Rectifier query

[wally.gribbles]

Hi all, I'm back again with another (probably quite simple) question...

I picked up a so-so, incomplete DIY amp a little while ago that somebody else made. It's ugly, real ugly--what I'm using it for, more than anything else, is a good source of spare parts

Anyway, I have traced the circuit to discover that the rectifier on the existing circuit is as illustrated in the attachment. Frankly, I'm stumped as to how it manages to work...

That is, with the two diodes wired in as they are, it would seem that this should only be a half-wave rectifier, but it's clearly not since the B+ voltage is 315V (measured). And what are the capacitors I've labelled C1 (with .01M Z5U printed on them) supposed to be doing?

I'm OK with how the B+ supplies work, I'm just a bit confused as to why this rectifier is configured the way it is, and how it manages to work. Any insight would be greatly appreciated!

(As you may have guessed, I'm not an electrical engineer )

Many thanks in advance.

[hermanv]

This is a half wave voltage doubler. The lower 100uF capacitor is charged to the peak winding voltage by the lower diode on the first half of the sine wave. The upper capacitor is also charged to the peak winding voltage but this time by the upper diode on the second half of the sine wave. The output voltage is the sum of the two capacitor voltages and double the transformer voltage.

This kind of design has poor ripple (60Hz instead of 120HZ like a full wave bridge) and poor output regulation (double the winding Ohms and double the capacitor droop).

[wally.gribbles]

Ah, I see--so it's the C1 and 100uF caps that are causing the B+ voltage to get to where it is, and the design itself that is causing some the problems I've already discovered on this unit (the poor output regulation was immediately obvious).

Thanks!

[SpreadSpectrum]

C1's purpose is to absorb diode switching noise.

[SpreadSpectrum]

I am going to have to respectfully disagree with hermanv. This is a full wave voltage doubler and I use them all the time and like them. The thing you have to remember with them is that they double the voltage capability of the transformer but they also halve the current capability. (you can't get something for nothing)

[SpreadSpectrum]

Furthermore, if you are charging the top capacitor on one half of the AC cycle and charging the bottom capacitor on the other half of the AC cycle, you have to get 120Hz ripple. Correct me if I am wrong.

[Eli Duttman]

SpreadSpectrum is correct. That topology is known as a full wave voltage doubler. Some justification for the name exists. The ripple freq. is 2X that of the AC mains and both halves of the sine wave are used.

Properly executed, "full wave" doubler B+ PSUs are quite good. Some of the best units ever made employ the topology. The H/K Cit. 2, Marantz 8B, and Fisher 500C are examples.

LARGE caps. are appropriate in the doubler stack. Large 1st filter caps. imply a small conduction angle and a good deal of ripple overtone energy. Follow the doubler stack with a low DCR inductance and more capacitance. Think "extended" CLC filter. A H/K Cit. 2 overhauled ala Jim McShane kicks serious butt. Jim replaces the OEM doubler caps. with low ESR 820 muF. Panasonic 'lytics. The rest of the B+ PSU is beefed up too.

FWIW, we used something quite similar, at a lower current, in "El Cheapo" and (unsurprisingly) obtained highly satisfactory performance.

TANSTAAFL applies and the available DC current is approx. 1/4 the AC RMS current of the rectifier winding.

[wally.gribbles]

Yeah, a google search revealed that I have a full-wave doubler here, not a half wave, but that's OK.

However, the poor voltage regulation is a very real thing--I worked out the transfer function for the whole thing, and compared the actual (measured) results with the computed ones (assuming a true DC input signal) to find some reasonably large discrepancies. This is now easily explained by both the 100 Hz ripple current (2x AC mains frequency) and the undersized caps. The two factors combined effectively mean that the filtering is less than what it needs to be.

Either way, I have once again learned a bundle--best forums ever

As well, I'm planning on doing things differently in the upcoming build anyway--full wave rectification via an RCA 82, with the input coming from a purpose built transformer (no voltage doubler or anything).

[hermanv]

Yep, you guy's are right (damn ) it is a full wave and as a result the ripple is 120Hz (ripple from lower Capacitor interleaves with ripple from upper capacitor)

[wally.gribbles]

@hermanv: Not your day today, as 2x 50Hz = 100Hz