I had a wifi dongle with the main chip completely absent. I took it apart and was initially impressed how they could do wifi with so few components. There was an empty QFN footprint, perhaps that is for an option. But there seem to be no semiconductors actually fitted at all. Hmm.... Got a same model replacement next day and it worked great.
Although they very likely did bed-of-nails, so that's not as impressive or truthful as it sounds.
My Trinitron monitor has test points on its boards, including the single side phenolic PCB in the power supply.
Tim
I notice you aren't questioning the EMC part...
If the transmitter house is so poorly constructed that there's induction-cooking levels of ground loop, you're f***ed one way or another. If the current isn't going through the shield, the induced voltage is going into your even-more-sensitive inputs (and outputs).
The only possible hope you have of reducing interference in the signal is with shielding, probably requiring multiple layers at that. But if that's melting things, you're SOL either way.
What, you treat that case like it's a posterchild advocate for cutting shields. That's the dummest thing I've ever heard!
Tim
Sigh. It was balanced, transformer coupled audio at 150 ohms impedance. Any RF picked up on the conductors is out of phase and cancels. When you've built a radio or TV station, you can lecture me.
"SPC", statistical process control. Some of our customers are insisting that we do that. To comply with what they want, we'd have to stop 100% testing our products.
Right.
Try doing that B2B.
All of my turntables have had both left and right shields disconnected at the turntable connector, with a separate (green) wire to ground the chassis to the amplifier.
They're only required to be the same size up to certain sized conductors (6 or 8, can't remember - big enough that I don't care).
Hello? Tektronix service? That 20 GHz scope doesn't work; send us another one.
Ok, so it was isolated. Meaning whoever had the idea to short a shield across the isolation barrier was an idiot to begin with. That's not you or me misinterpreting shielding practice, that's bad design.
Well, I've designed a 50kW transmitter, crammed into a box a fraction of the size of a commercial radio transmitter of the same rating. With low voltage, high speed digital logic mere inches away. Does that count?
Tim
It was wired according to the OEM specs, and all within a pair of six foot relay racks that were bolted together.
Not unless you installed it, and the antenna system. I'm sure that the lead length of the ICs was a very tiny percentage of a wavelength, as well. How about 195 KW of Analog UHF TV? Or a site with 5, 50 KW FM transmitters into a curtain antenna that is 1200 feet up a 1700 foot tower? There was around 10 MW EIRP from that one site. Have you ever been to the WLW site with the 500 KW 700 KHz transmitter with their Blaw Knox tower?
Anyone can cram a transmitter into a smaller case. Try to maintain one like that when it has to run 24/7/365.25 and you will curse the fool who designed it. Modern AM transmitters are modular, and under microprocessor control. Lots of redundancy. Till the backplane fails, or the dozens of power supplies catch fire. Harris and others are proficient in designing modules with a short production life. They don't really want anything field serviced. They use too many components that are EOL soon after the equipment ships. One station I removed a transmitter from had a pile of bad Harris modules. Enough to build another transmitter and all unrepairable.
I've had 980 KHz AM riding in the shielding of 75 ohm video cable at a small radio & TV station. There was more RF than Video. That mess required a complete rewiring, and a new ground system. I pulled out over a mile of audio, video and telephone cable from that site.
Have you ever needed an Oxyacetelene tourch to heat the end of a ground rod red hot to drive it through of premafrost?
No, I'm staying on topic and you aren't.
USB is important; audio isn't.
Sounds like nobody here can afford a DVM.
I checked two different USB hubs and one memory stick.
All had pin 4 of the USB connector hard connected to the connector shell.
The hubs had their power barrel connector negative side grounded ditto.
NexStar USB terabyte hard drive: everything is case ground, including USB shell and pin 4, optional barrel power connector ground, and SATA connector shell.
One USB mouse measures 5 ohms from shell to pin 4.
When in doubt, ground it out.
Every time you slap a ferrite clamp around a shielded cable, you're 'lifting' the shield connection, at RF. This is definitely something to try during an EMC test. More important, extra parts might be avoided by cutting of connections that make ground loops in sensitive signal wiring. There's more than one way to lift the shield connection, stay aware of all the options.
On the other hand, I've worked with products that consisted of tested boxes plugged into a tested power strip. Then, the test protocol went through full AC fault tests of the ensemble. The customer wasn't allowed to plug or reseat anything in the power strip, of course.
Did you check the shell of the cartridge (which is also a shield)? It usually has a connection, sometimes a removable strap.
Why can't you use statistical testing on 100% of your units? Those statistics would be very reliable.
How does that matter? The RCA connectors had no shield connections.
If the only thing you do is USB, you're right. Shield at both ends and have a boring life. Real life is more interesting.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required