Isolation transformer confusion

[pwnell]

I have frequently read that it is important to use an isolation transformer when servicing something like valve amps where the voltages are lethal.  So I understand how they work (providing galvanic isolation), I understand how a typical home is wired (at the breaker board the neutral and ground are tied together but not downstream leaving a potential difference between N and G and L and G), but what I do not understand is how an isolation transformer helps at all when the valve amp already has a transformer. 

First, the transformer in the valve amp is identical in operation to an isolation transformer (bar perhaps the faraday shield) as far as galvanic isolation is concerned - at least, that is my understanding.  So (considering the DUT - the valve amp) except for the short path of the live wire from the IEC terminal to the primary leads of the transformer, the rest of the circuit on the secondary side sees no benefit - right?  Further, since the secondary side of the amp has the center tap of the windings tied to signal ground and also tied to mains earth (in most of the more modern designs), and since the mains earth is passed through an isolation transformer (at least, in my Tripp-Lite it is), then the secondary has a direct connection to mains earth and the isolation transformer is then pointless?  As in - I cannot connect the ground of my oscilloscope to the B+ line of the power supply output?

Or am I missing something?

[wilfred]

Speaking as a fellow beginner where isolation transformers are concerned, you are missing that valve radios were frequently live chassis without a transformer.
Youtuber "Mr Carlsons Lab" did a good albeit somewhat long-winded video on the subject.

I'll also be interested in what answers you get because I feel I know just enough to be dangerous.

[vk6zgo]

Speaking as a fellow beginner where isolation transformers are concerned, you are missing that valve radios were frequently live chassis without a transformer.
Youtuber "Mr Carlsons Lab" did a good albeit somewhat long-winded video on the subject.

I'll also be interested in what answers you get because I feel I know just enough to be dangerous.

Transformerless valve radios were "as rare as  hen's teeth" in Australia, for several reasons, apart from the obvious safety aspect.

(1) Oz valve (tube) factories were overwhelmingly set up for 6 & to a lesser extent 12v types, so they had no desire to set up a new line making various voltage valves for series filament use.
Such valves were hence imported, & more costly.

As most radio manufacturers were either the same company as the valve one, or closely allied with them, that was reflected in their designs.

(2) Although there were still some DC mains supply networks in country towns, they varied between 110 v ones & 220v ones, making simple designs like the "All American Five" less attractive.
The Brit style ones for 220v DC were far more complex, making them unattractive.
In the end, supply of radios for such places became a "niche" market.

(3) Transformers were relatively cheap to manufacture, & despite comments about their weight from posters who live in "transformerless" countries, a "mantel set" transformer could be comfortably held in one hand.

(4) Transformers allowed the use of higher HT voltages, as well as fullwave rectification which requires less stringent filtering.

I don't think transformerless valve amplifiers were common in any country, as a HiFi valve amplifier required a very large audio output transformer, which would cost several times the price of a power transformer, so there was no advantage in eliminating the latter.

[SpecialK]

It's all down to how your PE, chassis, and DC ground are tied together and what happens when you attach you oscilloscope crock tip on the wrong point in circuit.

So say you'd like to look at the signal across a interstage coupling capacitor.  So you attach the oscilloscope probe tip on one side at random and then connect the ground clip to the other side.  In a bright flash you've maybe directly shorted 200V to ground and melted the crock tip.  How is this possible?  Even though your power transformer gives you galvanic isolation, you've bonded primary and secondary together with the typical earthing strategy.

The oscilloscope's exposed BNC connectors are tied to earth.  Maybe not an issue if you use a Fluke Scopemeter or other battery powered scope. Multimeter probing is much safer. 

[Jwillis]

Here's a great tutorial on the in's and out's of isolation transformers for diagnosing and servicing tube amplifiers. He starts by trying to build one and then goes through to types of isolation transformers pros and cons then where to connect oscilloscope probes. Best to watch the whole video.
I recommend to always use a current limiter when servicing any vacuum tube device.   

Isolation Transformers: Internal Design and Safe Use with Oscilloscopes

[vk6zgo]

It's all down to how your PE, chassis, and DC ground are tied together and what happens when you attach you oscilloscope crock tip on the wrong point in circuit.

So say you'd like to look at the signal across a interstage coupling capacitor.  So you attach the oscilloscope probe tip on one side at random and then connect the ground clip to the other side.  In a bright flash you've maybe directly shorted 200V to ground and melted the crock tip.  How is this possible?  Even though your power transformer gives you galvanic isolation, you've bonded primary and secondary together with the typical earthing strategy.

The oscilloscope's exposed BNC connectors are tied to earth.  Maybe not an issue if you use a Fluke Scopemeter or other battery powered scope. Multimeter probing is much safer.

Your example is a quite unusual one.
Normally, the user would observe one side of the coupling capacitor w.r.t common (normally chassis).
They would then look at the other side to see if the signal was still present.

If they were fortunate enough to have a two channel scope & two probes, & really wanted look at the signal across the capacitor, they could probe both sides, & set the 'scope to "A-B", which will provide the required display.

An easy way to check what potential a point is at before probing is to measure between croc clip & that point with a DMM.

Analog power supplies are not complex, & it is quite easy to determine where the DC ground (common) is connected.
You then know that is the part of the device where you must place your probe ground croc clip.

If you are not particularly fussy about displayed signal integrity, & are basically chasing a signal which has "disappeared" between "point A" & "point B", you can just remove or tie back the probe ground lead, trusting that the chassis is returned to power Mains ground.

[Jwillis]

  Tube amps consist of AC signals over high voltage DC . The coupling capacitors are there to block the high voltage DC and pass the AC signal from one stage to the next . If you get no signal through the coupling capacitor it's either no good , to small of a value , the tube is faulty or not connected correctly  . So you won't be measuring any signal on the  high DC voltage side of a coupling capacitors or the output transformer with an oscilloscope . It's completely unnecessary for one , not to mention dangerous . Especially if that DC voltage exceeds the maximum voltage rating of your oscilloscope .  If you connect all your ground clips to the common ground of the Amplifier you can find all the information you need from the low side of all coupling capacitors alone . Even if your checking that the power filter is delivering  clean DC without a differential probe you can see that clearly when your low voltage signal is clean without AC modulation from the supply . Or if your still unsure connect a 100nF capacitor to the end of you probe to block any high voltage DC . 100nF is plenty to pass all of the AC signal to your oscilloscope yet block the high voltage DC.
  Even tone control will be low voltage because nobody wants to mess around with potentiometers with 100 to a 1000 Volts  passing through it . If your design is correct you don't need to measure any high voltage with anything other than a multi-meter.
  The only way to totally float your tube amp is to remove the earth ground entirely . But as soon as your connect the oscilloscope ground , the amp is no longer floating , so connecting the ground of your second probe anywhere other than where the first one is connected can be lethal because of the potential voltage difference  . Regardless if you have a isolation transformer or not .   

[vk6zgo]

  Tube amps consist of AC signals over high voltage DC . The coupling capacitors are there to block the high voltage DC and pass the AC signal from one stage to the next . If you get no signal through the coupling capacitor it's either no good , to small of a value , the tube is faulty or not connected correctly  . So you won't be measuring any signal on the  high DC voltage side of a coupling capacitors or the output transformer with an oscilloscope . It's completely unnecessary for one , not to mention dangerous . Especially if that DC voltage exceeds the maximum voltage rating of your oscilloscope .  If you connect all your ground clips to the common ground of the Amplifier you can find all the information you need from the low side of all coupling capacitors alone . Even if your checking that the power filter is delivering  clean DC without a differential probe you can see that clearly when your low voltage signal is clean without AC modulation from the supply . Or if your still unsure connect a 100nF capacitor to the end of you probe to block any high voltage DC . 100nF is plenty to pass all of the AC signal to your oscilloscope yet block the high voltage DC.
  Even tone control will be low voltage because nobody wants to mess around with potentiometers with 100 to a 1000 Volts  passing through it . If your design is correct you don't need to measure any high voltage with anything other than a multi-meter.
  The only way to totally float your tube amp is to remove the earth ground entirely . But as soon as your connect the oscilloscope ground , the amp is no longer floating , so connecting the ground of your second probe anywhere other than where the first one is connected can be lethal because of the potential voltage difference  . Regardless if you have a isolation transformer or not .   

Amazingly, people worked on, & probed tube equipment with 'scope probes for decades, myself included.
We actually knew how tube equipment worked, & our work practices reflected this.

The reason why coupling capacitors are used is because the signal on a tube anode is indeed ac superimposed upon a DC level.

The next tube stage requires a fairly low negative voltage between grid & cathode to bias the tube into the correct operating state.
Unlike transistors, a tube with no bias voltage applied between those electrodes will be turned on, & conduct reasonably heavily.

If the grid to cathode voltage becomes positive, the tube will conduct very heavily, probably to saturation.

Without coupling capacitors, tube stages would need to be hung "like beads on a necklace" across a very much higher HT voltage than is used in normal practice.

The ac signal is present at the anode of the previous amplifier stage, & can be observed there with a normal Oscilloscope using a X10 probe, either on "ac coupled", or in many  'scopes using "DC coupling", because they are provided with a variable "DC offset" function.

With those 'scopes having very low voltage rated input caps there may be a problem  on " ac coupled" IF you are looking at the raw HT supply rail, but in an audio amp, all the stages before the output stage have anode load resistors which will limit the current available to damage  the cap &/or input stage.

This is certainly not a problem with classic analog 'scopes!

Oh, by the way, in your original example, any damaging current through the ground clip will again, be limited by the anode load resistor.

i appreciate you may have limited experience with tube circuitry, & have bought "hook, line & sinker" into the "Ooo, scary!" narrative, but that is not a reason to try to refute information from those with that experience & theoretical knowledge..

[bob91343]

The number of times I have used isolation transformers is small.  Back in the day when I didn't have one, I would measure between chassis of my test gear, or other convenient ground, to the chassis of the DUT.  If that reading was close to power line voltage, I reversed the power plug and saw that the voltage became very small.  Then I didn't worry about it.

That only needed to be a consideration when working on cheap radios and other devices.  For the usual run of decent quality gear it was unnecessary.

The most recent incident related to an old Boonton Q meter that used a constant voltage transformer in the power supply.  Thinking how power line voltages have become so stable these days I eliminated that part of the unit, since it had failed anyway.  Surprise came when I felt a tingle when connecting to the Q meter.  I then learned that the constant voltage section also provided line isolation.  So I backtracked and repaired the supply and all has been well since.

The lesson here is to know what you are doing.  Don't rely on rules of thumb or standard practice or the seat of your pants.  As Louis Armstrong said (on a different subject), If you have to ask, you will never know.

[bdunham7]

The number of times I have used isolation transformers is small.  Back in the day when I didn't have one, I would measure between chassis of my test gear, or other convenient ground, to the chassis of the DUT.  If that reading was close to power line voltage, I reversed the power plug and saw that the voltage became very small.  Then I didn't worry about it.

That's a great plan until you turn the unit off and discover  that the chassis is now live again.  Or, you have a full-wave direct mains rectified system with the chassis tied to the negative DC rail.

[TimFox]

In North America, where the original poster lives, radio sets and televisions with vacuum tubes often lacked power transformers for 110 VAC mains.  HiFi audio amplifiers usually needed them.  Isolation transformers were common on test benches for radio repairs.  An additional feature of 5-tube transformer-less radios was their use on DC mains, which lasted until surprisingly late in some US cities.  I own a WWII merchant-marine short-wave receiver from EH Scott that was “AC-DC” due to DC mains aboard ship.
However, audio amplifiers with power transformers usually had two-prong power plugs, so the chassis was not necessarily grounded unless you connected the chassis to your oscilloscope, etc.