only rated at 300V, even something typically requiring a modest 220-250V HT could have this rise to 350-400V during warmup allowing for peak + regulation.

Any more details on the application?

[M0FYA Andy]

They'll probably work for ever, whereas electrolytics do have a finite life, often surprisingly short according to their spec, although doubtless much better in practice.
It's the first time I've heard anyone regard low leakage as a disadvantage! But a good power supply design will have a bleeder-resistor across the output anyway.

Andy

[turretslug]

Originally Posted by M0FYA Andy

It's the first time I've heard anyone regard low leakage as a disadvantage!

That's what risk-assessment culture does to one- if only so that it's possible to breath a sigh of relief and mutter, "Told you so...."!

Generally, I find that if you do a quick bit of factor-weighting arithmetic on quoted lifetime hours for electrolytics, the expected hours become gratifyingly greatly extended. Again, it'll be the CYA culture, the worryingly short lifetimes often quoted may be maximum ripple at maximum temperature or otherwise pushing things. It's a bit like finding "Side effects may include death" on a packet of aspirin.

[M0FYA Andy]

Perhaps you come from a similar background to mine, I worked on safety-critical aircraft flight controls throughout my career, so for any design I always concentrate on 'how will it fail?' - making it work correctly is the easy bit!

As we know, in consumer electronics components are usually operated at maximum ripple at maximum temperature.....

Andy

[Neil Purling]

The role was to be a TRF radio & I was going to use a transformer from a Pye P75 that only produces smoothed HT of 175V. I was concerned about RF getting into places it ought not to be and introducing instability.

What value do you suggest for a bleed resistor?

[Argus25]

Use something like 500k to 1Meg ohm and be aware, if there is no other source of current drawn from the HT once the tube heaters/filaments go off, it will take a minute for them to discharge to a safe level.Your circuit though might have some other bleed resistance which could help. Typically that could be a voltage divider on a screen grid, but you might be using triodes.

I have used many of these Russian PIO capacitors, they are excellent quality and very good value for money. They are superior in every respect to an electrolytic capacitor.The only reasons they would have not been chosen over an electrolytic in a commercially made apparatus is the cost and the physical size would in many cases prohibit it. If those parameters are not an issue, then the PIO capacitor is a much better electronic part than an electrolytic cap, by a country mile.

Also, if you were concerned about RF across the supply, with electros you would have to bypass them with 0.1uF caps, with these you won't have to bother.

[M0FYA Andy]

I think I'd have suggested something a bit less, maybe 100 - 200k, but it's not at all critical.

[kalee20]

If you're worried about RF getting where it shouldn't via the power supply, local decoupling with 0.1μF and 1kΩ, positioned right AT the relevant stage, would be better than a single capacitor, even if perfect, but connected with wires several inches long!

[turretslug]

Possibly I sounded a tads negative in post 2- with just 175V HT, by all means use them if there's space, as well as everything else mentioned it'll add a bit of early radio authenticity!

[ms660]

Time constant/voltage table in here:

http://www.electronics-tutorials.ws/rc/rc_2.html

Lawrence.

[Diabolical Artificer]

A little bit more sophisticated bleeder R calculator here - http://www.engineersedge.com/instrum...ator_12909.htm I usually use 100k.

Andy.

[Radio Wrangler]

Bleeders are high value resistors and many types drift high or go open,

so never never rely on the bleeders doing their job. Always check twice!

David

[ms660]

Good old AVO....

Lawrence.

[M0FYA Andy]

Originally Posted by Radio Wrangler

Bleeders are high value resistors and many types drift high or go open,

so never never rely on the bleeders doing their job. Always check twice!

David

Put two or three in parallel if you're feeling really cautious!

[julie_m]

How about adding a high-brightness LED in series with the bleed resistor? Some of the real eyeball-stabbing blue or white ones will still show down to a few tens of microamps, so it won't extiguish fully until the voltage is down to a safe level. It needn't normally be visible except when the set is being worked on, if you are concerned about appearance.

[M0FYA Andy]

That could lull you into a false sense of security if the resistor went o/c.......
I guess the procedure would be check the LED is lit with the set on before switching off and confirm it extinguishes after x seconds.

[Radio Wrangler]

Some people did the same thing with little hivac neon bulbs, but as said, you have to watch them go out to know things are working.

A big solid earthing cable and clip is always reassuring. The neon/led just tells you that you aren't going to get molten copper in your face when you put it on. This seems picky on lower HT, no-one criticises this level of care with 2kV HT on a big amplifier, but the value comes in at any voltage enough to hurt you. I was working on such an amplifier and had the EHT clipped, but I got a flash off a 250v screen supply with no bleeder that I'd forgotten about.

For years I had some 4kV 32uF photoflash PIO capacitors in the attic. The terminals were kept wired. Air currents can charge these things over years and leave them with a very nasty surprise.

David

[Argus25]

There is no perfect value bleeder, I tend not to want to waste energy in them, if forced to use them, so I pick a higher range value like 500k rather than 100k but the discharge time is longer.

It pays to use a physically larger part than required, even though a small 1/4w calculates as OK, use a part such as a 2w metal film resistor. These are far less likely to go open circuit and modern 2w resistors are only the size of 1/4 w ones from the past and will easily fit between the capacitor's solder lugs.

[Argus25]

I have attached a photo of an amp I built a while back with some similar PIO capacitors, on the left you can see a 1Meg 2W metal film resistor used as a bleeder resistor on the power supply filter cap, and its relative size.

For many of my restorations now I've come to use the 2W metal film type to replace the old 1/4 and 1/2W carbon composition resistors. Because they will probably last forever in the under-stressed application and because of the physical size being similar to the original lower power rated types, they give a similar look under the chassis. Modern 1/4W resistors just don't look right under the chassis for valve gear.