This article was first published in The Ringing World, issue 5523, March 3 2017, page 196, and revised January 2023.
In November 2015, I sang in a concert in the Gustav Adolphus Memorial Church in Nuremburg. During the concert, the 3.4 tonne steel tenor rang the curfew, a lovely and captivating sound. The experience inspired this article, which seeks to explain why British steel bells have extremely poor tone when some German steel bells sound the equal of bronze bells.
Naylor Vickers – English steel bells
In the UK, almost 1,000 steel bells were produced by Naylor Vickers in Sheffield in the second half of the nineteenth century. The article Colonel Tom and his cast steel bells by David Cawley gives some historical background. A 1972 paper by Ulrich Troitzsch brought to my attention by Heino Strobel which I have recently translated from the original German, shows that the moulding technology used by Naylor Vickers was originally invented by the proprietors of Bochumer Verein and licensed to Naylor Vickers. The two companies collaborated closely on its exploitation. The key British patent under which the steel bells were produced was taken out on behalf of Bochumer Verein by a German chemist living in London, Ewald Riepe, who described how to construct moulds capable of withstanding the high temperature of molten steel. Another page describes in detail the collaboration between Bochumer Verein and Naylor Vickers, and gives details of the patents.
Steel bells in the UK have a reputation for poor sound. A.H. Cocks in Church Bells of Buckinghamshire (1897) says of steel bells “Let us hope that nothing will deter the process of disintegration which hastens the removal of their miserable remains from their towers”. Chris Dalton in Bells and Belfries of Dorset (2005) describes the steel five at Toller Whelme as “this weird and frankly ear-splitting installation.”
Bells have many frequencies (partials) in the sound. Among the most important to the musical effect are:
- the hum, prime and nominal, which in a true-harmonic bell are in exact octaves
- the octave nominal, which with the nominal and other higher-frequency partials forms the strike pitch in the listener’s ear an octave below the nominal
- the tierce, which is usually a minor third above the strike pitch.
The intervals of partials are given relative to the nominal, due to its importance in creating the strike pitch of a bell. Cents are used to measure these musical intervals. One cent is 1/100th of a semitone, so there are 1,200 cents to the octave of twelve semitones.
The peal of six at Waddesdon in Buckinghamshire are typical Naylor Vickers bells. In the table below, the partials of the tenor at Waddesdon are compared with the average of many true-harmonic bells. Negative numbers are intervals down, so the true-harmonic hum is two octaves below the nominal:
|Hum||-2091||-2400 cents||Over three semitones sharp|
|Prime||-1150||-1200 cents||Half a semitone sharp|
|Nominal||A little flat of F#|
|Octave Nom.||1328||1237 cents||Nearly a semitone sharp|
Both the hum and the octave nominal are exceptionally sharp; they are in the sharpest 1% for these partials in my collection of 5,000 bell recordings. The sharp octave nominal means that the listener’s ear creates an ambiguous strike pitch, giving the twangy sound typical of steel bells. The sharp octave nominal arises because Naylor Vickers bells are very thin, particularly in the sound bow (the rim of the bell).
In addition to the deficiencies of the individual bells, the Waddesdon bells are not in tune with one another. Their strike pitches have discrepancies of up to half a semitone from the musical scale defined by the note of the tenor. Some Naylor Vickers bells are so out of tune with one another that two bells in the peal sound almost the same note, for example at Toller Whelme and Ryde, IOW. I have recordings of nearly 150 Naylor Vickers bells and all show the same deficiencies, both within the bells and across the peal. Bronze bells with similar discrepancies in their partials also have an unpleasant or surprising tone. Examples of thin bronze bells with sharp upper partials include the second at Little Somerford in Wiltshire, and the fourth at St Edward, Cambridge. The bells at Castleton in Derbyshire have very sharp hums like those found in Naylor Vickers bells.
Naylor Vickers appear to have had no interest in what their bells sounded like and there is no evidence they ever tried to tune them. The ‘Colonel Tom’ article explains that the practice was to cast bells speculatively, and fit a set together as a peal for a particular order with roughly the right strike pitches. The peal of steel bells at Hale (previously at Bootle) were tuned by the Merseyside Bell Group in the 1980s. The nominals were corrected so the strike pitches are in tone across the peal, and other partials were also improved. However, removing metal on the tuning machine further sharpens the upper partials which worsens the twangy sound.
Naylor Vickers cast some big bells, and I have recordings of three of them: the bell at the Italian Church in Hatton Gardens in London (said to weight 87 cwt); the bell at Faneuil Hall in Boston, Massachusetts; and the bell at St Ignatius Church in San Francisco, California (originally cast as a fire station bell). This last bell is rumoured to be 7 tons in weight, but a contemporary account in The Illustrated London News gives the weight as 2 tons 12 cwt, and there is no reason to doubt this. The Hatton Gardens and Boston bells sound awful – the former has been described as sounding like a large frying pan. The upper partial frequencies are so chaotic that no strike pitch is formed, and it is difficult to identify the named partials as their spacing is so unusual. The San Francisco bell is better.
The cause of the poor sound of Naylor Vickers bells is their shape, both the overall profile and the lack of wall thickness at various points, especially in the rim or soundbow. Rust is sometimes blamed, but steel which is kept dry and relatively free from atmospheric pollution loses little mass to corrosion once the initial layer of rust has formed. There are many examples of German steel bells from the 1940s to 1960s which retain their tuning even though visibly rusty. I have just completed an investigation into the effect of rust on steel bells.
Bochumer Verein, a steel founder based in Bochum in the Ruhr (West Germany), cast steel bells from around 1850 to 1970, and were the original inventors of the moulding process, for which Riepe obtained patents on their behalf. In this time they cast 18,000 church bells and 20,000 bells for other purposes. The bulk of their production took place after the two world wars, when there was a huge demand for replacement bells and shortages of both bronze and money. Many recordings of German bells are available on YouTube, which makes it easy to study their partials. I have 149 recordings of German steel bells, almost all by Bochumer, dating from 1858 to 1963.
The partials in early Bochumer bells are as disorganised as in Naylor Vickers bells, but in the 1910s and 1920s they settled on designs which gave reasonable consistency in partial frequencies. Hums are generally two or three semitones sharp of the octave (sharper than an old-style bronze bell), with a lightly flat prime. Smaller Bochumer bells of this period often sound no better than Naylor Vickers bells, but the larger ones can be good.
One example is the bell they cast for the 1936 Berlin Olympics, weighing 9.6 tonnes. The bell featured heavily in promotion for the games, and toured Germany before installation at the stadium. It still exists, but after the stadium tower was burnt by the Soviet army the bell was used for target practice and then buried in rubble for ten years. It has cracks and a shell-hole through the waist. After watching many videos of the opening ceremony I found a 2-second clip in which the Olympic bell is clearly identifiable from its decoration, and the partials can be analysed. The bell has an accurate seventh hum, and a prime which is close to the octave. The upper partials are normally arranged.
There are many other good heavy Bochumer bells from the period before the second war. They sound similar to bronze old-style bells, with a slightly steely tone if the upper partials are sharp. In the 1930s, Bochumer began to experiment with other profiles, but all information about bell production at Bochumer was destroyed in the war.
In the 1940s Bochumer began to produce true-harmonic bells, with the hum and prime in exact octaves with the strike, and the upper partials conventionally arranged. This change was as dramatic as that achieved by Taylors in 1895/6. Tuning marks are clearly visible inside the bells, and they had discovered how to control the partial frequencies via shape and tuning. Provided the upper partials are not too sharp, these bells in my opinion sound as good as the best bronze bells. Bochumer continued to cast true-harmonic bells up to the end of production in 1970, and also cast bells with sixth hums, seventh hums, and major third tierces.
There are many examples of good bells from 1947 onwards – one fine example is the five (tenor 2.6 tonnes) at Dortmund St Petrikirche cast in 1963. Bochumer’s heaviest bell still rung is the 14-tonne true-harmonic bell of 1949 at Neustadt, which sounds very similar to the Nottingham Council House hour bell (Taylor 1928), but a fraction deeper in pitch. One sad example is the 1947 peal of five, tenor 5.4 tonnes, at Scheyern, which were replaced in 2009 with fourteen bronze bells cast by Perner with a heavier tenor. The biggest Perner bell is poor in tone. The two largest Bochumer bells hang in frames on the ground, one at Scheyern and one at Mariastein in the Tirol, sounding fine in comparison but slowly rusting away. As a commenter says on YouTube about one of the steel bells, “The best bell of Scheyern now rings alone”.
Steel vs bronze
It is said that steel bells weigh less than bronze bells with the same strike pitch. However my investigation into the weights, diameters and nominals of bronze and steel bells suggests that a steel bell with the same nominal as a bronze bell will weigh about 95% of the weight of the bronze bell. It’s diameter will be 11% bigger than the bronze bell. Walter Schäfer’s list of notable German bells shows confirms that stell and bronze bells of the same weight sound broadly the same note.
Could steel bells be produced today with the same tonal quality that Bochumer achieved in the last twenty years of production? The higher melting point of steel (1370 deg C as compared with 915 deg C for bronze) and the need to avoid decarbonisation means that different furnaces, moulds and casting techniques are required than for bronze bells. The founder would also need to experiment with designs and tuning. The economics of today’s bellfounding and steel industries suggests that it is unlikely that anyone will try.
My thanks are due to Tim Jackson, Nick Bowden, the Bell Historian’s group, and all the bell enthusiasts who post videos to YouTube for recordings and helpful insights.
Great Bookham, Surrey
References and links
Walter Schäfer’s list of notable German bells, including a considerable number of steel bells
Colonel Tom and his cast steel bells, The Ringing World 2000 pp. 1259-1263
Amateurs steel themselves for the task, The retuning of Hale bells, The Ringing World 1984 pp. 161 and 167-169
Illustrated London News, 7th Jan.1860, page 12 an account and picture of the casting of the San Francisco bell
An article on bell production by Bochumer (in German)
A Wikipedia article (in German) with a photograph of Bochumer’s biggest bell (15 tonnes) and information on their experiments with profiles