Input jack grounding

[Steve Conner]

I don't like isolated input jacks, because they allow common-mode RF picked up on the shield of your guitar cord to romp straight into your amp's innards.

Re catalin's question: I think the circuit on the right makes more sense.

[trem]

I don't like isolated input jacks, because they allow common-mode RF picked up on the shield of your guitar cord to romp straight into your amp's innards.

Innards.?
Pardon my ignorance...
...but is there a difference (RF wise) between a non-isolated jack, and an isolated jack that uses a ground wire to ground the sleeve.?
Thank You

[Steve Conner]

Yes, the self-inductance of the ground wire, and the mutual inductance between it and other wires in the "innards".

If the ground wire ran halfway round the chassis, those inductances could be quite considerable. If it was a short wire going to a stud on the chassis right next to the jack, there would probably be hardly any difference.

[trem]

Yes, the self-inductance of the ground wire, and the mutual inductance between it and other wires in the "innards".

If the ground wire ran halfway round the chassis, those inductances could be quite considerable. If it was a short wire going to a stud on the chassis right next to the jack, there would probably be hardly any difference.

10-4 Thanks.
I have never had much success with any kind of Star Ground. So yeah, for me, it would be a "Cliff" jack with a short wire to chassis ground.
Thanks Again

[catalin gramada]

Thw star point is usually a grounding lug that is bolted to an aluminum chassis or soldered to a steel chassis. In which case you would use an isolated input jack and ground it at the star only. Nothing should ever have two paths to ground.

Hello
maybe was not pretty clear cause my english, sorry.
My local stars was done directly to the - pole of each capacitors and those are ;bolted; to a heavy copper rail which linked each to another into a bus. Unfortunetely there are some inch (2-3) of wire from each local ground point on the board to the capacitor pole. In this case the input can run on the grounding point of the first stage on the board(where are tied all components for each stage which runs to the ground) ,or directly to the - pole cap which deserve both triode section (stage1 &2) this is the end of mine rail bus. I put two drawings to ilustrate both situation.
Sorry for my confused english
Regards
Catalin

[Gingertube]

I don't like isolated input jacks, because they allow common-mode RF picked up on the shield of your guitar cord to romp straight into your amp's innards.

Re catalin's question: I think the circuit on the right makes more sense.

For a start I use bussed 0V with connection to the 0V bus follows the schematic, that is Rg1 of the input tube is the first connection and the power supply main cap is the last connection at the other end of the 0V buss. That keeps the large cap chraging currents away from the sensitive input stages.
Then you use an INSULATED input jack, shielded cable from the jack with the shield connected at that input tube Rg1 point.
AND
The ground side of the input jack is connected at RF to chassis right near at the jack via a 10nF ceramic cap. That makes sure the input cable shield is at chassis potential RF wise and extends the chassis RF shielding action the entire length of the input cable.

Doing it this way means that the actual point that you tie the 0V buss to chassis (either directly or via a back to back diode parallel with 10 Ohms and 100 to 220nF cap) can be anywhere along the 0V buss length.
I usually use one of 3 points:
For zero global feedback amp I use the input tube Rg1 point.
For an amp with global feedback I tie the output tranny secondary common to the point along the buss which coincides with where the feedback is applied and use that point for the chassis tie as well
Occassionally I have used the Rg1 point of the Reverb Recovery stage, sometimes that keeps the reverb quieter.

Cheers,
Ian

[R.G.]

...
Although "star" grounding may seem like a good idea, it's not.

It may not be a good idea if you don't implement it well, but then there are lots of those. The whole area of grounding requires that either (1) you know all the currents that are traveling down every ground wire and how they interact, either by being smart and determined to calculate and find them or by being determined and lucky, or (2) you use some system which separates the ground wires so the currents can't interact. Star grounding - done right, which is difficult - is the logical conclusion of (2), since currents can't interact if they don't share a wire.

It is also possible to do it wrong and route the wires so you get either electromagnetic/antenna pickup on the ground wires. You do have to do it right.

Connecting the PS and signal ground to the same place always results in more hum.

That does not follow. The only way non-interacting grounds can be beaten for hum is if you somehow manage to use different hums to cancel one another. It can happen, and that's the basis of the hum-cancel pot in amps and other tube equipment. It inserts some filament AC hum and you get to choose a variable amount in one of two polarities. But it only works if (a) the hum it's cancelling is power-line frequency, not ripple return and (b) the hum it's cancelling is not phase shifted relative to the AC power line.

I don't like isolated input jacks, because they allow common-mode RF picked up on the shield of your guitar cord to romp straight into your amp's innards.

Try an isolated jack which is connected to the chassis by a 0.01 or 0.001 ceramic from signal ground right to the chassis at the jack location, perhaps with a ring terminal. You're right, the shield leads RF in, but the cap "grounds" the RF to the chassis/shield for RF frequencies only. This setup shows the differences in how "ground" acts at low/power line frequencies and at RF, where Faraday shields and low-inductance planes are the only way to get low ground impedances.

"Ground" is used for at least three different things. (1) Signal voltage or power supply reference voltage, both of which imply almost zero current, (2) shielding from radiated electromagnetics and (3) what I call "sewer ground", the return to the power supply common point. If you mix these up, you get mixed results. You have to know what currents flow in a wire to know how it will act in a grounding scheme.

[GainFreak]

I always use insulated jacks everywhere and always ground the amp at PSU main filter caps to the chassis via back to back diodes/cap/resistor.
My PCBs' ground is an equivalent of a straight bus wire that goes to the PSU. Speaker ground goes to power tubes PCB ground and from there to the PSU. Every time this results in a dead quiet amp. I don't know if it's a matter of luck but this works for me every time, no audible RF so far.
Considering this kind of design can somebody explain to me how one would separate the "power supply ground", the "preamp ground" and the "signal audio ground" since my amp ground from tube #1 to the PSU caps can be considered as one bus wire ground? I've also seen other well known amps using the same technique and where all stage decoupling caps are situated close to the main caps, not at the stages they decouple.
Also since most amps implement first gain stage grid stopper and/or ferrite beads is it still necessary to use caps to ground the insulated input jack to the chassis?

[soundguruman]

I always use insulated jacks everywhere and always ground the amp at PSU main filter caps to the chassis via back to back diodes/cap/resistor.
My PCBs' ground is an equivalent of a straight bus wire that goes to the PSU. Speaker ground goes to power tubes PCB ground and from there to the PSU. Every time this results in a dead quiet amp. I don't know if it's a matter of luck but this works for me every time, no audible RF so far.
Considering this kind of design can somebody explain to me how one would separate the "power supply ground", the "preamp ground" and the "signal audio ground" since my amp ground from tube #1 to the PSU caps can be considered as one bus wire ground? I've also seen other well known amps using the same technique and where all stage decoupling caps are situated close to the main caps, not at the stages they decouple.
Also since most amps implement first gain stage grid stopper and/or ferrite beads is it still necessary to use caps to ground the insulated input jack to the chassis?

It's better when the power supply (filters) ground is separated from the preamp ground.
including the bias filters, Marshall guys.
Through the resistance of the chassis.
There is too much 60 cycle ingress when it is all bussed together.
Some manufacturers are using a 10 ohm resistor, or other methods...to separate the grounds.

[Chuck H]

OK... Let's just do away with all the crap. Any misunderstanding has certainly been clarified in this post (and several previous). The isolated input jack is fine if the ground point is near the input and the sleeve is included. Star ground works in a perfect world, but not always in the real world. That can be argued, but it's true. My hybrid grounding scheme...:

Iso jacks with metal sleeves all around.

I use a three star layout. All the preamp and associated filter grounds to point one. All the PI, reverb, effects loops, trem, presence, etc. to point two and all the PSU main filters to point three. If the earth ground is near the input this should minimize modulation and RF chanelling via the shields. Am I off here??? Any extra hoopajupe with diodes and/or 10R resistors seems pedantic at this point. Doesn't it? R.G. raised an excellent point about recognizing ground function. It's either a 0V reference, a current source or a shield. I do actually run all my ground leads separate. I maintain a neat appearance by placing the ground points under the board. Overall this has been working for me on all fronts and the amps are quiet. I would suggest that anyone interested pay special attention to R.G.'s indication of the differeent purposes for ground/0V. It all makes better sense once that is considered.

[soundguruman]

OK... Let's just do away with all the crap. Any misunderstanding has certainly been clarified in this post (and several previous). The isolated input jack is fine if the ground point is near the input and the sleeve is included. Star ground works in a perfect world, but not always in the real world. That can be argued, but it's true. My hybrid grounding scheme...:

Iso jacks with metal sleeves all around.

I use a three star layout. All the preamp and associated filter grounds to point one. All the PI, reverb, effects loops, trem, presence, etc. to point two and all the PSU main filters to point three. If the earth ground is near the input this should minimize modulation and RF chanelling via the shields. Am I off here??? Any extra hoopajupe with diodes and/or 10R resistors seems pedantic at this point. Doesn't it? R.G. raised an excellent point about recognizing ground function. It's either a 0V reference, a current source or a shield. I do actually run all my ground leads separate. I maintain a neat appearance by placing the ground points under the board. Overall this has been working for me on all fronts and the amps are quiet. I would suggest that anyone interested pay special attention to R.G.'s indication of the differeent purposes for ground/0V. It all makes better sense once that is considered.

That's a good way to do it, as long as the grounds are separated, through the resistance of the chassis.
The chassis may be 6, 8, 10 ohms or even more. Depends on the material. that's why you don't need the resistor.
In smaller high gain amps, manufacturers have been using the resistor method.
Sometimes they use a diode, in an effort to 1/2 wave rectify the AC on the chassis. This cuts the hum in half...
Sorry, there is not 0 volts on ground, or the chassis, even though theoretically it should be...
There is always some small amount of AC on the chassis, or on your ground conductor. And that AC is amplified.

It's really not a load of crap, it's true. But it's important to understand this stuff, especially that star grounding the power supply and audio to the same point is NOT the best method. That's a really important lesson for amp builders.
In a low gain amp, it's hardly an issue. But when you start trying to obtain severe overdrive gain, it really becomes a problem.

[GainFreak]

For some hardcore grounding/RF protection check out Marshall AVT20/50/150 schematics where all inputs/outputs are coupled to the chassis via earth washers ans small capacitors. The diodes combo is also there:

http://music-electronics-forum.com/t13445/

However if you have a powder coated chassis this becomes a nuisance or you have to include a thin "chassis plate" contacting all pots/inputs/outputs or prevent powder coating around those areas which somehow defeats the purpose of powder coating the chassis.

[soundguruman]

For some hardcore grounding/RF protection check out Marshall AVT20/50/150 schematics where all inputs/outputs are coupled to the chassis via earth washers ans small capacitors. The diodes combo is also there:

http://music-electronics-forum.com/t13445/

However if you have a powder coated chassis this becomes a nuisance or you have to include a thin "chassis plate" contacting all pots/inputs/outputs or prevent powder coating around those areas which somehow defeats the purpose of powder coating the chassis.

** If you guys are working on a P to P wired Marshall Head, like a Super Lead...
Move the bias filter cap ground. Take it off the preamp ground buss, and move it over to where the power transformer is grounded.
This will cut the 60 cycle hum IN HALF.
You can thank me later...

[GainFreak]

I'm mostly into high gain PCB amps where one should be more careful but thanks for the tip.

[J M Fahey]

As said before, I use "cascading" grounding, solving all problems for *one* gain stage, then the next, and so on, in which case it's easy to use just a simple ground wire linking said stage to the next, and so on.
So basically no *need* for Star grounding.
*But*, high current returns, specially if AC carrying (100Hz charging pulses or speaker returns) get their own wire.
So I guess it's a "pragmatic" approach, solved on a case by case basis, without preconceived ideas.
My amp wiring ends up being *very* simple and uncluttered.
A couple extra points I didn't mention before:
1) I always use Cliff type insulated jacks (or insulate metallic ones).
*Maybe*, if situation merits, I'll ground it to chassis 1/8" away, why not, but let *me* decide it, not the jack .
And it also lets me ground where I want, either to a jack washer , the PCB, a pot case, whatever works best.
2) by my preferred type of construction, preamp chassis and power amp one are not in physical contact, so I can treat them as separate problems to solve (easier this way, less variables) and then join them by a single ground wire, applied where it works best.
3) this one is , I think, the most important of them all: the first gain stage input (be it the first grid , gate, or, usually, the first Op Amp +In, lives 1 inch away from the input jack.
*Worst* case: no more than 2", and inputs are *always* RF bypassed, both to ground *and* to feedback point, which is not the same.
4) I always use thick aluminum chassis or panels, treated and painted on the outside, but naked on the inside.
Thanks God I have always lived in electrically "dirty" places, which have forced me to do something abouit them ... or else.
Sort of a mixed blessing, but as they say: "what doesn't kill you , makes you stronger."
In fact, I have won bids supplying sound equipment to, say, Theaters or Churches (plus a Yoga Meditation Centre, the Planetarium, Recording Studios and the Buenos Aires Female Prison) by consistently demonstrating *practical* lower hum, noise, and interference compared to other bidders, many of them "famous brands".