pl400 ground loops and hum

[mlucitt]

Agreed, I was a little confused between in-rush and floating ground in your comment. The problem I see with the thermistor between Star Ground and Chassis Ground is that this is a variable resistance based on current and heat. I prefer a set amount of float for a varying voltage and (hopefully) very low current that would never be high enough to increase the cold value of the thermistor (so it's a resistor).

The Aleph schematic is adding to the confusion because there is an in-rush thermistor and a floating ground thermistor... (see what I did there?)

In the case of a catastophic failure in the amplifier, the current will have to flow through the floating ground thermistor for it to function. If the failure is on the primary side of the transformer, that thermistor is not in the circuit, but the other one is. If the failure is on the secondary side of the transformer, likely the component causing the failure (output transistor) is already damaged before the thermistor reaches its max current carrying capability. Think of a washer that falls between B+ and Chassis Ground in the schematic above. Because of the 10 Ohm thermistor, this will cause 4.8 Amps to flow through the circuit initially, not enough to break the 6 Amp fuse (thanks to the 6 Amp in-rush thermistor), but enough to cause substantial damage to the circuit as the floating ground thermistor heats up and cools down due to the current flow. A 2.2 Ohm resistor will fail immediately (21.8 Amps), open the link between Star Ground and Chassis Ground, and cause the 6 Amp line fuse to melt. Hopefully, the 15 Amp power panel circuit breaker trips.

Did I get that right? It was a stream of consciousness thing. I could be completely wrong, but this is still a good discussion.

 

[laatsch55]

Yes it is...love this stuff.

 

[mondialfan]

I was worried that the aleph schematic might throw a wrench into my explanation with the two thermistors being shown . But after thinking about it I realized it also highlights the two different uses for them.

In your failure scenario within a few secs of the fault current beginning to flow the 10ohm floating ground thermistor heats up and drops to a fraction of its cold resistance which will blow the fuse. The thermistors have a very steep drop off in resistance versus temp. Its an exponential curve with the steep part of the curve on the cold temp end of the chart. You might heat a trace or a wire up during that period but the fuse should blow before it begins destroying things.

If on the other hand a fixed resistor is in place between the star and chassis grounds and it fails quickly due to a spike event then the resistor may show no outward signs of failure. Many folks would repair the original problem that blew the fuse and any subsequent visible damage but may never realize the resistor in the ground loop has opened.

There's also the concern that the fixed resistor might open up before a slow-acting fuse which is what many amps call for. If you were using a smaller say 1/4w or 1/2w resistor it may go first.

I agree with you on the fixed value of the resistance between the grounds being the better choice in normal operation. However in normal operation there's not enough current flowing through the isolation thermistor for it to change much if any in resistance. A few mA of current will not heat them up.

If I were to use a fixed resistor it would be a larger wirewound. Mondial used a 50W chassis mounted wirewound for this purpose in the Aragon 2004/4004 amps. Later Aragon 8XXX series used anti-parallel diodes with a fixed resistor in parallel with them to isolate star and chassis grounds.

 

[mlucitt]

The Mondial and Aragon amps were certainly units to be reckoned with.

If you are dealing with a quality thermistor, I agree their specs are impressive and the stated performance can be a good design attribute.

However...

I like to build impressive reliable amplifiers, and selecting parts that improve performance while maintaining circuit integrity, is part of the challenge. I don't build amps to have them fail or blow up. If that happens on the bench or during a maximum SPL passage of Humble Pie "I Don't Need No Doctor (Live)" then I will have done something wrong and back to the bench it goes for a re-design, de-smoking, and/or re-build (that has not happened yet).

There should be no spike events between Chassis Ground and Star Ground if standard electronics practices have been observed and quality parts have been utilized in the assembly. I would be curious to know if Aragon/Mondial ever had an amplifier that used the fixed resistor and experienced a failure that tested the capabilities of that device?

 

[mondialfan]

The Mondial and Aragon amps were certainly units to be reckoned with.

If you are dealing with a quality thermistor, I agree their specs are impressive and the stated performance can be a good design attribute.

However...

I like to build impressive reliable amplifiers, and selecting parts that improve performance while maintaining circuit integrity, is part of the challenge. I don't build amps to have them fail or blow up. If that happens on the bench or during a maximum SPL passage of Humble Pie "I Don't Need No Doctor (Live)" then I will have done something wrong and back to the bench it goes for a re-design, de-smoking, and/or re-build (that has not happened yet).

There should be no spike events between Chassis Ground and Star Ground if standard electronics practices have been observed and quality parts have been utilized in the assembly. I would be curious to know if Aragon/Mondial ever had an amplifier that used the fixed resistor and experienced a failure that tested the capabilities of that device?

I just mentioned the spike event in reference to your dropped washer example.

I think we're all on the same page as far as our desire to build quality electronics. However no matter how bulletproof and perfect as we think they are there can still be faults lurking. It might take hundreds of operational heating and cooling cycles for a problem we overlooked to arise but when it does we want the component to fail in as safe a manner as possible and not put the end user at risk.