Spooky Mass W4 Gremlins HELP!

[bjornson]

This is a long one so stay with me...
I've been putting together a passive 3 way split based around Mass W4 connectors.
It has a direct out and 2 outputs with a shared switchable ground lift.
The stage box is mounted in an SKB (as of yet shielded) case in a 46 x12 configuration.
There are 3 W4's mounted on the front. I have a 250' W4 with 25' fan for the FOH, a 50' fan for MON and a 50' fan for record split.
I took it out for the virgin test run and found a few inputs were making barely audible crackling noises. Almost like digital clock hash. I could hear it in my phones but FOH and MON world couldn't due to audience masking.
Back at the shop I ran through the snake with my trusty Swizz army tester(the one with the 3x3 matrix of LEDs) and found some inputs that showed shorts between pins. I pulled out my fluke and on continuity (transistor test), it didn't show any shorts. Then I measured resistance between individual conductors with fans connected. Most connectors showed 1.5 megohm or greater resistance between any two (open air) pins on a connector. None showed a direct short.
One of the noisy channels that had showed a short between pins 2 and 3 using the Swizz tester showed a resistance of 450K ohms between pins 2 and 3 using the fluke.
My questions now become:
1: What the hell is causing the popcorn? Are the long runs of wire acting as Satans capacitor?
Should I of finished shielding the stage box before the test run?
(I shop tested the SKB box with a beta 87 while a fluorescent ballast was sitting on it... hum free)
2: At what resistance does the Swizz tester decide "it's a short" ?
(I was going to start plugging resisters into the tester but I'm done for the night.)
3: What is an acceptable resistance between pins down 250' of W4?
Do i need to be concerned about 450 kohms of resistance?

I never had any of these types of problems with my ghetto elco passive split.
While the client was very pleased with the product, I'm keenly aware that something is amiss....
I await the knowledge of the great pumpkin.......

[TMI]

According to the W4 manual, 2.5 Gigaohms is the minimum insulation resistance for the connector itself. You do not mention the wire you are using but it also has a similar spec. (10 Gigaohms for Mogami) 450k ohms is not a reasonable result for a complete assembly.


Think of it in terms of the ratio between a 150 ohm microphone source impedance and the 450k leakage-about 3000 to 1, seems like a lot but expressed in db noise vs microphone level it is not.

Leakage between 2&3 should not be as big a deal as between 2 or 3 and 1 for obvious reasons but it indicates a problem that must be addressed.

The leakage may not be stable. After all it is not a low noise metal film resistor and it may be mechanically variable and have signifigant reactive impedance that also varies with vibration.

When you splay the recording, stage, monitor and FOH lines out you have a pretty good longwire antenna that is not duplicated in the shop on a reel.

The isolation you enjoy at DC may not be as effective at blocking the hi frequency changes in neutral or ground voltage from the stage, monitor, FOH, and recording AC supplies. Transformer specs are downgraded by cable and box construction.

I suspect you have nicked the conductor insulation while stripping the outer jacket on some pairs and this is where your excessive DC leakage is. Tugging on the individual wires within the connector should wildly vary the reading and/or loosening the strain relief. Disconnect the cables from the splitter and verify it IS a cable problem.

Use a sweep generator with a 150 ohm balanced source impedance and the DB function of your fluke or a quality noise meter as a test rig. These days, the best generator and meter you have may be your recording rig fitted with a first class set of transformers and some test software. For relative RF susceptibility, a spark gap and a scope or your rig is your best bet.

Your Swizz uses TEST TONES, not DC but they appear to be limited to 1Khz or 440hz. As a result, it WILL show continuity through the splitter, so be careful what you are measuring. Also, 450k could be a dry skin reading or a very tight strain relief.

Popularity of W4 connectors aside, there is a substantial area where:

a) Pairs are unshielded

b) Twisting them is difficult

c) They are subject to leakage to the other pairs through normal mode coupling.

d) They are not protected from drain currents through normal mode coupling.

If they really wanted to make an audio connector, since the pinout is standardized, it would have made sense to place a guard ring around each pair so the shield is continued through the connector. The tighter the pin spacing, the WORSE the isolation performance. Technically, going from ELCOs to a W4 is a downgrade.

Here is some reading material if you want to bone up on what is going on. It might be easier to hire a tech to sort it out.


http://www.ebtechaudio.com/swizzeq.pdf

http://www.ebtechaudio.com/swizzman.pdf

http://www.jensen-transformers.com/an/an005.pdf

http://www.whirlwindusa.com/ftp/Conn...l/w4manual.pdf

ftp://ftp.mogami-wire.co.jp/pub/mit/catalog.pdf

Sincerely

[bjornson]

Tom,
Thanks!
looks like I have some reading to do.
I have isolated the 450 kohm reading to a pair in a whirlwind assembled new fan out. I'll have a look under the hood!