Tube Amps / Music Electronics
For current discussions, please visit Music Electronics Forum. New: visit Schematic Hell!
The sunn still shines online!

ampage archive

Vintage threads from the first ten years

Search for:  Mode:  

Diodes before Rectifier

11/8/2005 9:15 PM
Smitty Diodes before Rectifier
I normally don't work on amps with PCBs, but I had a reissue Fender Super Reverb on the bench today. I noticed that the two HT windings had silicon diodes in series before pins 4 and 6 of the the 5AR4.  
Why would Fender do this? Does this make the amp solid state rectified? You still get sag, though. Right? You still get the switching noise from the diode. Right?  
Does it increase PIV? Is that the big killer of tube rectifiers these days?
11/8/2005 9:25 PM
It's a safety measure. If the rectifier tube shorts, the silicon diodes take over and prevent burning out the power transformer and the filter caps.  
It's a ***HIGHLY***RECOMMENDED***MODIFICATION*** on all tube-rectified amplifiers.  
Unless the tube rectifier is kaput, you do not get turnoff noise from the solid state diodes. It does NOT make the amp solid state rectified unless the rectifier tube shorts. You still get the same sag as with the tube normally.  
In fact, the SS diodes have remarkably little effect unless the rectifier tube does die. Then the amp gets a bit louder as the SS diodes have a lower voltage drop.
11/10/2005 1:04 PM
Jack Koochin

Not sure how this works....  
If the rectifier tube shorts out, it will likely blow a fuse. If the tube fails due to filament burn-out or something, the tube won't conduct.  
So how do the diodes help out with this?  
11/10/2005 3:38 PM
Think of it like this Jack:  
If I wanted to, I could remove the tube rectifier entirely and replace it with two appropriately rated diodes. (I'm assuming the typical two-diode centertapped B+ transformer here.) Each diode goes in series with one end of the B+ winding, and the junction of the two diodes connects to the filter cap. All fine, many amps work this way.  
Now I break the connection between the two diodes and the filter cap and insert a tube rectifier, with its anodes paralleled, cathode to the filter cap. If properly fed with heater power, this works too, but now the tube's internal resistance causes some voltage loss, so the voltage on the first cap is lower. Some people have actually done this to add sag to a SS rectified amp.  
In the previous example the two SS diode cathodes are connected, and then connected to the paralleled anodes on the tube rectifier. If I instead connect only one diode to one anode, the other diode to the other anode instead of paralleling the SS cathodes and vacuum anodes, now it still works, because I have a SS diode and 1/2 of the tube rectifier in series on each half cycle of AC. This works just like the tube rectifier, sag and all, but I lose 0.7V to the SS rectifiers.  
Now I remove the SS diodes entirely. It **still** works, because the tube rectifier is doing all the rectification. This is the vacuum rectifier we all know and love for its sag.  
But back up one step, to the example with one SS diode in series with each vacuum diode anode. As a practical matter, it's impossible to tell this setup from the all-vacuum model by the output voltage only. But what happens if the vacuum diode shorts on (a) both sides or (b) one side.  
For (a), the shorted rectifier tube now shorts both its internal rectification *and* both SS diodes together, and produces the example we started with, two SS diodes doing all the work. The DC voltage on the first filter cap moves up maybe 50V as the SS diodes don't have the vacuum rectifier's internal resistance. The amp gets fractionally louder and sags less as long as the filter caps can take the bit of extra voltage.  
For (b) the same thing happens, but you now have SS on one half cycle and vacuum rectification on the other, so you get a fair amount of 60Hz ripple, not the 120Hz ripple you expected. But the amp still works.  
If one end of the 5V heater winding is shorted to the B+ in this mess, OK, it still works.  
If both ends of the 5V heater winding get shorted together, then it probably burns out your power tranformer because the low voltage/high current does not increase the AC line current enough to blow the AC line fuse. To protect that, you need a 5V filament winding fuse - which I also highly recommend. PT's are expensive.  
By the way, shorts on all heater windings carry the distinct possibility of killing the PT without blowing a line fuse. A heater fuse is good preventative, because it can be sized to protect just the heater winding, not the whole transformer. That lets you get the fuse much closer to the desired value.  
One SS diode in series with each anode of a vacuum rectifier protects the power transformer, filter caps, probably the choke from a shorted rectifier tube. A rectifier fuse ices the cake. With the diodes in place, your rectifier tube shorting, causes a change in tone, but you finish the gig, and you may not even notice. Without them, the amp dies. Some amps have caught fire from shorted rectifiers.  
Will a line fuse protect the amp from a shorted rectifier? Maybe. Maybe not. I've seen it both ways. Fuses cost about $0.25. 1N4007's are 0.04 each. I'd spend $0.16 to not ever put my AC line fuse's protection qualities to the test.  
Naturally, if there is no heater supply to the vacuum rectifier or if the rectifier tube has a burned out filament, there is no amp operation because the rectifier tube does not conduct. But the failure is contained to the rectifier tube.  
Does that explanation help? Please correct me if I'm missing something.  
I recommend 1200V / 1A or more FREDs for the diodes. They're under $5 each. You can use one or more 1N4007s for them, but 1KV is tight on voltage rating, and you need to parallel them with 1M's to equalize reverse leakage if you series them.
11/11/2005 12:43 PM
Jack Koochin
Good stuff, RG
Thanks for taking the time to sort that out for me. Very cool. Ya know, if you're not careful, I just might learn something around here!  
11/10/2005 1:17 PM
jaysg what's a good choice
"It's a **HIGHLY***RECOMMENDED***MODIFICATION*** on all tube-rectified amplifiers."
for these diodes? Just wondering whether 1N4007's are okay or whether something beefier should be used.
11/10/2005 3:59 PM
I like FREDs. The trick is finding them small enough to be cheap.  
1KV diodes like the 1N4007 are OK for smaller amps, marginal for big ones. If you have a 450V B+, that means that the off-state diode is withstanding twice that, about 900V. I don't like the odds.  
Two 1N4007's in series are good for 2KV, so you can use two in series, but you really need parallel swamping resistors to force the reverse voltage to equalize, not divide in proportion to the leakage currents as they'd otherwise do. 1M to 10M resistors should do the trick. Use two 1N4007's in series on each vacuum rectifier anode, and parallel each one with ... um, maybe 4.7M.  
The 1200V FREDS give you some margin for use with only one of them in each side of the transformer. These used to be quite expensive, but Digikey now sells the IXYS 1200V/11A FRED for $1.59 each.  
It's debatable whether you need FREDs or not. They only come into play when the tube is hosed up, so their recovery time doesn't matter so much.

  Page 1 of 3 Next> Last Page>>