UF5408 vs 1N4007

[Dmoney]

last night i was reading about swapping out 1N4007 diodes in amp power supplies for UF5408's.


any of the builders here experimented with this. i've read mixed reports

[HTH AMPS]

what did YOU think the differences were?

[Dmoney]

i haven't done it yet.
so i can't tell you.

i was wondering if anybody else had.

[HTH AMPS]

I personally think that the top end is a little smoother with UF diodes.  This is not a good thing imo.  The tone with the diode switching noise (all that top-end cr@p) is part of what we're used to hearing and associate with those great tones on records we dig.

I just use 1N4007s (1A) or 1N5408s (3A) where required and find my amps to sound fine.

[Keven]

by switching noise, do they mean the noise that happens when the DC current receives AC current from the negative and positive phase? what they call DC ripple right? i remember reading power supply filter caps tended to cure that little bugger a little by making AC closer to DC with little ripple (depending on the charge of the capacitor)

so as to the topic of 1n4007 vs UF5408, i guess the UF's turn on faster and limit voltage better than the 1n's which i recall reading are slower to turn on so there's something happening when the phases switch over that takes up more time than it should.

and you're telling me that a more efficient diode actually lets less high end through. sacrebleu, tone is definitely not electrically correct!

i wonder though, if the UF diodes turn on faster and thus kinda reduce ripple (do they?), i guess bigger power filtering caps also reduce ripple and also reduce high end, or is this a different kind of animal since the current is pretty much DC by the time it reaches the power supply caps?

sorry to hijack, but i felt this contributes to the nerdy conversation :D

[Dmoney]

you can have filter caps that are too high in value and I believe the result is ghost notes. you need to work out what you need and balance it out for filter caps.

someone who knows more can clarify that.

i thought faster switching diodes reduces sag in the power supply. The effects on tone ive read about seem to be pretty minimal. ie not a night and day difference. I've got some on order anyway, didnt cost much. gonna try them in my sommatone since its turret board built. easily to test this kind of thing on.

 

[jpfamps]

Ther UF diodes actually turn off faster than normal diodes, and have really been developed for use in switch mode power supplies.

There are two main switching noise issues to consdider with silicon diodes in rectifers:

Firstly, a small voltage is need across the diode for the diode to conduct. This is typically 0.7V for a 1N4007 diode. The consequence of this that there is a region around the zero crossing point of the input waveform where there is no conduction. This causes harmonics extending into RF to be induced on the power rails AND back into the transformer. These harmonics can be coupled to other windings, eg the filament supply. UF diodes still have a required forward voltage drop, so won't improve this situation.

Secondly, when reverse biased it takes some time for the charge carriers in the diode junction to swept out, so there will be a small reverse current. This can be thought of a a very small capacitor in parallel with the diode. UF diodes switch off faster (ie have lower capacitance) so in theory would actually cause more hash on the HT rail!

The UFs low capacitance is vital in a switching mode power supply when rectifying say 50kHz and above, but doen't actually help in a linera supply.

The easiest way to reduce switching noise is wire a cheap 10-100 nF 1kV ceramic cap in parallel with each diode in the rectifier ie increasing the capacitance of the diode.

For lower voltage DC supplies, eg DC filament, I use Schottky diodes which have the lowest forward voltage drop and soft recovery characteristics.