Diode Rectifier

[isaac]

Hey guys, what is the difference in these two rectifiers that would make one superior to the other?
http://bb.steelguitarforum.com/userp...3_layout_1.jpg
http://metroamp.com/wiki/images/1/17/SSRectifier.jpg
Any help is appreciated.
Thanks,
Isaac

[JHow]

They look similar to me. Two phase, full wave with silicon diodes. The gurus will probably explain why fender went with three in a row (greater voltage rating?). Best practice on switch is in hot line after fuse, not neutral line as in the metro amp.

[Tom Phillips]

+1.
JHow nailed it.
Back in the day when Fender did their original design the Peak Inverse Voltage (PIV) rating of the available silicon diodes was lower than is available today so they used three in series.
As long as the single diode doesn't blow up there is no affect on the sound vs. the three in series.
Tom

[isaac]

I was hoping that the six diode rect. would produce more voltage. Not sure though, what can gurus tell me. Guru, love that word. ;-)

[isaac]

Well, Dumble amps have the standard three+three, could there be even more to it. Dumble is a smart guy, and he built most of his famous amps, (as far as i know) in the 80's or newer, so could there actually be a difference or did he just not bother to change the design from the twin?
Ok thanks.
-Isaac

[Tom Phillips]

I was hoping that the six diode rect. would produce more voltage...)

There is a fixed voltage drop of approximately 0.6V through each diode. Therefore, placing additional diodes in series actually produces less output voltage. However, at tube amp power supply voltages, the difference is insignificant.

[isaac]

Ok well thanks.
-Isaac

[Tom Phillips]

Depending on the PIV rating of the diodes you use and the voltage of your transformer winding there can still be a reason to use 2 or more in series. It's just that it's to make sure the diodes survive. Not to produce a better sounding amp.

[Enzo]

The two supplies are the same. The extra diodes don't change anything. The extra couple diodes per side will drop about an extra volt, so instead of 475v B+, you'd have 474v. Considering that a 2v change in your mains voltage - which happens all the time - would change your B+ by 7 or 8 volts, you would never know the difference.


As Tom said, three in series creates a higher reverse voltage rating, so they didn;t have to go find special high voltage diodes that were expensive. Nowdays 1000v diodes are a penny. Your extra diodes don;t make more voltage, they just can HANDLE more voltage. Like using a heavier extension cord doesn't make your clock radio louder.

Dumble did it? I am sure Dumble is a smart guy. But that is what we call an Argument from Authority, and is a form of logical fallacy. We really cannot say "Dumble did it, so it must be a good idea." The reverse may be fine - it is a good idea, so Dumble did it (too). DUmble may have done it for the same reasons Fender did, to make sure he had enough reverse voltage protection without using expensive parts. On the other hand, he may have done it because it looked "right". Or because people expect it.

[R.G.]

Nota Bene:
When sticking things in series to get more voltage capability, (as sometimes done with diodes or capacitors) the dynamic voltages and leakages matter.

If you have three 100V diodes, can you be sure that three of them will hold off 300V?
No. The actual voltage across the diodes will divide by the ratio of their leakage resistances, which are not specified, and may not be stable over time and temperature. If one of them leaks a bit, it's possible to get more than 100V across one of the others, at which time it may short. This leaves the remaining two 100V diodes to hold off 300V - and they will die trying to do it.

Same argument with caps. The leakiest device puts more voltage on the others in the string, kills the strong one(s), and then may kill the weak remaining ones.

Same argument with dynamic voltage. When several diodes are trying to withstand quickly varying voltages, the ones with high capacitance take longer to charge, and leave a higher voltage for the low-capacitance ones. The leaky ones for DC may not be the same ones with high capacitance for AC.

The right way to do this is with high-value resistors to equalize the voltages across diodes and capacitors for DC, and low-value caps in parallel with diodes to swamp the internal diode capacitances to make them not able to leave some devices low and some high on quickly changing voltages.

[isaac]

I wasn't saying he did it, so it must be good, I was just wondering if someone here would have a good idea as to whether it would make sense for him to do it, besides by following example. But thanks for your explanation and time.
The amp is sounding really nice,(the one I'm trying to do this stuff in, although this is just Pwr Supply, Still thought you gusy might like to know.)

[Dave H]

Three diodes were used to handle the Peak Inverse Voltage (PIV). For that type of rectifier the PIV is higher than you might think. The capacitor charges up to the peak voltage on the positive half cycle then on the negative half cycle the voltage swings to minus the peak voltage so the diode has to withstand twice the peak voltage. For a 500V B+ supply that’s 1000V so even a 1000V PIV 1N4007 diode is marginal.

[R.G.]

Yep.

[Steve Conner]

I believe with modern diodes, the breakdown characteristics are predictable enough that it's OK to connect them in series. I've seen many high voltage rectifiers that are made of a bunch of diode chips connected in series internally.

Really you're supposed to use "avalanche rated" diodes.

[R.G.]

I believe with modern diodes, the breakdown characteristics are predictable enough that it's OK to connect them in series.

I **want** to believe...

I've seen many high voltage rectifiers that are made of a bunch of diode chips connected in series internally.

Me, too. However those were originally selected parts that were tested and binned with no more than X difference between them. As you note, maybe today's diodes are so good and so uniform it's not necessary. I have trouble believing, having seen the lengths to which part manufacturers will go to improve "yield".

Really you're supposed to use "avalanche rated" diodes.

Yep.

But all the amp techs I know of will reach in and grab several 1N4007s of unknown vintage and manufacturers and wire them up, given the intense pressures of time and money they're working under. There may be some who will stock the right stuff. Enzo would.