Do I really need a tube rectifier? Lazy amp tech's journey

[beasleybodyshop]

So i've got a few amp builds going right now, a 50w AB763 with no tremolo and a Matchless/Vox AC30 type thing going.

Both circuits call for a tube recto, but at this point I really don't want to punch a hole for an octal socket, mount, and then buy a (subjectively) expensive tube rectifier. that 5V winding could be used for something fun, like relay power supply.

Am I *really* losing out if I just sub in a pair of 1N4007/1N5408/etc and use an appropriately sized inline resistor to simulate the voltage drop of the recto?

All manner of hate and/or encouragement is appreciated.

[martin manning]

Am I *really* losing out if I just sub in a pair of 1N4007/1N5408/etc and use an appropriately sized inline resistor to simulate the voltage drop of the recto?

I'm voting no. If you want to experiment a bit, you could add a switch for sag/no sag that modifies the bias voltage for the B+ change and see if there is all that much in the sag.

[didit]

Am I *really* losing out if I just sub in a pair of 1N4007/1N5408/etc and use an appropriately sized inline resistor to simulate the voltage drop of the recto?

I'm voting no. If you want to experiment a bit, you could add a switch for sag/no sag that modifies the bias voltage for the B+ change and see if there is all that much in the sag.

Seconding this nomination ... er, add my vote ... er, whatever.

When of any value it's quite specific. Personal experience recommends only when seeking (im)perfect replication, and only particular designs. Tube plate resistance does flex slightly under fairly high and varying load, and this does impact the amp's power supply impedance. I do, for example, believe it's part of the Trainwreck Rocket recipe. Typically though it's just a bunch of more in all the wrong ways: cost, build effort, heat dissipation, failure probability, etc.

best ..

[WhopperPlate]

The only way for you to know what you prefer is to try both . There will be differences. There will be more similarities than differences . Will you hear it? Will you feel it? For brevity, there are too many variables outside the amp alone to mention within a discussion .

Having said that , if you have the diodes/resistors , and don’t want to spend more than that, then don’t . It will sound fine.

[pdf64]

The slow ramp up of HT voltage, the elimination of HT surge voltage / current, provided by a GZ34 is a wondrous thing.
Obviously some people don’t appreciate it, or at least don’t think it’s worth the additional infrastructure etc required to support it.
That’s fine, suitable solid state diodes are available. But note that if the 2 phase winding voltage is over 600V, a single pair of 1kV diodes isn’t sufficient.

[R.G.]

If you use bare solid state rectifiers, yes, there are noticeable differences. If you use a few simple circuit dodges I suspect that you will not - and that almost no one else would - hear any difference.

The differences include:
- lower voltage drop in the rectifiers; silicon rectifiers have about 0.5 to 1V drop at tube amp currents. Vacuum rectifiers drop 30-50V, so the B+ is that much lower. That's often audible.
- higher forward resistance in tube rectifiers versus silicon; this makes the tube rectifier version sag more on high power draw, and this "sag" is described as a kind of compression on transients that some people are craz... er, prefer very, very much
- ordinary silicon diodes have a "slam-off" shutoff on each conduction; this transient has been shown to excite RF ringing in the wires to/from the rectifiers, which in turn gets detected by the circuits and turned into a buzzy, thin hum at 120Hz.
- slow warm-up allowing the B+ to rise slowly at turn-on; some people think this prevents cathode stripping of the other tubes

The cure for these differences lie in
- insert a power resistor in series with the solid state rectifiers to fake the internal plate resistance of the tube rectifier; this has proven to be very, very convincing for both the desired B+ voltage and for sag that can be adjusted
- use fast, soft recovery silicon diodes (look up "FREDs"); this prevents the slam-off RF quarks and thin 120Hz buzz
- ignore the cathode stripping stuff; cathode stripping doesn't occur at the voltages guitar amps commonly use; If you run into a True Believer in cathode stripping, I have other circuit techniques that would prevent it, but it largely isn't a problem in guitar amps.

[Leo_Gnardo]

I wouldn't go tube rectifier in a build. Unreliability, compared to silicon rectifiers, that's enough of a convincer for me. Been replacing way too many GZ34 over the last 20 years or so.

JJ's were OK for a while then QC on those went to blazes. Maybe they're better now but I got stung and not anxious to repeat the punishment.

Ruby offered a fairly reliable GZ34, custom made for them since early 2000's but I can't get my hands on them anymore, don't know what happened.

I've been installing PSVane lately. So far so good but I have less than a year's experience with them. No immediate failures thank hevvins, also none of them make buzzing noises straight out of the tube as some others sometimes do. Only used a few so far so it's not much of a data base.

All in all, I'd recommend BUILD RELIABILITY INTO YOUR AMP and that means no tube rect. RG's good suggestion to add resistors between trans windings and tube plates - know that's exactly what you find inside Weber Copper Cap rectifier tube simulators. "People pay good money for those" but you can do it yourself for a pittance.

[beasleybodyshop]

Thanks guys. I will order some FRED diodes and some large power resistors to sub in. Too tired of replacing bad rectos! At least when they fail they make a cool light show.

[pdf64]

... Too tired of replacing bad rectos! At least when they fail they make a cool light show.

That’s nuts, it’s the simplest thing to eliminate that failure mode.
Just re-arrange hot switching standby switches (or just leave them permanently closed), most importantly fit back up silicon diodes and they’ll probably last until the cathode wears out.

https://www.premierguitar.com/the-immor ... lifier-mod
1k5V rated diodes, or series pairs of 1N4007, may be a belt and braces safer bet. 1kV rated diodes are typically working right up to their limit when used singly.

My thinking is that voltage / current surges at start up with solid state rectification are not ideal for achieving a long working life from many other components, principally the reservoir cap and, if applicable, the valve used in the DCCF circuit (g-k flashover if standby isn’t used, Ik surge if it is).

A GZ34 provides a lovely gentle wake up to everything downstream, all it asks in return is 10W heater power and a little consideration at the design stage

When the minimum series supply resistance limits are respected, the sag from a GZ34 should be little different to that from a solid state rectifier. Chart III p3 provides the formula https://frank.pocnet.net/sheets/093/5/5AR4.pdf

[JD0x0]

IMO, tube rectifiers are somewhat obsolete. Since we have good diode rectification, the only things you'd really be losing with a tube rectifier is the slow ramp up of voltages, and the 'Sag.'

The 'sag' can easily be addressed with a 'Sag resistor' in which a resistor is added inline with the B+ to simulate the voltage drop of a tube rectifier. Now, admittedly the voltage drop curve is slightly different between the tube rectifier and a resistor, but I honestly would be surprised if anyone could even tell the difference in practice. Another advantage with a 'sag resistor' is you can wire them to a switch, so you can select different levels of 'sag' with the flip of a switch.

The voltage ramp up of a tube rectifier is nice as it allows you to be more conservative with your mains filter cap voltages, since the unloaded B+ will not be hitting the caps, but it's also not difficult to just rate your filter caps appropriately. There's also ways to create that slow 'ramp up' without a tube rectifier, but it will require more parts to make that happen.

But essentially, there's no real magic or mojo to a tube rectifier. They're expensive, especially these days and they're a pretty common point for failure. I generally avoid them if I can.

[beasleybodyshop]

My typical standby switch wrinkle is to wire a 220K 2W and .047uF across the standby switch lugs to mitigate switch thump. Don't think it does anything for slow inrush though.

[pdf64]

The issue with a valve rectifier and the standby switch depends on whether it’s a hot switching arrangement, ie the switch, when open, it interrupts the reservoir cap’s charging current. If it does then it’s hot switching and will probably result in the peak anode current limit of ghe rectifier being exceeded momentarily when the switch is closed, due to the inrush current drawn by the reservoir cap etc.
That’s what generally causes the lightning storm in the rectifier.
http://www.valvewizard.co.uk/standby.html

I’m not sure that whether the switch thumps or not when flipped is super relevant.

Can anyone report on valve rectifier failure in the scenario of no hot switching standby and back up solid state diodes fitted in series with its anodes?

[R.G.]

In terms of eliminating inrush, a simple current clamp can be more effective than relying on heater warm-up. A power MOSFET and a few other parts let you just set the peak current that can happen. As long as this is set just a bit higher than the peak rectifier currents in full-power operation, It can't have any effect on normal operation, but will eliminate the high inrush currents at turn-on and/or standby.

As already noted, having a standby at all is probably not needed, but if you want one, it can be made non-damaging; expecially in the case of flipping standby off and on repeatedly.