How to recognise a failed capacitor

[Arpangel]

Been poking around inside this TLA2120 compressor, one of the capacitor tops has gone a dark copper colour, all other similar ones are the usual silver.
I'm assuming this needs replacing?
Alongside the capacitor is a track labelled "HT" could this be part of the high voltage valve circuitry? in which case I’m reluctant to investigate further.
Normally I would just replace the capacitor, but I'm not confident about discharging it, as it may be high voltage.

[Folderol]

A safe way to do this is with a testmeter set to the 10A current range. Attach a resistor about 10k to the positive probe of your meter and the negative meter lead to ground, then hold the meter probe with the resistor against the hot side of the cap for a few seconds.
If you don't have a meter you can use a plain crock clip lead if you're careful.

[Arpangel]

A safe way to do this is with a testmeter set to the 10A current range. Attach a resistor about 10k to the positive probe of your meter and the negative meter lead to ground, then hold the meter probe with the resistor against the hot side of the cap for a few seconds.
If you don't have a meter you can use a plain crock clip lead if you're careful.

Thank's Will, so shorting it out with a screwdriver wasn’t a good idea then.

[Folderol]

A safe way to do this is with a testmeter set to the 10A current range. Attach a resistor about 10k to the positive probe of your meter and the negative meter lead to ground, then hold the meter probe with the resistor against the hot side of the cap for a few seconds.
If you don't have a meter you can use a plain crock clip lead if you're careful.

Thank's Will, so shorting it out with a screwdriver wasn’t a good idea then.

People often do that, and most times get away with it, but under some circumstances it can make a cap explode, alternatively it might 'just' eat a hole in the screwdriver blade, and send a pulse around the PCB that pops a few more delicate components.

[Arpangel]

I don’t think I’ve got the confidence to deal with this Will, ill just put the lid back on and rock on.

[ef37a]

WTGR, none of the above!
The compressor has the ubiquitous 12AX7 in it and so simply clip a croc lead first to chassis then pin one of the valve (first pin on the left)
The valve will have an anode load resistor of about 100k and thus HT* will safely be drained to zero.

However, the horror stories you hear about capacitors murdering hordes weeks after the gear has been switched off are nonsense. Yes, for maybe a few hours the BIG caps in a valve power amp will hold a lethal charge but not days.

Really big, 1kW say, solid state amps have dangerous amounts of energy stored in filter caps but this will drain naturally at switch off.

*Any competent design should incorporate drain Rs across the main capacitors. That many designers fail to do this shows them to be tight-arsed bastards!

Dave.

[alexis]

This thread is scary as all get out.

The thought never occurred to me that counting to ten (or twenty at most) before opening a piece of equipment may not be sufficient to prevent death.

[James Perrett]

This thread is scary as all get out.

The thought never occurred to me that counting to ten (or twenty at most) before opening a piece of equipment may not be sufficient to prevent death.

Whenever I open a piece of gear I don't go touching any terminals/tracks/conductive parts until I'm sure that everything is discharged. There's no need to wait if you are methodical, careful and know what you are doing.

However, if all your knowledge comes from the internet or YouTube then I'd stop, assume everything you've learned is wrong, and send the gear to someone who really knows what they are doing. While there are certain great Youtubers, they are in the minority and you really need to be pretty knowledgeable already to work out who to believe and who to ignore.

[Hugh Robjohns]

The thought never occurred to me that counting to ten (or twenty at most) before opening a piece of equipment may not be sufficient to prevent death.

I think it's fair to say that if the thought never occurred, you really shouldn't be opening equipment up at all!

[James Perrett]

Been poking around inside this TLA2120 compressor, one of the capacitor tops has gone a dark copper colour, all other similar ones are the usual silver.
I'm assuming this needs replacing?

Can you put up a picture somewhere? It is possible that they've just used a different type of capacitor in that position. Bad capacitors tend to have slightly domed tops or signs of leaking electrolyte.

[alexis]

The thought never occurred to me that counting to ten (or twenty at most) before opening a piece of equipment may not be sufficient to prevent death.

I think it's fair to say that if the thought never occurred, you really shouldn't be opening equipment up at all!

Oh, I don't!

I just the other day worked up the courage to unscrew a mic capsule for the first time. As I really didn't come away much smarter, I'll probably never do that again, especially after reading this thread!

[ef37a]

Been poking around inside this TLA2120 compressor, one of the capacitor tops has gone a dark copper colour, all other similar ones are the usual silver.
I'm assuming this needs replacing?

Can you put up a picture somewhere? It is possible that they've just used a different type of capacitor in that position. Bad capacitors tend to have slightly domed tops or signs of leaking electrolyte.

My thought as well James, could be a different temperature range? A Tantalum for a special, low leakage circuit? Piccy please.

Now one never wants to be in any way dismissal of any safety advice but as one who has dived into countless hundreds of valve chassis* the idea that these capacitors are always there waiting to kill you is a bit, well, daft! (bit like ribbons and phantom power. Any solid evidence?)

Even if you had a say 50uF cap charged to 300V and you 'copped' it. Yes! It would bloody hurt! But a healthy person would survive. Now, Tony and I are FAR from healthy so DO take the greatest care mate!

What really annoys me is that the whole "lethal caps" idea should never have started? If people had just fitted a tupenny-ha'penny drain resistor in everything as a matter of course the "myth" would never have arisen.

*I was never blase' but you do get used to stuff. The things that made me VERY nervous however were microwave PSUs. ~3kV with an amp or so behind it. That takes NO prisoners.

Dave.

[Hugh Robjohns]

Now one never wants to be in any way dismissal of any safety advice...

...he says, dismissing it...

They're not 'waiting'. That implies they have somewhere else to go.

The point is, there is a serious risk of harm to the unwary and ill-informed. Not something to belittle, especially in print!

...bit like ribbons and phantom power. Any solid evidence?

There certainly is for vintage ribbons with centre-tapped transformers.

[Folderol]

Something that is very frequently overlooked about electric shocks is that even a comparatively 'mild' one can result in significant injury and damage - simply due to the persons involuntary reaction. I still have a scar on my wrist from the mid 1960s due to a sharp metal corner that I caught in the back of an ancient telly while trying to adjust the scan coils and getting a bite. The telly was also fixed... permanently.

[Martin Walker]

And here's a reminder that 'bad' capacitors are in no way a new phenomenon - here I am discussing them way back in my February 2008 PC Notes column

https://www.soundonsound.com/techniques/pc-notes-0

And I also agree with James, that bulging tops and leaking electrolyte are the two main symptoms to look out for - a different coloured top cap CAN simply mean the manufacturer used a different supplier for a particular cap value.

Martin

[ef37a]

Now one never wants to be in any way dismissal of any safety advice...

...he says, dismissing it...

They're not 'waiting'. That implies they have somewhere else to go.

The point is, there is a serious risk of harm to the unwary and ill-informed. Not something to belittle, especially in print!

...bit like ribbons and phantom power. Any solid evidence?

There certainly is for vintage ribbons with centre-tapped transformers.

Well I tried very hard Will NOT to make light of the possible dangers but obviously failed. My point is that yes, capacitors hold a charge but not forever (well, CRTs do, for months and I do know of an incident with one of those)

I did explain a procedure that would discharge the caps in valve kit in a way that is safe to both the person and the equipment. But, yes again, if you are not sure what you are doing keep out of high voltage equipment even when it is off.

And come off it Hugh! I have read several times where you have explained that phantom power is nowhere near the destructive force it is often made out to be. You have recounted more than once that BBC mic XLR panels had 48V on them all the time and Coles ribbons never suffered. Very old mics are a special case and of course, cable faults do happen. Still no evidence.

The TV that scarred you Will, was it perchance a 17" Ecko? They had a metal clip top centre of the back rail to earth* the foil on the cardboard back...JUST where your hand would go if you copped a packet!

*That was back when manufacturers GAF about RFI emissions!

Dave.

[Nazard]

In this month's PE mag, one of the regular writers was repairing a friends SMPSU that charged an electric motorbike battery. He said he might have been 'complacent' as it had been a long time since he had had an electric shock.

But he managed to discharge 350V across both arms from a capacitor that resulted in the metal handled Swann Morton scalpel he was using being thrown and embedded into the wooden skirting board. Next, in discharging the main paralleled capacitors, by shorting across the terminals, he blew the tip off his Bahco Ergo pliers.

Anyway, he did leave a useful list of 6 tips to follow...............

[Folderol]

Could be. Not sure, it was quite an old one (even then) very square. I was trying to get the scan coils in the right position for a horizontal image, and looking at a mirror (as usual). I don't know exactly what it was that bit me, but is sliced across the back on my wrist. The telly then committed suicide by sliding off the (somewhat rickety) bench and landing face down.

[Tomás Mulcahy]

And I also agree with James, that bulging tops and leaking electrolyte are the two main symptoms to look out for - a different coloured top cap CAN simply mean the manufacturer used a different supplier for a particular cap value.

I also agree. In addition a decent capacitor tester can be bought for €5. It is surprising how long they can last. Heat is the main killer. Despite that, I tested the PSU in a JX8P recently. None of the caps were bad and all voltages were in spec if a tad noisy. I replaced the caps anyyway since I was already replacing the original under specified rectifiers and adding heatsinks. Similarly, a c.1980 Yamaha CP80 had only one bad cap out of around 40 (if you count the ones in the pickup).

But he managed to discharge 350V across both arms... Next, in discharging the main paralleled capacitors, by shorting across the terminals...

Not once but twice? That is a fool.

[Arpangel]

I’ll try and get a picture of this cap, but don’t worry, I'm not going anywhere near anything that I’m not sure about. In fact, as I said, I’ll leave it to someone who knows what they are doing, who wants to take it on and have a go at it.
I'm shocked at how little information there is about this compressor, it's almost unheard of by most people, and almost nothing on the net.
I rang Tony Larking Audio and they didn’t seem very interested.

[Hugh Robjohns]

It seems your world view is that if it hasn't happened to you, or Doug Self didn't write a book about it, it doesn't exist!

And come off it Hugh! I have read several times where you have explained that phantom power is nowhere near the destructive force it is often made out to be.

This is true.

You have recounted more than once that BBC mic XLR panels had 48V on them all the time and Coles ribbons never suffered.

This is also true.

Very old mics are a special case and of course, cable faults do happen. Still no evidence.

You wanted evidence of ribbons being destroyed. You didn't say only modern ribbons connected with good cables...

Modern ribbons are reliably robust when it comes to phantom. Vintage ones and duff cables can cause damage. Most ribbons are damaged through misuse and thoughtless placement.

... and none of this is relevant to the potential dangers of uneducated rummaging around inside equipment — especially valve equipment.

[ef37a]

I think if I had replied to you in that way Hugh you would have at the very least sent me a stern email.

I did not say stored charge in capacitors do not pose a risk to the unwary.

Yes, in my experience people don't get shocked by remnant capacitor charge and in over ten years reading several electronics based forums I am yet to read of anyone who has. None of that means people should ignore the warnings.

Dave.

[Arpangel]

It seems your world view is that if it hasn't happened to you, or Doug Self didn't write a book about it, it doesn't exist!

If only you knew how much I wish that statement were true, being exactly the opposite has been the cause of a lot of problems in my life.



This compressor, it's all got a bit out of hand, it does actually "work" but I'm paranoid it’s going to deteriorate soon, discoloured caps, and weird things happening now and again, like the left channel level changes when I alter the gain on the right channel? the left goes down, the right goes up, although the stereo field seems in-tact when they are linked.
It's all a bit weird, and it’s a shame, as just passing audio through it has a nice effect on everything.

[ajay_m]

Microwave oven transformers are lethal. A bunch of people have gotten into the Darwin Awards by (mis)using them to burn patterns into wood. 33 people dead so far and counting.
https://en.wikipedia.org/wiki/Fractal_burning
It is pretty much the only appliance that, if it develops a fault, I would just replace, without any attempt to open it up and look at what's wrong. Just too dangerous.

Most people (and I am one of them) who ever accidentally got themselves across 230V and lived, end up with a very, very, healthy respect for high voltages. If you are working on mains-powered equipment, using an isolating transformer ensures that if you do accidentally touch a live wire, you should be ok, assuming there's no alternative ground path for the neutral side. Otherwise knowing how to use a good CAT 4 rated DVM safely is essential. Note that a lot of Chinese DVMs do not meet safety standards. In the event you accidentally do something silly like put it in 'amps' mode across 230V, the result could be catastrophic, as the protective fuses won't be genuine and may even explode.

At the very least, those little neon screwdrivers are a very quick and handy way of determining whether that wire is carrying a potentially lethal voltage, assuming that there's a ground reference of course.

You can get standards-compliant DVMs from reputable suppliers like Mouser or RS for not a lot of money. But repairing anything where there are high voltages not just on the mains side but on the circuit side is something that in my opinion should only be undertaken if you are confident you have the skills and training not to put yourself in danger.

[Hugh Robjohns]

But repairing anything where there are high voltages not just on the mains side but on the circuit side is something that in my opinion should only be undertaken if you are confident you have the skills and training not to put yourself in danger.