Yellow sheet mod

So I was chatting with Jim McShane yesterday over email (super nice guy BTW) about some options for my rectifier tube .

While we were chatting, he asked if I had done the yellow sheet mod yet and if not he highly recommends it with some UF4007 instead of 1N4007 Diodes. When I replied that I had indeed done the mod but had used 1N's he said it wasn't a big deal but i should change them out for UF at some point.

So my question is, what's the difference??

UF are "ultra fast" versions of the 1N. Both can provide the same levels of protection (based on the voltage rating), but some say the UF's sound better.

I'm sure there's two camps on that, just like the "do rectifiers sound different" argument. Flip a coin, eh.

PS ... I used 1N's on my ST120, mostly because I got a pile of em in the stash.

I used the 1N4007 Diodes on my m-125s and no problems after over two years of use.

I've had issues with UF (ultra-fast) diodes in other amps causing power supply hash noise. The abrupt turn off of UF diodes resonated with the Ls and Cs of the power supply to cause noise spikes. Problem solved when slower diodes were substituted. UFs are an overkill with slow commutation 50/ 60 Hz rectification. The 1Ns are perfectly fine with these type power supplies.

arledgsc wrote:I've had issues with UF (ultra-fast) diodes in other amps causing power supply hash noise.  The abrupt turn off of UF diodes resonated with the Ls and Cs of the power supply to cause noise spikes.  Problem solved when slower diodes were substituted.  UFs are an overkill with slow commutation 50/ 60 Hz rectification.  The 1Ns are perfectly fine with these type power supplies.  

Yes, they are.

There comes a time when the point of diminishing returns is met and exceeded - and when exceeded can cause unforeseen and unintended results.

In the "yellow sheet" mod the diodes actually do the job of rectification - which means they do turn on and off in normal service. The 1N**** series diodes produce PN "burst" noise on turn off due to movement of the minority carriers ("holes") across the PN junction. This movement creates a noise burst which - even if the frequency is outside the audio band - can intermodulate with other frequencies and create noise. The UF diodes have FAR less PN burst noise than the UF series. I've used 100s of UFs in a wide range of products and have yet to have any issues arise - that's not to say they couldn't under the right conditions.

If you are concerned that the turn off is too abrupt and can cause disturbances in the transformer secondary you can snub the UF diodes. The late John Camille wrote about this topic in Sound Practices issue #7 in an excellent article entitled "Taming Diode Noise".

My own preference is for fast/soft recovery diodes. This is based on the Audio Amateur article by Rick Miller comparing diode noise, etc. It was published in January of 1994. You can find more on the topic if you do a search for this application note - it's General Semiconductor's Quick Note 108 - "High Speed Rectifier Applications in High End Audio"

I highly recommend both Rick Miller and John Camille's articles!

Sadly, the diodes rated best in Rick's research are now all NOS only with very small supplies left.

You might find this thread interesting:

https://www.audioasylum.com/audio/tweaks/messages/5/55927.html

I do not believe that the use of higher performance diodes in the "yellow sheet" application falls under the heading of diminishing returns. The cost of diodes better than the 1N**** series is minuscule.

Jim:

Not by cost, by function. Agreed that the incremental cost of the typical small component between option a or b is minuscule, or less. I use 1N4007 diodes most of the time when an -004, or less would do. After which, not so much.

Keep in mind that even the slowest silicon diode is many dozens of thousands of times faster than 50/60 Hz. But, we do many things because we can, not because we must, or even should.

Peter W. wrote:Jim:

Not by cost, by function. Agreed that the incremental cost of the typical small component between option a or b is minuscule, or less. I use 1N4007 diodes most of the time when an -004, or less would do. After which, not so much.

Keep in mind that even the slowest silicon diode is many dozens of thousands of times faster than 50/60 Hz. But, we do many things because we can, not because we must, or even should.

But if you read the articles I linked to you'll find empirical evidence of why better diodes make sense. It's not just the electronic equivalent of "a little chicken soup can't hurt". It is not an academic exercise, it is a practical application based on demonstrated merit.

Jim McShane wrote:
My own preference is for fast/soft recovery diodes.

Yes, fast/ soft recovery is the type used to fix the guitar amp. Philips BYM26E which are no longer in production. I tired a lot of rectifiers and for some reason this amp was really picky on diodes. Its power supply was lightly filtered compared the ST120. But the fast/ soft recovery diodes work well at 50/60 Hz.