3

I've got an old (vintage!) GenRad Variac (below) that has two prong plug and two prong socket. I want to modify it to add a safety earth pin to both input and output. In doing so, I opened it up and have realized the circuit runs through the metal chassis itself, and I'm confused about what I'm looking at and would love to get a bit more clarity.

Have done a bunch of googling about chassis ground and hot chassis radios and such and I suspect this is just an old design, but I'm confused about why it appears in a Variac. Photos below showing the internal configuration, but the circuit is simple and as expected-- a 2PDT on-off switch connects both incoming lines to the transformer taps, and the other side of that is the output socket with a circuit breaker. The transformer mounting core and wiper mechanism is not isolated from the chassis (they are in contact at the front panel), so there's (multimeter-verified) continuity between the outer chassis and all taps of the transformer, the outputs and (when switched on) the inputs. (Pic showing continuity measurement added below.)

When I set the output to e.g. 50V, the potential between the outputs is 50V-- ie the Variac seems to work. When I measure the potential between each output and the chassis screws, one is ~50V and one is ~110V.

When I measure the potential from the chassis screws to real earth (wall socket), it's very low, < 1V, and when I touch the switch/case it doesn't shock me.

  1. This isn't a radio or TV, so there should be no RF interference concern here, so why isn't the circuit totally isolated from the chassis in the first place? Is this a form of safety intent? IS this design safe?

  2. I can replace the plug and socket with 3 prongs and connect the earth pins directly. I obviously would like to "ground" the chassis properly but I obviously can't here since that will allow current to flow to earth in normal operation and trip the GFCI-- right?

I'm trying to make sure I (a) understand what I'm looking at (b) it's working as designed and is, in fact, safe and (c) that I shouldn't mess with the chassis grounding when adding the third pin connection.

Thanks for insight or pointers.

enter image description here

enter image description here

enter image description here

enter image description here

enter image description here

BZo
  • 1,575
  • 1
  • 11
  • 22

2 Answers2

5

I don't see any connections to the case. There should not be any connections to the case.

Assuming there isn't, you should run the ground wire to the case and to the output receptacle. You can drill a hole in the case and attach both to the a screw with a lockwasher and nut.

Measurements made with a multimeter on a voltage range are not reliable since the multimeter is very high impedance and will pick up voltage from capacitive coupling. On a resistance range you can measure between input/output and case and there should be no continuity even on the highest range. Obviously you would do this with power off and unplugged.

You should not connect to either side of the output. A variac is an autotransformer and does not provide isolation. It looks like they are switching both sides of the mains. I think that's acceptable practice today (someone will correct me if not) but you should keep the neutral side (white wire) to the neutral side of the new socket. Modern receptacles usually have brass for the hot and silver color (eg. nickel plate) for the neutral screws. And green for the ground.

Edit: Well that ohmmeter photo you added is fairly convincing. I have to assume that the plug is polarized with the neutral pin wider- that would be a death trap of a product if not, and even so not all outlets are wired correctly

Okay, looking at some old data (this thing is perhaps 70+ years old) the diagrams actually show earth connected to one side of the line. This was sometimes done in radios of the all-American-five vacuum tube vintage but usually there was little exposed metal on them.

enter image description here

I would be inclined to find that connection and remove it, and ground the case.

Spehro Pefhany
  • 376,485
  • 21
  • 320
  • 842
  • Thanks. With the power off, unplugged, I attached the meter measuring resistance between each input plug prong and the exposed metal switch on the case. When the switch is off, infinite resistance. When the switch is on, I see 1-2 ohms. I'm not expert so I might be misunderstanding you, but that seems like continuity to the case. I added a photo to the question showing this configuration in case that helps. – BZo Mar 28 '21 at 17:37
  • Thanks again, for the info & ideas. The plug is not, in fact, polarized-- and it seems original. (The *socket* on the Variac is polarized, interestingly.) In an effort to figure out where the chassis connection actually happens, I removed the transformer core, and it sure looks to me like somewhere inside the sealed internals the wiring is connected to the chunky front aluminum mounting plate which gets bolted directly to the chassis. (See new photo). I'm getting a bit out of my depth as to what the ground-to-line stuff is doing above, but I don't know if I can disconnect this? :-| – BZo Mar 28 '21 at 18:05
  • Can you find a deliberate connection to the case or the frame of the variac? – Spehro Pefhany Mar 28 '21 at 18:05
  • The whole mount panel of the core is in contact with the chassis. My expectation would be to see no continuity between the taps and that mount panel (because there's some other evidence of isolation on the coil side) but there seems to be some internal connection within that block? (I have added a new pic showing this best I can) – BZo Mar 28 '21 at 18:07
  • Any indication of damage, like burnt smell? – Spehro Pefhany Mar 28 '21 at 18:08
  • Not as far as I can tell. Smells normal & no visible charring that I can see anywhere. I added yet another photo showing the rear coil side. There some sort of doped plastic insulating mat between the coils and the backside of the mount plate, and the inner tubular aluminum core (which runs through to the front) doesn't contact anything at rear when assembled (the rotating spindle from the front knob is not conductive). So if you're just eyeballing it, it sure *looks* like it's not supposed to have continuity to the chassis. But it also doesn't appear damaged at all. ¯\_(ツ)_/¯ – BZo Mar 28 '21 at 18:19
  • The fact the connection board markings say you can ground either end of the winding 2/4 indicates that it's intended to be isolated. Also the non-polarized plug which would be an unacceptable safety hazard even in 1950. Sorry. – Spehro Pefhany Mar 28 '21 at 18:27
  • Thanks, good catch. I guess what I'm taking away from this enlightening discussion is that not only can I not earth the chassis, I shouldn't use this thing at all, because even if I add a polarized plug something has gone wrong. There are no moving parts in the core which seems tightly bonded together, and so assuming there's no intentional connection-- maybe that doped fabric insulating mat has somehow decayed over time and become conductive? I remain somewhat confused as to how it even works (produces the expected output) if it is really damaged in some fundamental way like this. – BZo Mar 28 '21 at 18:37
  • In a word, "yes". A single short would not affect the output voltage, just the electrocution risk. – Spehro Pefhany Mar 28 '21 at 18:39
  • 1
    Well I'd rather keep that to a minimum, so points taken. :) I appreciate your helping me understand this. – BZo Mar 28 '21 at 18:40
  • You may be interested in what I discovered after having another go at this: https://electronics.stackexchange.com/questions/557212/whats-the-right-replacement-for-vintage-glass-cloth-insulation-in-variac – BZo Mar 30 '21 at 02:35
4

IS this design safe?

If this variac has only a two prong plug, and there is a connection between one of the prongs and the chassis, then the chassis is at neutral, not at ground.

IMHO, this is NOT safe. Hopefully you have ground fault circuit breakers or their equivalent in whichever country you are in. If the neutral wire should become disconnected, while the hot wire is connected, someone could get a nasty shock.

Math Keeps Me Busy
  • 18,947
  • 3
  • 19
  • 65