Widowmaker Amps & Isolation Transformers

[robrob]

I'm trying to nail down my understanding of the dangers of "widowmaker amps" so could you review the below and let me know what I have wrong.

I have yet to see an amplifier schematic that shows either power cord (mains) wire connected directly to the chassis like the 1940s All American Five radio receiver below. This radio is wired as a true widowmaker but the chassis is insulated from the user to prevent shock.

Does anyone have direct experience with a widowmaker with one of the mains wires connected directly to the chassis?



This Silvertone 1430 below is a "widowmaker" but it's metal chassis and input jack ground are isolated from the mains by the C3 RFI AC Ground Cap and 68k V1 cathode resistor. The C3 Ground Cap seems to function just like a "death cap" going from mains to chassis.



With the power cord backwards the chassis would be hot with mains voltage but max current would be limited by the C3 RFI cap and 68k cathode resistor to about 4.2ma with 125v mains voltage.

We could install a three prong power cord in this amp and connect the chassis to the safety ground but if you plugged into an outlet that had its hot and neutral reversed you'd short hot to the chassis safety ground and pop a circuit breaker (best case anyway). This is one reason why electrical code doesn't allow you to connect neutral and ground anywhere beyond the building's service entrance.

Adding an isolation transformer to the 1430 above removes the mains neutral ground reference from the amp circuit. We could actually touch either isolation transformer secondary lead and not be shocked because the secondary isn't referenced to ground--the voltage is from secondary lead-to-lead, not lead to ground.

After an isolation transformer is installed we can safely and legally add a safety chassis ground.

[Tom Phillips]

...Again, this amp would not function with its power plug plugged in backwards because the rectifier's plate would be grounded to the mains neutral...

The amp would still function because, with only the two wire connection at the power plug, the amp still sees the full AC line voltage across the input. The chassis is floating with respect to the outside world. The amp doesn't "care" if one or the other side of the AC line power is connected to earth outside the chassis. As you pointed out, this all changes if one attempts to convert to a three wire power cord. A key point is that the line power is AC. In the olden days, when that 5 tube radio was in production, there was the possibility to run the thing on 117 VDC. In that case you needed to get the polarity correct or it would not work.

[Enzo]

First, the circuit stands on its own. You are wrong that turning the plug over can't work. You are right that the rectifier plate would be directly connected to earth through the power cord, but so what? Nothing else on the chassis is connected to earth. All it knows is there is 120vAC between the power cord prongs, it has no idea which side is earthed elsewhere. The chassis is not grounded. There is still 120v in the circuit.

In the Silvertone, if the neutral goes to the rectifier plate, then 120v hot is on the B- common. The chassis then has 120v on it. Note the schematic has the cap parallel the 68k resistor. SO the chassis is not directly connected to mains, but who wants to grab a 0.05uf cap or a 68k resistor plugged into the mains?

In the radio the mains is directly connected to chassis, if you invert the plug, chassis has 120vAC on it. Yes the radio has plastic knobs and housing so you cannot normally touch chassis. But how about this: Your table radio sits on the kitchen counter. Plug is in "wrong". 120v on the chassis. The chassis is held in the plastic case by metal screws through the bottom. Mom is wiping down the counter, picks up the radio and her fingers slide beneath it and touch a screw. God hopes she doesn;t also lean on the metal sink or touch a pipe/faucet. Or her fingers are safe, but she sits the radio on the edge of the sink to wipe where it was. Screw grounds to metal sink, blows main breaker, and scars sink with burn mark.

The cap and resistor are not cathode circuit for V1, V1 is a grid leak biased stage. The cap and resistor are linking the cold front end with the hot power amp section in terms of signal.

The 50C5 is cathode biased, not grid leak. I myself would not call R6 a grid leak resistor, I would call it a grid return or grid reference.

Yes, an isolation transformer solves all of this.

[robrob]

Thanks Tom and Enzo, I didn't consider that a rectifier with a grounded plate but cathode at high negative voltage would function normally but your explanation makes sense.

Enzo, I had read about the All American Five's hot bottom screws' shock hazard.

[Enzo]

BAck in the 1950s it was common household knowledge that if you got shocks off a radio, TV, or other appliance, you turned the plug over. That also usually served to reduce hum, hence the ground polarity switch on old Fenders.

Another bottom screw scenario. Radio sitting there innocently. Mom has a roast thawing on the counter or some other reason a puddle forms. If the puddle runs under the radio, it can make that puddle live.

[Tom Phillips]

I’ve been pondering why the designers of the Silvertone 1430 used the small transformer to power the 12AU6 heater.

1. The 12AU6 heater operates at the same current as the 50C5 and the 35W4 so it would have been possible to run all three heaters in series and then drop the appropriate amount of excess line voltage across the R9 (R9 value adjusted as required)
2. Since they used a transformer anyway why not make it a proper PT? They must have been really trying to hold down the production parts cost. The delta cost in production quantities would have been minimal.

We will probably never know the answer. It just seems like a goofy design given that it was already an unsafe circuit. Anyone know a good logical reason they would do that? Maybe something to do with different heater warmup times?

[R.G.]

Rework of an older design, possibly.

Rework-design-work is an ugly, frustrating business. The watchword is to change it "a little" to add some new something or fix some old something. The budget to do this is negative - if you're looking at a design again, you're expected to cost-reduce it while you're in there, by at least enough to pay your salary on the time spent.

In perfectly logical fashion, the best, most experienced engineers run from this like someone is offering to dust sprinkles of Ebola virus cultures on their coffee. Whomever can't escape is saddled with the task, and again, understandably, gets the hot potato off his/her desk as quickly as possible, no matter how ugly the fix turns out. If there happens to be an off-the-shelf part to do a quick, if ugly, fix, then while the boss isn't looking, the cans of porcine lipstick are trotted out, applied to the task, and it's shipped. With some luck, you'll get nibbles on your resume before there's any customer experience.

There is an alternate pathway, where (say, in this case) the bosses' nephew runs a distributing business that has this unfortunate surplus of nearly free tiny transformers that would just ...beautifully... apply lipstick to this particular pig. Directed design outcomes are such a lovely process.

If ...all... of the experienced help manages to be deathly ill the day this task is handed out, a semi-tech-literate baby-boss-ling will sometimes try to curry favor by just jumping in and doing the work him/her self. This leads to odd tech decisions. I once heard a guy trying to defend his choices in a power supply design by noting that proper filter cap selection makes the secondary inductance of the power transformer resonate at the power line frequency. I was asked to leave the conference room, as continuous howling laughter was deemed to be not in the proper team spirit.

[Enzo]

They didn't consider that unsafe at the time. It was just common practice.

But it is a good puzzle. What is the transformer anyway? It can't be a 120/12, because it doesn't see 120v on the primary. Possibly it removes some heater to cathode voltage issue, if the 12AU6 had mains on its heater. And that would also be dependent on which way the plug was in the wall. Note they did reference it to the chassis ground rather than the mains"ground".

To me it smacks of a problem that arose after the design was final, and they fixed it with this.

[robrob]

Why is a widowmaker like the Silvertone above considered more dangerous than an amp with a power transformer, a two prong power cord and a death cap?

Is it because the widowmaker can deliver a full mains shock at mains voltage and amperage while a "normal" amp is limited to the amperage the power transformer can deliver?

[nosaj]

Why is a widowmaker like the Silvertone above considered more dangerous than an amp with a power transformer, a two prong power cord and a death cap?

Is it because the widowmaker can deliver a full mains shock at mains voltage and amperage while a "normal" amp is limited to the amperage the power transformer can deliver?

Ok take a widow maker turn the plug both directions with your tongue on the strings, then do it with an amp with a power transformer. shock = no shock
It offers a form of isolation.

As far as my understanding the deathcap when it is good just prevents mains voltage from appearing on the chassis.
Granted these opinions have not been evaluated by minds in the know.

nosaj

[R.G.]

What he said. To elaborate a bit, it's to do with the way the world is wired. AC power distribution networks reference themselves to the planet as a whole by using grounding rods (yep, that's where the term comes from) to establish a fixed 0V reference. In some cases, power is distributed on one wire, using the planet itself as a return wire, thereby saving half the wire cost. But AC power typically has one wire "hot" and the other "neutral". The neutral is grounded at the breaker panel and power pole.

When you plug into an AC mains outlet, one prong is the line voltage, the other is neutral, and the neutral is connected through wires to EVERY OTHER DEVICE WITH AN ELECTRICAL NEUTRAL as well as the plumbing pipes and the ground itself, especially if it's wet ground. So the widowmaker chassis is attached to the AC "hot" conductor 50% of the time when it's plugged in randomly. It's not just the voltage or amperage it can deliver, it's that your personal body is likely to to be the path for that current as you touch random other things that have not been deadly to touch at other times. It's a trap.

The isolated supply, on the other hand, isolates you from both the AC mains hot and neutral, and for three-wire equipment wraps the AC mains hot wire in a metal chassis that is itself hard-wired to the AC mains safety ground. Equally important, the plug goes into the wall the same way every time, so there's not the 50% of the time that the plug is the wrong/deadly way. The secondary of the transformer isolated power supply is connected to the safety ground too, but it's the same way round every time, and there is no way other than some component failure, generally wire insulation, for it to get to the chassis, and if it does, this causes a fuse opening. So does a wire shorting to the chassis. You can still get killed by contacting the high voltages made by the transformer isolated supply, but those are carefully hidden inside the chassis and enclosure, so you have to have the chassis open and the mains cord plugged in, and stick your hands inside the chassis while the power is on. There are many labels on the equipment telling you NOT to open the chassis because you could get dead doing that.

The bottom line is that the transformer isolated supply makes a good, solid advancement in keeping you from accidentally killing yourself. There are still ways for you to kill yourself left, but fewer of them, and you'r less likely to be able to find them accidentally.

[R.G.]

The AC power lines make a pretty good antenna for certain radio bands. They carry that RF hash into the amp, where it's amplified and heard as either odd radio stations, or as "hum", as much RF hash is synchronous with line frequency, and gets heard by the ear as "hum".

In a two-wire-only chassis, one wire coming in is the AC mains, the other is neutral. Neutral is attached to a safety ground wire back at the breakers, so the neutral conductor has much less length and exposure to RF. If the chassis is connected to neutral FOR RF SIGNALS ONLY, NOT POWER LINE FREQUENCIES then much of the RF is avoided. The death cap picks which of the incoming AC lines is tied to chassis, and one of them will always be quieter.

The problem with the death cap is when it dies. They used to be simply high voltage ceramic caps, and when they died, they died shorted. Now you have converted a safe(r) transformer unit into a hot chassis widowmaker in 50% of the AC plug and "line reverse" switch settings. Hence the name "death cap". A single failure of a part under voltage stress can kill you. There are modern standards for across-the-line and line-to-ground capacitors that are safer. Better yet, use a third safety wire ground and don't put in a death cap. Hence, three wire conversions.

[robrob]

R.G., I wired my barn for 240V 200 amp service so I understand hot, neutral and safety ground wiring and how neutral and ground are tied together at the service entrance.

The Silvertone widowmaker above has it's chassis isolated from the mains power by the AC Ground Cap which functions just like a death cap so its chassis can only pass a couple of millamps if it is plugged in "backwards"--just like an amp with a power transformer, two prong cord and a death cap.

Has anyone ever seen a power transformerless guitar amp that really directly ties a mains connection to the chassis? If an amp did have its chassis tied directly to a mains connection the "widomaker" name would be appropriate.

[g1]

The silvertone in your example has from chassis a 68K resistor directly to one side of the AC line. That is nowhere near the same thing as a 2 prong amp with an isolated power transformer and a 'death cap'.

[robrob]

That 68k resistor will pass less than 2 milliamps at 125 volts so it has less impact than the death cap's 2.4 milliamps at 60Hz and 125v.