Chassis at 60VAC, otherwise just fine

[DaJMasta]

I'm refurbishing a 70s era strobe tuner and I've fixed a broken bit, added a mic element (it had only an input), and adjusted each of the control pots to get the thing properly calibrated and it all works as intended.  I took it to the basement (concrete slab floor) for a last shot of contact cleaner into the motor which still makes mild scratching sounds and noticed a slight tingle in my finger after touching the metal front plate.   Upstairs, with wood floors, I don't notice any tingle, but sure enough, when I reference it to mains earth, the chassis is at 60VAC, and the multimeter reads 1 microamp (1 least significant digit) over the same span.  It's spot on 60Hz as well, same as the mains.

The tuner in question has an old-style two prong unpolarized connector and the circuit does not have 60V on it anywhere intentionally (50V and 25V from the transformer into a 20V regulated and a 400V boost for the neon lights in the front).  I've read this can be related to polarization of the plug, but switching it around gives the same potential, and I've read that it can be related to leaky caps, but I can't seem to get a good fix on them in circuit.  they are 30 year old electros, but they aren't leaking or bulged best I can tell - about .8Mohm across the two.  When powered up they're at 50VDC referenced to the circuit, but are at 60VAC referenced to earth.... a little baffling to me, but I think I'm just missing a bit of the theory.


Anyways, the metal front panel really can't be at potential because it's the primary interface and people will be touching it while it's in use.  What should I be looking for as a possible culprit?  I can certainly go to town on the caps, but they're large, nestled in the middle of the whole thing, and have several tabs soldered to things, so they're going to be a pain to replace or remove for testing.  Schematic can be found here: http://medpants.com/images/Projects/StrobeRefurb/Peterson400Schematic.jpg

[Tom45]

I would start by measuring the DC resistance from each of the plug prongs to the metal chassis. Obviously you should get an open circuit for each of the plug prongs.

Next I'd test the two 0.01 uf caps on the transformer primary.

[DaJMasta]

Open circuit from the prongs to chassis, open circuit across both ceramics on the primary side.  Each measures 17nF in circuit, about 400 ohms ESR, and the transformer primary is about 12.5 ohms.  From what I can tell, that stuff is in order.  Would make sense that 60V would be related to primary side, though, given the ~120 V input

[DaJMasta]

It occourred to me that I could just desolder the positive side of the caps and not have to deal with all the ground tabs going to various things and made some measurements.

Both caps measure 2300 uF or so, but one measures 4M instead of open circuit, and one measures 1.6M. 


Well, replacing them looks particularly bad, but actually 3000uF 50V caps shouldn't be too pricey, the old ones are just pill bottled sized...

I think I'll even go for 63V ones since one of the rails is supposed to be 50V.

[Richard Crowley]

It sounds like you have leaky bypass/filter capacitors between the power mains input and the chassis. This is not a rare symptom.  It seems like an EXCELLENT excuse to chop off that antique (and dangerous) 2-wire power cord and put a modern 3-pin power cable on it with a proper green-wire safety ground connected to the chassis. That would kill your leakage dead before it kills you.  I am very fond of putting standard IEC power sockets on old gear so that I can use common power cords like used by every desktop computer on the planet.

[DaJMasta]

An IEC would be great, but this unit has a cord coming out the front plate, and has a wooden case that the whole thing resides in while operating, so there's not really sufficient space on the front panel and not a good place to route it out otherwise.  I agree a 3 wire would be nice, to be sure I'll ask, is there any potential problems with tieing the current chassis ground (primary and secondary sides) directly to earth?  There's no need for isolation so I don't expect there to be an issue, but I haven't added an earth to an old design before.

[wraper]

There should be no fault. You will get the same thing with most of the more or less modern DVD players, TVs, Laptops (when charged) and such if they don't use ground wire. Half of the mains voltage is because there could be a capacitor connected from each of the mains wires to the case. However considering the current measured is tiny 1 uA (guess you measure it at AC range), that could be just leakage through the capacitance of the linear transformer. Modern devices will have about 0.1 - 0.3 mA leakage current.

[amyk]

There should be no fault. You will get the same thing with most of the more or less modern DVD players, TVs, Laptops (when charged) and such if they don't use ground wire. Half of the mains voltage is because there could be a capacitor connected from each of the mains wires to the case. However considering the current measured is tiny 1 uA (guess you measure it at AC range), that could be just leakage through the capacitance of the linear transformer. Modern devices will have about 0.1 - 0.3 mA leakage current.

Indeed, 1uA is far below what the safety standards require for leakage current limits. I agree that it's probably leakage through the transformer, and the 0.01u caps are open, since they would otherwise contribute (an unsafe!) ~450uA of leakage current.

http://www.aplomb.nl/SMPS_leakage/Doc_ie.html

[vk6zgo]

On the schematic,following the mains wiring from the power plug,you come across two 0.01 uF capacitors --one from each side of the Mains to "Ground".

In a normal (US) home wiring system,one of the incoming two conductors will,if measured,w.r.t. the building Protective Earth (PE),read 120v RMS,(this will be the Active ,or "hot" conductor),& the other one,very nearly zero volts.(this will be the Neutral,or "cold" conductor).

The two equal value capacitors are connected in series between 120v RMS &  "near as dammit" zero volts w.r.t the PE.

The junction of these two capacitors is according to the schematic connected to "ground",which in this case is the equipment chassis,& is not connected to PE.

It is thus, free to take up whatever voltage w.r.t. PE that is present at the junction of the two capacitors.
The Capacitive reactance of each capacitor is the same,& the capacitive voltage divider so set up will divide the 120v RMS equally ,so the chassis will be at 60v RMS w.r.t PE.

As the value of Xc of each cap is around 300kOhms,there is no real danger,but you will get a bit of a "zap" from it.
The best answer is to modify the device to use a 3 conductor power cable,as was suggested by an earlier poster.

[helius]

The above post is correct: there is no danger whatsoever from the 60 VAC measured on the chassis. The reason that the equipment needs a Protective Earth connection is in case of a fault or insulation failure developing, with a low-impedance path from Live to the chassis: the PE connection completes the fault path back to the source and blows the panel fuse or circuit breaker. Grounding provides safety in fault conditions, not normal conditions.

[DaJMasta]

Interesting that it's probably leakage from the transformer, and it sounds like an earth lead is the right solution beyond just sort of dealing with it.


Is there a good rule of thumb for checking large electros in the future?  I was a bit alarmed by the removed-from-circuit 1.6Mohm resistance of one of them, but after measuring some other known-good electros, It doesn't seem unreasonable or even low.  Is it a few dozen Kohms where they start to be a replacement candidate, a few hundred?


Well, I went to measure them one more time to be sure, and the capacitance range on my multimeter made one lead bubble when it was charging up to measure.... so they're definitely being replaced.  Regardless of the measurements, I don't want my caps leaking from the leads with just a volt or two across them   

[helius]

Were you testing the leakage with a DMM on resistance mode, or with the cap charged using a DC power supply? There is a huge difference.
Regular DMMs aren't very accurate for testing leakage. But it sounds like you found one way to find it (were the leads the right way when it was bubbling out?)

[vk6zgo]

Interesting that it's probably leakage from the transformer.

 

Did you actually read my posting,or the one from helius agreeing with it?

The chance of "leakage from the transformer" causing the chassis to take up a voltage exactly half of the supply voltage is right up there with that of me "flying to the moon on gossamer wings",whereas it is precisely what would happen due to the "capacitive voltage divider" effect I described.

If you completely disconnect the transformer primary,you will still see 60v RMS between chassis & the Mains Protective Earth (PE).

[DaJMasta]

So then every design with these primary side capacitors tied to a floating chassis point will exhibit this problem? It seems strange to me just blame the caps in question when they measured as they should (in value and continuity) and I wouldn't assume that a design should just be fine like this.... the transformer leakage was a reference to an earlier post and was something that I hadn't considered and seemed plausible enough given my level of understanding (which as you can probably see is incomplete).



Anyways, the capacitor measurements (value and resistance) were done with a DMM, and I had the leads on the right sides for testing, I've got an LCR meter but it wasn't ranging to pick up the cap in question and I didn't spend much time fiddling with it.... with later measurements with the DMM, it could have actually just been bad contact points.  I haven't used a DC source to test capacitor leakage before, is it something like applying a voltage with a meter inline and measuring current on the meter?

[madires]

Those are standard filter caps and they do what they are intended for. The real problem is that the chassis has no ground/earth connection to the wall socket. So the two filter caps act as a voltage divider (120V / 2 = 60V). And since Neutral is connected to ground/earth at the mains distribution box, you see 60V AC between the chassis and earth. The solution is very simple, swap the 2-wire power cord with a 3-wire one (this implies a plug with ground pin).

[helius]

I haven't used a DC source to test capacitor leakage before, is it something like applying a voltage with a meter inline and measuring current on the meter?

Yes, that's it. Many LCR meters even have external power supply points for that.

The solution is very simple, swap the 2-wire power cord with a 3-wire one (this implies a plug with ground pin).

Good solution, but wrong problem: there is no reason to "fix" the 60V AC on the chassis. What needs to be fixed is the possibility that a wiring fault could actually make the chassis live without blowing any fuses.

[madires]

The solution is very simple, swap the 2-wire power cord with a 3-wire one (this implies a plug with ground pin).

Good solution, but wrong problem: there is no reason to "fix" the 60V AC on the chassis. What needs to be fixed is the possibility that a wiring fault could actually make the chassis live without blowing any fuses.

60V AC on the chassis is no problem? Sorry, the original design is flawed. You put an X class cap between Line and Neutral in case of an ungrounded device, not two Y class caps between Line and chassis, and Neutral and chassis. That implies a grounded chassis.

[wraper]

60V AC on the chassis is no problem? Sorry, the original design is flawed. You put an X class cap between Line and Neutral in case of an ungrounded device, not two Y class caps between Line and chassis, and Neutral and chassis. That implies a grounded chassis.

You put both X and Y capacitors, they don't replace one another. And if the chassis are not grounded (double insulated device), Y capacitors are still used in most cases. Y capacitors do not imply grounded chassis.

[madires]

If I got that right, the device we're talking about got a metal front plate which is not isolated.

[wraper]

If I got that right, the device we're talking about got a metal front plate which is not isolated.

Double insulated device does not mean insulated case or insulated front panel, it may have all metal case connected to the mains voltage via Y capacitors. Double insulated just means that insulation fail in single point shall not cause complete failure of insulation.

[DaJMasta]

If I got that right, the device we're talking about got a metal front plate which is not isolated.

For reference, yes this is the case.  The chassis which holds the electronics includes the faceplace with user-accessible adjustments, then the whole thing is normally encased in a wood road case.  Looks like this:


Interestingly, though, this unit has a three prong power cable.  While there's certainly a couple of screws that would make good earth tie points for the chassis, I also don't see any evidence of there ever being an earth wire going to any of them.... wonder if there was a change in the way they were shipped out over the years or if mine had the cable replaced at some point.

In fact, that pic may be of a later model, as the circular dark area above the two dials isn't present in mine and is an internal mic port in the 420 version.  Mine looks exactly like this: https://upload.wikimedia.org/wikipedia/commons/e/e4/Peterson400Tuner.png , and that shows a two wire cable.

[wraper]

In fact, that pic may be of a later model, as the circular dark area above the two dials isn't present in mine and is an internal mic port in the 420 version.  Mine looks exactly like this: https://upload.wikimedia.org/wikipedia/commons/e/e4/Peterson400Tuner.png , and that shows a two wire cable.

That cable does not look like original as it is pretty loose in the hole with a huge gap.

[madires]

Are the two filter caps simple ceramics? If so, please replace them with proper Y class caps.

[Seekonk]

That looks absolutely space age compared to my Srobe-A-Conn.

[vk6zgo]

Are the two filter caps simple ceramics? If so, please replace them with proper Y class caps.

There is nothing magic about Y class capacitors---they are simply capacitors which have been verified to meet all the specifications of the  "Y" classification.
A bit like the CAT standards for DMMs.

The fact that a device hasn't been checked by  a particular authority to correspond with some EU standard doesn't mean it is inferior.
The capacitors are shown on the schematic as 1kV DC rated.

That is probably the working voltage,(vW)but lets be conservative & say it is  peak voltage,(vP),that implies (from memory),a working voltage of around 750v DC.
Again,from memory,it was common to use 600vW at 250v RMS,(354vp),so these caps should be well within ratings at 120v RMS.(170vp)

Of,course,at the moment,they are loafing along with 60v RMS (85 vp)  across them.

It was common to use capacitors between chassis & both sides of the Mains input in quite a lot of US equipment.
They probably assumed that most people would not get between the chassis & a good earth,& that it was only nuisance value,anyhow.

Equipment designed for the US split phase 240v system can be a bit of a trap---as such capacitors hold the chassis at  "virtual earth " potential due to the "balanced" nature of the supply,so despite the higher voltage,the chassis is not "hot".

These are compatible  with the Oz/EU 230v systems,but the chassis is now at the midpoint of a capacitive voltage divider across 230v RMS.
I remember buying a Drake SSR HF radio receiver,& excitedly setting it up with a makeshift antenna & earth,only to see a spark when I connected the latter.
The 2 wire cord was very quickly replaced with a 3 wire one!

Even in this case,there was no real chance of electrocution,but the zaps would be decidedly unpleasant.

Domestic "mantel" Radios in Australia mainly used 2 wire leads for many years,but they didn't have anything connected to the Mains input wiring except the "on-off" switch & the power transformer primary,so "zaps" did not happen.

A piece of equipment which is used in a workshop environment like the OP's should definitely have an earthed chassis,however,as the likelihood of rough treatment causing an internal short between Active & chassis is greater.

As it is a fixed cord,it would be simple to fit a 3 wire one.