Application Note AN-14

How to bias the 45B tube

Last Update:

The idea behind the 45B

The 45B tube resulted from the many conversations I had with Roger Modjeski, who unfortunately passed away in 2019. Roger spend his entire life and career on designing tube amplifiers. He had an outstanding memory for details, and visited many tube factories around the globe. He pointed out to me, the original RCA 45 tube was intended to use at positive grid voltage. It can be seen if you study the details. Yet, this can burn the grid, it is a risky method. His desire for a 45 tube with higher plate dissipation was always present. He consistently based his amplifiers on the principle of highest voltage and lowest current. He was exceeding the EL84 data sheet maximum voltage, because as he said, this was just their commercial choice. To his opinion, as long as plate current is reduced, the higher voltage is possible. Let me point out here, almost ALL of his EL84 amplifiers, made ever since, still run on the first set of EL84. So after all those decades, it proved he was right. He always thought the historic data sheets were already at that time intentionally made the use of the tubes look easier as it really is. The main "trick" with that kind of presentation, is to pretend the tubes were made for lower voltage. That makes the amplifier design look easier because lower voltage is preferred by many. But for getting maximum tube life, you need to work at low current and high voltage. To Roger's opinion that was so, and I fully agree with him on this.

What does this it mean, if we go this way of higher voltage and not higher current? An interesting situation we can see here with the original 45 tube by RCA,. it was already positioned in a high voltage, lower current range by RCA! We really have to stand still with this. It is much different from the 2A3, which was "downgraded" to 250V (Plate to heater) operation. Yet 45 and 2A3 have close to identical curves. So they knew this at RCA, and they did it at least right with the 45. Yet the 2A3, at the time the "higher power tube for everyone" was specified at 250V, and RCA never specified it at a higher voltage operating point. This sure would have been possible to do, but it would demystify the "250V" tube. Here you can see some of Rogers presentations.

The idea was born, for a 45 tube with higher plate dissipation, and higher possible working voltage. It was in the year 2015, together with Roger Modjeski.

The RCA 45 tube was designed for speakers which produce sufficient loudness at 2 Watt. For this, two classical operating points have been in use ever since, and these can not be improved. The 1.6 Watt circuit was intended for USA radios, which typically run at 250V DC Supply Voltage. The 2 Watt circuit is running at 275 Volts, and was recommended for HiFi use. The tube output power is always specified before the transformer. Given a normal transformer loss is 6...11%, the 2 Watt circuit delivers 1.8Watt at the loudspeakers, at low distortion. Designed 85 years ago, this circuit was the perfect balance of some difficult choices, such as output power, distortion, damping factor and lifetime of the tube. After a careful study of the old RCA data sheets, using today's computer simulation software, we have to praise respectfully the good work by the RCA company. Meaning those working points were chosen perfectly, and distortion data was published totally correct by RCA. Much unlike the original Western Electric 300B data sheets of the 1930's, with it's 17 Watt Output power working point, which is simply not realistic. You will never get 17 Watt out of it with that distortion number.

The 45B tube

The classic 45 was designed for speaker systems needing less than 2Watt per channel. Such speakers exist, but many other still excellent speaker systems, require up to 4 Watt per channel. Here is where the 45B comes in. The 45B can be used such to keep the sound character of the 45 tube, but do so at higher power. For this, some carefully chosen operating points are recommended below. These will all come near to the original 45 sound, in terms of harmonics development and damping factor, using similar transformer impedance too. For this reason you will see comparable distortion figures as with the genuine working points, only with the 45B that will achieved at higher output power. Additionally, some 45B working points are recommended with relatively high transformer impedance of 9k2, which aim for lowest possible distortion, lower as the classical 45.

The 45B tube will NOT give higher output power when the amplifier is not modified. However the 45B can be used normally instead of a standard 45, and will behave identical. So the modification of the amplifier can be done at a later time if required.

Following Roger Modjeski, the best result comes from increasing the voltage and not so much from increasing the current. As a side effect, this allows us to use the existing SE transformers. With great care, Roger set up the 45B working points for this.

Distortion in percent or Decibel (dB)

In electronics, distortion can be expressed as percentage of the signal, or in dB. Using dB, is more related to human hearing, meaning a similar difference in dB, is experienced the same. This is why many technicians prefer dB. (See also Note6). As as rule of thumb it can be said, a 3dB change can be heard by anyone, but below that needs a trained ear.

The 26dB rule of thumb: -26dB of 2nd harmonics is the level at which most human ears just do not hear this distortion. So at -25dB or more, you will experience this as minimal distortion. Whereas harmonics added with a levels of -20dB or more, sounds not pleasant any more.

 

With SE amplifiers, the harmonics will be mainly 2nd harmonics. The Ratio of 2nd. harmonic to 3rd. harmonic is an amazing factor 350 with the 45 tube, and the 4th harmonic or higher is virtually not present at all. So we need to look only at the 2nd harmonic.

Distortion and (modern) Amplifier Design

Historical bias settings give us a 1.6 Watt operating point, and a 2 Watt operating point. The 1.6 Watt circuit has the advantage of a somewhat lower distortion, but it needs really a loudspeaker system which requires no more than 1.6 Watt. The advantage of this circuit seems only the low heat development in the tube, and lower working voltage. It has to be said that the lifetime of historical (NOS) 45 tubes is probably much higher in the 1.6 Watt circuit, as these tubes are very small dimensions, and they get quite hot in the 2Watt circuit. Yet EML 45 has better cooling, and lifetime is not reduced at maximum dissipation.

Unlike in the 1930's, high voltage capacitors were expensive, and had short life time. So a higher working voltage was an issue in the early days of HiFi, but today capacitors are low cost and reliable. This makes the 2 Watt circuit the first choice today. When the output signal of the 2Watt circuit is only 1.6 Watt, distortion drops even 2dB below that of the 1.6 Watt circuit, so -30dB is achieved with the 2Watt circuit at 1.6 Watt only.

Still, with loudspeakers requiring more than 2 Watt, the standard 45 tube reaches it's limits, and we know it does become a problem for many owners of loudspeakers. Initially 2Watt may seem sufficient, but after some time this may appear not so.

The EML45B gives more options. This tube can be used anywhere between 6Watt and 22Watt plate dissipation. In case plate dissipation is below 6 Watt, like in driver circuit, the classical 45 or 45M would be the better choice.

Transformers: Several working points in the table below will have just 10% higher DC current compared to historical bias points. This allows in many cases to use Single Ended 45 transformers that may already be available. With some other working points, up to 55mA it will probably require another transformer.

Low cost transformers

Please keep in mind, transformers have all kind of loss, and efficiency is for from ideal with some manufacturers. Given what it costs to produce the few Watts, of triode amplifier, you would not want to waste 10% of the available power, and 50 Euro on the transformer. Moreover, transformers with more loss than others, have this (problem) for some reasons, and that is seldom a good idea to use such. Like when the core is relatively small, of course such a transformer costs less. Such a core is driven more into saturation, and that gives always more distortion. In addition, with transformers, by definition when loss comes in, this causes also frequency dependant phase errors.

For all transformers, the forms of loss are three: Copper loss, iron loss, and stray loss.

All together, loss can add up to 6...12% depending on the quality. A low cost transformer will typically have a smaller core. That is already negative by itself.. However, in order to get the same number of windings on such a core, the wire needs to be thinner. So there will be higher copper loss, additional to the core loss. For this reason, only finest transformers will be better than 94% efficiency, whereas some others can be as low as 89% efficiency. However there should be awareness, that the SE transformer can loose 5...11% of the output power, depending on it's quality. With loss, tube and the power supply need to generate this signal initially, but instead of speaker cone movement, this generates only heat in the transformer.

A trade off can be observed when comparing the Lundahl LL1623 with LL1663, in the below table. Now Lundahl offers those types both, because one way or another, users will always want lower cost products. Though frankly, LL11663 a fine transformer, and better than many others on the market. Yet, LL1623 is an ultimate products. So you will see the typical difference in loss. Since loss stands for copper loss + iron loss + stray loss, all together, each with it's own disadvantage, it is a good idea to look at the total loss as a first thing.

Transformer Effiency, some examples. This data is collected by computer simulation, using all parameters.
Lundahl 1623-060	94,4%
Bartolucci 17 (6k)	94,3%
Black Art 417 (6k)	93,3%
Tango XE-5SNF-S	92,1%
Lundahl LL1663-040	91,6%
Hammond T16S8SE	91,9%
Electra Print MK5KB	91,4%
Sowther SE01	89,4%

 

Ultimate transformers

To avoid copper loss, the wire needs to be thicker. In order to get the same amount of winding on the core, the core must be larger size. On top of that, the wire length will increase by the larger core, which needs to compensate that too. So low loss transformers can become very large dimensions. For SE transformers, the first consideration is a good frequency range. When there is a trade off between anything and else and frequency behavior, the frequency will have to come first. So SE transformers are often surprisingly large, just to get a few Watts out of them. As an example, a transformer like the LL1623-060 as used in the below table, uses the same core dimensions, as their 250 Watt mains transformer. Yet LL1623 produces only a few Watt Single Ended with the 45B tube..

Output power

There is great confusion about how many output power a classical 45 amplifier delivers. In the historical data sheets is always given 1.6 Watt or 2.0 Watt, but this means at the tube itself, so before the transformer. After the transformer only 1.4 or 1.6 Watt is left.

Conclusion

At the same distortion level, the EML45B is capable of delivering more than twice the output power, compared to classical EML45. The historical bias point delivers 2 Watt at the tube, or 1.8 Watt at the transformer output with -27dB harmonics. The EML45 B at -27dB harmonics can deliver 4.4 Watt at the transformer output. This is a factor 2.2.

Part2

Link to: Data EML45B Sheet