Ground question – finding the correct star ground point.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

pepe

Active member
Joined
Aug 31, 2006
Messages
40
Location
Halmstad, Sweden
I’ve been reading a lot of topics here about how to grounding and shielding properly, but can not stop wondering. Why can not the case itself be considered as the common point in a star grounding scheme? I find it strange that everyone seems to be using wires to a specific point in the case, usually the IEC connecter, and not uses the shortest path to the case itself. I would be very thankful if someone could give me an explanation about this.
 
Where dynamic resistance between powering voltage and ground is minimal is the best point. I.e. on the final filter cap. or on points from which feedback to a voltage regulator is taken (better if they are the same).
 
In some designs, like a few power amps I know about, the chassis is used as a low impedance power ground. However for high dynamic range audio paths (like say in a power amp) the voltage differences across the chassis caused by current flows are too high for it to be a clean reference. Where you make your single ground point is fairly arbitrary, and works OK as long as you are consistent about keeping signal paths sorted, so you could use one point on the chassis if so desired.

JR
 
The case can be the star ground, and sometimes is, but all the grounds eventually find their way to houses ground.

The problem with just grounding to the case all over willy nilly is you do not know what path those electrons are going to take to get to the houses ground.

Pour a glass of water out on your desk, some of it will flow off right into your keyboard some goes off the back of the desk where it is hard to clean up and the rest slowly moves about, pools and snakes about all over.

The electrons do the same thing in the chassis, you can not predict which way they will go, they will eventually get to ground, but how will they get there. They might take a nice direct route that will have a very low impedance and make everything happy, or they might take a very round about high impedance route causing all kinda of problems like ground loops.

In some things like guitar amps which are high impedance you can get away with having lots of chassis grounds, but the fewer the better. In low impedance devices this can cause real trouble though.

adam
 
[quote author="JohnRoberts"]Where you make your single ground point is fairly arbitrary[/quote]
Important words.

Please ignore anything you may have read which says that you HAVE to use somewhere close to the input sockets or other such advice. It is almost certainly founded on either suspicion or misunderstanding.

Pick a spot. -ONE spot. That'll work fine.

Now, run wires individually to that spot, and don't "secondary-hub" them elsewhere. Make a star. As long as you're not dumping vast chugs of AC current down thin wires, you will be fine.

That's it. No mystery.

Keith
 
While on grounding, please allow me to bring this one up:

I'm making a rack in which various D&R modules, FX etc can be slotted in. The original thing used a separate power-TX & regulation for each module, but for new ones that I'll be making I'd prefer to prevent adding multiple power-TXs.

http://www.twin-x.com/groupdiy/albums/userpics/DR-6RU.jpg

So obviously a common +/-15 V supply would be nice. I expect this will work fine, also done that a few times before. It's just that all these (unbalanced) modules can be patched here & there in the total setup...
so I expect lots of groundloops.
I know this still doesn't have to be a problem (mixers with inserts etc do this all the time & seem to get away with it), but still, it's not as neat as it could be.


One of the things I've been thinking of is using one common big power-TX and use separate regulator boards for each module. The 'idea' would be to AC-couple each one like for instance done here http://www.jlmaudio.com/ACDCVer2sch.pdf and keeping their grounds floating w.r.t. each other *.
Patching audio will let them meet again (once).

It'll require a few big caps for each module (but one), but that'd be OK. At 1000uF the voltage drop is still OK.

Bad idea ? Other suggestions ? Overkill ?

Thanks,

Peter


'*' maybe I need to provide some higher-Z paths to keep the electrolytes properly biased
 
[quote author="SSLtech"][quote author="JohnRoberts"]Where you make your single ground point is fairly arbitrary[/quote]
Important words.
[/quote]

One condition: no asymmetrical in/outs (hum), or electronically balanced (RFI). Unfortunately UL requires grounding through power outlet.

Please ignore anything you may have read which says that you HAVE to use somewhere close to the input sockets or other such advice. It is almost certainly founded on either suspicion or misunderstanding.

It worked fine for Hi-fi amps with asymmetrical inputs when grounding through power outlets was not required (I mean grounding of the chassis, not selecting the common ground point for the unit)

Pick a spot. -ONE spot. That'll work fine.

Not so simple.

Now, run wires individually to that spot, and don't "secondary-hub" them elsewhere. Make a star. As long as you're not dumping vast chugs of AC current down thin wires, you will be fine.

That's it. No mystery.

Keith

"secondary-hub" works well when it is the straight amplifier with the single signal path, and RC power filters go from output to input stages, and it has a transformer balanced input. In such case the chassis may be grounded near the filter capacitor from which output stage takes it's power.

As I said, always think of current through all ground wires, including power outlet and cable shields, to be happy.
 
[quote author="clintrubber"]

One of the things I've been thinking of is using one common big power-TX and use separate regulator boards for each module. The 'idea' would be to AC-couple each one like for instance done here http://www.jlmaudio.com/ACDCVer2sch.pdf and keeping their grounds floating w.r.t. each other *.
Patching audio will let them meet again (once).
[/quote]

...or using many small transformers in the single 1U rack, like in BrickRack(R) module.
 
[quote author="Wavebourn"]One condition: no asymmetrical in/outs (hum), or electronically balanced (RFI). Unfortunately UL requires grounding through power outlet.[/quote]
Audio ground does NOT have to be connected to power ground for UL approval, so that's an unneccesary complication, which serves only to confuse the issue.
[quote author="Wavebourn"]It worked fine for Hi-fi amps with asymmetrical inputs when grounding through power outlets was not required (I mean grounding of the chassis, not selecting the common ground point for the unit.[/quote]
Well, that's by default. -Think about it. The only ground connections are the input grounds (ignoring the speaker and antenna grounds), so there really isn't any OPTION there...
[quote author="Wavebourn"]Not so simple.[/quote]
Seems pretty simple to me.
[quote author="Wavebourn"]"secondary-hub" works well when it is the straight amplifier with the single signal path, and RC power filters go from output to input stages, and it has a transformer balanced input. In such case the chassis may be grounded near the filter capacitor from which output stage takes it's power.[/quote]
No doubt it Con work, with hub-grounding, but if you're not sure about how to calculate the current flow, (this is a question which most commonty seems to vex many people who are at a comparatively early stage of DIY learning, after all) then single-point-star grounding with NO hubs is a readily-grasped, VERY reliable approach. I'm not claiming it's the ONLY one, but it is the EASIEST one.

Again, I'm not suggesting that anything written by wavebourn is technically wrong, but -critically- neither is anything which I wrote, either.

Now, which is simpler?

Keith
 
[quote author="SSLtech"][quote author="Wavebourn"]One condition: no asymmetrical in/outs (hum), or electronically balanced (RFI). Unfortunately UL requires grounding through power outlet.[/quote]
Audio ground does NOT have to be connected to power ground for UL approval, so that's an unneccesary complication, which serves only to confuse the issue.
[/quote]

Thanks for good news, I always thought that chassis must be connected to the power ground. Please confirm before I opened all my boxes to cut all power ground wires. :cool:

Again, I'm not suggesting that anything written by wavebourn is technically wrong, but -critically- neither is anything which I wrote, either.

Now, which is simpler?

Keith

Simpler is to write down everything that you assume instead of expecting from readers to assume the same and read between your lines. :cool:
 
[quote author="Wavebourn"][quote author="SSLtech"][quote author="Wavebourn"]One condition: no asymmetrical in/outs (hum), or electronically balanced (RFI). Unfortunately UL requires grounding through power outlet.[/quote]
Audio ground does NOT have to be connected to power ground for UL approval, so that's an unneccesary complication, which serves only to confuse the issue.
[/quote]

Thanks for good news, I always thought that chassis must be connected to the power ground. Please confirm before I opened all my boxes to cut all power ground wires. :cool:
[/quote]
You are quite right. Chassis must be connected to power ground unless in cases of non-metallic or double-insulated chassis.

However, I wasn't referring to incoming AC power grounds. I'm talking about 0V secondary potential, and Audio ground.

Underwriter's Laboratory approval (as far as I know anyway) has no such insistence about tying them to chassis OR to the power ground.

Big difference.

Keith
 
Bad news, Keith! :roll:

But thanks anyway for the fast correction, I did not cut yet that wires I always want to cut, but always afraid to be punished... :grin:
 
We seem to be making this a little more complex than it needs to be. My understanding of regulations and it depends on type of gear and target market, but 3 wire grounded products, such as those used in install markets draw a distinction based on how the input terminals are labelled. If it's called ground, it pretty much needs to pass ground bonding test (X amps with < Y voltage drop). If it's just labelled as - , lo, or whatever, it doesn't need to provide a low impedance path to ground.

If designing products for UL approval, research their regs for your classification of product. Free advice off the web is worth every penny.

JR
 
[quote author="JohnRoberts"]

If designing products for UL approval, research their regs for your classification of product. Free advice off the web is worth every penny.
[/quote]

We have to live with them, no matter do we like that or not... So grounding in the randomly selected point is no good, it creates a ground loop that will cause hum and RFI of the level depending on what point was selected randomly... That's why I prefer transformers over electronic balancing.
 
[quote author="Wavebourn"]So grounding in the randomly selected point is no good, it creates a ground loop that will cause hum and RFI of the level depending on what point was selected randomly... That's why I prefer transformers over electronic balancing.[/quote]
I know I keep coming back to some of these issues, but I simply cannot accept this as a stated 'fact'.

How exactly does grounding to a randomly-selected point definately create a ground loop?

There are too many people with too sketchy a grasp of the issues reading this and taking this stuff as fact.

I have to say that I disagree with the above assertion as written.

There may be a language issue or something, but essentially, in plain english the above statement is misleading at best, and wrong at worst.

Sorry 'bout that, but it's how I feel.

Keith
 
[quote author="SSLtech"][quote author="Wavebourn"]So grounding in the randomly selected point is no good, it creates a ground loop that will cause hum and RFI of the level depending on what point was selected randomly... That's why I prefer transformers over electronic balancing.[/quote]
I know I keep coming back to some of these issues, but I simply cannot accept this as a stated 'fact'.

How exactly does grounding to a randomly-selected point definately create a ground loop?

There are too many people with too sketchy a grasp of the issues reading this and taking this stuff as fact.

I have to say that I disagree with the above assertion as written.

There may be a language issue or something, but essentially, in plain english the above statement is misleading at best, and wrong at worst.

Sorry 'bout that, but it's how I feel.

Keith[/quote]


May be it is a language related issue, so let's draw an interlingual picture.
If you take one ground wire and connect it through a power outlet, another wires through your ins/outs, there will be 2 ground wires at least (actually, more) and any current through the "earth" in power outlet will be reflected on your internal wires. That's why I always run my notebook on batteries when recording live concerts... Ground points in the gear I have to connect together were selected "randomly", indeed...

I preach some new topics:

"Keith, I wired everything like you said, but when I connect my mic pres, compressors, gates, equalizers, to my A/D converter I always hear strange beeps and boiling eggs. Help!"

If they ground one point, but randomly selected the point of minimal dynamic resistance in a power supply, the case is the best.
 
Some clarity can be had from distinguishing between shielding and grounding. Whereas a star ground is useful and intuitive in the audio realm, one must also concurrently address the rf domain. A ground every few inches is how most rf circuits are made, just to work at all.

This is the opposite of audio grounding where you would , by doing that, create multiple return points at various resistances.

So, I would say that it would be desirable to make the star ground sink point near the case grounds' interception of pin 1 on the xlr's and the power leads' green wire.

This would send incoming rf currents into the chassis, which acts just like a shielded cable does, in that it uses opposing currents on each side to achieve cancellation.

No?
 
It is what I am doing: connecting "zero ohm/zero volt" of the power supply close to that green wire I can't cut unfortunately... :evil:

Speaking of RF, again it is more complicated than looks: indeed, currents are taken in calculation, also inductances and capacitances can't be neglected! (like in audio, when you think of possible oscillations).
 
But if you route the rf present on the incoming cables "around the outside" of the unit, whilst retaining good star ground hygiene inside, you are properly addressing both.

Keep the rf currents out of the unit's inside, and use a good audio ground scheme.
 
[quote author="Wavebourn"][quote author="clintrubber"]

One of the things I've been thinking of is using one common big power-TX and use separate regulator boards for each module. The 'idea' would be to AC-couple each one like for instance done here http://www.jlmaudio.com/ACDCVer2sch.pdf and keeping their grounds floating w.r.t. each other *.
Patching audio will let them meet again (once).
[/quote]

...or using many small transformers in the single 1U rack, like in BrickRack(R) module.[/quote]
You mean using the 1U bottom strip you saw in my pic for that ?
(I found this BrickRack-pic:
ps8.jpg
nice idea to get rid of clutter)

But multiple TXs was what I actually want to avoid (have a beefy suited powerTX, no multiple smaller ones for now). The cost & hassle for multiple wouldn't be less than the 2-electrolyte AC-coupling idea of above, but I wonder why we haven't seen that before.


BTW, apologies for jumping into this thread - I thought the conclusion of the original question was reached, so thought it'd be OK to add some related question.

Regards,

Peter
 

Latest posts

Back
Top