Dual Bridge PS - Does Phase of Secondary Matter?

[stratus46]

sorry to revive this old thread.

Just to say that i modded my diy power supply for my MOTU 24i/o interface from center tapped to dual rectifiers ( the trafo has dual secondaries) and the lm3*7 regulators get clearly less hot. All the rest remais the same.

I'll take it as an advantage.

RMAA readings on channel one in the MOTU remained the same after the mod.

Of course the LM3xx regulators run cooler. You reduced the Voltage into them from the extra diodes but you did not reduce the total dissipation. You just shifted some into the bridges which likely have better heatsinking. You would have achieved the same cooling of the LMs by adding a diode in series to achieve the same Voltage reduction.

G²

 

[DF96]

I'm interested to hear your opinions on the validity of these reasons.

First, the dual bridge approach prevents the circulation of DC in the secondary windings when the load current on the positive and negative rails is not equal. Transformers, especially toroids, don't like DC flowing in their windings, as it can push them toward core saturation, degrade transformer performance, and result in buzzing (this is also a concern for DC in the primaries due to residual DC on the mains).

Any DC in the secondary should be balanced by corresponding current in the primary. This is because the current is caused by transformer action. DC current caused by applying DC voltage creates difficulties, as then it is not balanced by current in another winding.

Secondly, the high peak rectification currents involving the loop consisting of the secondary, rectifiers and reservoir capacitor, those currents are resolved locally without passing through ground.

These currents should never pass through ground anyway, if the correct grounding scheme has been used. However, many people seem to connect their charging pulses through their star ground then wonder why they have buzz. The PSU ground should be isolated, except for a single connection from the earthy end of the last smoothing cap to the star - this wire carries DC plus a little unavoidable ripple.

Thirdly, it has been stated that EMI from the mains to the secondary has less opportunity to get to the circuits on the secondary side because the transformer windings are only "connected" to the secondary circuits during the brief rectifier on times. In the single-bridge rectifier scheme, the transformer is always connected to the circuits on the secondary side via the grounded center tap.

Is EMI modulated at 100/120Hz better or worse than unmodulated EMI? If you add caps across the diodes then this point disappears.

 

[Elvee]

Originally Posted by Bob Cordell
I'm interested to hear your opinions on the validity of these reasons.

First, the dual bridge approach prevents the circulation of DC in the secondary windings when the load current on the positive and negative rails is not equal. Transformers, especially toroids, don't like DC flowing in their windings, as it can push them toward core saturation, degrade transformer performance, and result in buzzing (this is also a concern for DC in the primaries due to residual DC on the mains).

Any DC in the secondary should be balanced by corresponding current in the primary. This is because the current is caused by transformer action. DC current caused by applying DC voltage creates difficulties, as then it is not balanced by current in another winding.

There isn't even the need to invoke the primary: the transformer's core never sees DC bias, because the balancing occurs between the two secondaries, even when dissimilar currents are drawn on the positive and negative sides.

 

[gmphadte]

The transformer core noise is eliminated in the dual winding dual bridge circuit, in case there is a little difference in the secondary currents.(written verbatim)

It is written in one thread here, by Nelson Pass.

Gajanan Phadte

 

[djmiddelkoop]

Thirdly, it has been stated that EMI from the mains to the secondary has less opportunity to get to the circuits on the secondary side because the transformer windings are only "connected" to the secondary circuits during the brief rectifier on times. In the single-bridge rectifier scheme, the transformer is always connected to the circuits on the secondary side via the grounded center tap.

Is EMI modulated at 100/120Hz better or worse than unmodulated EMI? If you add caps across the diodes then this point disappears.

No ! Bob means mains garbage, usually HF, can only enter through the bridge rectifier diode when conducting. A cap across the diode doesn't help, it just increases the coupling of mains garbage into the circuit. When using soft recovery diodes a cap is not needed anyway.

 

[gmphadte]

HF occurs when the diodes do not conduct.

Gajanan Phadte

 

[DF96]

Given that EMI is inevitable, and we can't directly hear EMI, is it better or worse to modulate it - which makes it more likely to be audible via subsequent rectification somewhere in the circuit.

There are two quite separate issues here:
1. incoming RF on the mains
Pure RF won't be heard although if there is enough of it it could shift bias points. In reality it will usually be modulated but do we have to add more modulation? Caps across the diodes allow in more RF, but with less modulation. Which is better? I don't know.
2. locally generated HF from diode charge storage
A cap or appropriate diode can greatly reduce this.

 

[digitrax]

Given that EMI is inevitable, and we can't directly hear EMI, is it better or worse to modulate it - which makes it more likely to be audible via subsequent rectification somewhere in the circuit.

There are two quite separate issues here:
1. incoming RF on the mains
Pure RF won't be heard although if there is enough of it it could shift bias points. In reality it will usually be modulated but do we have to add more modulation? Caps across the diodes allow in more RF, but with less modulation. Which is better? I don't know.
2. locally generated HF from diode charge storage
A cap or appropriate diode can greatly reduce this.

Sorry for the late replay, but my experience has been this:
In fully balanced circuits it is fine left alone (although I've never known a properly tuned snubber to hurt matters).

Just to clear up the two misconceptions I keep seeing: "ringing" in the rectifier is cause by stored charges released when the diode shuts off and is caused by the interaction of the rectifier's "snap back" (recovery) with the transformer's leakage inductance (and winding capacitance). It is not the capacitor across the rectifier(s) that dampens this, but rather the resistor in series with that capacitor. (In fact, you could use a resistor alone, but it would greatly increase power dissipation in the snubber circuit – hence, the cap).

As far as power supply ripple, RF "hash" and the like....
Some experimentation came about when, around a year ago, I was trying to determine whether to feed dual bridges (form separate secondaries) in or out of phase. i.e. Is it better to have additive total ripple (but with the positive and negative halves 180° out of phase from each other) or a supply wherein the ripple is in phase, making the difference between V- and V+ constant, but swinging up and down a few millivolts relative to ground?

The answer depended entirely upon the circuit topology. Where "balanced throughout" was employed, having the secondaries feeding the rec bridges in phase (such that the resulting ripple on the DC outs was essentially mirrored between V- and V+ as they're opposing polarities) was preferential. In circuits whose largest gain stages were single-ended, the opposite was true.

(Of course, YMMV).

 

[nigelwright7557]

OP your lower 10,000uf is wrong way round.