bear said:
A potential barrier for dirt; with a resistor, anything goes through, with diodes not so much. But, true enough, if there's tonnes of RFI around, the diodes will efficiently demodulate the AM signals and inject them into the grounds.
Look who are talking about bombshell words. As I remember your favorite comment is BS, isn't it?
dimitri said:
No, you need to provide some serious technical contributions here to deserve the flag. Otherwise, this is my favorite comment:
Hmmm.. Well I ask because after some correspondence with Jneutron
I have spent a little more time with the notion of the chassis/earth to
circuit ground connection. He indicated that NEC requires this connection
to be able to reliably clear a 20A breaker, and checking out the characteristics
of your average 35A bridge versus the breaker it looks like a horse race.
Currently we use a Thermistor for this connection which gives us some
meaningful resistance until lots of current starts to flow. It will clear our
two pole 15A breaker. Nevertheless we continue to look at it.
Apart from the larger capacitance of big bridges between chassis and
circuit ground, is there a reason to use something as small as 1N4007's?
😎
I have spent a little more time with the notion of the chassis/earth to
circuit ground connection. He indicated that NEC requires this connection
to be able to reliably clear a 20A breaker, and checking out the characteristics
of your average 35A bridge versus the breaker it looks like a horse race.
Currently we use a Thermistor for this connection which gives us some
meaningful resistance until lots of current starts to flow. It will clear our
two pole 15A breaker. Nevertheless we continue to look at it.
Apart from the larger capacitance of big bridges between chassis and
circuit ground, is there a reason to use something as small as 1N4007's?
😎
Nelson Pass said:
Fits on a PCB 🙂 Yes, large capacitances between ground(s) and the safety ground pretty much defeats the whole purpose of the barrier (for this particular case).
Perhaps 35A bridges would apply to a large power amp, but for such things with 250mA fuses...
Anyway, you may want to look into low voltage power zeners. Zeners have the unique property to always short when subject to a large destructive current. Which would afterwards break your 20A panel fuse for sure.
Sure, a thermistor does just fine, but I don't like showing to the mains dirt any significant resistive path to my circuits. Are there any special designed thermistors for this function?
Nelson Pass said:
This one?
http://canada.newark.com/jsp/search/productdetail.jsp?sku=81F3387&_requestid=60154
Seems like it has 2.5ohm at 25 degrees centigrades. Is this really effective?
Thanks Nelson!
janneman said:
Thanx for explanation Jan, but what would happened with the active ground impedance in HF when opamp/buffer is out of juice? It will start to rise for sure, but is it important? I guess one have to use faster buffer than the circuit it powers, but how much faster?
aparatusonitus said:
Depends on the topology. If you for instance have the feedback resistor of a stage return to an active ground, and that ground impedance rises at high frequencies, that decreases the gain of the feedback stage. So it's a matter of gettting it all just right. IOW, standard engineering, and that's what we do, right 😉
Jan Didden
"http://canada.newark.com/jsp/search...requestid=60154
Seems like it has 2.5ohm at 25 degrees centigrades. Is this really effective?"
The CL60 is more like 10 ohms, it all depends on your needs.
Yes, they are effective. As you know, any ohm helps kill ground loops.
😎
Seems like it has 2.5ohm at 25 degrees centigrades. Is this really effective?"
The CL60 is more like 10 ohms, it all depends on your needs.
Yes, they are effective. As you know, any ohm helps kill ground loops.
😎
syn08 said:
Why not back to back diodes between grounds with a ferrite choke in serie and an eventual resistor to damp resonnance
The best of two worlds ?
Gives you plenty of impedance at RF so less risk/amount of audio rectification in the diodes.
JPV
JPV said:
The problem is with RFI leaking through and along the I/O connectors (not grounded straight to the case).
Otherwise, yes, a ferrite would be good. The input filter helps as well.
I once helped a pro audio guy install a small mixing consule in a TV studio, DIRECTLY under the 100KW TV transmitting tower.
Once done, there were no buzzes or other noises from the RFI, even though it must have been measurable in volts, not uV!
I was serioiusly curious, as living about 2 mils from this tower, I could get 50 hz frame buzz into phonostages. After some prodding, he told me the trick: All the shielded cables entering the mixer had their shields earthed directly to the case, at the EXACT point of entry.
No conductor that was externally accessable by RFI was allowed any path inside the casing. Signal or powersupply.
Worked for him, and apparently ALWAYS worked.
Regards, Allen
www.vacuumstate.com
Once done, there were no buzzes or other noises from the RFI, even though it must have been measurable in volts, not uV!
I was serioiusly curious, as living about 2 mils from this tower, I could get 50 hz frame buzz into phonostages. After some prodding, he told me the trick: All the shielded cables entering the mixer had their shields earthed directly to the case, at the EXACT point of entry.
No conductor that was externally accessable by RFI was allowed any path inside the casing. Signal or powersupply.
Worked for him, and apparently ALWAYS worked.
Regards, Allen
www.vacuumstate.com
dimitri said:
That's exactly what HP uses for the BNC I/O in their RF network analyzers. Certainly a good solution, but unfortunately also very expensive and not readily available.
Allen Wright said:
That's the best solution for avoiding the evil RFI. Unfortunately, this also creates an internal massive ground loop.
If, for example, you are hosting the transformer in the same case, this loop will happily capture the stray magnetic fields and transform them in a very disctinctive hum, usually worse than that induced by the RFI. Not to mention the NEC requirement to connect the earth ground to the case.
Bottom line, there isn't any silver bullet; any solution has to take into account the specific conditions. Me, I prefer the live with an unspecified risk of RFI issues (can't be that bad as long my stuff is passing the cell phone test) and focus on avoiding ground loops.
For power amps, I also separate the signal ground by the power ground via a small (10ohm) resistor, connected as close as possible to the input. Of course, it's the signal ground that gets to the case ground. From this perspective, the power supply is always floating. Chosing the signal and power ground points is though not always easy or obvious; for example, in a LME49810 IC power amp, where should pin 3 go? The right answer is: to the power ground. These decisions have to be made while laying out the PCBs.
Allen, your forgot to add that not only ins/outs, but even mixing amps inside of modern professional consoles are symmetrical. Best consoles still have very expensive input/output transformers and RF filters between them and XLR sockets.
Syn08, what is ground loop inside of the console? Ground in console is an extremely tricky thing, because of multiple ins/outs and inserts.
Syn08, what is ground loop inside of the console? Ground in console is an extremely tricky thing, because of multiple ins/outs and inserts.
- Status
- Not open for further replies.