star grounding

[josh]

There seems to be some conflicting information -from what I'm reading anyway- about "proper" grounding schemes.

What I have is my Input and Output coming straight from the board crimped together into one line, then I have both the actual input jacks (pin 1's) crimped together into one line, then I have both the actual output jacks (pin 1's) crimped together into one line, then the power line grouded.

They are terminated at a ring terminal at the same location. A "Star"...

Is this the correct way to do things always? It seems to be a simple way to keep track of your grouding elements and I am ready to start wiring my G9 project this evening.

 

[rafafredd]

I think everyone has their own way on doing it. I never ground he outputs. Also, I use a low value resistor from "circuit ground - the star" to "case or safety ground", like 10ohms or so... This seems to shut up any hum/noise problems for me. I´ve tried many schems before finding my own way.

 

[josh]

[quote author="rafafredd"]I think everyone has their own way on doing it. I never ground he outputs.[/quote]

Why? Does it not matter?

[quote author="rafafredd"]Also, I use a low value resistor from "circuit ground - the star" to "case or safety ground", like 10ohms or so... This seems to shut up any hum/noise problems for me. I´ve tried many schems before finding my own way.[/quote]

Is there another location to ground the board besides the ones coming out of the Input/Output clips?

 

[rafafredd]

The outputs shields are most of the time grounded on the other side of the cable, I mean on the other equipment ground you are connecting it to. If you ground it on this side also, you are just contributing for one more ground loop in your system.

Well, about grounding the unit, that´s what I do:

first filter cap very close to the star, conected to it by a fat copper wire or bar, or anything alike... Same thing with the rectifier I do.

Then, get separate as possible ground points in the circuit to your star ground. Most of the time, if it´s tube, you can use one ground wire for each tube stage. If it´s SS and pcb, most boards have their star point already in the board, so, just connect this star in the board to the star near the cap. All other grounds, like phantom and any other audio grounds you might have outside the board, like transformers or networks, you should also connect to this star near the first filetr cap.

Then, create a point in the chassis for you safety ground. Some people do it near the inputs, I do it near the IEC connector and I do not have any problems. Now, connect the safety pin on the IEC to this point and also connect shields to this point, including transformer shields and cable shields. In fact, the shields could also be connected directly to the chassis in any point, so sometimes I creat a ground lug in the chassis for each shield. It really doesn´t matter at all for me.

Now, you connect your star near the rectifier and the first filter cap to this point in the chassis where safety ground id, but do it throught a 10ohms resistor. Some people like to use fat 10watts types. I don´t hear a difference. Most of the time I even use 2w carbon types, just because I do have a bunch of these. Also, some likes to use a choke instead of a resistor, and I´ve even seen a choke in parallel with a low value resistor.

Heaters and relays PSUs could be kept separate from your audio circuit ground, and you could use separate star point for those, and then connect to the safety point throught another resistor, but I generally just connect it to the safety point all together, with no extra star.

Phantom grounds should be kept together with the main audio circuit.

For bigger power amp circuits, there are better schems, and more caution should used.

I´ve read many grounding articles, and there´s no easy way. You have to experiment for yopurself. I don´t know if I do it the right way or the best way, but that said, my units are really quiet, and seems to be safe. So... I keep doing it his way. I suffered many moons with noise and hum until I got my way of doing it, and really, I got enought of this. Now, I´have some TUBE mic preamps with 85dB gain, and still quiet :green:

By the way:

http://www.aikenamps.com/StarGround.html
http://www.geofex.com/Article_Folders/stargnd/stargnd.htm

 

[josh]

Thank you for taking the time.....

:thumb:

much appreciated.

 

[adamasd]

Why? Does it not matter?

There is no one right grounding scheme, one method may work great in project fjldk but fail miserably in project ijcd.

There is alot on grounding in books and on the web, and a good chunk of it is conflicting. People who do not know better (and some that do) like to spout on about how a certain ground scheme is the only way to go if you want good results.

You pretty much just need to pick a method and stick with it untill it fails you, Trouble shoot it and find out why it failed and continue on. After awhile you will learn what to do when. Learning how to trouble shoot a ground problem will be a much worthwhile use of your time then reading about "proper grounding technique'.

adam

 

[Brian Roth]

AES48-2005 covers this topic. Link to here:

http://www.aes.org/publications/standards/

and scroll down about 60% of the page. You can download the standard for free, although it can't be printed.

Basically, it says that XLR pin 1 (or, I assume, the sleeve of a TRS jack) has to "instantly nail-down" to the chassis at the entry point.

Bri

 

[PRR]

Dennis Bohn and his crew have written intellegently on the subject, though more in "interface" terms than in-the-box terms. It's the same problem only different.

http://www.rane.com/library.html
http://www.rane.com/note151.html

I think Dennis was one of the agitators for AES48.

 

[Jazzy_Pidjay]

For the Phantom Supply on a preamp
it seems a little bit more complicated to me.

Ideally, i would ground the Pin1 (shield) of the microphone cable
right at the input XLR to the chassis, because the signal is balanced
between Pin2 & 3.

But when using Phantom this Pin1 become the 0v of the 48v supply
and so, i would prefer ground it at the star, like for B+ or Heater

So i still don't really know what to do with this 48v ground question.

How the guys here do that ?

thanks

 

[damiafix]

I've the same problem. Who is the best way for a correct grounding procedure?

Best Regards, Damiano

 

[JohnRoberts]

Pin one needs to be a low impedance to chassis ground to collect any and all garbage dumped into it, while not corrupting signal grounds.

Phantom doesn't much care about the voltage at the current return, only that it be close to 0V and finds it's way home to the PS.

JR

 

[owel]

AES48-2005 covers this topic. Basically, it says that XLR pin 1 (or, I assume, the sleeve of a TRS jack) has to "instantly nail-down" to the chassis at the entry point.

Bingo! That is how I do it. It's a simple matter of bridging with a solder the Pin1 of the XLR to the case solder tab of the Neutrik jack. Then the Neutrik jack makes contact with the aluminum/steel case, installed securely via nut and bolts.

Safety ground from IEC AC inlet goes to a ring terminal and screwed to the case too.

That's all.

 

[Jazzy_Pidjay]

Yes, i do it too

So, the 0v of the 48v PSU is just grounded
to the star ground, and we don't care about it.

In this way, it is slighlty different from the Heater or B+
where we want to create a different 'current loop' for each
part of the circuit, until they reach the star ground.

Thanks for the infos !

 

[CJ]

40 yrs of diy and i just discover these, life is not fair.

 

[flaheu]

Can not have access to the link :sad:

 

[Jazzy_Pidjay]

Same for me.

 

[sodderboy]

The owel method is good, but be careful when using Neutrik XLR's. They are painted, and the shell does not always connect well to chassis (often powder coat or anodized metal), or the mating connector shell for that matter.
And when using bolt-together chassis like Par Metal, it is good to grind paint or anodize from both chassis sides around any screw holes. That insures that a six-piece chassis is actually one complete metal box.
Mike

 

[CJ]

ok, these are the guys that got nuked when we upgraded to ver 3 at ipower:

sorry about that, I am a victim of the science age: