>In my playing around I have found in two cases
>(A very small percentage!)
>where there appears to be diode action at very
>low levels on a mechanical contact.
Yes ! Everyone has experienced a loose connecting cable.
>The problem is, how to make the system impervious to
>that ground loop current.
Hardwire the suspect connection.
>(A very small percentage!)
>where there appears to be diode action at very
>low levels on a mechanical contact.
Yes ! Everyone has experienced a loose connecting cable.
>The problem is, how to make the system impervious to
>that ground loop current.
Hardwire the suspect connection.
Did not you see Cuprox Rectifiers?
They were used in AVO meters before Germanium diodes become available.
I hate RCA connectors... It is the dumbest thing of the 21'th century to keep them.
It's not the fact that non linear contact resistance occurs.
It's the fact that your test setup has problems with contact integrity even though the absolute value of the resistance of the primary connection is in the single digit milliohm range, any contact nonlinearity is an order or two of magnitude lower than that if it's a really poor contact, much lower if good (maybe two orders), the characteristic impedance of the cable is 75 to perhaps 300 ohms, 6 to 10 orders of magnitude larger than any hypothesized non linearity, the impedance of the load is what, 10K?, which would be 7 to 11 orders of magnitude higher that any hypothesized non linearity..
Your ground loop on the other hand, is a combination of resistances in the ten to 100 milliohm range. Meaning, the ground loop impedance is anywhere from 1 to 2 orders of magnitude larger than any hypothesized contact non linearity (perhaps 3 to 4). Contact non linearity is well within the realm of ground loop conductivity..
I measure in the low micro-ohm range, there's off the shelf equipment available for that. Nano-ohms are a tad harder to do. Been there, done that, but not easy at all..
That is one possibility, but it certainly makes for a repeatability a problem. Gold plated connectors get leached..any wiping of solder for repeat causes exposure of tin/cu intermetallic, which passivates...and then requires either RMA or RA flux to break through...conversation will also head towards the "which solder sounds better" area, a discussion with no fruit to bear...and it is an option not available to the consumer.
But it is a good way to troubleshoot a test rig.
Even I knew JC was kidding there...
And rca's...well, they served a purpose, and they were good enough at that purpose to last all these decades...but I agree, I don't like em. Not with present day equipment incapable of ignoring ground loop current. I hate hum...
j
Last edited:
I have that problem as well. I need to insulate at the 5kV level, but the wires run 24 awg give or take.. when the wire is soldered, it'll wick up the strands, making the strand bunch stiff. Because the insulation is stiff due to voltage requirements, any movement tends to flex the wire at the point where the solder has wicked up to. For serious hi-rel wire attached to wire, or wire to terminal, I have the techs remove enough insulation for two wraps of stranding that will not wick the solder and remain flexible.
BTW, you mentioned control of supply Z. Are you measuring this through any length of twisted pair? Would your application benefit from making a cable with 4 to 8 ohm scharacteristic impedance? That'd drop the series inductance and give you distributed capacitance, lowering your high frequency z..
jn
jn, one thing to note about John's setup- it gives different results on different days. When he showed me what he was doing, we could pretty consistently get repeatable results when we swapped cables back and forth (i.e., the cables showed differences, and the same differences each time we swapped them out), but the results were different than those he had gotten previously. That's suggestive of electronics rather than contacts.
And as a general comment, we're still talking about differences at least a thousand times lower than anything remotely audible.
And as a general comment, we're still talking about differences at least a thousand times lower than anything remotely audible.
I measure at the feedback node (Kelvin sense) at the output terminal. Any realistic piece of wire beyond that has a much higher resistance than the output impedance of my PS itself. So what is the limiting factor is now is the speed of the feedback loop. I'll send you a pm.
If I may come back to the double coax you fabricated and showed pictures of. You made it for ps cable, but what about the following.
Many pages ago in this thread I floated the idea that we should not treat all grounds equally since they serve different purposes. So, my idea was to use safety earth for shielding purposes only. This would place all the electronics and the interconnects in one single Faraday cage. Within that cage, we could - using Class II regulations - safely create the supply voltages and a new earth, fully dedicated to signal only. This signal earth should subsequently be treated with the same care as the signal itself.
But, to do that, you need interconnects that have separate shielding and signal earths. XLR with twisted pair inside a shield would do, as would the double coax you produced. Which cable you think would work best? It's only a thought experiment right now.
vac
The cable to cable variations in the short time frame suggests ground loop.
The day to day does indeed look like electronics.
He has a two by two set of variables.
It is significant that neither of these have been brought up when using the AP. Is this because it is not there, or because it was not reported.
Which is the general criteria for argument in high end audio, no??
Seriously, it is a test setup which is supposed to suppress reliance on contact resistance, geometry, et al...
In home systems with unbalanced lines and 3 prong IEC's, there has been no (high powered) attempt at supression of ground loop/proximity based problems... other that code violations.
It's like looking for your keys at the lampost..
jn
No problem. Others would be better helping you if the impedance is silicon limited, I can help for wiring and geometry issues.
Do I assume the design is IP for now?
Almost worth going balanced, no?
It sounds like a reasonable idea, but it doesn't eliminate induction coupled ground currents, just IR. Might be good enough for almost everybody..
jn
There are folks that believe in alien abductions with anal probes. Just as much evidence and plausibility.
I'll plug Bill Waslo's Diffmaker demonstration again.
Audio DiffMaker example files
Scroll to bottom.
John (N),
Once cost me a few bucks to replace all the signal processing in an arena where the equipment manufacturer got the ground scheme wrong.
These are to me very common issues. The problem used to be the consultants would specify a ground scheme that actually did low hum, but guaranteed it would always be present. But we have gone over this before.
The issue with J.C. is that different cables did behave differently. Even though he may not have been looking at the issue he thought he was, there is an issue there.
It took me three years to get good wire measurements and four more still refining it. I can now get consistent results even measuring solder joints! I absolutely will not be publishing this! If anyone can hear the difference between properly soldered joints using a quality solder, then we should all reevaluate the concept of sanity!
Wrong question. I prefer balanced by default, especially when they interconnect boxes connected to power outlet.
well...I hate( oh no!)-> I disrespect! designers still building class I audio/entertaining/computer equipment. RCA connectors are consumer class connectors and in this enviroment hum could be easily fought by just comply to modern safety standards and classII designs!
- Status
- Not open for further replies.