Chassis Ground / Signal Ground / Ground Loops

[TriodeLuvr]

I'm posting this to avoid derailing another recent discussion where the topic of grounding came up as a side issue. In that thread, I mentioned that all my recent projects - amps, line stages, phonoe preamps - isolate signal ground from the chassis. The chassis is only connected to the third wire earth ground. This eliminates hum from ground loops. Member @SS904 then noted he had done the same in an amplifier, but he added a 10 Ohm resistor and .1uF Y2 cap from the cap bank ground plane to chassis. When I asked about the purpose of those parts (which to me appear only to reintroduce the possibility of 3-wire ground loops), he said the path through those parts is "enough to break the ground loop and still provide a measure of safety beyond simply floating signal."

So, my question for @SS904 (or anyone else) is this: What potential safety issues might occur when signal ground is completely isolated from the chassis? Why is it important to maintain some type of connection between signal ground and the chassis (third wire ground)? I'm just trying to understand why this approach might be advisable.

Jack

 

[SONDEKNZ]

Great question!

 

[gadget73]

The cap at least provides a path to ground for any RFI that might be coupled in to the signal. Some things just are weirdly touchy about that. I recently found that I had high frequency noise in one of my amp builds because I skipped the cap across the lower resistor in the heater elevation circuit. Tossing an 0.047 film cap in there completely eliminated it. Cap was on the bench from some other experiment, there was zero thought put into the value. That particular amp has a grounded cord, and currently does have a direct connection between B- / signal ground and chassis but is built in a way that I could remove or alter it very easily.

Possibly the resistor is helpful in the case of static buildup for whatever reason, or as a path in case something plugged into the amp has leakage to ground. I've also seen high amp bridge rectifiers with a bypass cap used for this purpose with the idea that any potential over the forward voltage of the bridge would conduct to ground, but below that it has isolation.

 

[SS904]

My thoughts concerning this are it's better than isolating the signal ground inside a grounded metal chassis or having no mains ground with signal ground connected to the chassis. You hope the fuse blows if anything happens.

Considering most vintage and vintage type amps we are in contact with use or used two prong plugs with the signal connected to the metal chassis and it worked, I don't know this is something universally needed. Putting 10 ohms between signal ground and chassis ground has proven to be enough resistance to break a potential low level ground loop through the three prong power cords and other chassis from effecting the signal path, but likely not enough resistance to prevent something hot in signal from having a path to chassis mains ground. Again, you're hoping the fuse blows too.

In my own case I built my current mono blocks and a few others amps with power ground near the IEC connector and signal to a chassis star ground. I didn't initially have a problem. At some point, and I'm not sure where, I developed a nasty ground loop that could only be remedied by adding cheaters to everything except my pre amp. Signal ground was though the RCAs only for the power amps with a path to mains ground only in the pre amp. Was fine, but always in my head that I needed to fix it.

I came across this idea a few times. This forum included, but what made me try it was primarily the article linked. I did not use a bridge rectifier in my loop breakers and I choose to put it on the signal side, with transformer center tap connected to the first filter cap being my signal ground reference point. It is working and I feel better about it, but always willing to listen to input. As a side benefit, I no longer get a thump if I turn my pre amp off too soon after turning off the power amps.

Earthing (Grounding) Your Hi-Fi - Tricks and Techniques

 

[loudnoises]

I had to school myself on grounding schemes a while ago. Star grounds vs chassis grounds, it's actually pretty cool and wonderfully logical. Usually. How they connect and where seems to be the secret sauce.

I'm coming across a very similar issue right now with a phono preamp I'm setting up. The board is from an Onkyo and I'm driving it solo with none of the rest of the components. So the grounding isn't straightforward.

The board will get it's power ground at the +/- in, which according to the overall schematic goes to center tap. At no point can I find any actual connection to chassis ground on the schematic as it's an old unpolarized 8015. The rca signal ground DOES connect to chassis but via a .047uf cap. The ground from the cartridge seems to go to chassis in the original unit. So i need to work out where the two ground schemes connect, and where that rca ground to chassis cap is in the original unit.

Edit: just found it. Ground to chassis is the power input to the board! Yay!

 

[nerdorama]

I first saw a discussion of the "loop breaker" circuit here. I've built amps with the signal common floating without a case ground connection and also using a loop breaker. I always connect chassis to power ground. Both ways have prevented problems with ground loop hum.

Earthing (Grounding) Your Hi-Fi - Tricks and Techniques

Earthing (Grounding) Your Hi-Fi. Detailed explanations of the safety and sonic aspects of proper earthing (grounding) techniques

www.sound-au.com

 

[TriodeLuvr]

Well, after reading the above comments, I'm still not hearing a reason why a loop breaker is beneficial for safety reasons. @gadget73, you mentioned the possibility of static buildup. In that case, wouldn't a 100K (or even higher value) resistor between signal ground and earth be sufficient to drain it off? You also said "...as a path in case something plugged into the amp has leakage to ground." Sorry, but I don't know which ground you're referring to (signal or earth). Can you give me an example of this and how it might be hazardous without the loop breaker? I'm assuming of course that all enclosures are earthed through the 3-wire ground.

The subject of noise is a different issue and generally unrelated to safety. I will say though that the ground wire in house wiring is much too long to serve as a bypass for RF. It would need to be very short, say no longer than five or six feet to earth, in order to be effective for this purpose. In fact, at frequencies where the length of the ground wire is resonant, an impedance inversion can occur that actually isolates the chassis from earth. It's also worth noting that the connection to earth created by a ground rod or other scheme will almost always exhibit much higher DC resistance than the 10Ω resistor in the loop breaker. Perhaps this is why the loop breaker works (for noise).

Considering most vintage and vintage type amps we are in contact with use or used two prong plugs with the signal connected to the metal chassis and it worked, I don't know this is something universally needed.

The two prong system works from the standpoint of noise because there's no third wire ground connecting one or more of the units together. It's not safe though, due to the normal leakage that occurs between various components and the multiple metal chassis. Once the modern ground wire is added to all the units, the issue of ground loops has to be managed. That requires either the addition of loop breakers or isolating the active circuitry from the grounded enclosures.

Again, just to stay on topic, if all the enclosures in a system are connected to the ground wire, and the circuitry inside the enclosures is isolated (no loop breaker), is there a safety risk?

Jack

 

[SS904]

I suppose I over-think things, but in my head I wonder where would the power go in a completely isolated signal path if something on the signal path were to short to signal ground, even if it were a very short lived problem? RCA shells? Speaker terminal common?

For myself I did this for ground loop elimination first, and as I said the first step for me was signal path isolation. I added the loop breaker as a precaution, even if it's not technically needed.

 

[gadget73]

I need to do this with my mono amps. I don't seem to have much problem with grounded stereo amps but the two monos with two grounded cords is not a happy thing. Unhook one or the other and its fine. Those are buss ground, single tie from chassis to circuit so its not that hard to modify. I know I've said I need to do this a bunch of times, they just haven't made their way back to the bench. Maybe I'll try the cap and resistor first, I don't have a pair of bridges on hand to do it but I've definitely got resistors.

 

[Mr Boochie]

I wonder where would the power go in a completely isolated signal path if something on the signal path were to short to signal ground.

Electricity tends to flow along the path of least resistance. If you are part of the path of least resistance, it's going through you.

I've gotten "stung" while playing guitar by poorly grounded amps or frayed cables. Singers have gotten stung by hot mics.

 

[gadget73]

Well, after reading the above comments, I'm still not hearing a reason why a loop breaker is beneficial for safety reasons. @gadget73, you mentioned the possibility of static buildup. In that case, wouldn't a 100K (or even higher value) resistor between signal ground and earth be sufficient to drain it off? You also said "...as a path in case something plugged into the amp has leakage to ground." Sorry, but I don't know which ground you're referring to (signal or earth). Can you give me an example of this and how it might be hazardous without the loop breaker? I'm assuming of course that all enclosures are earthed through the 3-wire ground.

100k would, the value for that is not critical, but it needs something other than jus a cap I think.

As for the leakage, I was thinking if some upstream un-grounded thing has mains leakage to the signal ground, ideally that ought to get dumped to an earth ground somewhere. Most of my stuff is not grounded though so thats largely not a concern. Some of my amps do have grounded cords, but the one hooked up right now does not.

 

[loudnoises]

Electricity tends to flow along the path of least resistance. If you are part of the path of least resistance, it's going through you.

I've gotten "stung" playing guitar by bad guitar amps or cables. Singers have gotten stung by hot mics.

Back in about 96 my buddy's band played a dive bar up the hill from ball square. Mcsomthings i think. Anyways the wiring was so bad in this place that the lead guitarist was getting blue arcs between his tuning pegs and the metal shield on the sm58. No di, mic hanging in front of cab. Made for an interesting show without touching the mics, as he would also spark the mic.

Just think of the path required for this voltage to find it's way home.

 

[SS904]

Electricity tends to flow along the path of least resistance. If you are part of the path of least resistance, it's going through you.

I've gotten "stung" while playing guitar by poorly grounded amps or frayed cables. Singers have gotten stung by hot mics.

Pretty much my argument for adding these loop breakers rather than run a completely isolated signal path.

 

[Tom Bavis]

Ground current can create millivolts of voltage drop across millohms of ground wire resistance. if a preamp chassis has millivolts of difference from the amp chassis, that difference can be amplified as hum. When there are multiple grounded devices connected together, there's really controlling where those ground currents flow. A 10 Ohm resistor is many times the resistance of the ground wire(s), so can only carry a tiny fraction of that current. Ground currents can be induced by the magnetic field of a transformer, will be extremely bad if any two through bolts have continuity to the core. But even a shielded transformer can induce some current into the chassis. Depending on the locations of the transformer and safety ground, this could even couple through interconnects to other grounded devices. This a very low impedance source. There are other sources of ground current, capacitors from line to chassis, transformer primary to core and secondary, but they are a very high impedance and seldom cause trouble. Disconnect all the circuits from the chassis and this ground current can still exist without a problem. Likewise, grounding all the circuit(s) to a single point won't create a loop. And grounding though 10 Ohm resistor is low enough that capacitive pickup will be minimal.

Safety ground or double insulation is required for all modern products, and in my opinion, home builders should try to do this as well. Larger power transformers may not qualify as double insulation, so power amplifiers should have a 3-wire cord for safety ground. Smaller transformers with separate bobbins for primary and secondary will qualify as double insulation. Most sources will have 2-wire cords; a preamp can as well. If a 3-wire cord is used on the preamp, isolation of the circuit from the chassis may be needed (not always). Remember that safety ground is only for primary AC supply, has no safety relevance to any circuits isolated by a transformer. But a ground reference to the chassis is needed to prevent hum pickup. It needs to be low in impedance compared to the noise source. In Gadget's example, the heaters needed to be grounded with an impedance that's low compared to the source (perhaps 350VAC from the HV secondary through 50? pF to the heater winding).

I've done safety testing in in industry. The ground connection is required to test less than 0.1 Ohm after carrying 30 amps AC for one minute. So a ground fault is guaranteed to trip an upstream fuse or breaker. A separate bolt or stud with a crimped wire connection is generally required - solder connection is allowed for an approved device like an IEC connector. If that's grounded through a PC board, it will have to pass that 30A test.

 

[Selmerdave]

This is a topic I have been meaning to revisit, in my 300B monoblocks I disconnected circuit ground from chassis and got horrible loud noise, connecting through a resistor and cap in parallel did not make any different results. Connecting (signal) ground bus to chassis in one place did get rid of it. I would love to know why and how to get different results.

 

[TriodeLuvr]

I wonder where would the power go in a completely isolated signal path if something on the signal path were to short to signal ground, even if it were a very short lived problem? RCA shells? Speaker terminal common?

Signal ground and power supply ground are the same. A short between the signal path and signal ground might cause a loss of audio, burn a dropping resistor and/or blow the fuse, depending on exactly where the short occurs and whether DC is present at that point of the signal path. The RCA shells should be at power supply ground, so a short to those will behave in this way. Same for speaker terminal grounds, which should b e within a few ohms of power supply ground, depending on exactly how the output circuitry is configured. This is not a safety issue as far as I can see.

Jack

 

[TriodeLuvr]

Electricity tends to flow along the path of least resistance. If you are part of the path of least resistance, it's going through you.

That's the point of connecting the chassis to earth ground with a three-wire power cord. My question about safety assumes the chassis is grounded.

Jack

 

[TriodeLuvr]

I was thinking if some upstream un-grounded thing has mains leakage to the signal ground, ideally that ought to get dumped to an earth ground somewhere.

Could that occur other than a primary-secondary short in a power transformer? I think that would blow a fuse. What other type of mains leakage could take place that wouldn't be handled by the grounded chassis?

Jack

 

[GordonW]

Could that occur other than a primary-secondary short in a power transformer? I think that would blow a fuse. What other type of mains leakage could take place that wouldn't be handled by the grounded chassis?

Jack

If something upstream had a "death cap", in theory, the described behavior could happen.

I have also heard of a situation, where stray capacitance inside a power transformer, somehow managed to induce AC voltage on an ungrounded preamp chassis. Really odd and unusual behavior, but apparently not impossible.

Regards,
Gordon.

 

[gadget73]

capacitive coupling is mostly what I was thinking, but failing insulation could also do it. Could leak enough to cause problems without being enough to blow the fuse. A lot of things have fuses that are large enough that you could get tens or hundreds of milliamps of leakage without it blowing.

A GFCI would kick out for that sort of fault though.