Grounding Scheme Questions

[Willabe]

(The start of this thread can be found here;  http://el34world.com/Forum/index.php?topic=18825.0)

For the last week we have had at least 3 or 4 members posting questions on grounding in several different threads.

I'm going to try and start this thread on grounding to open a group thread on it and answer some of the grounding questions. If it goes well and I believe it will, then I will edit it down and we can put this thread in the favorites category for future reference with all the information and links in 1 place.

Please add any good links you know of on grounding schemes.   

1st, as I've posted several times now in the last few days;

Look at our host Doug, Sluckey's and Tubenit posts on their builds. They have a lot of builds that they have posted wonderful documentation and pictures with wonderful layouts and lead dress, study/copy them as they are proven builds.

You will find Doug's builds in here;

http://el34world.com/schematics.htm

And Sluckey's builds in here;

http://home.comcast.net/~seluckey/amps/index.htm

Tubenits amps can be found in here;

http://el34world.com/Forum/index.php?board=17.0

Look in here for some more info on grounding;
 
http://www.valvewizard.co.uk/Grounding.html

http://www.aikenamps.com/index.php/grounding

http://www.geofex.com/Article_Folders/stargnd/stargnd.htm#what_causes_hum

Kevin O'Connor's TUT 3 (chapter 2) book has excellent information with drawings on what he calls 'galactic grounding' along with excellent information on lead dress and chassis layout.

I am NOT trying to say don't ask questions, as that's what this forum is about, but I've seen Tubenit post more than a few times to encourage members to do research/read up on a subject then after digesting the information they found and read post questions about what they still don't understand.

On something like grounding, since it's a fairly big subject with different views, it can be helpful to try and get as much of an over view instead of asking questions 1 at a time, piece meal. Because many of these questions, their
answers don't exist in a vacuum, they are connected to other things in the layout, lead dress and wiring of the amp.       
                 

                       Brad   

[dude]

Nice post, great info, a keeper

Thanks,
al

[Willabe]

There's more than 1 way to skin a cat but I'm posting this to show some of the grounding principles.

Here's my take on 'galactic grounding' on a Champ, I did add a TMB TS.

The idea behind separate 'ground stars' is to bring all of the grounds together at the B+ filter caps ground that is supplying that circuit. Then you run a ground wire to the next 'star' ground.

The 1st B+ filter cap ground star has the most (ripple) current flowing through it and as such is a very important ground loop to keep separate from the other ground loop stars/points. (See note in drawing below.)

In the amps layout you also put the B+ filter cap as close as you can within reason to the circuit it's supplying as apposed to putting all/most of the B+ filter caps together on the PT side of the chassis then running +/- wires to their related circuits.   

This way, circuits grounds together with their B+ filter cap and locating that B+ filter cap close to the circuit, you keep the ground loops small and they can circulate from B+ to minus by them selves and not induce a voltage because of the current in that loop to modulate another circuit. These 'stars' make up the 'galaxy', hence the term 'galactic grounding'.

Also, look at where the OT ground is run to, it's because of the -FB loop. If there is no -FB loop the OT secondary would be grounded at G*1. It's going to the Champs output tubes driver tube, in a PP (push/pull) amp it would go to the PI (phase inverter) tube if the amp has -FB.

The link for Merlin on grounding has more drawings and info on this as do the other links I posted above and KOC in TUT 3 covers this as well in great detail.

Edit; Fixed drawing and removed MV.  Brad   

[Willabe]

It's my thought that isolation of inputs is almost never needed. Fender, and nearly every other mfg have had many, many years of amps that run quiet, with their jacks installed in 3/8" holes, in normal contact with the chassis. some have a second ground contact via the sleeve lug, some not. All are very quiet. As an example, I have a 70's super reverb, that had a bad hum when I got it. One of the 100ohm filament circuit balance resistors had burned open. Replacing that with a pair of new ones stopped the hum, and I've never heard a whisper of hum from it since. (The hissy Carbon comp resistors are another story)

I see guys with concern about "ground loops". I believe ground loops are a concern when they are large, like a common ground on inputs of two amps that also are grounded via power cords. That could create quite a huge antenna that picks up 60hz in walls around it. A loop of a few inches inside a chassis has never been an issue for me. Several feet of cords are a large antenna, and a low value resistance as well.

I have had to isolate a few jacks, where I've replaced marshall type jacks with switchcraft. Possibly, it was not needed in most of those cases, but I was attempting to copy the original hookup. I have amps that are point to point, with each terminal strip providing it's own lug for grounding of that stage, or using a center post of the sockets as a grounding point, with a wire running from there to any convenient chassis solder point, or a screw. These amps are also very quiet unless something else is wrong. Fenders very likely didn't have a hum from the factory, and I've never seen factory isolated jacks in one.

Just my two cents,
Jim

[trobbins]

A problem with the Champ schematic is the ambiguous reference to the first main filter cap grounding.  There is a note about connecting the negative of that cap to CT, and then to the G*1 star, whereas the main schematic shows the CT going direct to G*1 (which is not as appropriate).

Also the screen filter cap should preferably go to G*1.  Presently it goes to G*2, which is the driver stage which is filtered by a cap that should go to G*3, as that stage is powered from 'C'.

There is also benefit in taking the heater CT to G*3, not G*1, as heater induced hum is typically dominated by the input stage.

[Willabe]

Your right, my mistake. 

See above, fixed/changed drawing. Better?


                  Brad   

[Willabe]

There is also benefit in taking the heater CT to G*3, not G*1, as heater induced hum is typically dominated by the input stage.

Could you explain a little more?

I was under the understanding (KOC TUT books and Merlin) that PT secondaries CT's that are on the same PT and it's primary, any hum/noise present on 1 secondary can/will be induced into the other secondaries? So the B+ CT and the 6.3acv heater CT should be grounded together.

Aiken runs the heater CT to the B+ 1st filter cap ground also saying that there's no current flowing through the heater CT.

 
                   Brad   


Edit; Added Aiken's comment from his grounding article on hit web site.  Brad
 

[Willabe]

I copied this from this thread,  http://el34world.com/Forum/index.php?topic=18810.0

Quote from: hesamadman on June 16, 2015, 09:57:01 pm>

An exception would be the caps before and after a choke, wouldnt ya say? I would thing a dual cap there would be ok but im hoping you can tell me one way or the other.I prefer to use a dual cap in this position, even if there is a resistor instead of a choke. I consider screens to be part of the power circuit and as such should use the same ground point. I always ground the PT HV CT, 1st, and 2nd filter caps, and PA tube cathodes to the same ground point so using a dual cap in this position makes plenty of sense to me. Doing this keeps all the high current, 'dirty' DC from actually flowing in the chassis.

I also sometimes use another dual can for other parts of the circuit if those grounds would be connected to the same point.

I do prefer to use a single cap for the first preamp tube and it gets grounded near the input jacks.

I've had good results with this but I'm also not afraid to vary it a bit if the layout or other factors call for a different scheme.

[Willabe]

One thing that the information in the links on grounding that I posted above is getting away from using the chassis as a ground wire, random grounding to chassis instead of a 'wired' ground scheme.

Getting rid of some/most of these random chassis grounds eliminates possible ground loop problems. Like Barney Fife would say to Andy, "Nip it in the bud Andy, nip it, nip it, nip it."   

In the 'galactic grounding' you end up with only 2 ground connections to the chassis, 1 is for the amps circuit, 2 is the power cord 3rd wire safety ground. This is not to say you have to go this far to have a quite amp grounding scheme.


                              Brad   

[trobbins]

Once the principal high current loops are managed, wrt star grounding, then the nuances of the other grounding minutae are likely to be negligible for most amps.

Compromises come in, such as with respect to multiple can caps.

The drawing is still ambiguous for G*2.  The screen filter cap ground and G*2 is shown in two star points.  My take is that G*2 should be the star at the output valve cathode ground point (which is not specifically labelled).

Certainly rectifier noise from the HT-0-HT winding can couple capacitively to a simple heater winding, with the heater CT being the return path - and so I think the logic is to make that return path direct to the power supply end of a distributed star arrangement. 

The capacitive coupling of the heater voltage to the input stage grid is often a dominant hum entry path (coupling to cathode is usually a lesser issue unless the cathode resistance is high and unbypassed), which is normally managed by the CT or humdinger balance, or a humdinger pot.  The 'voltage signal' that is capacitively coupled to the input stage grid is the sum signal from each end of the heater and the input stage 0V.  Those two signals (each end of the heater) are hopefully neutralised by the symmetry of the humdinger/CT.  If the CT is star connected back at the power supply, then each 'heater voltage' signal is the series of the voltage signals of the heater winding, plus the 'heater CT to input stage 0V' voltage signal (ie. along the grounding wiring between star points).  Ie. the heater voltage signal being neutralised may have additional signals added to it (depending on the noise/signal voltage along the ground wiring).  This result is mostly likely so negligible that it has never tangibly mattered where the CT connection is made.

[Willabe]

Compromises come in, such as with respect to multiple can caps.

You might be stuck using cap cans in an old build or a clone build but you don't have to use them in a new build. That's why KOC and Merlin both recommend not using can caps unless you use, lets say both 20uF caps in a can as a single 40uF.

The drawing is still ambiguous for G*2.  The screen filter cap ground and G*2 is shown in two star points.  My take is that G*2 should be the star at the output valve cathode ground point (which is not specifically labelled).

I disagree, the power tube > MV/grid leak/K R/K bypass cap/screen grid B+ node cap ground are all clearly brought together in a single star and it is labeled. I then drew a line down from that star so I could label it. I did that to get past the OT ground wire that is going to the 12AX7 driver ground star. I think it's pretty clear, please feel free to post a clearer drawing.

Thanks for explaining about the heater CT.   

[trobbins]

Sorry Brad, my comment on the schematic was related to those new to amp building, and who may get confused when trying to understand the term 'star' and G*2, and see two 'star' nodes - one labelled G*2, and another unlabelled star node just above it with another group of wires going to it.

[Willabe]

I don't have a better idea at this time on how to fix that.

If you or any one else does please post what you come up with.


                    Brad   

[sluckey]

I think your G*1 and G*2 should be one and the same.

[Willabe]

They are for the most part. There's only a short wire separating them.

They often are both brought together as 1 single star. The thought behind separating them with a short single wire is to allow the 1st B+ filter caps B+/CT and ground (which have the most ripple current) to circulate by them self. The wire going to the second B+ node/star isolates this noisy ground loop. Same with all the B+ node/star grounds, they are isolated from each other then connect to each other down the line with a single ground wire.

It may not be necessary, Merlin and Kevin O'Connor prefer it.

Here's from Merlin and illustrates the star grounding better then what I drew up. The link for Merlin's web site is posted above and is where I got the diagrams I posted below. There you will find all the text on his take on grounding schemes explaining his diagrams.


                           Brad         


Much Thanks to fellow forum member Merlin for the excellent diagrams!   

[Willabe]

Here's how/where to ground the -FB loop.

[trobbins]

The good aspect to Merlin's diagrams is the visual association of key parts (valve, main filter caps, and main dropper resistors) to what many people see under the hood - it removes some of the misinterpretation that can creep in to schematics, and adds to the awareness of layout, especially as so many are wedded to using tag boards and the commercial catch phrase "point to point wiring". 

Switchmode IC manufacturers were one of the leading providers of reference pcb layout designs, because so many were incorrectly interpreting, or not appreciating the influence of star grounding of functional circuit stages and circuits with high current pulses.

[Willabe]

Yes, Merlin's diagrams on star grounding are very good!