1
\$\begingroup\$

I have troubles understanding how do metal cases behave in some cases.

Case 1:
I made passive attenuation box for amplifier(basically L-pad resistor), simple configuration:
Amplifier => attenuation box => speaker It has 1/4 jack sockets on input and output like this one
enter image description here
And pot resistor with metal on it's case
enter image description here
Everything works fine, but as soon as I touch case or any metal part of pot with my hand it starts screaming from a speaker high-pitch sound. Why does it do that? Other commercial equipment uses the same jacks with metal cases and it has no problems with touching the case. And I can't ground it obviously because it is passive and input is floating from transformer.

Case 2:
On other powered devices without ground connection(example - guitar FX pedals) you also have metal case with metal jacks and it doesn't have any problems with touching either metal case or pots.

In both of these cases I don't see what's the trick here, at first I though I wire something in a wrong way, but there is no way to wire socket in such a way so that sleeve would not touch case - since the outer plane of socket touches case, obviously sleeve of connector will touch it no matter how you wire stuff internally.

So basically the questions are - how to properly wire metal parts that touch the case in order to make case out of the circuit.
I really tried to google that by myself, but it seems that most of people have no troubles at all with it, maybe I miss very basic thing.

Thank you for reading such a long question.

\$\endgroup\$
5
  • \$\begingroup\$ This does seem unusual - at speaker voltage levels you shouldn't be able to hear human noise. Perhaps you're acting as either an antenna or your capacitance and either causing feed back into the amplifier somehow - can you give details of the amp? \$\endgroup\$ Commented Sep 23, 2014 at 22:20
  • \$\begingroup\$ It's 100 Watts tube guitar amplifier, output voltage is about 25 Volts RMS the highest. It is quite probable that I'm an antenna in this circuit, but still the question is on how does commercial equipment goes around that problem. And yes - the noise is notable even not on max volume(actually even on very low volume it screams loud enough) \$\endgroup\$ Commented Sep 23, 2014 at 22:31
  • \$\begingroup\$ I would question why you're doing this at all. Amplifiers have volume controls: why do you need another one? especially one that wastes power? \$\endgroup\$
    – user207421
    Commented Sep 24, 2014 at 1:36
  • \$\begingroup\$ @EJP, just google "tube amp attenuators", not a hard thing to find. To be short it has to do with tubes perform poorly on low output volume, hence you crank it loud and then lower it down. \$\endgroup\$ Commented Sep 24, 2014 at 2:38
  • \$\begingroup\$ @ScienceSamovar Tubes don't perform poorly at low output. It is more probably the reverse: tubes performing poorly at high output, which the musician may want, so he cranks it up and attenuates. \$\endgroup\$
    – user207421
    Commented Apr 19, 2015 at 1:10

5 Answers 5

1
\$\begingroup\$

It's a common problem any time you get a high power output signal anywhere near the instrument level input signal. It's an induction/radiation problem, and it basically causes feedback. Always keep your instrument and its cabling away from your speaker wiring.

\$\endgroup\$
1
  • \$\begingroup\$ yeah, it turned out to be feedback problem form guitar pick-ups, I didn't expect that O_o Basically, electrical signal travels from output attenuator metal casing through your hands and through air to pickups. That's the real problem, done experiment. I marked your post as and answer, though I still want to ask you - what to do about it? Do I leave casing energized like that or what? \$\endgroup\$ Commented Sep 28, 2014 at 7:24
1
\$\begingroup\$

The simple answer is to use a plastic enclosure. The next simplest answer is to buy parts that don't make electrical connections to the case. You may have to look at data sheets to figure this out. With some connectors, like BNC connectors, there may be a description of "Isolated" in the part name somewhere.

Other approaches would really require you to look at the data sheet for each part as you buy it, know what electrical connections get made to the case, and be ready to handle them. For some parts, insulating hardware for case mounts is available. For others, no such luck, and you have to jury rig your own. If every connection to the case is Ground, for example, you should be OK (but there may be electrical safety concerns).

For best safety, Plastic enclosures and no metal parts that can be touched by a user is the way to go (double insulation).

\$\endgroup\$
3
  • \$\begingroup\$ Unfortunately in some cases you just can not have plastic case because the device gets really hot and will melt the plastic. The solutions you described is quite obvious, but the interesting thing is that there are many devices with metal cases and metal jacks installed and they somehow figured out how to make things work :( \$\endgroup\$ Commented Sep 23, 2014 at 22:27
  • \$\begingroup\$ Then, the answer is easy. If your devices get hot enough to melt plastic, you shouldn't get near enough to them to be able to touch the case. \$\endgroup\$ Commented Sep 24, 2014 at 0:03
  • \$\begingroup\$ The problem is not me touching the case, the problem is metal case acting as an antenna even though not a good one, but it is still not a good practice to have something in the signal that doesn't belong there. That's not an appropriate answer, if manufacturers would write like that in manual "don't touch the case, it makes devices useless" I doubt they will sell even one of them :) \$\endgroup\$ Commented Sep 24, 2014 at 0:11
1
\$\begingroup\$

You're upsetting the load seen by the amplifier's output stage, probably by adding shunt capacitance, so you're causing an oscillation, at the resonant frequency of the speaker inductance and the total shunt capacitance (or some other frequency derived more complexly).

Solution: isolate the case completely from all connections and components. That potentiometer looks like it may carry connections through to the metalwork.

Another possible solution may be to add a Zobel network across the output inside the amplifier, typically 10R in series with 1nF. Tube amplifiers generally don't have these: solid state ones almost invariably do, and some also have a small series inductor in the output path.

\$\endgroup\$
0
\$\begingroup\$

This is perhaps a comment, but I hate plastic enclosures.

Why can't you ground the metal box? If the whole thing is floating with a lot of gain, then it should sing when you touch it... (maybe it would help with the florescent lights turned off?)

Ground is both your friend and enemy. The good news is that if you make a nice metal box there is both an inside ground and outside ground. (above a certain frequency... and only electrically, not magnetically, unless it is a thick steel thing.)

Getting it all right is a learning experience.... I could tell a story about a steel lock washer on the wrong side.

Edit: I should add that I have seen EMI leaking down metal shafts when you touch them.
(our solution is a trade secret, and would cost at least a beer. :^)

\$\endgroup\$
9
  • \$\begingroup\$ Of course, it's the easy way, but then you have a power cable going to the wall just for the wire, which just looks stupid and adds one more wire. And the case here is - why do I need to make plastic case/ground wire/plastic connectors/etc. when there are tons of equipment that use all metal stuff and feel great? :) Just first links from google: suhr.com/jim-kelley-power-attenuator.html amptone.com/powerattenuatorfaq.htm nutsvolts.com/index.php?/blog/post/the_lost_world_of_tubes It is clear that it is possible to do so, question is how? \$\endgroup\$ Commented Sep 23, 2014 at 23:14
  • \$\begingroup\$ Just to be sure -the answer is not "painting the case", because paint gets ripped off and inside of a drill holes will very likely have a contact with metal anyway. Not to mention that these things may travel a lot and 40% of paint is gone after 4-5 years of heavy usage. \$\endgroup\$ Commented Sep 23, 2014 at 23:17
  • \$\begingroup\$ No, it does not plug into a wall, it is passive, just attenuator(resistor in a box if you prefere), it does not need any power supply of any kind. And even if it does, commercial equipment(like FX pedals mentioned in question) do not use any ground connection at all either, so there is no point of talking about grounding it through outlet. \$\endgroup\$ Commented Sep 23, 2014 at 23:25
  • \$\begingroup\$ connector is connected to output of transformer of an amplifier, there is no ground in there either. Transformer is floating by definition. \$\endgroup\$ Commented Sep 23, 2014 at 23:32
  • \$\begingroup\$ @ScienceSamovar Do you have the amplifier's schematic? Sometimes one side of the secondary winding is grounded and sometimes not. \$\endgroup\$
    – venny
    Commented Sep 23, 2014 at 23:34
-2
\$\begingroup\$

The sleeve of a phone plug and the outer conductor of a COAX connector should be connected to the ground/case of the equipment. If you have a poorly-made 1/4" phone plug cable that interchanges the shield/outer connection and signal/center connection you would likely have the problem you describe.

The case of your accessory should be connected to the case/ground of the equipment it is used with.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Speaker cables have no shield. \$\endgroup\$
    – venny
    Commented Sep 24, 2014 at 2:25
  • \$\begingroup\$ This device is connected to an output transformer which may not even have a well-defined ground. \$\endgroup\$
    – user207421
    Commented Sep 24, 2014 at 2:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.