Grounding Input, floating output?

[Ethan]

I seem to recall some sort of discussion on grounding where someone stated that one should ground the input XLR to chassis and not the output. Or something to that effect. I tried searching the archives, but there's just too many sporatic blurbs about grounding without much direction or consistent thought on the subject.

(Perhaps this could turn into a META for grounding, as it seems to be simple yet very complicated.)
[BTW, maybe I should create a reference "sub-forum" with JUST all the meta threads--whaddya think?]

Anyhow, back to my original question... Can anyone explain the reasoning behind that?

Thanks,
E

 

[CJ]

Grounding is a tough subject because it usually just involves trying this and that til you get the lowest hum/noise. Closer to voodoo than most DIY topics.
Houses are wired different, studios different, lights, e fields, mag fields, noise everywhere you look.

Gound lift switches everwhere, I say.

Another thing I have wondered about, if you rack something, isn't it now grounded with everything else in that rack via the chassis and rack frame?

 

[Ethan]

[quote author="cjenrick"]Another thing I have wondered about, if you rack something, isn't it now grounded with everything else in that rack via the chassis and rack frame?[/quote]
Yeah, Exactly!

And another thing I've wondered, if the chassis isn't grounded to an actual "Earth" (which often times it isn't) isn't the whole damn thing just "floating" anyway?

 

[SmG]

I did a quick search using the term 'ground loops', and came up with this thread, which may have been what you were thinking of.

The argument about leaving a screen connection off all relates to preventing a ground loop between two chassis that are grounded at slightly different potentials when their signal lines are connected. The argument is that the cable screen only needs to be connected at one point to be effective as a screen, and that ideally this would be at an input - simply because of the almighty hum that would potentially result from an unscreened connection being made - and also, it would make it rather hard to unbalance the input if you needed to, and there wasn't a ground connection present.

[quote author="cjenrick"]
Another thing I have wondered about, if you rack something, isn't it now grounded with everything else in that rack via the chassis and rack frame?[/quote]
It most certainly should be - but to effect cable screening even within the rack, you would need to connect a screen at one end. The reason for doing this inside the chassis is to prevent the need for non-standard leads.

 

[SmG]

[quote author="Ethan"]
And another thing I've wondered, if the chassis isn't grounded to an actual "Earth" (which often times it isn't) isn't the whole damn thing just "floating" anyway?[/quote]
Yes, but it's potentially not a good idea (please excuse the pun).

 

[Bjorn Zetterlund]

Definately a loaded topic where you'll find loads of opinions and differing experiences based on what people have dealt with and how they have made things work.

It is also a topic with no simple answers, and where the answers will differ depending on what your situation is, and one person's experience may not translate well to someone else in a different country with different gear etc. Anyway, here's one angle;

In any studio or live situation, you have to pretty much assume that individual pieces of equipment will be sitting at different potentials...simply because that is the imperfect world that we live in. The potential difference (PD) between two pieces of gear can be anything from barely a millivolt to a few volts...to lethal voltages if you have faulty equipment, but then you have bigger problems than hum to worry about! If screens are connected at both ends, then there will be a current flowing between any two devices. If each pin 1 on each device is connected directly to mains ground with a separate wire, this current flow shouldn't be a problem...as long as you have good line drivers/receivers and ideally starquad cable, or at least tightly twisted pair.

You get into **** when the pin 1's are connected to a circuit board where the track also runs through parts of the internal circuitry of the device. Any current running through this path will create a voltage through the resistance of the PCB trace, and this voltage will effectively be superimposed on the audio signal if this ground trace is also used as signal reference ground for different parts of the circuit. This hum is actually injected before or after the balancing stages, meaning that no amount of balancing of the actual transmission line will do anything to get rid of it! This is the real problem, and to be safe, and because we can never really trust our manufacturers, we cut one end of the screen to stop this current. For short runs, cutting the screen at one end is no big deal. However, for long runs it will degrade the RF shielding at the cut end as the resistance of the screen starts to show; basically the screen is not as good at screening if it is connected to ground through a resistor. On a 100 meter multicore, the screen may measure about 10 Ohms or so. Most devices will have some form of RF filtering on the input, and so there is a case for cutting the screen at the input rather than the output, as the output will typically not have any form of RF protection.


I'm sure there will be lots of good posts on this thread, and I look forward to reading all of them!

Bjorn

 

[thermionic]

For anyone reading that is unsure, I think a clear distinction needs to be made between fitting a lift resistor to the signal-ground, and not the mains ground :wink:

I haven't made a hobby of measuring the dc-resistance between case and signal ground on commercial studio kit, but a quick scan about shows that some is enough to trigger a bleep and other kit doesn't.

Pwr amps I've made have the signal-gnd floated from case-mains-gnd - I guess they're on the end of the chain but they have absolutely no hum. A preamp I've been fiddling with lately has to have the case connected to signal gnd otherwise it hums.

It would be interesting to hear why some oems float sig-gnd entirely, yet others do not offer a gnd-lift switch. GML's compresssor has a terminal strip at the back offering 3 discrete gnd terminals - 1-signal / 2-case / 3-relay, belt-and-braces I guess (and like).

Justin

 

[CJ]

Yeah, and Ollie over at Funkenwerk has a soft ground switch on his stereo V72, whatever that does, I do not know.
I think grounding only one end of a shielded cable is to avoid a ground loop.

 

[Bjorn Zetterlund]

When it comes to grounding issues, I think we need to be absolutely clear of our definitions of grounding and lifting ground etc, or else this thread will just become another source of confusion rather than a thread that may help explain things.

The world of balancing is already full of confusion. For instance, what is a fully floating output? How about a fully floating input? Or fully balanced input/output? If audio and mains ground are connected with a 1k resistor, is that a lift? How about 100R? Or 47k?

Bjorn

 

[daArry]

Hey

another great thread in progress So...

You get into **** when the pin 1's are connected to a circuit board

would this then be a bad idea?:

 

[Bjorn Zetterlund]

dunno!

A PCB trace for your grounds (pin 1's or sleeves) doesn't have to be a bad idea. If you still observe 'star ground' practices for your circuit reference grounds, and route your (pin 1's or sleeves) to this same point, you should still be ok. The point is that you need to make sure that any current flowing between (pin 1's or sleeves) to mains ground will not affect your circuitry. As long as all of your circuit stages take their ground from a single point, then at least in theory it shouldn't even matter if there are also ground currents flowing through this point. If any (pin 1's or sleeves) are connected on your board, and if this board has only one connection to earth mains, this is invariably what will happen, as some voltage will be created along this one connection to earth. So your whole device is jumping up and down with this noise voltage; but with good balanced connections, this shouldn't matter as this is just the kinda thing they are made for...and this is where you do get your 100 dB of common mode rejection (ok, more like 40 dB if you're lucky!)

If all interconnex were either on trs jacks or XLR, then we could all be happy, but then how do we deal with the unbalanced stuff? Ultimately there isn't even a standard for how to connect a balanced output to an unbalanced input...should pin 3 be grounded or left floating?

Bjorn

 

[Kev]

[quote author="cjenrick"]Grounding is a tough subject because it usually just involves trying this and that til you get the lowest hum/noise. [/quote]

correct
and that's where the subject ends

.... :shock:

apart from safety aspects that is.

In countries that have a true ground and where countries have rules to how power should be handled at the box. These rules should be obeyed. The ground will enter the box via the IEC inlet and have the proper boot then go straight to a solid chassis connection point.

This is all about people safety and has nothing to do with audio.

As far as audio ground is concerned then .... anything goes. ... :shock:

AC ground
DC ground
Dirty earth
Clean earth
signal earth
digital ground
ground loop
0 volts ??
.........
arrgghhhh!!


I have amp racks that are ALL wood and NO rack connection to each other. FULLY FLOATING inputs.

I have DOUBLE insulated gear.

Many of my DIY units have NO connection or audio earth to chassis. I mostly DON"T connect pin 1 to the XLR body.


I could go on. :roll:


Understand the rules
then do your best to apply them
and
on occasion you may find you need to break them.

 

[CJ]

Wood racks! Very interesting! And where the heck have you been? Boxing kangeroos?
And when's "Dad" gonna post? :razz:

 

[Kev]

[quote author="cjenrick"]And where the heck have you been? [/quote]

very near by ... as always

just very busy and a little conscious of my number of posts. :roll:

 

[owel]

Yeah, Grounding is a voodoo subject. What may work on one DIY gear, may not work on another. It's trial and error, trying out all the tricks in the book.

If all else fail, stick a transformer between the gear.... that will get rid of hum caused by ground loops. Even the cheap Radio Shack Hum eliminator works. Or go the EBTech route, or your own DIY box full of audio transformers.

 

[CJ]

Speaking of grounding and cables etc, what do you guys use on your projects as far as XLR's and 1/4 inch jacks are concearnerd? I mean do you follow certain rules like all mic level stuff is XLR, (obviously), and all line level stuff is 1/4 inch phone jacks? If so, how do you keep from grounding your 1/4 inch jacks? I use plastic jacks and a fish paper gasket to lift the 1/4 inch jack from ground, then put in a ground lift switch if desired. But with the metal jacks, you can not isolate them?
Do any of you use XLR jacks and cables when running from, say a micpre to an eq or compressor? How about going into a mixer, what kind of cables/jacks? Does anybody use XLR all the way thru to the board?
Thanks! I'm pretty green in this area.
Any part numbers for 1/4 inch jacks would be helpful also. Those plastic jacks don't strike me as being partucullary bullettproof.
cj

 

[owel]

I use those Neutrik XLR/TRS combo jacks for my inputs. That way, I can plug a mic, or 1/4" jack.

Also, I have the XLR/TRS combo jack both in the front panel, and at the rear.... they're just connected in parrallel. It will be nice to add a switch so I can switch between front inputs, or rear inputs as the active source.

For outputs, I use those 1/4" plastic jacks from Allelectronics.com.

Pin1 of XLR and ground of 1/4" jack are wired to the PCB main ground trace/area. Then from the PCB, the power supply DC gnd wire gets grounded to a screw terminal near the IEC connector. The Gnd line of 110Vac is also connected to this screw terminal, which also grounds the chassis.

 

[BYacey]

The best 1/4" jacks in my opinion are the Switchcraft "12" series. To isolate them you can get a fibre washer with a shoulder and a phenolic washer for the other side of the panel. These jacks will give trouble free service for a lifetime or more.
Bill

 

[Svart]

all i can say is that when i hook all the grounds together in my insert patchbay i start getting more and more hum. so now i leave everything isolated. pretty much just a large star ground with the console as center point for audio ground(all boxes have power ground to chassis for safety). that seems to work the best for me.

 

[Guest]

[quote author="daArry"]Hey

another great thread in progress So...

You get into **** when the pin 1's are connected to a circuit board

would this then be a bad idea?:

[/quote]

***** about swich impedances.
Glass tube with contact and mercury inside which switches may help.

xvlk