• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Tube vs. Semiconducter rectification

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all..
Got to wondering ..why do most amps still use tube rectification in their designs...all the literature I'm reading, somewhere it says " as capacitors at that time were not available with that voltage rating"...seems to me that us designers would embrace the benefits of the new stuff. I know of the RF, harmonics & phase shifting of the coil & cap arrangment & the warm-up issue but I think a pure high mikes of caps and attention to the warm-up issue & startup surge are easy fixes.
__________________________________Rick.........
 
Ex-Moderator
Joined 2004
This topic can invoke strong views. My thoughts on this:

Tube rectifiers are quieter than most SS rectifiers, some of which tend to produce 'hash' or switching noises. Tube rectifiers also warm up and start conducting slowly, especially the indirectly heated variety, which many people prefer because they consider it to be kinder on the tubes in the amp and the smoothing caps. These are the main reasons why tube rectifiers have retained their popularity in 'HiFi' amps.

However, tube rectifiers have significant foward voltage drop, which increases with increasing load current, like a resistor but less linear. This causes sag, which guitar players may like but 'HiFi' afficianados do not. With amps in classes other than A, this can be a problem. To an extent, it can be countered by using choke input filtering and big reservoir capacitors but you need a higher initial voltage for this to work. The use of TV damper vacuum diodes as rectifiers can reduce the voltage drop still further but will not eliminate it completely. The most effective cure for the sag of tube rectifiers, in my opinion, is voltage regulation. However, like choke input filtering, this solution requires a higher voltage to allow the regulator to do its job, since it can only work by reducing the output voltage.

With modern SS rectifiers such as HEXFRED, Ultrafast recovery and Schottky diodes, the hash problem can be overcome but the fast warm up problem is still there. Another cure for hash, which doesn't give too much voltage drop and does give a slow warm up, is a hybrid bridge using SS diodes in one half and TV damper diodes in the other.
 
why do most amps still use tube rectification in their designs.

In decending order of importance:

1. Fashion. But that's true of far too many design choices in audio. The extra glowy bottles are attractive, and fashion will trump engineering 9 times out of 10.

2. Noise. What Ray said, except the noise thing is vastly overblown. Yes, you can measure all sorts of differences within the rectification/filtering loop, but if the grounding is done properly, most of those differences go away when you measure the output of the supply (note the excellent noise measurements on my preamp, which has only silicon in the DC path). However, the tube rectifier, with its high internal impedance, tends to help out when the builder is sloppy about grounding. You can get the same benefit (perhaps more) by merely adding a series resistor to the SS diodes.

3. Noise again: because tube rectifiers are quite weedy as regards their tolerable pulse currents, they enforce design choices that minimize ripple currents. Of course, the designer can do exactly the same thing with SS, but it's so tempting to use that big ol' filter cap....

4. Warmup and surge. There actually is a real advantage here in 1kV and higher supplies, where cathode stripping is a possibility. At the voltages used in more common circuits, the tubes are run well below their Vao ratings, and I have yet to see any evidence of increased lifetime with gradual or delayed B+ turn-on (this excludes special oddballs like 6528). Surge suppression is a valid point, but can be done without compromising supply impedance by the use of an NTC resistor on the primary side of the power transformer.

As you might gather, I don't like to use tube rectifiers except in very special circumstances. Where I have heard differences in properly designed and constructed circuits (that's an important caveat!), the tube versions were distinctly inferior as one would expect with the higher source impedance and degraded regulation.
 
SY said:


...

1. Fashion. But that's true of far too many design choices in audio. The extra glowy bottles are attractive, and fashion will trump engineering 9 times out of 10.

...

SY,

Guilty as charged. I once actually selected a tube rectifier based on the shape of the bottle. :eek: (The specs were more than good enough so I went for looks. :D )

However I've also now built a few amps with SS rectification and think they sound fine as well.

It is nice to see that I'm not in danger of cathode stripping since all my amps are decidedly well under 1kV. Do you recommend a B+ delay circuit for "lower voltage", under 1000V or do you think they are not needed?
 
Hi all,

I'm mostly doing restauration job, on old amps; and then have the habit to work reusing original circuity:
If I found a tube rectifier, I will reuse it with some "elegant" ameloration if possible (like NTC and others).
If I found a SS rectifier, I stay in the same topology replacing old SS diodes by generic TV FRED (i Bought a lot of cheap BYV96E giving me satifaction) adding caps in // for peaks, NTC on primary ... usualy this give a good upgrade to sixties SS supply.

Ray said:

Another cure for hash, which doesn't give too much voltage drop and does give a slow warm up, is a hybrid bridge using SS diodes in one half and TV damper diodes in the other.

I'm very curious about that !
could you give me please some examples of that topology (generic schematic)?
wich kinds of tubes are usable ? (caracteristics/models?) I have on hand 2X2, PYxx, EYxx, DYxx, UYxx

I have to add that:
- I personnaly have nothing against slow warm up (In the case it adds only "psychologic peace of mind" it won't do something bad either (especialy cathode stripping)).
- I often like sag, sure I play guitar, but I like that """kind"" of compression" even on old HiFi tube gear (I own old Thorens / dynacord stuff that use tube rectifiers and sounds very pleasing to my ears (and are able to reproduce square waves like nothing else I saw for now).
 
The only caution is that low rectifier impedance - at least at the PSU Designer II and Spice simulation level - doesn't appear to be a universal panacea. So much depends on the type of supply and the quality of components used. For example, a strong potential for supply ringing exists when solid state rectifiers (or low loss tubes such as an 83 (<<<not recommended, major toxicity alert!!)) are combined with low DCR inductors and transformer secondaries. The trick apppears to be tuning, tuning and tuning.
 
I tend to use snubbers on the secondaries of my power transformers(about 0.1uf to 0.47uf at at least 3x the secondary voltage). I don't see any ringing on the scope but my scope ain't that great.
I got into this habit in my gainclone days and remember a slight softening of the sound.

Shoog
 
Administrator
Joined 2004
Paid Member
Hi SY,
Thank you for your first post there. I am in full agreement with you on all points.

I will restore a unit with the original power supply unless there is a compelling reason to do something else. There should be a term for "an unreasonable fear of solid state devices" that afflicts some audiophiles.

Hi rdf,
I have a bunch of 83's. I do my best not to drop them. In case any runs out, I have a 1 oz bottle of mercury around here somewhere for refills. (can't figure out how to pour the air out, more keeps coming back in) It's in a small plastic bottle. Have you seen it?

-Chris
 
Here is my answer. A solid state regulator with positive feedback that brings up B+ when tubes start consuming current. Much better than tube rectifier with a choke. If was initially designed for a vintage radio receiver, but may be used for a power amp as well with slight modifications. Input is "floating", i.e. negative legs of a filter cap neither rectifier must not be grounded!!!

bc348bplus.jpg


bc348filament.jpg


04.jpg
 
bembel said:
I'm very curious about that !
could you give me please some examples of that topology (generic schematic)?
wich kinds of tubes are usable ? (caracteristics/models?) I have on hand 2X2, PYxx, EYxx, DYxx, UYxx

There's a couple examples of a hybrid bridge rectifier on this page, and another one on the power supply of my amp. I use the 12AX4 & 12D4 family of TV damper diodes for my PSU, but there's many other types which can be used depending on how much power you need.
 
It seems very promising Wavebourn,
but can you give us a few informations about ratings ( I mean 12v6@xA, -12@xA (CCSink OK or forbiden ?), 230@xx-xxmA)
Can you give also an equivalent for 2SK727 ? or more generic for other transistors except TIP.
How much ripple ? etc.

Thanks
 
thx Aerius,
You seem to like motors caps as filters, I was thinking about that, any drawbacks ? (whats the comparison with electrolitics ?)
How about the soft warm-up feature ? And how TV damper (tube" diodes respond under "low voltage" (only 300-400V) ?

PS: you like chokes too ! ;-)
 
aerius said:


There's a couple examples of a hybrid bridge rectifier on this page, and another one on the power supply of my amp. I use the 12AX4 & 12D4 family of TV damper diodes for my PSU, but there's many other types which can be used depending on how much power you need.


I don't know what the point of the "Recitifer bypassed by diodes" would be...seems to me to be the worst of both worlds...the only good point would be that if you blow the rectifier tube, it turns into a typical all-diode bridge. You even lose the slow B+ ramp-up that a tube rectifier provides.

I'm currently using a hybrid design in my Aikido-based preamp, which is still a work-in-progress.

Better Diagram: Lundahl's Site
 
SY said:
In decending order of importance:

You left out the most important reason:

0: they (can) sound better.

Believe it or not, there are a lot of people (including myself) that have tried a variety of rectifiers - from SS to vacuum to mercury to thyratrons - and concluded that SS can't complete when it come to letting the music through. It could be these experiences are mainly with SE amps that do not have a high PSRR. Perhaps with PP it doesn't matter ...

Feel free to try to figure out *why* tube rectifiers sound better. I'd be thrilled to know, but knowing or not doesn't change the fact of reason 0.

-- Dave
 
bembel said:
It seems very promising Wavebourn,
but can you give us a few informations about ratings ( I mean 12v6@xA, -12@xA (CCSink OK or forbiden ?), 230@xx-xxmA)
Can you give also an equivalent for 2SK727 ? or more generic for other transistors except TIP.
How much ripple ? etc.

Thanks

-12V is for biasing purpose only, use as less as possible from it. It may be more voltage, if you use a different Zener diode between collector and base of the transistor, but remember that both B+ and bias voltages are products of the same HV rectifier, so if you for example need 400V B+ and -50V bias you need more than 450V rectified. More, I mean 450 + ripples + about 5V headroom for the FET.
The 100 Ohm resistor after the FET is to limit inrush and accidental shortcut currents, if you need say 200W of output power you should use it of less value.

Output voltage is defined by a voltage divider between bias and B+ rails, you may use Ohm's Law to calculate proper values (assume near 1V on the base of the transistor).

Output power is limited by a power FET. You may use some IR devices, for example 30A 800V types that can dissipate 150W with heatsink, in this case you may get enough of power to fire up your OTL multitube superamp.

Speaking of specs, I did not measure them, I just made it for myself and was highly satisfied.

As soon as I put the schematic in public domain, feel free to ise it without worry to step on some patentshit. It starts raising up B+ very slowly when tubes are cold, but as soon as rhey start consuming current they charge capacitor in the bias regulator faster bringing the whole thing up when tubes are warm and ready to play soft, hard,, sweet, bright music. ;)
 
Believe it or not, there are a lot of people (including myself) that have tried a variety of rectifiers - from SS to vacuum to mercury to thyratrons - and concluded that SS can't complete when it come to letting the music through. It could be these experiences are mainly with SE amps that do not have a high PSRR. Perhaps with PP it doesn't matter ...


Black magic ???
maybe breaking a few 83 on the listening room floor will help for reason 0
(I'm joking of course, thought my old neighbour is really boilling lead (not mercury phew!) for real black magic)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.