ADVISORY: Radial caps and electrocution danger

[xtian]

Friends, I have been learning about and building high voltage circuits for nearly 10 years (thanks in large part to you all!). An important part of what we learn is SAFETY. But somehow I never read or heard any warning about the aluminum bodies of radial filter caps.

So, here's the shocker (literally): the aluminum case of electrolytic capacitors is not isolated from the circuit.

Normally, this does not present a hazard, because the caps' negative leads are connected to ground, thus to the chassis, thus to earth, and there's no danger to you if you touch the cap. (The exception is in reverse-biased bias circuits, where the positive lead is connected to ground).

However, when you use PAIRS of radial caps in series (very common for B+ reservoirs), the B+ voltage is split evenly across the two caps, and half the voltage is present on the body of the cap furthest from ground.

This information (that the cap body is not isolated) does not appear in typical data sheets. I did find mention of this in a Nichicon document: http://www.nichicon.co.jp/english/produ ... uminum.pdf

2.1.1(7) Aluminum electrolytic capacitors must be electrically isolated as follows:
The aluminum case and the cathode foil are connected by the unstable resistance of a naturally formed oxide layer inside the aluminum case and the electrolyte.

2.1.1(8)Outer sleeve of the capacitor is not guaranteed as an electrical insulator. Do not use a standard sleeve on a capacitor in applications that require the electrical insulation. When the application requires special insulation, please contact our sales office for details.

So, any time you see something like this, with pairs of caps with exposed aluminum bodies...BEWARE!

[martin manning]

A useful PSA, I suppose, since stacked filters are becoming more common. The same is true of multi-cap cans.

CDE:
Electric Shock Hazard
Aluminum electrolytic capacitors rated above 35 VDC may con- tain lethal energy at lethal voltage, and should be handled as such, ensuring that the capacitors are discharged before han- dling. Be aware that the metal case will usually be approximately at the same potential as the negative terminal, so an electric shock can occur even when only one terminal is touched, if the case is also contacted. Capacitors can build voltage if they are not kept discharged. See the subject of Dielectric Absorption (DA) in a previous section of this document.

ELNA:
10. Insulation between the capacitor case and thecathode terminal.
• The capacitor case and the cathode terminal are connected through the electrolyte which has uncertain resistance. If a complete insulation of the case is necessary, add an insulator at assembly.

Rubycon:
6)Insulation
・Electrically isolate Aluminum Electrolytic Capacitors as follows:
1Among aluminum can, negative terminal, positive terminal and circuit pattern.
2Among auxiliary terminals on snap-in type, positive terminal, negative terminal and circuit pattern.
・The sleeve of Aluminum Electrolytic Capacitor is not recognized as an insulator, so that standard capacitors should not be used to a circuit requiring electric insulation. Please consult us for capacitors with insulating sleeve.

Nichicon:
(7) Aluminum electrolytic capacitors must be electrically isolated as follows:
The aluminum case and the cathode foil are connected by the unstable resistance of a naturally formed oxide layer inside the aluminum case and the electrolyte.
q(a) Case and negative terminal (except axial leaded part such as JIS configuration 02 type)
(b) Case and positive terminal
(c) Case and circuit pattern
w(a) Auxialiary terminal of can type such as JIS style
symbol 693, 694 or 695 and negative and
positive terminal, including the circuit pattern. eCase and both terminals of a bi-polarized capacitor.

[tubeswell]

Admins, please stickie this thread. TIA

[Synchu]

Good and important points. I found this (fortunately) during my pedal building days, burning various parts a circuit (few times) by shorting a radial cap to the enclosure...

Niki

[R.G.]

Good thing to become aware of ... again.

It was once taken for granted that the outer aluminum can was tied to the insides of the cap, except in specially constructed caps. It was once also common for there to be a plastic disc covering the top/open end of the cap to prevent accidental shocks by touching the uncovered part of the top. I suspect that stopped when hair-spray-can caps could no longer put vents on the top/terminal side, and instead the tops were expected to rip to vent the cap if needed.

Yep - the cans are hot, and as far as I know, always were.

[pompeiisneaks]

Agreed on the sticky and done.

Great info, I didn't realize this either. Glad you just got a shock and nothing else xtian.

~Phil

[demonstratedspatulas]

I have actually been shocked by this, and not even while I was working on my amp, just while playing it.

Every time I work on an amp, I am extremely careful, but I never really used caution when playing through an amp because I never expected something like this to happen.

This was about 10 years ago, first amp I ever built. I built an amp that used caps in series for the plate and screen reservoir caps. (2x 350VDC caps in series to reach 700VDC rating) I was using cap clamps as it seemed like a good mounting solution, especially for a point to point amp.

The amp was plugged in and I was playing through it, the chassis was sitting on my desk in the user friendly orientation, the guts weren't exposed. I reached over to adjust a control, and my thumb grazed the exposed foil on the top of one of the caps. My left hand was on the fretboard. 250 VDC went through my heart. The shock lasted maybe 1/4 of a second. It was painful for sure but not excruciating. I set down my guitar and couldn't believe what had just happened, holy shit, that went through my heart. In the first minute or so I didn't even realize what had shocked me. It soon dawned on me that the foil on the top of one of the caps was sitting at 250 VDC.

I feel lucky to be alive.

[LightningPhil]

Good point. Not considered it as its not a thing in any of my amps. But good to be cautious of.

Once while working on a pulsed power circuit, contentedly started to tweak it before realising it was turned on. Had 1kJ at 800V. Very carefully extracted hand from chassis, turned it off and went for a coffee. Quite happy I survived.

So a few thoughts. Good to make sure things are off before inserting hands. On new amps, there should be little need to use series. Is anyone doing this? And lastly, note to self. Label inobvious hazards.

[xtian]

On new amps, there should be little need to use series. Is anyone doing this?

Oh, yes, you'll find many examples of modern amps with series caps, stacked to make higher voltage tolerance, including our "precision power supply" boards for the later ODS models.

[LightningPhil]

But, but, shudder...

Electrolitic caps are inaccurate, leaky, explody beasts.

While I've ended up using up to 12 in series, that was because they were free. Also, it wasn't 12 single caps in series. More 24 trays of 68 large capacitors for each of the 12 layers. It was an ancient physics experiment PSU and took 3 weekend's to reform them. 50 something of them blew but as individuality fused, all was ok. It's those memories making me twitch about series capacitors.

There's no need now as high voltage caps are available. Electrolitic to 800V and metalised film for higher.

[xtian]

There's no need now as high voltage caps are available. Electrolitic to 800V and metalised film for higher.

Maybe? But cost is pretty important to me. As an example, say you need a ~100uF reservoir for a PSU that surges to 600v on startup. Here I can get a pair of 220uF/350v caps for under USD$10:

https://www.mouser.com/ProductDetail/Ni ... J6SHwvo%3D

What you got?

[xtian]

Actually, this 100uF/800v film cap is pretty nice, and only $19.14:

https://www.mouser.com/ProductDetail/KE ... 2BZQ%3D%3D

[LightningPhil]

Nice find!

[Murrayatuptown]

I've always been aware & spooked by exposed electrolytic tops, because I worked on radio gear (Heathkit, etc.) first, and series electrolytics were common as a way to avoid & pricey large oil-filled capacitors. So I always put heatshrink or tape over them...in case I forgot, or someone in a future time didn't know.

Now I'm obsessed with using metallized polypropylene when I can fit them. In perspective, it's easier than the oil-filled rectangular drawn can types. People would give them to you when I was a teen...scarce now.

Murray

[Murrayatuptown]

Newark/Farnell has been offering promotional prices on select Kemet C4AQ metallized PP for at least a year...they rotate which ones get the special pricing.
Today I found 60 uF 650 or 800 V (I forget-the other was 55 uF at the opposite voltage.) $5.04 for one and $5.09 for the other. Yes, they are relatively large.

They are generations better than electrolytics in many parameters.

If you give yourself a minimum of 20% overvoltage headroom, you don't have to derate 105 C to 85 C. The internal losses are so low at 50-120 Hz the lifespan at 105 C is probably at least 100000 hours.
Due to the extremely low losses, hot spot temperatures internally will likely never get anywhere near 105C. Dissipation factor on the order of 0.1% vs. 20%.
For all practical purposes they should outlast the lifetime of your amp and your generation. No drying out, no reforming.

Just BIG. Not expensive anymore thanks to economy of scale from HV inverters in automotive industry.

Kemet bought Nissei-Denki, who bought Arcotronics. Decades of experience optimizing power film capacitors.