equipment grounding

[Frank Raia]

This may be a dumb question but.. what is the proper way to ground sound systems?

Do you cut shielding before power amps?

Do you connect ground straps between processors and amps?

I have seen many different ways and want to know how you guys recommend

[Harold Hallikainen]

I recall an AES standard that recommends the shield be connect to the chassis at both ends of the cable, but I do not have a copy of that standard. https://www.ranecommercial.com/legacy/note151.html endorses this method with the shield being considered an extension of the cabinet. If chassis ground on the different units is at a different potential, this can lead to current through the shield. However, all equipment SHOULD have the chassis well grounded to a safety ground so the difference between the voltages on the chassis is minimal. If the pieces of equipment are far apart, there is more likely to be a "disagreement" as to what ground potential is. In those cases, unshielded balanced lines are commonly used with a transformer coupled input at the receive end (This is the approach used in equalized telephone lines for broadcast use).

[Frank Raia]

thank you Harold for your reply. What you said makes sense and is what I thought. But I have seen instances where they cut the shield at the amplifier. I guess they did that because they had a HUM they could not rectify. I just wanted to find out if it was common practice or not. Thank you for the link, great stuff!!

[Steve Guttag]

If your goal is to have the quietest electrical system for sound, you'd want/need to isolate the sound system. This is trickier than it sounds since most everything it touches can contaminate the ground and all it takes is another ground reference at a different potential to create the loop and noise. But, if you isolate off your conduits (either by stubbing off or by using PVC to land them) and then run a separate technical ground for your sound (you still have to ground but you are getting a direct, and uncontaminated ground) to the grounding point for the building (where the ground rod ties into the main building feed). You can get a very quiet system. All conduit has to be grounded but that goes to building ground, not sound. All cables that are outside of the sound rack have to be evaluated for what they touch at the other end to determine if that would contaminate the ground. For analog one can float the ground at the "foreign" end. For analog audio, by lifting the shield...for higher frequency signals (Digital and the like) you can use a capacitor on the foreign end to allow high frequencies to have a nearby ground to drain into but not contaminate the ground at low frequencies.

It is a lot of work and easily thwarted.

The other method is to embrace the building ground with many tie points so that nothing has a significant potential difference. It isn't as good as the completely isolated method but it won't burn you so much. With analog, most will float the shield at one end. The most consistent form is to float it on the sending end as the receiving end, more often than not, has a good ground reference. In cinema, for example, the sound head in a film projector has a questionable ground, at best (it is often isolated by isolation bushing for vibration). For microphones, they have no means to provide a ground so the mic is depending on the receiving end to provide the ground. You want zero current travel on a ground wire...that is the way to keep it at a zero potential difference.

The Rane Note that Harold referenced is a good one but there are times in a tricky system where you need to get a bit of Zen to see what is happening and why the noise is forming. Starting with good practice and understanding what you are trying to accomplish will go a long way to having a less noisy system.

[Frank Raia]

Thank you Steve, much appreciated insight.

[Harold Hallikainen]

It's also helpful to run reasonably high audio levels (0 dB output trim out of the sound processor). When I started at USL, I was told that they wanted the amplifier input gains at maximum so an amplifier swap was quick (just set the gain to maximum). But that required the output trims on the sound processor to be set to -15 to -20 dB. I could hear a click in the speakers when someone turned off the fan in the bathroom. My fix there was to put fixed attenuators right at the amplifier inputs. The amplifier inputs could still be at maximum gain, but the sound processor outputs were around 0 dB (and adjusted to yield the 85 dB(C) at the reference position. This was all in our engineering conference/demo room.

[Steve Guttag]

Yeah, most don't get that another name for amplifier attenuators are "sensitivity" controls. If you want a good gain structure, you won't want those at full, necessarily. Matching gains is not so tricky that a label or marker can't suffice. Cinema, for the most part, once Dolby became the norm, went for a 300mV reference (-8.2dB). So, if you are coming out of a USL at .775V (0dB) you are going to need some attenuation on the amplifier input. You'll get a quieter system, if the noise was coming in via the wire between the processor and amplifier (a relatively short wire).

Where the clicks and pops came in were, partly, due to the fact that cinema was very slow to move towards a balanced system so the ground was the return conductor. Any noise that got onto the ground got into the sound system. For Dolby, that didn't change until the CP650...so it was late in the game. Panastereo, on the CSP1200 was already using balanced outputs (and lower noise opamps...one of the reasons I used Panastereo's in screening rooms was it was about 10dB quieter than a CP65 or CP650...it also integrated better owing to its RS232).

Christie, using the chassis for their negative on their lamps didn't help anything. Most every booth that used Christie lamphouses/consoles got to hear the igniter strike the lamp at the start of each show. They still do it in the digital world too. Having your sound on a separate technical ground can keep the noise off of sound since the noise has to drain all of the way down to the building entrance and then come back up rather than continuing to mother-earth. But, failing that, keeping the projector's ground and the sound system grounds separated until the load center will help. When wiring up anyone's lamphouse/rectifier, I tried to keep the grounding and power separate from the rest of the system. While there is a lot of touching metal, having the dedicated ground wire for each component having a solid, wired connection to the load center went a long way to keeping noise out (as well as other anomalies). I've seen basement reader LEDs flash when lamps strike where everything shares the same power/ground in a console/pedestal/column.

[Harold Hallikainen]

All the USL stuff I worked on had balanced analog audio in and out with a reference level of 300 mV RMS to give lots of headroom (The output drivers used a +/- 12V power supply, so they would clip at around +/- 10 V. 10V peak is about 7V RMS, so we had 20 log(7/0.3) = 27 dB headroom.

Xenon igniters are nasty! We had a fair number of IR panels fail due to lamp strikes. These were the old IRC-21 or IRC-23 that used an external 24 VAC transformer. The incoming AC went to a bridge rectifier using 400 volt diodes. When a lamp was struck, a nice spike would come down the power line, be capacitively coupled from the primary to the secondary, and blow out the diodes. The later external switching power supply eliminated this problem.

About 25 years ago, I designed a few dental curing lights for Denmat. One of them, the Saphire, uses a xenon lamp. The first units had a lamp power supply where one side of the lamp was grounded. It was a lot of work to get it to not crash the microcontroller. Later ones had a bipolar output where one end of the lamp would swing positive, and the other would swing negative. Keeping the wires close to each other minimized the capacitive coupling from the leads into the microcontroller. So, as in audio, balance lines are good!

[Steve Guttag]

Yeah, you came into USL relatively late. Everything (or just about everything) before the JSD-80 was unbalanced (JS5-95, JS105-195, JS205-295 and the JSX-1000).

Strong should have been a bit better on lamp stikes getting into the sound owning to the fact that they would generate their 120V via a step-down transformer (and it was a full transformer, not autoformer)...so the 120V used for the lamphouse/igniter was decoupled from mains power. You then had a dedicated ground wire from the lamphouse that should have been a good connection through the rectifier and back to the load center. That should be a favorable ground, for noise.

I think I saw more of the older power cubes literally burn up (you could tell the length of service by how brown they got) than fail due to lamp ignition but I didn't too many Christie systems. Of the later power supplies, they themselves are failure prone, as power supplies. The voltages are such that there are not many choices out there.

[Frank Raia]

My head is spinning !! So I guess I should bond all audio together and try to isolate that ground on its own. Or that would not work because the power is all on the same service ground?

[Carsten Kurz]

A Steve says, there are no simple ToDos that fits all installations. Do you have an actual ground issue? Are you rebuilding the booth? Sometimes adding just an auxiliary device may introduce issues.

I would usually go the default way and leave all shields connected as cables and equipment come out of the box. If then you have an issue, you need to investigate.

For temporary applications, I always have a range of audio transformers on site. They are the fastest and safest way to fix issues when there is no time for a fundamental rewiring.