[ntsc525]

Hi, everyone:

I believe this question applies to most transformer sets with AC line bypass capacitors between the line and chassis, that require a ground connection. The issue is AC "tingle" (leakage) between the chassis and ground, and introduction of noise when the chassis is grounded.

In the case of the 70 (and probably most sets from the same period), the value of the bypass condensors (#47 in the 70 schematics) is .015. When I did the first set, I used plain polyester caps at .015, but on the second one I had safety caps (X and Y), and combined a .01 and .0047 to keep the original values. These fit inside the block. When I pull the first chassis out again, I'll replace the yellow caps with safety ones.

On the first 70 I did, I decided to install a nice cloth covered three wire grounded power cord and ground the chassis internally so I wouldn't have to provide a ground connection to the lug on the chassis. I noticed the set picks up hum on some stations (especially my home transmitter, a Talking House prone to hum anyway), but didn't think much of it until I did my second set.

The second set has a nice original power cord that's still supple and intact, so I left it, and I built a small adapter to plug into a 3 prong outlet, tap the ground lug and run it as a separate ground wire to the chassis.

So now with all that background, we have this:

Radio on the bench with a short wire for an antenna, which is enough to pick up the So Cal major stations (KNX and KFI) plus a handful of weaker stations in between.

I noticed two things when I connect the ground cable: Some stations (KFI, whose antenna I can see from my house) get weaker, KNX has increased noise and hum, and on some of the weaker stations, they may get stronger, stay the same, or get more noise.

Also, making the connection draws sparks and makes noise on the radio. I get about 60-something AC volts, in the under 100 milliamp AC range, which is enough to "tickle" but not a major shock hazard.

Through experimentation, I determined that adding a .0022 uF capacitor between the ground wire and the lug improved reception, as it's a small enough value to provide fairly high resistance to 60 Hz but still reasonably pass 550 KHz. However, in this configuration, one would still get a small shock if they touched the chassis and ground, and they'd feel the same thing from the antenna lead, which will carry the same "tickle" voltage from the chassis through the first RF transformer.

So, my question is, do I leave the cap values as designed? Or, would an ideal bypass condenser value for the AC lines be smaller, say .01 to reduce the amount of leakage to ground?

I like the idea of a grounded plug on an AC set that may come into the hands of non-technical people, because it prevents people from feeling the leakage, and all they have to worry about is the long wire antenna. However, I don't like the noise it introduces. I thought about modding my set with the grounded plug to put .0022 uF between ground and the chassis, but that would expose people to the 60 V tingle which they may find objectionable. (They'll also feel the tingle from the antenna lug.)

If a set is going to have a grounded plug, I think the ground must be tied to the chassis, since the original design calls for grounding the chassis, but now I see what's really happening through the bypass caps.

I know some people object to the idea of a three wire plug on any antique radio, but electrically it is exactly the same thing as grounding the chassis to a water pipe which was the original design, and no doubt a good ground kept people from feeling the tingle from the chassis.

I'm no electrical engineer by any stretch, but I did discover this page for calculating a capacitor's impedance for a given frequency (https://www.allaboutcircuits.com/tools/c...alculator/), and at 60Hz, the .015 uF bypass caps provide about 177K Ohm "leakage" (impedence) to ground, which, I'm guessing is enough to feel a tickle from an ungrounded chassis.

Anyway, I'd like to hear opinions, because the ungrounded chassis is still out, and I'll have the grounded chassis out soon to move to a better cabinet, so I have an opportunity to make changes.

Thanks.

Dan

[morzh]

100 mA? 50 is enough to kill. Your values should not draw even near this.

[ntsc525]

I stand corrected. I was reading milliamps as amps because of where the meter put the decimal, and with another meter I confirmed I was off by a factor of 100. Today I got about .7 mA. Still enough to tickle, and still enough to make little tiny sparks if you drag the ground wire across the chassis.

As for the statement of how much current can kill, I will remind you my measurement was to ground, not through my body. I think my body draws only a few microamps, so even touching a full 15 Amp service, I'm still OK. (I wouldn't have made it out of Jr. High Electronics if that wasn't true!)

I've been shocked off the antenna lead before, on an unrestored radio because it had a bad ground connection and I assumed the line bypass caps were leaky. But now I think the tickle was normal if the chassis isn't grounded.

And just to be sure I'm seeing what I think I'm seeing, I assembled a couple of .01 caps in a Y configuration across the line and measured about 60 Volts to ground, and about .4 mA between the apex and ground. It was still enough to feel. The current going through my body wasn't enough to register because, as I've been told by my boss, I have high resistance! ;-)

So, I guess I've answered my own question about it being normal to feel a potential between the chassis and ground, and reducing the bypass caps to .01 doesn't help much, so the only thing left is to get opinions from others who have seen the issue with hum, and what solutions you've come up with for hum and noise. (Besides making a loop, which I understand would reduce static to some degree due to some "magic" I don't understand about propagation.) Right now, a .0022 uF cap between the chassis and ground seems the best compromise, providing a decent path for RF to ground with a high impedance at 60 Hz.

For my own use, I'm not as concerned about a few microamps of current as I am about the fact that if I connect the ground as designed, I get more hum and noise.

I guess for now, nothing to do but put 'er back in the case and use my longer wire antenna and see where that gets me.


Dan

[morzh]

As for the statement of how much current can kill, I will remind you my measurement was to ground, not through my body.  

Yes I know that. It'd be stupid otherwise. I just gave you the killing numver vs what you read on your meter which, given the caps values, is impossible.
10nF seems to be a good value; if want to have no tickle at all, about 2nF would still do the job.

[Steve Davis]

I always .01 safety caps. Unless you are barefoot and standing on a concrete floor, you shouldn't feel a shock when you touch the chassis. At least, I can't.

Steve

[David]

When working on radios I do not like feeling a tingle. Have a nice thick rubber mat in my work shop about 3/4 inch thick from tractor supply if its good for a horse its good for me. Less than 50 millamps can damage nerves!!!!! Please be careful. David

[morzh]

I always .01 safety caps. Unless you are barefoot and standing on a concrete floor, you shouldn't feel a shock when you touch the chassis. At least, I can't.

Steve

Steve,

you simply got desensitized.

[Electrothaumaturgist]

I blew the circuit breaker for the entire upstairs of my house in 1978, playing around in the backside of a working 1929 RCA Victor R-32 console.    That was the first and last time I ever got walloped to that degree. Something in me changed that day, as I made a resolve to never let that happen again, and that little incident may be the reason you all are graced with my presence here in this Phorum today!  

[morzh]

I do not know if I told this here....when I was.....10, I think, or even 9, I decided to make a soldering iron, having (!!!!) a bare nichrome wire wrapped around a metal rod with a wooden handle, connected to a plug. All bare, mind you - no insulation anywhere. The voltage was 220V. And then I plugged it in.
Blue pieces of the wire flew in total darkness.
By some God's miracle, I 1) did not get shocked, 2) did not get any of these pieces of the hot nichrome wire in my eyes, or anywhere on me, 3) did not start the fire.
Didn't even get scared.
I went up the fuse panel (2 fuses, you know, screw-ins, maybe 15A each) knowing full well the fuses have been replaced with what we called "bugs" (pieces of wire wrapped around the burnt fuse), and simply did exactly that (at 9 years I knew that much), and when my parents came home, they didn't find out a thing!

I was suspicious about non-insulated wires, so this only reinforced it, so I had become a bit more careful since.

[cpollock]

I experienced some arcing last night when I tried attached my scope ground to the chassis, so I was happy to see this post (I'm playing with a 38-7).     I figured it is was due the two 0.015 caps on the power line, which is exactly what is discussed above.   The caps form a voltage divider for the 60 Hz line frequency, and the impedance, as noted above, calculates to 176kohm, so you would expect, as Dan observed, to find the chassis about 60V above ground, and able to pass a peak current of a milliamp through a good conductor connected to ground (like my scope probe). 

This got me wondering, what is the purpose of those input capacitors?  Is it from an older time when the power lines were a lot noisier? Is it to shunt RF signals trying to leave the radio and go back to the power grid?  The latter possibility seems unlikely because the power transformer probably blocks most RF stuff, and also I doubt FCC regulations of RF emission even existed back in the 30s, at least not to the degree they exist today on things like laptops, so its doubtful Philco was trying to deal with that. 

Does anyone know the reasoning behind these things?  I can't figure out what advantage these caps offer when the "cost" is putting a high impedance 60 VAC on the chassis.  

I have seen some schematics posted in the Phorum that the cap was only connected between the two power lines, and not to ground. For shunting noise on a power line (if that is the goal) this probably is good enough.   I'm going to isolate the caps from the chassis  and see if I can sense any changes.  I'm also thinking, as discussed by Dan, of going to a smaller cap.  I just can't see much advantage to this capacitor.

I'd really enjoy learning more about the reasoning behind the "bakelite capacitors", if anyone knows the story.
Thanks,
Clif

[Steve Davis]

The purpose of these capacitor is to provide an RF ground path for the antenna circuit. If you look at the antenna transformer, you see that it is connected to chassis ground. The line capacitors connect the chassis to the power line, thus providing an RF ground through the power line.

Steve

[cpollock]

Thanks, that makes a lot of sense.  At 1MHz it is equivalent to a 1 ohm short to ground, so the size of the cap makes sense. It could probably be smaller, perhaps 1/3 that size and still function, but clearly it needs to be in the 0.01 uF range to serve as a short to ground for the AM band.

 If I were to use a polarized plug, where one side is always neutral (assuming the house is wired right), would it be sufficient to only put a cap to the neutral?  That would eliminate the "AC tingle" but still provide a path for RF to ground.

Or, on the other hand, if a current path through the second antenna terminal, such as running a line to ground ground, or using a loop antenna,   is provided, are the caps necessary?.

Clif

[Radioroslyn]

No because no matter which way the plug is inserted you've got the hot side of the line connected to the chassis thru one of the  two caps. If you connect the chassis to earth ground that will remove the tingle.
The cap that is going from the neutral side to the chassis provides the rf ground for the rf circuits in the set if you don't have it tied to an earth ground. The hot side cap provides the rf filtering to help clean up rf noise on the line.
As for the power transformer acting as filter inductor at rf not so much. There is this thing called distributive capacitance where the windings from pri to sec couple the remaining noise like a cap would. RF noise is insidious!!

GL

[EdHolland]

Well, If (and it is a substantial if) one could depend on the polarised plug always providing neutral on the same wire in the line cord the following idea could work in theory:

Connect a cap from neutral to chassis, and a separate cap across the line from live to neutral. Both must be safety types.

In this way, the capacitive divider effect would be avoided, no 60v on the chassis, and both lines would have an RF path to the chassis.

However, if the plug or socket wiring was incorrect, the capacitor would provide 120v on the chassis, worse than before - hence my "if" at the beginning of this post.

[morzh]

Well, it can tingle you, but won't really shock: 120V is through large capacitive resistance. But yes one could feel it.