Electrical Safety

A little learning is a dang'rous thing; drink deep, or taste not the Pierian spring: there shallow draughts intoxicate the brain, and drinking largely sobers us again.

-- Alexander Pope


If you are an experienced technician, it is important that you read this. If you are not an experienced technician, then it's IMPERATIVE that you read this. What you will learn here could save your life. This is especially true if your experience up till now has been exclusively with solid state. You can work for years with solid state and never have to deal with potentially lethal voltages. It is one thing to rig up an experimental circuit on a solderless prototype board powered by a 9V battery; VT equipment is a whole 'nother story. The VT, itself, is a high voltage, low current device. That means you will be dealing with voltages that could prove fatal if mishandled. This is not to say that one should allow oneself to become excessively fearful. That is also a great contributor to accidents. There is a big difference between fear and a healthy respect. It is the latter that one must cultivate before contemplating doing any repair, restoration, or development work with VTs. As odd as it may seem, this is one area that does not receive the necessary emphasis, even in the universities. You can go the entire four years of a bachelor degree program in electrical engineering without receiving even so much as a single lecture on electrical safety. As for myself, I had the advantage of having both a grandfather and the father of an early childhood best friend who were both electricians by profession. I learned this lesson starting at three years of age.

Just What is an Electrical Shock?

At one time or another, almost everyone has experienced this, after all, that zap you get from reaching for a metal doorknob after walking across a carpet on a cold winter day is a momentary shock from static electricity. Everything from tingles to a sharp burst of pain can be the result of poorly grounded, malfunctioning appliances, or carelessness while plugging in some appliance. It's a fact of modern life, it seems.

Any current flowing through anything that's not a superconductor generates heat. The level of heat generated depends on the current and the resistance. Tissues are more or less conductive, and will heat up during a shock. This can damage internal organs or cause burns that are not visibly obvious. This burning is the cause of long term injury or disability resulting from severe shocks, such as lightening strikes or from high tension power lines.

One of the major effects is the disruption of nerve signals. The nerves depend on small electrical signals to release neurotransmitters and in response to neurotransmitters. An electrical current will simply bury the nerve's natural electrical signals in noise. As a result, the nerves lose control over muscles. Current through muscles will cause contraction and involuntary movement. If the shock victim grabs onto a live conductor, the muscles of the hand will contract and clench even more tightly onto the conductor, making even better contact. No matter how desperately he wants to let go, he can not. Needless to say, this could prove fatal. Of course, the diaphragm and heart are muscles. Current of sufficient intensity can stop the lungs' air pump and/or cause the heart to stop beating. If rescued quickly enough, CPR may not even be necessary as the heart and lungs resume operation once the victim is separated from the current. As bad as that sounds, it's actually weaker currents that are more dangerous. Currents from 10 -- 100mA can desynchronise the heart's muscles, called fibrillation, that does nothing to push blood through the arteries. This condition is even more dangerous than simple paralysis. If this has happened, perhaps CPR administered without delay may help the victim to recover. Or it may require defibrillation, and an emergency response team to administer it. Conducting current across the chest is therefore especially bad. Arm-to-arm, or arm-to-foot (especially left hand to right foot) is to be avoided. It is for this reason that the suggestion to keep one hand in a pocket when working on a live circuit is frequently given. It's a good idea, so far as it goes. You must not neglect to consider that the feet may also provide a path to ground.

This possibility is often neglected. After all, don't shoes provide enough protection? Sure the rubber of the soles of sneakers is a good insulator. However, did you recently wear those sneeks in the rain? Take a walk through early morning dew? Do your feet sweat? All of these may compromise the insulating properties of footwear. Carpeting can also present a false sense of security. Concrete floors, even if they appear bone dry, contain a great deal of water, and concrete is conductive enough to permit lethal currents to flow. This is a fact that surprises many. Indeed, most folks use their basements with concrete floors for a workshop. If this is where you work, it would be a very good idea to put down an insulating mat before starting any work. Furthermore, don't leave it there when not working. Plastic mats will gather moisture underneath. This not only makes mildewy messes, but also compromises insulation.

So How Much Does it Take?

There is an old folk saying that it isn't the voltage that kills, but the current. This is true, so far as it goes, however current doesn't just happen. It takes a voltage and a ground connection to drive it. The resistances that a body can have varies tremendously. Sometimes, several hundred volts won't be enough to cause a lethal current; sometimes considerably less will do the trick. According to a WW II-era Army manual, anything above 48V should be considered potentially lethal. This assessment was based on the lowest known voltage to cause an electrocution. Granted, that doesn't happen too often, but still 48V is within the range for solid state equipment, so you may not be completely free of risk even if you stick to solid state designs. Here (Darwin Awards) is a case where someone did something foolish and managed to kill himself with a nine volt battery! Apparently, this really did happen. Granted, this is an unusual case, and things like this are highly unlikely. Let that be a warning to you: never slack off just because the voltage is "low". There is always the chance that it won't be low enough. Needless to say, the voltages in VT equipment deserve a great deal of respect.

Very low voltages can also represent a danger, but for other reasons. Any low voltage, high current source can easily bring a piece of jewelry to red heat before you get the chance to scream. This would include the 5.0V power supplies for computers, the power supply for just about any solid state amp, and the filament supplies for VTs -- especially the 5.0V and 6.3V transformers for power diodes and finals. Before working, be sure to remove all items of jewelry: rings, watches, necklaces. All of these could be welded to a current source, resulting in some nasty burns. Furthermore, gold has quite an affinity for solder, and solder splatters on gold jewelry aren't pretty, nor easy to remove. Of course, rings and watch bands also make excellent electrodes for conducting current into the body. Ditch the jewelry!

Another problem that may not seem related to electrical safety at first is the possibility for unexpected behaviours. This could include anything from sparks, to peculiar odors, smoking parts, unexpected loud crackles from speakers, resistors going up in flames, to exploding capacitors. The actual event itself may not be serious, but could cause one to react in surprise and thus contact a live conductor, or to pull a hot item into a lap, or pull away and cause nasty cuts against chassis or a metal cabinet. Especially dangerous is the possibility of an exploding electrolytic capacitor. These capacitors are filled with a water-based electrolyte. Even if it claims to be a "dry" electrolytic, it really isn't. All that means is that the electrolyte has been gelled so as to not leak so easily. If these are wired with reversed polarity, or are over-volted to the point of break down, they will conduct a heavy current that will boil the electrolyte and build up pressure sufficient to burst the container with explosive force. This is especially true of the older units. Modern electrolytics are equipped with rubber blow out stoppers. Still, expect the forceful ejection of hot and caustic material. Eye protection is highly advisable.

Be aware that electrolytics rescued from old equipment may have lost most of their dielectric. In that case, these won't behave as capacitors until the dielectric is reformed. Also, a 450V electrolytic that's been running at 200V isn't a 450V electrolytic anymore. Even if the proposed operating voltage seems within the nominal rating, it very well may not be safe. All old electrolytics should be reformed before return to service, and all connections should be checked and re-checked to make certain that the polarities are correct. When powering up for the first time, it is always a good idea to bring up the voltage gradually with a Variac, or failing that, connect a light bulb in series with the mains. Make certain that nothing is amiss at a lower voltage before running at full power. In any case, expect the unexpected.

If you are doing restoration work on old equipment, keep in mind that workers were far less safety conscious in those days. Almost every "All American Five" AM radio of that time used the "suicide box" (so named for a damn good reason!) topology: a 35Z5 or 35W4 half wave power diode connected with no power transformer to one side of the AC main, with the other side wired directly to the chassis. Now, you can get away with that for radios and TV sets where the only thing the end user is expected to connect is an antenna. Even if the chassis was inside a plastic or wood cabinet, the mounting screws were all too often exposed. Still there were hazards such as the loss of plastic knobs. Anyone using a pliers as a replacement was potentially exposed to the full mains voltage. This type of power supply would be perfectly acceptable IF the chassis were always connected to the cold side. Polarized plugs and wall sockets are relatively new, nor can it always be assumed that the polarity is correct. Wrong connections can be made at the wall socket and/or inside the equipment, especially if it's been worked on before (nearly always the case these days). . Even more inexcusable was that a good many guitar "practice amps" also used this topology. Electrocutions and near-electrocutions from hot guitars did happen. Connecting such an amp to computer sound cards or preamps can ruin other equipment. If you're going to be working on such equipment, an isolation transformer is a wise investment. If restoring such equipment, it's a very good idea to install an isolation transformer.

Power Supply Safety

Ideally, all equipment would come equipped with bleed down resistors to insure that all the large value filter capacitors will be discharged at power off. If you are designing and building your own equipment, there is no excuse for not including this essential safety measure. If not needed to improve voltage regulation, the bleed resistors can be allowed to draw a few milliamps. The power ratings of these resistors should also be quite generous, as burnt out bleed resistors give a false sense of security. Not including proper bleed resistors is inexcusable. As for vintage equipment, PS bleed resistors were not used with distressing frequency. The pencil pushers rationalized that the tubes would serve to draw down the voltage at power off since they continue to pull current while the cathodes are still hot. True enough, but do they stay hot long enough to get that voltage down to a safe level? That's more problematic, especially when the equipment is power cycled before the tubes heat up fully. That can easily leave dangerous voltages on the filter capacitors that can persist for unexpectedly long times. The A Number One lesson here is to never assume, and to always measure before doing anything to a "dead" circuit. This is also good advice even if you're working on solid state equipment and know that you'll be working with sublethal voltages. Transistors and ICs are notoriously easy to poof.

Should you come across equipment that doesn't have bleed resistors, then install them.

Should you find that the filter capacitors aren't discharged, don't short across them with a screwdriver. Yes, some texts do recommend doing this, but it's a bad idea. Large capacitors, or even smaller ones with high voltages, can pack enough energy to break internal fusible links, or the wires that connect the plates to the external connectors, or to instantly vaporize a steel screwdriver tip. If the PS filter includes chokes, the current impulse can induce some tremendous voltages that can flash over the insulation. In any case, something usually gets ruined. For discharging filter capacitors, prepare a 500 Ohm, 10W resistor with insulated alligator clips. Clip this across the charged capacitor(s) to soft discharge them. Once the voltage is down to a few volts, then short with clip leads, and leave these connected while working. Due to a phenomenon called "dielectric absorption", electrolytics and some PiO capacitors can "recharge" themselves. This can lead to unexpected voltages if a "discharged" capacitor is left with an open circuit.

Work Station Safety

The first rule is do not allow clutter to accumulate. You should have out just the tools you need to complete the task at hand. Items not directly job related just take up valuable space, and can lead to confusion as to what may be connected to a live circuit and what is not. It isn't too great an imposition to have to fetch something else when you need it, and to remove items that you're finished using. If you are that impatient or rushed, then you shouldn't be working in the first place! Find a better time.

Power for equipment undergoing testing/repair/restoration/development should have its own fuse box. The best kind of fuse box is the old-timey one with screw in fuses, and a knife switch to make/break the connection. Circuit breakers and other types of spring loaded switches can fail to open. If you live with others, be sure that you inform husbands/wives/partners/significant others where that main cut off is and how to operate it. Should something go wrong, and you become incapacitated by shock, they need to know how to interrupt the power immediately. If at all possible, it's a good idea to never work alone. You should also include a positive indication when the work station is powered up. You can rig a light near the fuse box to indicate a live work station, and it serves to draw attention to that all-important switch. Failing that, you can also wire the lights to work off the main cutoff only. That way, if the lights are on, the work station is live. If you are living with pets or small children, it's essential that they be excluded from the work area. You don't need the distraction of "Fluffy" jumping onto the work bench, or "Fido's" deciding he'd like a walk, when taking a measurement on a 1000V power supply or trimming the bias on a 600V power amp final. Children are especially nosey and lacking in common sense, and especially vulnerable to electrocution. They have a strong need to "help Daddy" (or Mommie as the case may be) but they are NO help here. In some situations, children can be distracted with pointless little "projects", but here isn't that place, and now isn't that time. Furthermore, you can't be paying full attention to what you are supposed to be doing if you need to keep one eye on little Sammy or Susie. Get them to understand that this area is strictly off-limits without supervision. Also, make certain that the main cutoff is well out of reach of small, curious hands.

The equipment being worked on should be plugged in only while a live test is on-going. It is never a good idea to rely on the power switch for power down. Some "suicide box" topologies did not fully disconnect even when "off". These switches can sometimes be defective, barely breaking the circuit, but can reconnect with the slightest disturbance. It is also possible to accidentally trip the switch at the wrong time. So always switch off and unplug. That way, you can be certain that the equipment is truly "off". Since this also requires two positive acts, you are less likely to forget that something is still on when it shouldn't be. Powering up for testing also requires two positive acts. This is important: always unplug. It's also a good idea to keep a good CO2 or Halion fire extinguisher on hand. That way, you can stop a small fire from becoming an inferno. Also, a CO2 extinguisher does not create a mess as do the nitrogen + powder extinguishers that are often sold for RVs and boats. Under no circumstance rely on any of those old fashioned liquid extinguishers, or those one-time "beer can" extinguishers. Make sure to inspect the extinguisher to see if it's maintaining pressure.

Equipment Safety

These days, you frequently have the three wire system, where the third connection goes to ground. Now, this third wire ground is no good for RF grounding, and barely usable for AF grounding, but is essential for equipment safety. It is quite possible to have a short circuit from a high voltage coil to a transformer core that puts hazardous voltages on a metal chassis without interfering with the normal operation. This situation could exist for years without notice. Another possibility is short circuits between the primary and secondary of VT output transformers. This, too, won't affect normal operation, and yet leave the speakers and cables with dangerous voltages. Making certain that all exposed metal is grounded through that third wire ground will prevent this from occurring. Should something short to the chassis, the third wire ground will blow fuses or open breakers. In any case, it keeps high voltages off exposed metal. Therefore, never neglect to include this essential ground. Yeah, I know that it's terribly inconvenient to have a three pronged plug and a two prong socket. All too many folks "solve" this little problem with a pair of dykes. That is a very bad idea. It is also a bad idea to not include this safety feature in the mistaken belief that you're doing someone a "favor".

One final word: do I need to point out that one should never mix booze (or more illicit happy stuff) with dangerous voltages? Also keep in mind that some OTC cold and flu remedies also make you just a bit stupid. It's best to avoid doing any electronic work when one is feeling out of sorts for whatever reason.

Multi-meter Safety


Typical Multi-meter

Typical Multi-meter

Here is a quite typical multimeter with the usual test probes. Such test probes are OK for use with solid state equipment and typical low voltages. However, this is unacceptable for work with VT equipment. Such probes all but force you to use both hands. Should something go wrong, or the test probe insulation have an unseen defect, you will take a shock across your chest -- the very thing we most want to avoid. The other problem with such probes is that they are liable to slip. The resulting short circuits can poof transistors, create unexpected sparks, or can even become welded firmly in place so that you can't interrupt the current before the magic smoke escapes.

Better Test Leads

Better Test Leads

Shown on the left is a typical test lead fitted with alligator (or crocodile) clips. These can either be made up quite easily, and are also sold in sets. On the right, we have a spring loaded "IC" test clip that has a small hooked end which can be slipped over one lead of a typical small signal transistor or DIP IC lead. They are also very useful for gripping fine leads of grid resistors or tube socket pins. Both types have a considerable advantage in that they can be clipped in place on a dead circuit. The measurement can be taken with the power on without having to handle anything that could potentially be "hot". The IC test clip has an additional advantage in that it will stay put if it's necessary to turn the chassis upright, which will almost always be the case when running with the finals in place. Many large signal VTs are meant to be run vertically only. This is especially true with directly heated types. Filament sag when hot can easily short circuit to the nearby grid wires if heated in a horizontal position. This almost always poofs the tube, and that can be a most expensive and easily avoidable mistake.

It should also be pointed out that, with the exception of low voltage circuits, these test leads should be considered to be uninsulated wire. The insulation will not stand up to most VT power supply voltages. This means that you should never touch them when running hot, nor should these leads come near a grounded chassis. May be they will flash over, and maybe they won't. Either way, don't count on it.

The modern digital multimeter is so cheap these days that it's well worth it to buy several. This is a great convenience since more than one meter may be connected to measure voltages at multiple points at once, or for monitoring both voltage and current. The less you have to handle during testing, the safer you will be. It's also helpful to have backups in case you do make a mistake and poof a meter. Nor is it necessary to lay out BIG BUX on some extra special deluxe multimeter. These expensive units don't work any better than the much less expensive versions anyway.

One final word concerning multimeters: be sure that you get extra fuses. It is all too common to blow the meter fuse by drawing excessive currents, or forgetting that the meter is set to read current when attempting to measure a voltage. Avoid the common mistake of simply wrapping a blown fuse in foil to get the meter back on line. I guarantee you: you will forget all about this "temporary" fix. At the very least, you lose a meter. At the very worst, well, you should know by now...

Copyright © 2009 All rights reserved