Shield vs no shield - tests w/ soundclips

[farren]

This is something I hadn't even considered doing till two hours ago, so it's possible I've made an oversight somewhere. I believe this is simple enough, however, that the only variables pertain to the shield. I will probably do more complex versions of this test in the future, hoping to find a difference between classic wiring and more modern wiring where the signal negative and shield only meet in one place. The difference between classic and treating signal negative and shield as separate things that should only come together at one place (usually from the 3rd volume pot lug to the back of a pot housing or common shield screw) is likely only to be evident when there are fluorescent lights and other such noise sources nearby, but that is something I've not yet tackled, so on to what I have.

The constants:

-Mesa/Boogie Mark V head in 90w mode, full power (~123v), channel 3 Mark IV mode, fixed volume, gain high at 3:00, FX loop bypassed, fan off.
-Mesa Stiletto 4x12 with V30s @ 8 Ohms.
-Single 10' George L cable straight into amp from jack.
-Switchcraft 1/4" TS jack mounted facing amp.
-Dimarzio Air Norton 7, known to be quiet in the guitar that I removed it from with a layer of insulating electrical tape on the bottom of the fiberglass base plate to prevent poles from touching shield.
-Stationary cardboard box with fully closable lid. Pickup fixed to bottom, facing up when box lid open. The box and thus the pickup sit on a plane with the amp, cab just underneath the head.
-Minimal active room electronics: only those required to record with playback disabled.
-LED-free PC running Audacity with LCD switched off while recording.
-Shure SM58, with globe removed, on boom close-miked halfway between center and edge of top right cone.
-Mic connected to Roland digital mixer via XLR, then 2x TS from mixer to TRS 1/8" Auzentech Forte X-fi soundcard line-in.
-Everything but amp are voltage regulated (the irony, I know).

Clip 1: Enclosure unshielded, 4-conductor wiring straight to jack with hot to tip, series link soldered and taped off, signal negative and bare shield wire to sleeve.

Clip 2: Enclosure shielded with aluminum on all sides, but with lid open, exposing inside from above only, 4-conductor wiring straight to jack as before, screw through shield with wire connecting shield to bare.

Clip 3: Enclosure fully shielded with aluminum, lid closed, 4-conductor wiring straight to jack as before, screw through shield with wire connecting shield to bare.

I won't comment on the three clips (I want you to do that ) except to say the results are more dramatic than I expected. I've done my best to eliminate sources of environmental noise, but only relative to my house. I'm in the city so there will be garbage in the air. If replicated, the results could be more or less dramatic depending on environmental noise in other settings.

Please note the Mark V has a ridiculous amount of circuity in a rather small chassis, and the high gain channel which I used is known to have a soft low frequency background hum to it normal playing conditions (ie. outside the studio). This is the secondary hum you may hear which is common to all three clips. But hey, it's hardly a 5150 in the noise department. My line power is also a little hot averaging 123v, but a modern tube amp should be able to cope with that, and based on the tube life I get, it does.

I recommend listening with volume high, or doing spectrum analysis.



"Copper would have been better."

Yes, and more expensive The foil looks sloppy, but is overlapped wherever necessary. The soldering is clean.

Thoughts? Criticism? Thoughtful criticism?

[farren]

My conclusion (hopefully you're only reading this after forming your own opinion of the clips, or the methodology): While this test fails in practical take-away in one regard--I have no idea how tone would be affected by shielding like this--I can say clip 2 and 3 are quieter than any of my current guitars with passives or any I've ever owned or played. Time to do some work... Also, to my ears and eyes, clip 2 and 3 are equally quiet. I can't hear any pickup-based noise in them at all, only the amp.

[newey]

Nice work, farren! +1 for applying the scientific method. I agree with your take on the results.

If you wish to extend the research, repositioning the box, with lid on and off, may affect the results. Noise is highly directional, as can be seen by simply moving a noisy guitar in various directions. It may be that the "box open" test is more susceptible in some directions than others, while the "box closed" is more uniformly quiet.

I don't know if you had seen this thread before or not:

D2o's How much benefit do you get from shielding?

Much of that thread is devoted to proving/disproving the notorious "Urthman hypothesis", one of the more contentious issues in the history of this Board, but D2o also did some shielding tests as well. In addition to his sound clips, he also posted graphics of the sound files as well.

[JohnH]

Good demo, +1 from me too.

When you play a standard ttype of guitar that only has a front pg shield, the rear is somewhat protected by your own body, which is grounded by touching the strings, unless you are playing something involving the chord E11th.

I had a guitar that I shielded with foil inside the cavity and the pg, and I was able to have it plugged in and turned up as I placed the completed pg assembly down onto the guitar.

Before doing this, on the benchtop, it was buzzing, and it reduced a bit as it was lowered over the screened cavity. The cavity was not itself grounded, but as it closed down and the pg foil connected to the cavity, it crackled, and buzzy noises were stopped. (It still played though)

J

[farren]

Thanks, guys. I certainly plan to do more testing, particularly with the alternate methods of grounding I mentioned--it just has to wait as first I have to stop slacking off and get a few of my guitars back in working order... Only having one electric with strings on it is no fun. Any given day I usually have the urge to play in three different tunings.

I checked out that thread, interesting stuff... I felt sorry for Urthman after reading it and hoped to read the original thread afterward to have a laugh at his expense, but it appears to have been pruned. Anyway, a thorough debunking.

I laughed a little at "standard type of guitar." I've actually never owned a guitar with a pickguard or even single-coils, so a standard type of guitar for me is two humbuckers I realize I'm from a different musical background than most around here, though. I'm sure the fruits of shielding were more apparent, and more rewarding, with that guitar.

By the way, if anyone is interested in this grounding idea of treating shield and signal negative differently, allowing them to meet only once, I read about it in this thread and it seems logical enough, though I may as well run it by you guys before I bother converting. There's an enormous amount of BS on that forum, like most (I read a long, long thread there full of people raving about switching to pure nickel strings and how much warmer it made the high strings--it took 5 or 6 pages for someone to mention "hey, you guys realize the high strings are still plain steel, right? It's just the wound strings that are nickel.") There's also something interesting about potentially added noise when wiring for two or more independent volume controls on page 3.

[newey]

By the way, if anyone is interested in this grounding idea of treating shield and signal negative differently, allowing them to meet only once, I read about it in this thread and it seems logical enough, though I may as well run it by you guys before I bother converting.

The original Guitarnuts site recommends "star-grounding" so as to eliminate ground loops. This is essentially the same thing as is being discussed by Caldwell on the LP forum- the shielding and the signal negative connect only immediately prior to the jack, and are separate up to that point.

I think star-grounding is a good practice for several reasons, including the difficulty of soldering to the backs of pots. But the idea that ground loops in passive guitar circuitry act as antennas for noise, while theoretically sound, is practically unproven, and is probably unlikely to be a major factor.

I emphasize "in a passive guitar circuit" because at some point in the LP thread Mr. Caldwell starts talking about ground loops in amps that he builds, and in effects pedals. Ground loops in powered circuits are a far different topic.

Notice, too, that he admits at the outset that any supposed effect from eliminating ground loops cannot be heard except in particularly noisy environments (noisy, that is, from an electrical standpoint). For example, he cites to a particular Coors sign where he gigged, implying that the sign caused excessive noise in his LP. He implies that his guitar became more quiet after rewiring, but we are not told of any true A-B comparison, before and after rewiring. This is, at best, what scientists call "anecdotal evidence", in other words, not really proof of anything.

You have already done things more scientifically than anything in his post. For example, we are not told if he tried unplugging the Coors sign to see if it was in fact a noise source. If not, we don't know what is actually contributing to the noise in that environment (i.e., it's an uncontrolled variable). And, I'd venture that his report of it being quieter after rewiring was done on a different day, thereby implicating his subjective memory of how noisy it was to begin with, and also confounding his baseline noise level, since the electrical environment could have changed in the meantime (new bulbs in a fluorescent lamp, compressor on a fridge not running on the second occasion, etc).

Now, we can always theorize that, in a certain environment, crazy noise is occurring which eliminating ground loops might theoretically cure- but proving that it happens in the real world is another matter entirely. Most noise is picked up by the long coils of wire in one's pickups. In addition to the confounding variables noted above, we also don't know that, by rewiring the guitar, Caldwell may have improved a shaky pickup connection that was picking up noise independent of any ground loops.

But the underlying theory is a bit shaky as well.

The ability of a loop of wire to act as an antenna is proportional to its length. The tiny lengths of wire between Vol and tone pots are unlikely to be picking up any noise. JohnH did some empirical testing a while back (and of course I can't find the link . . . ) where he connected the ground wires of his guitar to the heating vents in his house- thereby essentially making the entire ductwork of his home into one big ground loop- and reported no increase whatsoever in noise.

And, as ChrisK noted, the only ground loops which could possibly matter are those in the signal path. Any loops in the shielding, including the bare wire or braided shields of the pickups, cannot possibly be adding noise to the signal because they aren't carrying any signal. (Of course, "vintage" Gibson pickups had but two conductors, a "hot" and a braided shield, which was carrying the signal as well as shielding things- in those pickups a loop in the shield might matter)

So, by all means star-ground your guitar if you are in the process of rewiring it. But I wouldn't bother to take a functioning guitar apart to redo the grounding unless there was a problem- any purely theoretical advantage in eliminating supposed "ground loops" isn't worth the hassle or the possibility of screwing up something else while your in there . . .

[sumgai]

newey,

Apr 15, 2013 23:03:15 GMT -5 newey said:

JohnH did some empirical testing a while back (and of course I can't find the link . . . )

Would this perhaps be the missing link:

The Dreaded Tone Suck!



Note to John: could you please refresh the image in that posting? Thanks.




sumgai

[newey]

Thanks for finding that, sg!

Further on the subject, I had forgotten this little tidbit from ChrisK, found at the end of that thread:

If guitars were all that sensitive to ground loops, the method of using a stereo jack to switch the battery on in an active guitar (where the operational supply current flows thru the signal ground) would have fallen on its face once tried.

[farren]

Thanks, newey. I'm once again blown away by the wealth of knowledge dispensed. The battery bit really drives it home.

I won't bother testing traditional vs star grounding. At first I'd thought what Caldwell was proposing was a particular form of star grounding, but yeah, they're one in the same. I intended to test it with noisy stimuli like an EL-backlit keyboard I have (no Coors sign here ), but it seems unnecessary now. I knew a ground loop couldn't be a source of noise in normal situations. It's one of those things people mindlessly repeat everywhere like "you have a ground problem if you touch the strings and the noise gets quieter--this is your body grounding the strings." I've linked "Bucket O' Noise" countless times and it's always met with "lol moron ur wrong" hostility. Ahh the internet...

[JohnH]

Apr 16, 2013 1:34:38 GMT -5 sumgai said:

newey,

Apr 15, 2013 23:03:15 GMT -5 newey said:

JohnH did some empirical testing a while back (and of course I can't find the link . . . )

Would this perhaps be the missing link:

The Dreaded Tone Suck!



Note to John: could you please refresh the image in that posting? Thanks.




sumgai

I fixed the image link, which was about ground loops and related mythbusting.