| Arx |
I just finished my gainclone and am noticing a really weird phenomenon. The chassis feels like it's got a slight voltage running through it when you brush your hand against it. It only happens when the power supply is connected (it's external), but i've checked and there are no shorts to chassis, and checked each terminal to chassis with my vom on the megaohm range.
Any guesses what could be the cause? I was thinking perhaps it was in the magnetic field of the transformers, but it doesn't seem to matter if I move it a couple feet away.
-Nick |
|
|
| quasi |
If you can "feel" a voltage (apart from static electricity) then you have a potentially dangerous situation. You should check the voltage against an earth reference with your meter. It could be that you just need to earth your chassis or it could mean something else. So for peace of mind try to isolate what it could be.
Cheers |
|
|
| Arx |
I measure a couple volts when I've got a source connected, and around 20vac with nothing connected.
what I'm really wondering about is where the voltage is coming from.
you can hardly feel it, it's most obvious if you lightly drag your knuckles across the chassis.I'm not particularily concerned. I haven't got the mains supply shorted to the chassis or anything. It's running off a pair of transformers which aren't in the chassis, so it's isolated there, and I'm sure there's nothing shorted (and besides that, what could short
the question is merely, how do you get 20vac in a floating chassis that isn't shorted to anything.
could it be rf? if so, it feels like it's modulated at 60hz, since I can definitely feel a buzz.
I can also feel it at 2v, assuming my meter's not lying, which I think it likely is.
I wouldn't have noticed it at that low level, but since I know it's there I can feel it if I'm trying.
-Nick |
|
|
| quasi |
Most digital multimeters have too high an input impedance to accurately measure "stray voltages" because they are susceptible to RF and other electromagnetic inteference. Try putting a 47k resistor across the meter leads while you measure. This will give you a more accurate reading.
Cheers |
|
|
| Arx |
| quote: | Originally posted by quasi
This will give you a more accurate reading. |
Unless it is HF noise which I'm feeling ;)
Maybe if I can't figure it out otherwise I'll hook it to earth through a series resistor and cap and see if I can discern the frequency range. |
|
|
| Nordic |
| How is your starground connected to chassis? |
|
|
| pwillard |
Yes, it sounds like you have your MAINS wired incorrectly. You might want to recheck your 3 wires coming from the wall plug and make sure they terminate at the correct locations in your case.
PW |
|
|
| dnsey |
| Could be a leaky mains transformer, but if you're using a type which can develop such symptoms, you should also use a grounded chassis. |
|
|
| sangram |
The chassis should be shorted to mains ground. If you're not shorting *anything* to the chassis you should add a small tag on the back of the chassis and connect it to a good ground reference (such as the chassis of the CDplayer/preamp).
Sometimes a fully floating chassis may have some voltage on it due to induced voltage from a component somewhere else. Chips have a dielectric between them and the chassis. A bolt may be very close to a PCB track. There may be some voltage on the shield of an upstream component, with the sockets connected to chassis. |
|
|
| ralphs99 |
In countries with 220V mains...
For safety, a chassis made of metal or with metal parts containing components at mains potential must be protectively Earthed to the mains Earth. (with exceptions for double insulated equipment)
The green/yellow mains Earth wire must be strain relieved and securely connected to the chassis immediately it enters the chassis. The wire should be terminated in an eyelet by crimping and passed through a screw, nut and star washer combination. Make sure of good contact between the chassis and screw by removing any paint or insulating coatings on the chassis under the screw assembly.
You should also check that your mains outlet is Earthed. I have come across outlets and extension cords that have been mis-wired or are missing an Earth.
I'm not sure of the specific requirements for Canada, but following the European 220V requirements is good practice even when they are not strictly necessary eg the US market.
The cause of the symptoms you describe can be a number of things, most probably parasitic capacitances in the power transformer or Y-capacitors if a mains filter is employed. There are specific requirements for maximum allowable Earth leakage for various equipment types. If in doubt you can assume 0.5mA.
Cheers,
Ralph |
|
|
| jpg |
| Just curious: have you tried plugging in the mains cable the other way around (phase/neutral reversed)? |
|
|
| Arx |
I'm sure hooking an earth ground to the chassis would remove the voltage, but it still doesn't explain it.
As I mentioned in the original post, I'm using an outboard power supply. The mains is all correct, and so far as I know the transformers work correctly and don't leak. Even if they did, there should be no connection to the chassis. (unless maybe capacitively or something)
I'm using the basic lm3875 kit from audiosector.com It uses the insulated package chips, which are screwed to the aluminum chassis, and the board is so small that rather than screwing it down, I just hold it away from the chassis with some stick on rubber feet. (It would hang fine even without the feet.)
All of the connectors are xlr, with the bodys not connected to anything so there's no direct electrical path to the chassis there either.
I am 100% sure it's not mains in the chassis, though it does certainly feel as though it has a noticible buzz (60 Hz?)
My guess would be that the switching of the diodes in the supply may be throwing rf off the supply leads, since there's very little capacitance on the rectifier board. These rf pulses would occur at the line frequency.
Any thoughts?
And to be doubly clear. I'm not just trying to remove the voltage, but would like to know how it gets there. |
|
|
| SY |
Leaky power transformer, as dnsey said, is most likely. It does happen. Sometimes caps used to ground the power line at RF can get leaky, too. I know you said that your transformers don't leak, but (no offense), I don't believe that, unless you've got clear measurements indicating otherwise.
In any event, it is extremely unsafe to have an ungrounded metal chassis connected to circuitry. If you're using a separate supply with a multiconductor cable between the chasses, at least one of the conductors (preferably a nice thick one) should connect chassis to chassis, and the power supply chassis should have a SECURE connection to earth ground. If you have hum issues due to a ground loop between different pieces of equipment, you can float the signal ground away from the chassis via a small (10-100 ohm) resistor. |
|
|
| Nordic |
| Well I can now add, I just experienced first hand, where one rail of the live cable touched the chassis... gave me quite a fright as I wasn't expecting any shock working on the other side of the near empty case. Fisrt thing I was to test between the case and the pins on the power plug for continuity and confirmed the short, luckliy only one side of the cable shorted.... needless to say I have corrected the situation... thats what happens when you rush to get done around rolling power outages... |
|
|
| Arx |
| quote: | Originally posted by SY
Leaky power transformer, as dnsey said, is most likely. It does happen. |
Yes, but as I said before, the chassis isn't connected to anything. there's a 3 conductor connection between the power supply (which isn't even in a chassis yet, just set up to test) and the amp, via an xlr connector. (no chassis connection) Even if the transformers were leaking 120v straight through, how would it get to the chassis?
The inputs are also on xlr connectors, with the chassis not connected to anything.
Everyone can stop telling me it's dangerous now ;) The danger is fairly limited, since in order to actually put a dangerous voltage on the chassis, I'd need both a transformer failure passing 120v to the secondary side, and some sort of physical failure connecting that voltage to the chassis inside the amplifier.
Really, with the amount of tinkering I do, I'm probably safer not having any really good grounds at hand.
You ever think about why they don't ground the chassis on toasters?
-Nick
|
|
|
| gmphadte |
This phenomenon is common in table fans. U will only feel it only on a loose contact with the skin. If u hold the chasis, u will feel nothing. From this it is inferred that this is a coupling from ur transformer winding as there is nothing other than winding in a fan. Don't know much but it is quite common with the metallic grilled table fans.
Worried, than better earth the chasis.
Gajanan Phadte |
|
|
| Arx |
| quote: | Originally posted by gmphadte
This phenomenon is common in table fans. U will only feel it only on a loose contact with the skin. If u hold the chasis, u will feel nothing. From this it is inferred that this is a coupling from ur transformer winding as there is nothing other than winding in a fan. Don't know much but it is quite common with the metallic grilled table fans.
Worried, than better earth the chasis.
Gajanan Phadte |
I'm not worried about it at all. I think you're on the right track, though I'm not so sure it's coupling from the transformer, since I can move the amp a couple feet away from it and still feel the effect.
But it does sound like you're talking about the same thing I am at least. When I put my hand solidly on it, I can't feel it. only when lightly touching it.
-Nick |
|
|
| gmphadte |
That means insulation resistance between primary wdg to sec winding is low. U better meggar the trafo and find out. Beware, meggars output lethal DC voltage ranging from 500Volts DC.
Gajanan Phadte |
|
|
| owen |
Has any-one thought of the obvious....
That the insulation on the insulated chip may not be perfect..?
Try a quick test using a sil-pad to see if that cures the problem...
Owen |
|
|
|
