Indirectly- versus directly-heated rectifiers

Added 11-01-2015: The text below is primarily meant as a directive when designing valve amplifiers, not as a guide how to modify valve amplifiers. Most likely the choosen rectifier(s) will suit the amplifiers technical specifications. Be aware of that when changing these..

A while ago I posted on the website of Jimmy Auw about the use of valve rectifiers. More or less when to consider indirectly heated rectifiers or directly heated rectifiers. Maybe it is worthwile to dig into that a little deeper..

His post was about building an universal power supply and was considering several types of rectifiers. This is what I replied:

Interesting topic. Preheat rectifiers versus Cold start.. When using directly heated (dh) rectifiers, preheating them will result in maximum current demand from both capacitors and signal tubes when high voltage is applied. This results in a current peak and high voltage peak. Cold start (just one on/off switch) will take care of instant current demand of signal tubes and the voltage peak. When directly heated power tubes are used, operational temperature of the filaments will be reached at about the same time. So more hand-in-hand. Current draw from capacitors will remain at a too low cathode temperature. This was no problem at that time since only low-uF capacitors were available. When indirectly heated (idh) rectifiers are used, current draw from capacitors and dh power tubes will be present at a too low cathode temperature. Current draw from idh-power tubes will go hand-in-hand. My experience is that it is best to use dh-rectifiers in solo dh-tube circuits, and idh-rectifiers in solo idh-tube circuits. Use of type power tubes (dh vs idh) dictates more or less type of rectifier. So when a mix of idh-tubes and a dh-power tube is used, it is best to use dh-rectifier since these will demand highest current.

Since a lot of rectifiers are interchangeable and choice can be of importance in an amplifier, I will clarify my post a little bit and put down my recommendations.

Rectifiers
There are basically two types of rectifiers, directly heated and indirecly heated. A directly heated rectifier is operational as soon as the cathode has reached optimal temperature. Since the cathode is heated driectly, this can be as fast as 2 or 3 seconds. An indirectly heated rectifier is also operational as soon as the cathode reaches optimal temperature only the heater needs to heat the cathode, hence indirectly heated. This can take up to 10 to 15 seconds. This means that high voltage supply will be available faster when dh rectifiers are used.

Other valves..
The other valves in an amplifier are also directly- or indirectly-heated and require about the same time to reach operational temperature as the rectifiers do. Now what happens if these are mixed? To get an idea it is good to get some kind of understanding what is happening when a valve heats up.

Basic principles.
To emit the cathode needs to have a certain temperature and a plate voltage supply. There a two stages that are somehow critical:

• First is when the cathode has not reached it´s optimal temperature but the applied plate voltage forces it already to conduct.This causes so called cathode stripping, the "conductive layer" gets peeled off the cathode.
• Second is when the cathode is emitting but can not conduct due to the absense of a plate voltage. The emitted electrons can not find their destination and will fall back to the cathode, this is called cathode poisoning.

Both result in damaging the valve.
Of course current draw from the amplifier is what completes Ohm´s law. In general we have two kinds of current draw to take into account:

• "Static current draw", meaning the current draw of the passive components like capacitors. This static current draw is shortly present at amplifier switch on. Capacitors pose for a very brief moment a short circuit when high voltage is applied, the so called inrush current. The higher the capacitor value, the greater the inrush current. Because of this most valve rectifier data sheets mention a maximum value capacitor to be used as the first capacitor in the power supply since valve rectifiers can deliver modest current compared to diodes.
• "Dynamic current draw", the current draw from active components like valves. Of course this only applies when the valves´ cathodes are at operating temperature.

Put these facts together and you are presented with some kind of dilemma: No matter how you switch on an amplifier, problems will occur.. Luckily engineers knew that as well and designed most valves to withstand (for a short period) some of these shortcomings. But it leaves us with the question what is the best way to switch on an amplifier with a certain valve configuration? I´ve seen (and used) different methods to deal with this:

• Preheating all valves by switching on filament voltage before high voltage. This will protect somehow all the valves but the rectifier, since it will be confronted with both the static and dynamic current draw at switch on. Often a disturbing "thump"-like sound can be heared through the speakers at switch on. This can partly be solved by parallelling a high value resistor over the high voltage switch.
• Preheating all the valves but the rectifier, switch on the rectifiers filament voltage and high voltage simultaneously. This is safe for, again, all the valves but the rectifier since it will suffer from cathode stripping because current demand is present when its cathode is not yet at operating temperature.
• "Sacrificing" a cheap tv-damper diode by putting it in series with the rectifier´s cathode or the centertap of the high voltage transformer. This will protect all the valves at the cost of some additional heater supply and extra high voltage drop.

Recommendations please?
Here is my shortlist..

• First: Always respect the datasheets maximum value of the first capacitor in the power supply and avoid large capacitor values (>470µF) in the power supply.
• Best is to keep an amplifer overall indirectly heated or directly heated. This way you can get away with one power switch since heating time for all the valves is about the same
• If this is not possible for whatever reason, match the type of rectifier to the current demanding power valves. So when using indirectly heated power valves use an indirectly heated rectifier as well or vice verse.

If you decide to go for a directly heated rectifier with indirectly heated power valves, take into account that a much higher supply voltage will be present few seconds after switch on, since the rectifier is already conducting when the power valves are still heating up. This means that here is no dynamic current demand at that time so no voltage drop over resistors and/or chokes. This is especially the case if a choke input power supply is used with no effective bleeding resistor. Be sure that capacitors, resistors and valves can withstand this higher voltage!

Datasheet indicates that the EL3 can withstand 550 Volts on the plate in cold condition..