History of Reverse Action Design on the Autoharp
Menu:
Comparison of Wigand, Back and Millington/Young
Above, the three instruments are shown together in the same orientation, and the drawings roughly scaled to show the similarities and differences. In terms of difference; the speaking lengths of the strings appear consistent between Back and Millington/Young, the steeper bridging rake to the high treble reflects the simple straight design of the nut at the dead end (Back), whilst Millington has arched the dead end nut to enable a shallower rake. Both illustrate clearly that the remaining playing surface is limited, and requires optimising. Wigand’s harp depiction does not appear realistic (given the pitches called for relative to the speaking lengths of the strings). Keyboard size appears to be another discrepancy, and it is entirely possible that the inventors did, in fact, envisage different sizes of keyboard.
Despite the fact that all these three instruments appear to have been developed without knowledge of the others, overall the designs have striking similarities; the design concerns are similar and the presented interfaces are also very similar. Consequently, it is worth pausing at this point to compare them in more detail.
The most striking commonality in all three of the designs is the keyboard orientation with respect to the strings. The keyboard is oriented keys-facing-the-bass-side of the harp perpendicular to the strings. This is the most logical design platform from which to integrate a keyboard because the obvious lever mechanism, the keyboard itself, is placed in line with the damper bars and is available for a direct coupling. 
It is evident that this keyboard position must lead to a different playing position of the instrument than the one that we expect on the autoharp today and we can surmise that the instrument was not intended to be held in the lap, but instead, designed to be played as a table-top instrument, in the manner depicted in the photograph of Zimmermann in the previous chapter. Other contemporary pictures also depict players using this playing position.
Back describes playing the instrument thus:
“The strings, as in all instruments of this class, are vibrated with a plectrum, and it is usual to run the latter across all the strings, so that the open ones will be sounded while the damped or muffled ones will be mute”
Which supports this view of the table-top playing position for all these instruments.
Although this does not seem a very sophisticated way of playing the instrument, and would not easily allow for the complex combinations of pinch, tremolo and hammer that are part of today’s autoharp finger pick technique, there are advantages. The closest commercially produced autoharp to the Zimmermann original was the Victoria autoharp, (1888), which clearly features the pinch mechanism which results in damping on the sides of the strings; a Zimmermann idea.
I met a player of such an instrument, Keith Holmes, at the UKautoharps meeting at Mickleover in 2012, who did indeed play his instrument as depicted in the photographs.
Keith, in fact, played a replica instrument of the Victoria harp, which he had made himself (his original Victoria model had become too fragile to risk travel). Keith’s autoharp was as much of a curiosity to the UKautoharps members as my own and the playing position and extended chord possibility provided by the Victoria formulation produced quite different sounding music to the other autoharps present. It was focussed on melodic and chordal accompaniment, and not on rhythmic strumming. The strum action was pretty much as Back describes; and this particular player, at least, strummed only in one direction (bass to treble) with an emphasis on the final melody note, which was sounded at the top.
Indeed, a significant advantage of this system is that it is relatively easy to render a precise melody, distinct from the accompaniment, whilst the more flexible chord system adds to the harmonic complexity and provides increased ability to render non-chord tones without having to change to an alternative damped chord. Keith and I played some tunes together, and the resulting blend of the two formulations was attractive. I was sorry that there was not more time to experiment with this combination.
Given that table-top playing was the accepted playing position of the time, and also the most obvious integration orientation for a keyboard, it is evident why none of the three designers questioned this design. I also seriously considered this keyboard position before embarking on this project, but finally rejected it, as I was concerned it would not provide a sufficiently well balanced relationship between the two hands.
With regard to the keyboard designs of the three inventions, Back’s is undoubtedly the most attractive to a pianist, and is perhaps the only arrangement that would offer a comfortable and “convincing” keyboard feel to a pianist. Wigand and Millington/Young place the key pivot point at the end of the visible keyboard surface; this may appear to the eye as the obvious pivot point of a key and is indeed a design that is to be found on a lot of electronic keyboards; but it is not so on a piano. The pivot point of a piano key is much further back than the playing surface might suggest. This is not immediately apparent because most of this length is hidden behind the casing. Playing keyboards with a shorter pivot point tends to give a pianist the feeling that there is something “wrong” with the keyboard. Pianists commonly (and mistakenly in my opinion) attribute this wholly to the sprung (as opposed to weighted) mechanism. Shorter pivot points result in steeper angle of key drop, and I believe it is actually this aspect that is most disturbing in the feel of most cheap keyboards — not the fact that they are sprung. This was a key finding within my own prototype series, and led to a much improved keyboard feel from prototype 3 onwards.
All three inventors consider the issue of sustain — though the solution is simpler in Wigand and Millington/Young than the complex design suggested by Back (through a bar which simultaneously lifts the far end of all the keys, thus raising all of the damper bars together). I have considered such a device for inclusion in my own prototype series, and had even determined a position — in front of the keyboard to be operated by dipping the left wrist. However, I abandoned it as a primary consideration within the prototype series; I discovered when I imagined the sound and control, whilst away from the instrument, I could hear how an integrated sustain pedal device might enhance the sound, but this perception vanished upon actual musical engagement with any of the prototypes (even prototype 1) and practice sessions invariably led to the conclusion that the device was not necessary. This is a subject for further discussion within the prototype series.
In considering these patent documents, I wonder to what extent the prototypes described were realised and what problems might they have encountered? Certainly this hidden history is no longer obvious today — for the overwhelming majority of the time spent prototyping, I was unable to trace any evidence of instruments that have survived to the current day. The US Patent Office had, in 1870, dropped its requirement that a working model accompany a patent application, so a fully functioning instrument was by no means a certainty. The less than credible dimensions of the Wigand harp body do not inspire confidence that a functional instrument was produced, but overall all the mechanisms appear to be quite robust. I suspect that the feel of the Wigand and Millington/Young keyboards might have been quite heavy, but Young nonetheless successfully pursued his patent all the way across the Atlantic — this suggests that a working prototype was indeed a reality.
In March of 2015 I was approached by Katherine Rhoda hailing from the state of Maine in the USA. This communication included photographic evidence of an extant instrument called a Gondolin, apparently dating from 1910, a request for more information, and if possible, identification.
My immediate suspicion was that this was a Millington/Young instrument. Though considerably changed from the patent drawings, many of the same design concerns are apparent. The keyboard is tapered to optimise the playable string surface. In order to further assist this, the toe and dead ends of the instrument have effectively been reversed.
The keyboard depiction above the strings (in order to facilitate string distinction during playing) is another feature unique to the Millington/Young formulation and is also displayed in this instrument.
This design evidence is confirmed through the marking “Witten, Witten and Co” (shown in the plate above), which is the trading name cited in the Millington/Young patent.
A problem that I suspect all three inventors would have encountered as these mechanisms progressed to finished prototype is that of harmonic damping. Harmonic damping is a problem for autoharp and reverse action harps alike; it happens because damping occurs at points which are in simple whole number proportions to speaking string length; and because single damper bars are used for multiple strings, it is very difficult to calculate appropriate positions easily. Instead, autoharpists mitigate excessive harmonics using two practical strategies, firstly by repositioning the damper bars in different combinations away from the node points, and secondly by extending the damper out from the damper bar (often termed an “outrigger”). The direct couplings of keys to damper bars found in these three prototypes do not easily allow the possibility of repositioning and I consider this to be a factor in the final functionality. The issue of harmonic damping will be the subject of further discussion.