Repair Safety First

[Dave Curtis, dB AudioTech]

Originally Posted by mykey:
"the tendency of a capacitor to charge itself is referred to as the chemical charge. the capacitor will slowly charge itself up, even after sitting unused for years, and even when there are discharge resistors across it."

Originally Posted by R. G. :
"Actually, if you think about it for a minute, this is impossible. If capacitors could recharge themselves to any significant extent with appropriately sized discharge resistors across them, it would in effect constitute a source of neverending electricity."

That's mykey for you. Ask him about aluminum increasing resistance with age, or how a 10,000uF cap on a tube rectifier is a good thing. Massive grain of salt required. -DC

[Rob Mercure]

Mykey,

While I respect the time you've put in repairing amps - I've got around 37 years in and Enzo, Bruce, Ray and other have much more time while Steve and some others are "genuine" Phd Engineers - you somehow don't seem to consider your posts/responses before your send them.

Do you understand how a Ground Fault Circuit Interrupter works? It really doesn't give a rodent's rectum whether there even is a ground connection - what it does is compare the current in the hot conductor to the current in the neutral conductor and if they differer by more than a few mAs then the circuit "assumes" that the difference is flowing to ground - hopefully not through a person - and trips! So one can easily get their hand across the secondary of a PT and have nice brown smoke coming out of their ears while the hot and neutral currents remain the same - even if some of that current is leaking into the chassis ground if it's being drawn equally through the PT primary you're fried. And, of course, an isolation tranny makes the situation worse as the primary and neutral feeds to the iso will balance even if the secondary currents aren't doing so. Oh, and this also means that a GFCI protects a two wire - "double insulated" tool as well as the older metal cased ones.

Again, I'll credit you with your long experience in modifying amps - but some of use have run/are running out own businesses and service shops and can bring much practical experience to the table. I hope you can work with us to add to those experiences - ain't nobody on this forum, as best I can tell, got anything to prove. We've each got our strengths and weaknesses

Steve: Remember that you can never make anything "idiot proof" enough to stop a sufficiently determined idiot <grin>.

Rob

[mykey]

Originally Posted by R. G. :
"Actually, if you think about it for a minute, this is impossible. If capacitors could recharge themselves to any significant extent with appropriately sized discharge resistors across them, it would in effect constitute a source of neverending electricity."

capacitors definitely recharge themselves, even after years of standing idle. and the process is very slow, not a very good source of everlasting electricity. the point is that you can be shocked by one even when the power supply has been unplugged for a very long time, even when there are resistors across it, and plenty have been shocked this way.
coincidently I think in many cases a 10,000 mfd capacitor in a power supply is a great idea, and so do several other amp builders. perhaps you are referring to the minimum recommended capacitive filter to use with a rectifier tube. this does not rule out using a much larger capacitor. some would prefer eliminating the 60 cycle noise in amplifiers, reducing ripple in the power supply to the most minimum. or when building a DC heater supply for a high gain tube amp 10,000 is about the LEAST I use. just because you don't pay attention to 60 cycle noise does not mean that people who do care about reducing noise are stupid.
this also means that Jim Macintosh, who used 4700-10,000 mfd caps in his tube amps is also stupid, and that other manufacturers such as Levinson, Yamaha, Krull, Crown are all idiots. After all who would waste the money to filter a power supply enough to remove the 60 cycle noise? what idiots! Just think -60 cycle buzzzz in your amp does not matter anymore, because Dave Curtis said so.
And then while your at it, call Sprauge, Cornell Dubleer, and other capacitor makers, and tell them to stop making these wasteful unnecessary capacitors, because Dave said we don't need them anymore!
Reality is quality low noise amps have very big capacitors for very low noise, and if we want to build them that way, we do not let thick head know it alls tell us different. We want low noise and we will build for low noise.
We will continue to use DC heaters (like Macintosh did) and 10,000 mfd capacitors not because we are stupid, but because we are more intelligent, and we demand a higher standard than PV repairmen.

[Bruce / Mission Amps]

OH brother...

Mykey:
"...you can be shocked by one even when the power supply has been unplugged for a very long time, even when there are resistors across it, and plenty have been shocked this way."

I'd like to see some documentation of anyone getting shocked from a filter cap that mysteriously recharged itself... that is pure bunk.

Mykey:
"perhaps you are referring to the minimum recommended capacitive filter to use with a rectifier tube. this does not rule out using a much larger capacitor.
I think in many cases a 10,000 mfd capacitor in a power supply is a great idea, and so do several other amp builders."

Jeeze Louize... Not with tube rectifiers and a standard tube amp setup. Of course it rules out using absurdly high value filter caps.
And what is this "minimum recommended capacitive filter to use with a rectifier tube" statement mean?
Have you recently read a tube manual with respect to a tube rectifier and a capacitive filter vs inductive?


And speaking of the rectified B+ rail.... what 60Hz noise are you raving about?
You need to recheck out how FW or FWB rectifiers work.
By the way, the AC filament supply runs at 60Hz... but in 99% of other DC supplies, like the high voltage B+ rail from that rectifier tube we were discussing, ... it does not.

I think most of this is basic electronics 101 but maybe you are so far beyond most of the builders and home brewers here you've forgotten much of this.
Nah!!

[mykey]

I'd like to see some documentation of anyone getting shocked from a filter cap that mysteriously recharged itself... that is pure bunk.

the chemical charge of batteries and capacitors is a fact.
there is no mystery to it, the construction and dielectric chemicals in the capacitor will cause a charge to be accumulated, slowly.
a car battery is shipped dry. when installed the acid is added to it, no charging required, the battery will come to full voltage without charging. this is also an effect of the chemical charge.

[R.G.]

capacitors definitely recharge themselves, even after years of standing idle. and the process is very slow, not a very good source of everlasting electricity.

Sir: Caps don't charge themselves. They are an energy storage device. They can only put out what was put into them at one time or another. If they continued to leak out electricity more than we had put in, we'd call them batteries.

There are two mechanisms to longstanding charge: dielectric absorption and chemical dielectric change. However, these two mechanisms can only put out what was stored into them. That was the point of my illustration - to try to point out that you can't get out more than was put in. I think I shot too high.

There are two limits on this energy. (1) the total energy that can be stored that way and (2) how fast it can come back out. If (1) was even 1% of the total energy stored in the cap, we'd never use that kind of caps. The stored energy is **TRIVIAL** compared to the total energy in the cap. Not that it's not detectable or doesn't happen, but it's vanishingly small compared to the working energy of the cap. It can also only go in and come back out at some rate. The reason that caps can accumulate a charge after being discharged is that the energy stored in the dielectric by both mechanisms leaks back out into the "usable area" of the cap as stored voltage. If there is a resistor on the cap that can dissipate the stored dielectric energy as fast as it comes out of the dielectric, the stored energy in the cap and hence its voltage must remain zero. The resistors eat it all.

But I can see that you're not going to let this go. Show us some numbers. Give references and verifiable facts. I can do that for each of my points. Can you?

the point is that you can be shocked by one even when the power supply has been unplugged for a very long time, even when there are resistors across it, and plenty have been shocked this way.

Can you produce some references to (a) electrolytic caps as used in tube amps (b) with bleeder resistors across them all the time (c) shocking someone? Please let us go read them.

coincidently I think in many cases a 10,000 mfd capacitor in a power supply is a great idea, and so do several other amp builders.

Apropo of what? I like purple caps, I think they're a good idea too. The issue is always how do the work? I think a stack of 40 car batteries would be an even better source of B+ than a transformer/rectifier/filter, especially in terms of surge current and hum. That doesn't make it a particularly practical thing to do.

perhaps you are referring to the minimum recommended capacitive filter to use with a rectifier tube.

No, I wasn't. But then, there isn't a minimum recommended filter to use with a rectifier tube. There is a clearly stated MAXIMUM first filter cap for many rectifier tubes. That's there because bigger caps make for bigger pulse currents in rectification and big enough pulses overcomes the emission capability of the rectifier tube's cathode. But I've never seen a recommended minimum. Can you show us where a minimum is stated for a cross section of tubes, not a maximum?

this does not rule out using a much larger capacitor. some would prefer eliminating the 60 cycle noise in amplifiers, reducing ripple in the power supply to the most minimum.

No it doesn't. Not wanting to ruin rectifier tubes quickly is what rules out using much larger first caps. After the first filter cap and some isolating element like a resistor or choke, you can go for all you can afford. But the first filter cap is what determines what the pulse currents are and that is what needs to be limited to give decent rectifier tube life. SS diodes can use first caps as big as you can fit in and afford precisely because their repetitive pulse currents are huge compared to tube rectifiers.

Then there's the whole issue of ripple in power supplies. I bet the same "some" would be confused looking for 60Hz ripple in a full wave rectified power supply. That's because there isn't any. It's 120Hz, produced by the folding nature of the full wave rectifier. Power supply caps can't reduce 60 cycle ripple. They reduce 120 cycle ripple. The rectifier gets rid of the 60 cycle.

or when building a DC heater supply for a high gain tube amp 10,000 is about the LEAST I use.

Huh? When I'm building low voltage high current power supplies I use caps as determined by the math to get ripple as low as I need it to be. In many cases, it is cheaper to use smaller caps and electronic regulation than to brute force it with caps, but each to his own, I guess.

just because you don't pay attention to 60 cycle noise does not mean that people who do care about reducing noise are stupid.

Who says I don't pay attention to 60 cycle noise? I *do* know where 60 cycle comes from, and it's not the DC power supply. And you know, I don't recall saying that anyone was stupid. Can you play back my quote and show me where I said that?

this also means that Jim Macintosh, who used 4700-10,000 mfd caps in his tube amps is also stupid, and that other manufacturers such as Levinson, Yamaha, Krull, Crown are all idiots.

You think so? I believe those are your words, not mine.

Reality is quality low noise amps have very big capacitors for very low noise,

Did you mean "low power supply ripple"? Because big caps don't affect the 60 cycle noise at all. In fact, big caps don't reduce thermal noise either. Or rather, can you explain to us HOW big caps reduce noise as opposed to power supply ripple.

I can, in fact, tell you one mechanism that big caps have for CREATING more power supply noise. It's that rectifier-pulses thing again. Those big pulses are not sine waves; they have significant harmonic content and they can both radiate directly and excite RF resonances in the wiring and ESL of the caps to radiate bursts of RF that the rest of the amp picks up as line frequency buzz. The bigger the cap, the worse the buzz. What helps here is not big or little caps, but good wiring practice and snubbing.

and if we want to build them that way, we do not let thick head know it alls tell us different. We want low noise and we will build for low noise.

Just you inside, smart, determined guys like McIntosh, Levinson, yada, yada, and YOU???? That's a pretty elite club you're in there, fella. You're right - don't listen to anybody that has a different opinion, because you and the low noise mafia there have it all sewn up.

We will continue to use DC heaters (like Macintosh did) and 10,000 mfd capacitors not because we are stupid, but because we are more intelligent, and we demand a higher standard than PV repairmen.

Just you few smart determined guys, eh?

Every so often a guy hits the pockets of the internet that renews my sense of amazement. Most recently it was walters/markphaser. But I think we have a new contender.

[Amp Kat]

Every so often a guy hits the pockets of the internet that renews my sense of amazement. Most recently it was walters/markphaser. But I think we have a new contender.

That's why when I first saw the post I thought he was a troll and still not completely sure he's not but unlike the other two who I believe had some disorders this one still in Limbo.

The 10,000 uf seen in those Hi-Fi 's even though tubes, could also be for the low voltage supplies. I'm not sure I've ever seen a 10,000 uf cap over 250vdc and that is computer grade and as Bruce pointed out the tube rectifier isn't going to like anything over 33uf much less 10,000uf. I really enjoyed that math on the discharged cap scenario RG so it wasnt' to high for some of us. FWIW I've never seen over 12 volts on a resistor discharged setup and many times much lower and correct also that they won't charge back up to full plate supply voltage but enough to give you a good tingle if you go across them when no discharge resistors are used and powered down with the standby on or no HV.

[Enzo]

Crown makes tube amps? Wow, learn something every day.

[mykey]

Crown makes tube amps? Wow, learn something every day.

i think i was referring to the fact that several manufacturers use 10000-20000 mfd capacitors in amplifiers and that the reason these are used is to obtain quiet operation, peak demand reserve, maximum ripple rejection in the power supply. i do not recall that i ever said crown makes tube amplifiers.
i continue to maintain that there is no fault in reducing power supply ripple in any amp tube or not, by using 10000 mfd capacitors, particularly in DC heater supplies. While some may except loud 60 cycle buzz in guitar amps (example mesa boogie) I do not except the noise. Many recording studios and musicians appreciate this standard of low noise.
In fact i will give you another reason to make fun of my high standards, i have used 1 fared (100000000 mfd) capacitors in some mixing consoles, succeeding in reducing noise to immeasurable levels, after many before me had failed miserably. Of course the only ones who appreciate this are the studio engineers and owners those who love PVs will not.

[R.G.]

i think i was referring to the fact that several manufacturers use 10000-20000 mfd capacitors in amplifiers and that the reason these are used is to obtain quiet operation, peak demand reserve, maximum ripple rejection in the power supply. i do not recall that i ever said crown makes tube amplifiers.

Your sloppy use of reference and hyperbole implied it, and not just to me.

This is a largely tube amp forum, and the discussion was about tube amps before your reference. If you had said something like "in solid state amps, DC power supplies can use caps as big as 10,000uF" then we would have understood that you didn't mean that Crown makes solid state. We would have said something like "Well, duh." because we all know that these amps use big value caps as a given.

You're not the only one here that has a decade or two fixing amps, fella.

i continue to maintain that there is no fault in reducing power supply ripple in any amp tube or not, by using 10000 mfd capacitors,

If you next say "in the first filter position after a tube rectifier" which is what the context of tube amps implies, you mark yourself as someone who does not understand tube rectifiers. That's what your posts implied.

particularly in DC heater supplies.

That's fine. Well, duh. That's not a particularly good flag to wrap yourself in. It provides you no cover.

While some may except loud 60 cycle buzz in guitar amps (example mesa boogie) I do not except the noise. Many recording studios and musicians appreciate this standard of low noise.

You then make the leap that the people you're talking to don't know what they're talking about, and that the fact that they don't understand what your muddled train of references means implies that they have a love of 60 cycle hum. You flatly miss the item which was explained to you slowly and in detail that 60 cycle hum does not and can not come from small caps in power supplies, miss the fact fact that it's 120 cycle ripple that power supply caps reduce, and somehow turn this explanation of fact into a group of people demanding high levels of hum in recording studios.

That is, you're trying to deflect criticism by blaming us for having sloppy standards when in fact you completely missed the accurate technical explanation of the inaccuracies in your posts. This little trick works pretty well in sound bites for the newspapers, but it doesn't work at all for a technical discussion with people who know what they're doing. Give it up, act like a man and admit you muffed it.

You seem to not understand that further assertions of poor you, misunderstood by naive, dull, and misinformed us will only make you look worse in this forum. This particular group does know what it's talking about. You might possibly know what you're talking about and might even have something to contribute here, but you are being sufficiently defensive (and offensive, too) that you're tarnishing whatever value you might have here.

In fact i will give you another reason to make fun of my high standards, i have used 1 fared (100000000 mfd) capacitors in some mixing consoles, succeeding in reducing noise to immeasurable levels, after many before me had failed miserably. Of course the only ones who appreciate this are the studio engineers and owners those who love PVs will not.

Again, get off it. You're dealing with a knowledgeable audience and wrapping yourself in the flag of "my high standards" only makes you look silly.

Let it go, contribute where you can, and quit thinking of the rest of us as uninformed. The longer you don't do that the worse you look to the group here.

[Amp Kat]

Of course the only ones who appreciate this are the studio engineers and owners those who love PVs will not.

I worked for an Audio Co. for 6 yesrs and we visited many professional Studios. I've seen Engineers that know what their doing build low pass,high and band pass filters from scratch and have seen one pull his complete Studer down himself and recap every E-cap (bucket full) in the whole damn thing every five years. Didn't need 10,000uf caps to reduce noise just a simple change. You can only get DC so smooth to where it has no effect on noise at all. Peavey amps ? Why not ? They aren't used for recording only playback and good engineers put all of their money into the console,speakers and killer Mic-pres and products like LA1,Summitt and have Neuman U-87's and an awesome array of tube and dynamic mics. The least thing they need is noise reduction gear and in fact I talked to an Engineer the other day that told me one of his secrets was adding noise to the mix so I find what your saying hard to believe.

[bob p]

Wow. This thread got interesting while I was away for a few days.

I wanted to add something about the GFCI and isolated supply topic.

Do you understand how a Ground Fault Circuit Interrupter works? It really doesn't give a rodent's rectum whether there even is a ground connection - what it does is compare the current in the hot conductor to the current in the neutral conductor and if they differer by more than a few mAs then the circuit "assumes" that the difference is flowing to ground - hopefully not through a person - and trips! So one can easily get their hand across the secondary of a PT and have nice brown smoke coming out of their ears while the hot and neutral currents remain the same - even if some of that current is leaking into the chassis ground if it's being drawn equally through the PT primary you're fried. And, of course, an isolation tranny makes the situation worse as the primary and neutral feeds to the iso will balance even if the secondary currents aren't doing so. Oh, and this also means that a GFCI protects a two wire - "double insulated" tool as well as the older metal cased ones.

Rob is right about how a GFCI works -- very small hot/neutral current differences in the range of 5 mA can be detected, and the GFCI will open the circuit in a few milliseconds. That's why the National Electrical Code requires GFCI placement in outdoor and potentially wet locations. They interrupt the circuit when "harmless" 120VAC currents on the order of 5mA are detected. As I mentioned in an earlier post, 5mA falls into the macroshock category, and a properly functioning GFCI will open a 120 VAC circuit before the current level is enough to become threatening to someone who has an intact high-Z skin barrier.

To keep things in context, its important to note that when you think of a GFCI, you need to consider that its calibrated to open the circuit at currents that are safe with 120 VAC potentials, not rectified High VDC tube power supplies. The GFCI protects you at 120 volts. Step up the voltage and you're dealing with energy levels that are outside of the realm of the GFCI.

Another thing to consider is that a GFCI provides absolutely no protection against microshock. You won't find one in a hospital operating room; Even though the voltages there are 120 VAC and the potentials are in the range of function of a GFCI, because of the absence of high impedance barriers, the lethal currents are orders of magnitude lower than the sensitivity limit of the GFCI. In that type of environment a GFCI is not used because the patient would likely be dead long before a GFCI ever sensed a problem.

Back to the subject of isolated power systems, there seems to be a little confusion about this with respect to ground potential references. In a non-grounded system an isolation transformer does not increase the shock hazard -- it actually decreases it. In a true isolated power system, the room is not grounded. Its possible to stand in a puddle of saline solution and stick your tongue onto one of the Hot 1 or Hot 2 isolated high voltage lines without getting zapped. That's because the floor that's soaked in an electrolyte solution bears no continuity to ground. Because you're not contacting the circuit at two points you don't complete a circuit and there is no current flow.

When a faulty piece of equipment is plugged into an isolated power system it does not present a shock hazard. It simply converts the isolated power system into a conventional grounded power system, and thereby removes all of the protective features that the isolated power system would have otherwise provided. When this happens, there is a marked reduction in the impedance between Hot 1 and ground or Hot 2 and ground. The line isolation monitor would quickly detect this condition and sound an alarm to alert the staff that a defective / hazardous device is present in the circuit and is presenting an active electrocution risk. (that dangerous piece of equipment immediately gets disconnected and sent to the techs, who presumably know how to work on it without killing themselves.)

Alternatively, if you're not dealing with life support equipment and you're willing to tolerate circuit interruptions, the LIM in an isolated power environment could be replaced with a GFCI on the secondary side. In that case, the GFCI would trigger and open the circuit when it sensed asymmetric current flow. But you still have the problem that the GFCI is pretty useless, as it trips at currents that are way too high for patient safety. The key to understanding this situation is to understand how ungrounded/isolated power environments work. Admittedly, this is something that is out of the realm of what most people are familiar with in their day to day lives.

Going back to Rob's tube amp example, if you're working in a true (intact) isolated power environment, you can stick your tongue onto the B+ rail and you won't complete a circuit unless 1 of 2 things happens: a) you concurrently touch the grounded chassis, or b) there are already leakage currents in the defective amp, and they have converted the isolated system into a grounded system. In that case, all of the conventional rules of safety still apply.

[Euthymia]

Citations, please

....so little has been said about safely working on these amplifiers.

Many have categorized safety rules as a joke.

Are you referring to this forum? Here, safety is talked about often, with great seriousness and if someone shows evidence of inexperience in a post, they are warned about brushing up on safety issues before going any further.

....with so many killed by electrocution every year....

Yes, in the hundreds. What were they doing when it happened? Working on tube amps? Probably not.

a large portion of those killed are the pros who should know better and have let their guard down.

Statistically true, but the kind of "pros" they are is usually construction, utility, or farm workers. How many appliance repair people zap themselves?

There is a risk of electrocution when working on guitar amps. It does not need to be overstated. Doing so just makes the poster look pedantic.

[mykey]

I believe you would not need to to this with a Studer, but in huge consoles such as StudioMixer which have such pathetic decoupling, thin wires, and long distances it works tits! Also consider the tiny little power supplies these things come with that are installed internally.
You could tell a lot about the noise in a console just from the fact that the power supply is external.
You will also get a laugh out of this: modifying mixers so that the frequency response is 3-1000,000,000 hz. Sorry I think the FCC might be a little upset about that. we call this the "air sound". The engineer I recently did this for said it made his $4000 Levinson preamp sound like a joke.
So there, now you have some more rumors to spread about me.
I won't go into too much detail about the Marshall clones with $10,000 sets of custom made silver mica Cornell Dubleer capacitors, Teflon circuit boards with gold plated terminals, aircraft grade silver/ Teflon wire, 2 parts per million aerospace resistors, $425 sets of Bugle Boy output tubes (all grids matched and heaters too). Not just the plates and current guys, ALL the grids AND heaters.
That would be too much for you guys. You don't believe it? Call Richardson Electronics and Cornell Dubleer and ask them, go ahead! You think this is crazy? It's more than that, it's DERANGED.

[mykey]

some prefer to discharge capacitors by shorting them to ground (kablam)
some say that this wrecks the capacitors, and that these should be slowly discharged through a resistor over a period of time. Some like working with the capacitors charged and don't give a damn. What would you say is the safest for the capacitors? Have you ever heard this, shorting the cap to ground wrecks the capacitor?