# Rectifier question

Hi all,

I've been looking at rectifier circuits and my brain has blanked out. Please see attached image (and click it for full res):

Something is telling me that the top circuit "uses" both secondary windings for both halves of the AC cycle (i.e. the transformer is fully loaded) while the bottom circuit only "uses" one secondary at a time (i.e. one secondary for each polarity).

I've tried drawing it on paper and tracing the current flow for both halves of the AC cycle and for some reason, I just can't wrap my mind around it.

Am I nuts, or is the top circuit "twice as good" as the bottom circuit?

I know that the top circuit has two diode Vf drops while the bottom has only one... that's not what I'm asking. What I want to know is this: Are both circuits functionally identical, or is the top one "twice as good" as the bottom?

Thanks!

(click pic for full res)

I would think that the upper one would be able to supply nearly twice the current provided that the transformer core does not saturate and the two windings are equal.

I agree. I've never seen the bottom one that way. Only time I saw something like it was in a dual rail supply. Aka, a mirrored negative line was added (which loads the other half of the wave).

Only advantage of the bottom circuit is that the 0V is connected directly to the transformer. Sometimes you might need that. You could also do that with a single winding (and single diode) as well but with more ripple.

The top one is called a "full wave bridge rectifier"
The bottom one is called a "full wave centertap rectifier"

As was stated the top one uses the both secondary windings all the time, however will have two diode drops.

The bottom,will have more of a loss in the transformer because at any given time the IR drop in the secondary will be 2x that of the full wave bridge.

In higher power circuits the full wave bridge is usually chosen, however in low power circuits you will often see the half wave version.

Now to really complicate things, if you need +/- supply you will end up with both! It is easiest to think of the bottom circuit with two sets of diodes going in each direction, which it turns out to be the same as the top circuit with the center tap grounded (and capacitor moved)

The great advantage of a centre-tapped transformer is that a bridge will generate both +ve and -ve rails.

If that's useful in a project ...

Allan

The center tapped version was pretty much the only version used in the days of valve systems. Typically the two diodes were inside one valve with a common cathode and two anodes. As well as a capacitor on the output it was common to use a choke as smoothing and frequently the choke doubled as the electromagnet used in the loud speaker. Ah the good old days!

That is going back a bit! Those old radios also often had a 'hum bucking' potentiomer to try and cancel out the inevitable varying magnetic field and resulting 100 Hz hum in the sound... I think it was wired across the ac heater of the output valve and fed to the grid bias...

Those were indeed the days!

Allan

allanhurst:
That is going back a bit! Those old radios also often had a 'hum bucking' potentiomer to try and cancel out the inevitable varying magnetic field and resulting 100 Hz hum in the sound... I think it was wired across the ac heater of the output valve and fed to the grid bias...

Those were indeed the days!

Allan

I don't remember how the designers attempted to nullify the 100 Hz hum but it was just about impossible to get rid of. I used to like being given those sets since the speakers were a good source of enamelled copper wire.

Due_unto:
I would think that the upper one would be able to supply nearly twice the current provided that the transformer core does not saturate and the two windings are equal.

Firstly transformers saturate due to net amp-turns of both primary and secondary, which mostly cancel
out, so saturation is a complex function of several things, like winding leakage inductance as well
as current in particular windings.

The thermal limit of the transformer as a whole will not, in general, be twice as large when using
two secondaries at once - consider the extra primary current contributing to the thermal load.

In reality the conduction angle of the rectifier/smoothing capacitor load is important too.

But yes, bridge rectifier is more efficient. That's why you hardly see anything else after the valve
era of circuitry (rectifier valves were costly).

MarkT,

I'm not sure I follow you regarding saturation. I was under the belief that saturation was due to volt-seconds not amp turns.

I don't remember how the designers attempted to nullify the 100 Hz hum but it was just about impossible to get rid of. I used to like being given those sets since the speakers were a good source of enamelled copper wire.

Vandal!

but I did much the same , stripping old tele's and radios for parts. All one could do as a poor schoolboy with an electronics hobby back in the '60's .

Allan

stowite:
I don't remember how the designers attempted to nullify the 100 Hz hum but it was just about impossible to get rid of. I used to like being given those sets since the speakers were a good source of enamelled copper wire.

Old tube (valve) audio amps usually used a "5U4" type of rectifier (directly heated cathode... the filament was the cathode).

Because the filament had 5 volts AC across it, the AC hum added itself to the B+ supply (which, because of filter caps in the power supply, ended up being on the chassis ground instead). The 5 volt hum got added to the audio input via the ground.

The solution was to place a potentiometer of around 25 to 100 ohms or so across the rectifier filament, and connect the wiper to chassis ground.

This created an adjustable "center tap" for the rectifier filament, and adjusting it properly would balance the hum out (because when one side of the filament was going up to 2.5 volts, the other side was going down to -2.5 volts and being common mode, they cancelled out.

Another solution to this problem was to use a center tapped 5 volt winding for the rectifier filament and connect the center tap to chassis ground. Although this worked well, it wasn't adjustable and due to variations in the filament (specifically the thorium coating to enhance electron emission), the "electron conductivity" of the filament was not uniform across the whole length of the filament, so an imbalance in current from different parts of the filament to the plate(s) allowed hum to leak in and it couldn't be nulled away.

The solution was to place a potentiometer of around 25 to 100 ohms or so across the rectifier filament, and connect the wiper to chassis ground.

Surely this can't be right. With a directly heated cathode it'd be sitting at several hundred volts above ground, and this trick would ( nearly) short it out.

Allan

That's for the other filaments, not the rectifier filament which needs its own winding.

MarkT:
That's for the other filaments, not the rectifier filament which needs its own winding.

Yes. Usually a tube (valve) circuit uses 6.3 volt or 12.6 volt filament power (and these usually indirectly heat a cathode which is near ground potential). The transformer winding(s) for these are no big deal.

But, the rectifier (such as a 5U4 or 5Y3) uses a directly heated 5 volt filament. And, since the filament is the cathode, full B+ voltage is present on it. This requires a special transformer winding which is (1) 5 volts, not 6.3 and (2) very well insulated from the other transformer windings and the frame (to withstand the full B+ voltage).

By the way, one of my college labs included plotting the grid/plate characteristics of a 6C4 triode on graph paper, then comparing it with the red RCA tube manual (shows you how far back I go!)

One of the students had trouble getting any data from his tube. He forgot to power the filament! LOL!!!