DC block designs compared

[okw]

I've been playing around with various homemade DC-blocks, but results seems to be quite similar. I measure around 5Mohm between in/out for all blocks. I was expecting it to be much higher(?).
Are there any pros and cons between the various designs considering it's gonna be a part of a 240V 7A mains filter for amps?

[moffy]

I don't know why they are called DC blocks, any DC above 1 or 2 diode drops will pass right through except #4, that is 6 diode drops.

[tooki]

The caps in the current design will experience shorter lifespans, since they're polarized caps being exposed half the time to negative voltage, even if it is limited by the diodes.

See https://sound-au.com/articles/xfmr-dc.htm

I don't see the point in alternative 3. Why would you want a resistor in parallel to the DC-blocking caps, allowing through DC?

[tooki]

I don't know why they are called DC blocks, any DC above 1 or 2 diode drops will pass right through except #4, that is 6 diode drops.

Well, one should never be seeing multi-volt DC bias on mains AC, since that causes huge problems for the components of the grid. See the link I posted in the previous reply.

[moffy]

I don't know why they are called DC blocks, any DC above 1 or 2 diode drops will pass right through except #4, that is 6 diode drops.

Well, one should never be seeing multi-volt DC bias on mains AC, since that causes huge problems for the components of the grid. See the link I posted in the previous reply.

So they only block mV or as I stated 1 to 2 diode drops except for the last one.

[Gyro]

I don't know why they are called DC blocks, any DC above 1 or 2 diode drops will pass right through except #4, that is 6 diode drops.

Well, one should never be seeing multi-volt DC bias on mains AC, since that causes huge problems for the components of the grid. See the link I posted in the previous reply.

So they only block mV or as I stated 1 to 2 diode drops except for the last one.

They're intended to block small (<1.5V) DC offsets caused by certain half-wave / unbalanced electronic devices. Small DC offset can cause saturation issues in transformers, particularly torroidal ones.  Whether they are needed and whether such offsets are generally present is debatable, and have been debated here before.

[MariuszD]

I measure around 5Mohm between in/out for all blocks. I was expecting it to be much higher(?).

I don't see the point in such a measurement.

Check how the circuit affects the operation of transformers, and what the typical DC values in the grid are.

I checked, there is an impact on the volume of the toroidal transformer buzz that is audible when I'm not listening to music. Sometimes the buzz changed and I didn't know why. The cause was a DC.
For me, the change in sound is audible when the DC component exceeds 200mV, but usually it is below 100mV.

The grid has low impedance, making it difficult to generate a significant DC component. However, sometimes a diode was used to limit the power of the heaters, such a setup causes several hundred mV DC and loud transformer humming.

[okw]

I measure around 5Mohm between in/out for all blocks. I was expecting it to be much higher(?).

I don't see the point in such a measurement.

Wouldn't this tell me how much DC can pass? If the resistance at DC (which the DMM uses to measure resistance) is infinite or very very high, it means it's effectively blocking DC, right?

[MariuszD]

I measure around 5Mohm between in/out for all blocks. I was expecting it to be much higher(?).

I don't see the point in such a measurement.

Wouldn't this tell me how much DC can pass? If the resistance at DC (which the DMM uses to measure resistance) is infinite or very very high, it means it's effectively blocking DC, right?

You don't know at what voltage the resistance was measured.
You will receive some number that depends on the voltage you don't know.

[TimFox]

DMMs normally measure DC resistance by forcing the current (maybe 0.1 to 1 mA) and measuring the voltage (maybe < 0.2 to 2 V).

[floobydust]

There's no way to measure the asymmetry, the net DC? Just make a test jig.

Realistically, the power transformer noise is due to mains harmonics on the core & winding's vibration.
You'll notice it quieter at nighttime when industrial loads are not running and power quality improves. Growl I find is from high mains voltage working the core flux higher.

Electrolytic capacitors don't mind some reverse DC: "capacitors may withstand continuous application of 1.5V reverse voltage" ref. CDE knowles app guide

[ejeffrey]

As MariuszD said, the diode is non-linear so the measured "resistance" depends on the voltage.  You can most strikingly see this on a manual ranging DMM: cycle through the ranges and the measured value will change dramatically for each step.

In addition, 5 megaohms is probably just a measurement error.  The expected small signal resistance of a zero biased diode is approximately Isat / 26 mV.  Isat for big power rectifiers is typically 10s of nanoamps to a microamp even at room temperature.  The data sheet for the GBJ2510 shows a plot with reverse leakage of 100 nanoamp at 25C.  That would give a small signal resistance of 250 kOhm, or 500 kOhm for two in series.  On top of that, DMMs don't measure near zero offset, they usually operate with a voltage of at least 100 mV which will cause the reported resistance to be much lower.

[TimFox]

Measuring a relatively small DC offset on a large AC voltage does not work simply with a normal DMM, e.g. with a 200 mV DC full scale on 120 or 240 V rms AC.
One needs to carefully filter the voltage to attenuate the AC component to avoid over-driving the DC circuitry.  An old-style DC meter (e.g. Simpson 260) may work better for this measurement.