What's the best diode for pwramp supplyyy

[primalsea]

I've just got my small poweramp project to the it "works good stage" prior to further tweaks to milk more performance.
I bought a couple of broken Cambrdige Audio A1's, the very 1st ones. I've fixed one up so far and given it minor tweaks. Things like a current mirror in the LTP so its now balanced!, bigger XO, single retifier feeding 2 4700uf caps which then feed 4 more 4700uf caps through so .22ohm resistors, seperating the +/- supplies on each channel. I plan a few further things like changing the TIP3055 & 2955 OP devices to something more robust and removing the VI limiting and replacing it with relay based speaker protection. Also better rectifiers in the PS. I plan to make / mod 3 amps for my triamp speaker project. These are for use in my living room so a well thought out powersupply with clean rails rather than a brute force design with the biggest of everything is more important, the amp is only 56 watts something I will never ultilise fully in my home environment.

Anyway thats just some background if anyone is interested.

Looking at the current (no pun intended) PS rails on my scope I get the usual sawtooth type waveform with a ripple of about 60 or 70mV at idle. With the scope at .2us/div & 5mv/div the waveform is a bit jumpy. This may be just a feature on a unregulated supply when you zoom right in with your scope, I'm not sure. I was wondering if things would improve if I used "better diodes" for the rectifier. The current ones are 1N5304's with a 100n cap in parrallel.

Thing is Im not sure what would be best, just go for a large 35A metal case rectfier or something a bit fancier like some large soft recovery diodes?????

The caps were a bit of a compromise. 4 of them are huge LCR 4700uf 100V things that might be quite old. I bought them for 40 pence each so can't complain with performance to pounds ratio. The other 2 are 4700uf 40v caps that are smaller in size new old stock and only 70p each from a junk shop. Im aware that as they are old their ESR may have increased a bit but they are so physically huge I guess this was very low anyway, Also larger sized caps tend to have better ripple currents!

I also guess that he caps are not the cause of the wobbly wave form of the rails anyway but if I'm wrong set me striaght, wont you.

[sam9]

I think even a theoretically perfect diode will produce ripple. In addition to ripple, a rectifier diode (excluding Schotkys) produces HF noise when it turns off. The old fashioned way to reduce the HF noise is a snubber - see www.hagtech.com for an article on this. Others try exotic diodes to reduce the HF but I suspect that does no better than using appropriate snubbers.

As for ripple, ever bigger filter caps will reduce the ripple, but there are diminshing returns. A regulated supply can ("can" doesn't mean always) eliminate ripple but then there are issues with that, too. Possibly the least troublesome, but not very exotic or gee-whiz solution, is to combine substantial filter caps with an amp that has a high PSRR.

[costiss]

...ixys screw terminal 100A... i think its the absolute top...
otherwise generic 35a bridges will do the job well for long enough...

[jackinnj]

Originally posted by primalsea
I've just got my small poweramp project to the it "works good stage" prior to further tweaks to milk more performance.
I bought a couple of broken Cambrdige Audio A1's, the very 1st ones. I've fixed one up so far and given it minor tweaks. Things like a current mirror in the LTP so its now balanced!, bigger XO, single retifier feeding 2 4700uf caps which then feed 4 more 4700uf caps through so .22ohm resistors, seperating the +/- supplies on each channel. I plan a few further things like changing the TIP3055 & 2955 OP devices to something more robust and removing the VI limiting and replacing it with relay based speaker protection. Also better rectifiers in the PS. I plan to make / mod 3 amps for my triamp speaker project. These are for use in my living room so a well thought out powersupply with clean rails rather than a brute force design with the biggest of everything is more important, the amp is only 56 watts something I will never ultilise fully in my home environment.

Anyway thats just some background if anyone is interested.

Looking at the current (no pun intended) PS rails on my scope I get the usual sawtooth type waveform with a ripple of about 60 or 70mV at idle. With the scope at .2us/div & 5mv/div the waveform is a bit jumpy. This may be just a feature on a unregulated supply when you zoom right in with your scope, I'm not sure. I was wondering if things would improve if I used "better diodes" for the rectifier. The current ones are 1N5304's with a 100n cap in parrallel.

Thing is Im not sure what would be best, just go for a large 35A metal case rectfier or something a bit fancier like some large soft recovery diodes?????

The caps were a bit of a compromise. 4 of them are huge LCR 4700uf 100V things that might be quite old. I bought them for 40 pence each so can't complain with performance to pounds ratio. The other 2 are 4700uf 40v caps that are smaller in size new old stock and only 70p each from a junk shop. Im aware that as they are old their ESR may have increased a bit but they are so physically huge I guess this was very low anyway, Also larger sized caps tend to have better ripple currents!

I also guess that he caps are not the cause of the wobbly wave form of the rails anyway but if I'm wrong set me striaght, wont you.

With regard to Sam's comment on the snubber -- what you will see if you closely look at the waveform are small spikes residing on the triangular or sawtooth waveform. These are the result of the interaction of the diode capacitance and the transformer inductance. The Hagtech article is well worth the read (as are the application notes on www.onsemi.com wrt snubbers.) The correct snubber will reduce the size of these spikes -- they are the kind of annoyance which makes its way through your amplifier regardless of PSRR.


MUR860's seem to be the rage these days.

[primalsea]

Previous post:
__________________
i am currently using 60,000 uF of capacitance in my preamplifier..


I've just finished the psu of my preamp. It has 4 X 33,000uf caps feed by 2 X 10,000uf caps with 1.2 ohm resistors between the capacitor banks. The XO is about 5-0-5v, 200VA. With all that capacitance and resistance I get +/-7.8V. The waveform is very smooth and I am not using any regulators at all.

This may seem OTT for a preamp but I'm working on the idea that the smallest signals need the least amount of fluctuation from the PSU.

The 10,000, 40V caps I got for nothing and the 33,000uf, 10V, 20A ripple Computer Grade caps cost me 40 pence each from the same place I bought the LCR caps in my poweramp. The XO also cost me nothing so cost to performance ration is very good even if the PSU does not make much of a diiference compared to a smaller one!!!!!!!!!!!!

[darkfenriz]

aren't zeners in rectifier bridge a cure??

[Mr Evil]

Schottky diodes of course! More efficient, less noise... what's not to like?

[jacco vermeulen]

Originally posted by primalsea
changing the TIP3055 & 2955 OP devices to something more robust

I think these used to be referred to as "work horses".

[primalsea]

At about 70 pence each those output devices just can't be anything special!!