Dumb grounding question...

[Fred G.]

A dumb question about grounding lead dress - is it best to run ground wires as directly as possible to the ground point, or should they be routed along the crease between chassis wall and floor? Or does it even matter?

I usually just follow the original scheme as done by the manufacturer, but in this instance the amp has been tampered with a great deal and I need to redo the grounding scheme. The longest length of ground wire has been run along the "crease" at the chassis rear wall/floor back to the power supply, along with the filament wires, and that's apparently created a parasitic oscillation of some sort. When I lift the wire a bit, the amp starts ticking and making a bit of a commotion. I'd like to get this ground wire away from the filaments altogether, and am wondering whether to just run it the shortest distance possible to the ground point, or across the chassis to the front "crease" and back to the ground point (much more wire that way).

[Helmholtz]

Ground wires should always be as short as possible. That not only lowers resistance but - more importantly - lowers ground inductance.
If all ground impedances were zero, there would be no grounding/ ground loop problems.

But there are other valid grounding rules as well and sometimes one needs to compromise.

[TomCarlos]

Back in the late 70's, I was working as a Technician for the Audio Design group at National Semiconductor in Santa Clara. It was then when I had first exposure to audio chips like the LM386, LM383, LM1875, and others. I was helping to model the PC boards that we would use to demo these chips. It was then that I learned all about Low vs High signal ground. I was amazed to see differences in Total Harmonic Distortion readings when ground traces were moved ever so slightly. So I try to remember this then working on amps.

No, I really do not want to second guess the amp designers and their techniques for grounding. But learning about "Star Grounding" and standard grounding of various amp stages was very helpful. So that is yet again something to add to the checklist when trying to debug "noise."

And no, it was NOT a dumb question. Heck, I cannot remember the number of times I asked grounding questions here. I compiled all the responses and have them in a section of the big reference binder that I keep on my workbench.

[Fred G.]

Apparently, grounding is not an exact science. I have read numerous opinions on the Internet regarding it. "Star" grounding seems to have a bit of controversy surrounding it. I've had great results on builds and repairs using other methods, including haphazard grounding to the nearest ground point, having three separate ground points (AC mains, power supply, and preamp), and every other manner. I find it damned confusing.

[Chuck H]

I've never had a problem with ground schemes I've implemented. That isn't to say I've never had a problem with ground schemes. Sometimes with tampered amps or even manufacturers designs (that work well enough until you want to make a modification) there can be an issue with an existing ground scheme. It actually IS an exact science. But the rules are many and complex and I don't know all of them. Few do. Some know most of them. I would suggest reading any grounding related threads here where R.G has participated.

What works for me is simply to ground circuits relative to their power supply rail ground. And, as Helmholtz said, keep the leads as short as practical. I also run separate ground leads for every circuit. NO DAISY CHAINED GROUNDS. You actually CAN daisy chain some grounds, but I just don't do it because the variables can be questionable sometimes. It might make sense when working out a design for economy of labor, but there may be bugs to work out. Knowing what you can get away with and learning what you absolutely can't comes with more experience than I'll ever get. So I run individual ground leads from each circuit. That way there there's never any possibility of any signal sharing via ground leads. This ideal is significant for guitar amps because even vintage type designs have much more gain and fewer stability measures than most other types of amplifier. And...

Once you get your head around this method of grounding and use it a few times it becomes easier to spot likely problems with other, existing ground schemes. You'd think that just because a manufacturer did something a certain way that it should be alright. NOT SO! Sometimes it works for a while until the PS filter/decoupling caps start to age even a little. Sometimes an amp is close enough to instability that ANY change creates a risk.

Never trust to OEM or "stock" WRT ground schemes. Always ground circuits with their respective power supply filter. Keep ground leads as short as practical. Never daisy chain grounds.

If you DO run up against an anomalous circumstance, like reverb jacks that are chassis grounded wherever they happen to be on the chassis rather than WRT their PS filter, be prepared to retrofit isolation for the jacks and follow the grounding scheme I mentioned.

Learn to recognize anything different from what I've described. Because these could be potential problem areas requiring relocation of the grounds.

I consider myself fortunate to have only had to deal with a few ground related issues in my time. But I've read A LOT here and tried to implement the common wisdom on the matter of grounding. There may be lot's of opinions (mine included) and lot's of people who will tell you what worked for them on the same amp or in a similar circumstance. But you're much better off learning about grounding for yourself in the end.

[Fred G.]

Well, Chuck - I had some problems understanding what you were trying to say. Re what grounding points you are actually referring to - "Grounding circuits relative to their power supply rail ground" is wildly vague to me - I'm just not following that, at all. I can't physically visualize what you are trying to say.

Can you clarify?

I have found that multiple ground points to the chassis DO work. Having a separate mains ground to chassis, a separate power supply ground to chassis, and a separate preamp ground to chassis. What I'm hearing, lately, is "ground it from the input jack back to the power supply", which, to me, doesn't necessarily make a lot of sense in certain situations. None of my builds have ever involved such a grounding scheme. This particular amp appears to have implemented that scheme, and - it hasn't really worked worth a sh*t.

The more replies I get for this (here and on other sites) the more confused I get. Plus, much of it is contradictory to my own experience, and nobody seems to agree on it. Sheez, I'm confused. I wish I'd just plowed ahead and not asked, lol!

At this point, I'm just going to go ahead with THREE ground points to chassis - mains, power supply, and preamp. I guess, or maybe FOUR (?), counting the input jacks being connected to chassis...I can't get anyone to agree on any of this. Leo Fender didn't seem to worry too much about it. Who the hell am I?

All that aside, I appreciate your help and feedback.

[vintagekiki]

About Grounding and Shielding

http://www.geofex.com/Article_Folders/stargnd/stargnd.htm
Star Grounding in Tube Amplifiers

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

http://web.archive.org/web/20090414225956/http:/www.aikenamps.com/StarGround.html
Star Grounding

http://web.archive.org/web/20180318110158/http://www.rane.com/pdf/ranenotes/Grounding_&_Shielding_of_Audio_Devices.pdf
Grounding & Shielding Audio Devices

https://el34world.com/charts/grounds.htm
Tube Amplifier Chassis Grounding info

https://diyaudioprojects.com/Technical/Grounding-Shielding/
Grounding and Shielding for your DIY Audio Projects

http://www.paulrubyamps.com/info.html#Ground
Theory:
What is Earth and Signal ground?
What is Star ground?
How do I wire my filaments to avoid hum?

[Intripped]

http://www.valvewizard.co.uk/Grounding.pdf

This pdf, and this site in general, helped me a lot

[Helmholtz]

This pdf, and this site in general, helped me a lot

Yes, Merlin's book is excellent. IMO, the chapter on grounding explains about everything there is to know.


But do you have his permission to publish parts of the book?

[Intripped]

Mmm, this file is freely downloadable at the valvewizard site, and I've only linked the download page, directly from his site ...

Is it an infringement of something?

[Helmholtz]

Mmm, this file is freely downloadable at the valvewizard site, and I've only linked the download page, directly from his site ...

Is it an infringement of something?

I see. As Merlin provided the pdf himself, it should be alright.

[Fred G.]

OK, after a lot of reading, I think I see what you guys were driving at. In summary, the AC mains should have a dedicated ground point, as close to the AC input as possible.

The power supply and it's reservoir cap should have it's own ground, back on it's "own side" of the amp.

The preamp and it's filters should have it's own ground, either a start point or at the input jack, depending on whether the jack is grounded to the chassis or is instead isolated from the chassis.

Another question, now - how should ground wires be routed with respect to signal wires? Parallel, perpendicular, and avoiding proximity as much as possible? I'm not seeing much on this.

[Chuck H]

Another question, now - how should ground wires be routed with respect to signal wires? Parallel, perpendicular, and avoiding proximity as much as possible? I'm not seeing much on this.

You probably guessed that BECAUSE you're not seeing much on this that it's not a big issue. Yes, all leads carrying power or signal radiate. Yes this creates invisible circuits with no components in an amp that are strictly relative to lead dress and component location. BUT, Since ground leads are at 0V there is very little opportunity for induced signals to be large enough along a ground leads wire resistance to cause any problems. Because there IS resistance in leads you DO want to keep ground leads as short as practical. So, for the most part ground lead routing is a non issue. Typically you want to get them right against the chassis with short length. That's about it. I'm sure there are extremes where more consideration is needed but you can cross that bridge if/when it comes up.

A scenario for the Devil's advocate might be if you had a very long ground lead used for a very sensitive signal input circuit. In this case it might be possible to pick up enough of a discordant phase radiant field to cause a problem. I think the chances are remote. At least at audio frequencies. I don't know much about ultra sonic electronics but some of what I've read makes it seem pretty sensitive. And, of course, this is still an area where parasitic oscillation can happen. So just get ground leads close to the chassis quickly and keep them as short as practical.

[Fred G.]

Thank you Chuck, for clearing that up for me!