The SOLD KM2 Bubble Sextant

4 11 2013

In 1939, the SOLD-Libellen-oktant (SOLD level-octant), made by C Plath, went into service with the Luftwaffe (German Air Force). I have not been able to discover where the “SOLD” came from, but it is a fair bet that it is an acronym which includes the words “Sextant” and “Libelle” (Level). While it was used extensively by Luftwaffe reconnaissance aircraft especially, it is less well known that Naval  versions appeared,  with better protection from salt water spray and a self-contained power source in the form of a rechargeable NIFE battery.  The aircraft versions were powered via a flying lead. While the results of observations made with a bubble sextant at sea can be expected to be rather inaccurate, in some circumstances, on war patrol at night for example, or in poor weather when no horizon is visible, even a poor result can be better than none at all. Since this is a blog about the nautical sextant, it is with the nautical version that this post is concerned, specifically a KM2 (Kriegsmarine 2) of 1944. It differs otherwise only in minor details from the Luftwaffe version.

Please note that all illustrations may be enlarged by simply left- clicking on them. Return to the text by clicking on the back arrow.

Optical lay-out

Figure 1 shows the light path.  It is perhaps slightly easier to understand the path for night time observations of a star. The star is located by looking upwards through the partially reflecting mirror 5.  Light from the bulb 1 is reflected off mirrors 7 and 6, through a circular bubble level 2, off the reflective face of a pentaprism  3 into an objective lens 4, whose focal plane is at the bubble. The lens renders the rays parallel, so that the bubble then appears to be at infinity as it is observed reflected off the mirror 5. While keeping the star in view, the mirror 5 is  rotated to bring the image of the star and the image of the bubble together, when the angular altitude of the star above the horizontal may be read off a scale attached (indirectly) to the mirror 5. The radius of curvature of the bubble cell and the focal length of the objective lens are chosen to be the same, so that once the star and bubble images are coincident, they tend to move together when the sextant body is disturbed. For daytime observations of the sun, it is the bubble that is viewed through the mirror while the sun’s reflection off mirror 5 is observed indirectly. For this type of indirect observation, the mirror 7 is removed and a frosted window substituted, so that diffused daylight illuminates the bubble chamber, though daylight viewing with the bubble chamber illuminated by the lamp is possible too.

Figure 1: Light path.

Figure 1: Light path.

Averaging observations

Pitching and rolling of a ship imparts accelerations to the bubble chamber which for practical purposes may be considered as being random, and the bubble is seldom at rest. With early bubble sextants, a handful of observations were taken of the body and the altitudes and times averaged. However, random errors are reduced as the square root of the number of observations, so that, for example, to reduce the average error to a quarter, 16 observations have to be made and to reduce them to a fifth, 25 have to be made.

The next step was to provide some sort of drum on which marks could be made by the observer when he considered the bubble and body to be in coincidence, but it was soon pointed out that the bubble might well be considered to be at rest when it was in fact the subject of a large, but constant acceleration. The next step was to cause the marker to operate automatically at intervals of about a second, leaving the observer to maintain coincidence and to choose the median at the end of the observation period. A variety of clockwork averagers  giving a reading of the mean of the regular observations followed. The SOLD sextant, however, continuously  integrates rather than averages the observations. There is, however, probably very little practical difference in the results obtained by a British Mark IXA instrument, which averages 60 observations over 120 seconds and the SOLD which continuously integrates observation over a period of 40 to 200 seconds. The former is, however, more compact and of a better ergonomic design.

General arrangement

Figure 2: General arrangement, left side.

Figure 2: General arrangement, left side.

The optical parts, the battery and the integrator are sandwiched between two aluminium alloy plates joined together by pillars at intervals. Figure 2 shows the parts related to the outside of the left hand plate. At the front is a rectangular timing unit with its winding knob and starting lever. A shaft with a pinion on its end passes through the left hand plate to the integrator unit. A little below it and to the front is the daylight window which slides into a socket between the plates. This window can be withdrawn and a 45 degree mirror 7, Fig.1, substituted for night-time illumination of the bubble chamber. A slide allows a red filter to be placed in front of the bulb so that the bubble field appears to be red and night vision is preserved. The level unit control and air chamber lie outside the plat,e while the bubble chamber itself is held rigidly in the optical path between the plates. Provision is made to remove the level unit without having to dismantle the whole instrument. The left hand handle, behind the bubble control, has within it a rheostat to control the brightness of the bubble lighting, and below and to the rear of this can be seen part of the battery housing which lies between the plates. A slide to receive a 2-power telescope is to the rear of the top of the handle and above this, mounted between the plates, is a bracket containing two filters or shades to reduce the brightness of the sun or moon before their rays reach the observation mirror (5, Fig 1).

Figure 3: General arrangement, right side.

Figure 3: General arrangement, right side.

Figure 3 shows a view of the right hand side of the instrument. At the front is a rheostat used to vary the brightness of illumination of the scales. It is marked “Hell” (light) and “Dunkel” (dark). Behind this is the main drum, the periphery of which is divided into three lots of ten degrees and a vernier (not normally used) allows readings to one minute. There is a mechanism within the hum of the drum which limits it to two and two-thirds turns, so that altitudes up to 80 degrees may be measured. As the drum rotates, a linkage to the observation mirror causes it also to rotate. Once approximate coincidence between the bubble and body images has been obtained, the clutch above the drum, is engaged, linking the drum to the integrator mechanism. The values of further movements of the drum to maintain coincidence are then integrated. to give a mean value at the end of the observation period. Above and to the rear of the drum is a housing for the illuminating system of the scales. At the end of the observing period, the integrator lighting comes on and when the scale lamp switch button is depressed it goes out and the scales lighting comes on.

Figure 4: General arrangement  top view.

Figure 4: General arrangement top view.

Figure 4 shows some parts already described. There is an additional view of the shades bracket with the shades in the operating position. The main switch can be seen at the top of the handle with the on position (“Ein”) marked. The various scales, illustrated more clearly in Figure 5,  are shown, together with the window through which the integrator reading is viewed. The observing mirror is seen for the first time. Figure 5 shows a view of the top as if from a body at an altitude of about 30 degrees.

Figure 5: View as from a body at 30 degrees altitude.

Figure 5: View as from a body at 30 degrees altitude.

Figure 6 shows a view seen as if taking an indirect observation at about 30 degrees. The image of the bubble field can been seen in the objective lens, through the semi-reflective observation mirror.

Figure 6: View of indirect observation.

Figure 6: View of indirect observation.

Observation mirror controlling mechanism

The inner surface of the drum, shown face-on in Figure 3, has a spiral groove machined into it (Figure 7).

Figure 7: Drum spiral.

Figure 7: Drum spiral.

The drum rotates about a fixed conical axis (Figure 8) and a follower, seen on the right-hand side of the figure, engages in the spiral groove of the drum. A lever attaches the follower to the axis of the observing mirror, so that as the drum is rotated, the follower moves out along the spiral, thus rotating the mirror about its axis. A long helical spring prevents backlash by keeping the follower against the inner wall of the spiral groove.

Figure 8: Controlling mechanism

Figure 8: Controlling mechanism

Integrator controlling mechanism

When the clutch-engage lever (Figure 3)  is depressed a spring presses a conical peg into one of the holes drilled radially into the periphery of the drum at each whole number of degrees. This then links the drum to a radial lever arm that rotates about the drum axis, seen at about 1 o’clock in Figure 8. The teeth of a  sector attached to the lever and rotating about the same axis engage with a rack, one end of which carries a pin which engages with a fork attached to the integrator roller axis. Thus, as the drum is rotated back and forth to maintain coincidence between the observed body and the bubble, the integrator axis is also rotated. The spring-loaded forked device seen at about 5 o’clock in Figure 8 prevents backlash in both directions via a pin at the lower right of the sector. Figure 9  shows the mechanism in operation as if the drum had been rotated forwards. Rotation backwards would engage the other jaw of the fork.

Figure 9: Anti-backlash mechanism in operation

Figure 9: Anti-backlash mechanism in operation

Integrator mechanism

On the inner end of the integrator roller axis is a roller which is vertical when the drum is in the mid-way position with the clutch engaged (Figure 10). The axis runs in ball bearings while the roller itself rotates about plain conical bearings. A long horizontal spindle is pressed against the roller by a spring mechanism on the right hand plate. When the timer is started, its clockwork motor pinion engages with a rack attached to the carriage that carries the spindle and moves it from right to left. As long as the roller remains at right angles to the spindle, it simply rolls along it, but as soon as the roller is tilted, the resultant forces cause the spindle to rotate one way or the other, depending on which way the roller is tilted. The rate of spindle rotation is proportional to the amount of tilt.

Figure 10: Integrator mechanism.

Figure 10: Integrator mechanism.

Figure 11 shows the integrator in operation with the roller tilted. At the left end of the spindle is a small drum which shows the integrated minutes of deviation of the main drum from the mid position and a disc geared to this drum indicates the total number of degrees of deviation, offset by three degrees in order to avoid subtraction when arriving at the final result. After the observation, the reading of the tens of degrees scale, the whole number of degrees shown on the main drum and the degrees shown on the integrator disc are added to the minutes on the integrator drum. The time is of course taken from the mid point of the observation. The integrator read-out is shown in close up in Figure 17, below.

Figure 11: Integrator in operation.

Figure 11: Integrator in operation.

The level unit

I have given details of the construction and principles of the level unit in the preceding post (26 June 2012) so will not repeat them here, but will describe how to remove the unit from the SOLD. Figure 12 shows the location of the locking screw that secures the unit in the instrument. It is captive in a wedge that pushes down a short arm that in turn forces the level unit against a machined seat between the plates. Unscrewing it withdraws the wedge.

Figure 12: Bubble unit removal ,1.

Figure 12: Level unit removal ,1.

Figure 13 shows the wedge fully withdrawn, when the slotted head pin just below where the wires enter the plate is pushed upwards to release the arm from the level unit.

Figure 13: Bubble unit removal, 2.

Figure 13: Level unit removal, 2.

Figure 14 shows that two conical pins on the lever arm locate the level unit accurately in position.

Figure 14: Level unit removal completed.

Figure 14: Level unit removal completed.

Lighting system

A master switch is placed at the top of the left hand handle.

a) Level unit.

When the securing stirrup between the plates under the front of the instrument is swung forwards, the cover to the bubble lamp can be swung backwards to reveal the lamp in its socket and a slide (Figure 14). The slide has a plain and a red filter, so that if required, the bubble illumination can be made in red light, to preserve night vision. The bulb is 2.4 volts with a 5 mm bayonet fitting and the ground-glass exterior is rather delicate. The rheostat in the left handle controls the lighting intensity.

Figure 15: Bubble illumination lamp and filter slide.

Figure 15: Bubble illumination lamp and filter slide.

In full daylight, a ground-glass screen is inserted into the socket on the front of the instrument and this can be seen in place in Figure 15. For night-time illumination or in poor daylight, this fitting must be removed and replaced by one having a 45 degree mirror to divert light from the bulb to the bubble chamber (Figure 16 and Figure 1, above)

Figure 16: Fittings for bubble illumination.

Figure 16: Fittings for bubble illumination.

b) Integrator

While the master switch must be on to obtain lighting anywhere on the instrument, the bubble chamber will not be lit until the timing unit has been wound fully. This resets the integrator so that the minutes drum is zeroed and the degrees disc set to 3 degrees. The bubble chamber remains illuminated until the end of the observation, when the front of the integrator carriage operates a switch that cuts current to the bubble lamp and switches on the current to the integrator lamp. This switch is seen at the left of Figure 10 and is shown in close-up in Figure 17.

Figure 17: Integrator switch and read-out.

Figure 17: Integrator switch and read-out.

Removal of a thin sheet metal cover reveals the bulb and the switch wiring (Figure 17 a). The end of the bulb is painted black to limit stray light. The integrator lighting intensity is controlled by the rheostat on the right hand plate at the front of the instrument.

Figure 17 a: Integrator lamp

Figure 17 a: Integrator lamp

c) Other scale illumination

The tens of degrees scale and the degrees drum are lit by a lamp at the rear right of the instrument. The lamp housing is shown in close up in Figure 18. Pressing the red button switches off the integrator lighting and switches on the scale lighting.

Figure 17: Scales lamp housing.

Figure 18: Scales lamp housing.

Removal of the slot-headed screw reveals the lamp fitting and change-over switch (Figure 19).

Figure 19: Scales lamp fittings.

Figure 19: Scales lamp fittings.

d) Accumulator

This was a nickel-iron accumulator of two cells with a potassium hydroxide electrolyte and although the manual in giving instructions about charging refers to allowing the cells to gas after charging, the unit provided appears to be completely sealed (Figure 20), no bad thing in an aluminium alloy instrument.  Potassium hydroxide, as well as corroding aluminium with great ease, also generates potentially explosive hydrogen gas.

Figure 20: Exterior of battery.

Figure 20: Exterior of accumulator.

The unit appears to be a rechargeable hand torch, adapted for use in the sextant, as there is a socket for a bulb in the top and current is provided by a projection at the front and the metal switch at the rear, with internal contacts in the battery compartment. The switch plays no part in the instrument other than providing a path for current flow. The exterior of the battery housing is shown in Figure 21 and Figure 22 shows its interior. When the housing is closed, the rear contacts of the accumulator and the instrument come together, and the contact at the front of the lid makes the circuit from the switch on the accumulator via a curved brass strip only when the housing is securely closed.

Figure 21: Exterior of battery housing.

Figure 21: Exterior of accumulator housing.

Figure 22: Contacts inside battery housing.

Figure 22: Contacts inside accumulator housing.


The timer at the front left of the instrument is wound by depressing the knob (Eindrücken) and winding in the direction of the arrow up to a stop. This also returns the integrator carriage to its starting position and zeroes the integrator read-out, as well as preventing the integrator lamp from being accidentally lit. The timing period may be set to 40, 120 or 200 seconds, but only when the timer is running. It is started by depressing the trigger. The movement has a pin-pallet escapement and a monometallic balance wheel (Figure 20). The escape wheel has three set of teeth on its periphery.  A different set is moved into the path of the pallets for each timing period.

Figure 20: Timer movement

Figure 23: Timer movement


This is s short Galilean telescope of 2 power and has an objective aperture of about 28 mm, for use in indirect observations of fainter stars such as Polaris, increasing its apparent brightness by a factor of two. The recommended procedure for the inexperienced was to locate the star by direct observation and get approximate coincidence of star and bubble, before changing to indirect observation using the telescope. The telescope is focussed by rotating the objective mounting and is located in a dovetail slide on the left side of the instrument (Figure 24). It cannot be used for direct observations.

Figure 24: Telescope in place.

Figure 24: Telescope in place.


The exterior of the sheet- metal case is shown in Figure 25 and its interior in Figure 26 (These are courtesy of Alan W Heldman).

Figure 25: Exterior of case.

Figure 25: Exterior of case.

Figure 26: Interior of case.

Figure 26: Interior of case.

The two oval pockets are for accumulators, two of the large round pockets are for 110 volt blue incandescent lamps used for dropping the 110 volt DC ship-board current to one suitable for charging the accumulators, two of the large round pockets are for the ground- glass screen and mirror for bubble illumination (Figure 16), the rectangular slot is for the telescope, and four spare bulbs can be seen in place. The instruction manual is in a celluloid pocket at the front. The case also contains a charging adapter and lead, and a suspension strap, unlikely to be used aboard a ship.

The case measures 150 x 250 x 250 mm and weighs 2 kg. The instrument itself weighs 2.6 kg.

Operating instructions

The following is translation of an extract from the 1944 “Beschriebung und  Bedienungsvorschrift für den SOLD-Libellensextanten KM2″. I hope that those more expert in German than I will forgive any solecisms.

II Operating Instructions.

a)    Daytime observations

1)    Remove the instrument from the case with the left hand and insert the ground glass screen into the opening below the integrator.

2)    While holding the instrument with the top edge horizontal, find the bubble and adjust its size so it appears somewhat larger than the sun (about 1/3rd of the distance between the squared lines in the field of view). The bubble control of the level is on the left of the instrument.  Set the size of the bubble by tilting the front upwards a little so that the bubble sits over the triangle in the field of view. The triangle marks the correct place as well as giving a guide to the correct size. Return the instrument to horizontal and rotate the control back a little to the left to prevent further bubbles entering the bubble chamber.

3)    Transfer the device to the right hand, supporting it for better balance with the little finger under the level lamp housing. Wind the movement with the left hand by pressing in the winder and rewinding up to the stop. Set the running time by pressing down the starting lever and adjusting to the desired running time. Then rewind the clockwork.

4)    Point the instrument in the sun’s direction and rotate the right knob until the sun appears in the field of view. Engage the clutch when the sun and the bubble are together in the field of view. The inexperienced may find it easier to first locate the sun by the direct observation method.

5)    With a finger of the left hand, press down the clockwork starting lever and at the same time call out “Zero” to the note taker.

6)     While the clockwork is running one should try constantly to keep the bubble and the sun at the same height, i.e. to keep one alongside the other.

7)    If there is no note-taker at hand, then the time of observation is taken when the clockwork has run down. The observer then has to count off seconds from the observation time to obtain the time reading and deduct this from the reading time. The time is then recorded with the note “e” to indicate that it was the end point of the observation that was noted. The half of the time set on the clockwork should also be recorded at this time, as this will be deducted from the end time to obtain the averaged time of the observation.

b)    Night time observations.

1)    Remove the instrument from the case with the left hand and install the telescope and level lighting mirror.

2)    Remove the accumulator from the case and carefully install it in its housing. Open the lid only to 20o as there is otherwise a risk of damage to the hinge. Turn the switch in the left handle to the “Ein” (on) position.

3)    As for daytime observations (wind clockwork).

4)    The observer must make sure that the integrator lighting resistor is set to dark. The bubble field is made visible by rotating the resistor in the left-hand handle upwards with the thumb.

5)    As for (2) in daytime observations (bubble setting).

6)    View the star directly through the observation mirror and turn the right-hand hand wheel until the bubble and the star are in near alignment. Then engage the clutch.

7)    Operate the clockwork trigger and carry out the observation as under 5, 6 and 7 for daytime observations.

c)    Reading the average height.

Reading is as described on page 13. The basic reading on the degree drum is added to the integrator reading.

Figure 8: Reading the Integrator.

SOLD fig 8

                                                                                                                   1     Minutes drum

                                                                                                                   2     Degrees wheel

                                                                                                                   3     Degrees wheel index

Example: Clock time, end of observation 22h 11m 38s, running time of clockwork 40 sec, index error 0’. The reading gives:

Tens scale                                                     50 o

Degrees drum                                               8o

Integrator degree wheel in 4 division       4o

Minutes drum                                                 17’

Observed altitude                                         62o 17’ (see Fig 8a)

The relevant time of this observation is determined as follows:

Half time run                                     -20s

Clock time                        22h 11m 38s

Time =       22h 11m  18s

Next calculation as usual

Because of backlash, it is possible for the degrees wheel to indicate in the area of 2 degrees while the minutes wheel is already above zero (Fig 8,b). In this case, if the minutes wheel indication is past zero, take the higher degrees figure and if below zero, the lower.

Werner Luehmann has kindly provided the following comments and photographs of his Luftwaffe version of the SOLD:

As an owner of an “aircraft SOLD sextant” I can say that the only differences between the KM 2 an the aircraft version are: (1) left handle (also the orientation of the off/on switch is perpendicular), (2) accumulator (navy) versus a 3 Volts battery (air force) or optional a flying lead with in integral resistor to reduce the on board voltage, (3) the case (wooden box for aircraft type). Also, (4) the aircraft sextant is missing the “suspension bracket” (although there was a gear available to suspend it in the aircraft). There are neither obvious other differences in parts, nor any in “sealing”.

Figure A1: Exterior of wooden case.

Figure A1: Exterior of wooden case.

Figure A2: Luftwaffe instrument in case.

Figure A2: Luftwaffe instrument in case.

Figure A3: Instructions in lid.

Figure A3: Instructions in lid.

I wish also to acknowledge the kindness of Alan W Heldman in entrusting me with his SOLD sextant, for sending me a copy of the operating manual and for providing encouragement as I struggled with the translation. I am happy to provide interested readers with the whole translation, which includes the original illustrations. “Contact me” if you would like a pdf file of the translation. I make no guarantee as to its accuracy.

As this web site was started as an encouragement to readers to buy my book “The Nautical Sextant” I hope this post will act as further encouragement to potential readers; and I would also like to draw attention to my book “The Mariner’s Chronometer” More details may be found at

C Plath Bubble Horizon Attachment

26 06 2012

I have placed this post in the Aircraft bubble sextant category because, although the C Plath bubble horizon instructions imply that it can be used at sea with a nautical sextant, this is not the case. Bubble horizons are of use on land and in the air, where accelerations either do not occur or occur in a well-understood way. However, accelerations at sea are too great and unpredictable for them to be useful, except perhaps on a very large vessel in calm waters or in ice where no horizon is visible. In 1960, the German Navy was supplied with these attachments as were US submarines. It would be unsurprising if they were never used on aircraft, as much better, dedicated aeronautical sextants had long been in use (see Comment by Dr Andreas Philipp).

This post is preceded by  “A gummed up AN5851-1 averager”, “Bubble illumination of Mk V and AN 5851 bubble sextants” ,  ”Refilling Mark V/AN5851 bubble  chambers” ,  ”Overhaul of MkV/An5851 bubble chamber” ,  ”AN5851-1 : jammed shades carrousel” ,  ”A Byrd sextant restored” ,  ”Update on Byrd Aircraft Sextant”, “A nautical sextant bubble horizon” and “Sealing A10 vapour pressure bubble chambers.”

When C Plath resumed sextant production in the early 1950s, they applied their experience with the SOLD bubble sextant to designing an artificial horizon attachment for nautical sextants. They had been instrumental in producing a bubble sextant for Admiral Gago Coutinho of Portugal, who made several long over-water flights in the 1920s, but that sextant was an adaptation of a nautical sextant employing two linear spirit levels at right angles to each other, both of whose bubbles had simultaneously to be aligned with the observed body.  The SOLD sextant had a circular bubble, adjustable in size, and contained in a rigid bronze casting. The bubble unit used in the attachment for the post-war nautical sextant was based around a bubble unit almost identical in design and execution to the SOLD units. Figure 1a shows a general view of the units, while Figure 1b shows the bubble chambers from above, with the SOLD-type units on the right. Figure 1 c, from a war time manual shows a slightly modified design that is clearly the same design as the post-war unit.

Figure 1a : General view of bubble units

Figure 1b : View of bubble chambers from above.

Figure 1 c : Drawing of bubble unit (modified from SOLD manual).

When the adjusting screw is rotated anti-clockwise, the bellows are compressed and the pressure in A rises. It is transmitted via the drilled passage 1 to the bubble chamber and, if the bubble lies correctly within the triangle etched on the glass at the front of the unit, air is forced into the air chamber and the bubble size decreases.  Conversely, if the pressure in A is caused to fall by rotating the adjusting knob  clockwise, air is drawn from the air chamber, via passage 2, into the bubble chamber. The correct position for the bubble when observing is shown by a central square. In use, to adjust the bubble size, the sextant is tilted backwards to bring the bubble within the area marked by the triangle. After long storage, the bubble chamber may be nearly empty but nearly always this seems to be that the air and the fluid , alcohol in the SOLD, redistribute themselves, rather than being due to leakage. In this case, most of the fluid can be shaken out of the air chamber by repeatedly flicking the wrist with the air chamber upper-most. The air in the fluid chamber can then be displaced into the air chamber by patiently rotating the control knob slowly back and forth with the triangle up, until the bubble reaches the required size.

Figure 2 shows the ray path through the attachment. Rays from the observed body are deflected by the horizon mirror of the sextant  into a window in front and pass through a semi-reflective mirror into a 3 power Galilean telescope. The bubble formed in the bubble chamber above lies at the focus of a concave mirror below, so that rays forming the image of the bubble emerge from the mirror parallel, and the image of the bubble appears, like the observed body, to be at infinity. These parallel rays are deflected by the semi-reflective mirror (in fact a piece of plane, unsilvered glass) into the telescope. Thus, the image of the bubble is superimposed on that of the body and it lies to the observer simply to bring them into coincidence, a by-no-means-easy task.

Figure 2 : Ray path through attachment.

The bubble chamber is illuminated above by a bulb contained in a screw-on holder. Light from the bulb is diffused by a red diffusing screen (Figure 3) and its intensity varied by a variable resistance (“potentiometer” or “rheostat”) of about 10 ohms resistance. If the bulb is replaced by an LED, then the potentiometer will also need to be changed for one of about 1000 ohms, as described for my first post of September 2013.

Figure 3 : Bubble illumination.

Power for the lamp comes from the battery handle of the sextant. Current flows through the plug and lead to the wiper on the variable resistance (black wire, Figure 2) and thence to the brass ring (red wire) in the top of the attachment. The ring is insulated from the body of the attachment and, when the bulb holder is screwed home, another, brass ring, threaded for the bulb and insulated from the bulb holder, makes contact to carry the current to the bulb. The return current is carried from the central contact of the bulb to the body of the bulb holder, through the body of the attachment and thence, via the mounting fork, to the body of the sextant.

There is no provision for illumination of the bubble by daylight, so for sun observations, the shades of the sextant are dispensed with and a dark shade attached to the front window, so that the bubble can be seen at the same time as the sun. For star and planet observations, of course, no shades are used with the sextant or the attachment. Daytime moon observations may be possible  by removing the attachment’s shade and obscuring the horizon with all the sextant’s horizon shades and by using an index shade that just allows the moon to be seen, but I have no personal experience of this. The unit was originally supplied with only one shade.

I am grateful to Murray Peake for entrusting me with the overhaul of his attachment.


To expand on the paragraph after Figure 1c, when the bubble cannot be reduced in size, and small bubbles seem to enter the chamber at random:

Holding the unit upright, turn the bubble control fully anticlockwise (looking from above) until it will go no further. This will make the bubble big, as it will be sucking air out of the air chamber.
Hold the unit upright in one hand at arm’s length and swing it vigorously towards the floor as if a pendulum, but stop suddenly at its lowest point (this is a bit hard to describe, so I have added two photos; I am not slack-jawed in the first, I am saying “Take it now”). This is intended to force any fluid remaining in the air chamber into the bubble chamber and is much more violent than a flick of the wrist.
Tilt the unit backwards about 20 degrees and turn the bubble control slowly clockwise. The bubble may at first be so big as to appear invisible, but should then slowly disappear into the air chamber.
A tip : rather than having to attach the unit to a sextant in order to see the bubble as you manipulate it as above, unscrew the top lamp holder and illuminate the bubble chamber as necessary with something like a mini-mag-light.
Postscipt 1: Beginning of swing. Note position of unit in the hand.

Postscipt 1: Beginning of swing. Note position of unit in the hand.

Postscript 2: End of swing.

Postscript 2: End of swing.

Gummed-up AN 5851-1 bubble sextant averager.

9 12 2011

This post is preceded by “Bubble illumination of Mk V and AN 5851 bubble sextants” ,  ”Refilling Mark V/AN5851 bubble  chambers” ,  ”Overhaul of MkV/An5851 bubble chamber” ,  ”AN5851-1 : jammed shades carrousel” ,  ”A Byrd sextant restored” ,  ”Update on Byrd Aircraft Sextant”, “A nautical sextant bubble horizon” and “Sealing A10 vapour pressure bubble chambers.”

Bob Hauser asks a question in a different category that I think best deserves an answer in the form of a short blog. He asks

Recently acquired Bendix AN 5851 had stalled averager that could be wound up to stop as per directions but would then simply remain in that state when the release lever (“no. 3″) was pressed—-for lack of any better solvent/lubricant, I lavished Reel-X on the bull gears and down under those gears into the inner chronometer mechanism and gently turned the pawl driver clockwise by hand and repeated this 4 or 5 X manually until the averager ran on its own for the required 2 minutes …yes, it worked but for how long with that stuff in there before it gums up even worse? Reel-X is a solvent/lubricant that has about the viscosity of sewing machine oil and may wind up being the worst thing to admit in the chronometer like that…can you advise?

Generally, watches and clocks do not respond well to being flooded with lubricant for a variety of reasons: there is very little power at the end of a watch gear train, at the point where the rate at which the machine runs down is regulated by the escapement and balance wheel, so that even the surface tension of oil between the gear teeth or in the coils of the balance spring can bring the mechanism to a halt; the pivots, or bearings about which gears and other parts rotate, are provided with tiny oil wells (“sinks”) and the shafts are shaped at the end to keep the oil where it belongs. If the oil strays on to the plates of the mechanism, the oil that should be confined to the bearing tends to follow it; and  excess oil combines with fine dust and grit so that the bearings and pinions (those gears with relatively few teeth) eventually grind themselves to a halt.

So what is Bob to do? One could say “Sufficient unto the day is the evil thereof,” but ideally, the clockwork mechanism should be removed from the sextant, stripped down, cleaned, re-assembled and oiled in the correct places. This of course needs clock-repairing  knowledge and practice. While the official manual explains how to remove the mechanism, it involves removing substantial bits of the sextant first, with all the risk of introducing new problems or damaging or disturbing the optical system. I suggest that he simply expose the clockwork and try to do a bit of dry cleaning, removing visible Reel-X from the plates where they are accessible and from the gears that he can reach. He should pay special attention to the balance mechanism, removing any lubricant from between the coils of the hairspring and excess oil from the pivots. For the larger parts like the plates and large gears, small pieces of old cotton handkerchiefs applied with fine forceps are ideal, though he should take care not to leave stray threads behind. For smaller, more delicate parts, I suggest scraps of lens paper which, though it is not all that absorbent, it less likely than paper towels or handkerchieves to leave fibres behind.

How to access?


Figure 1 : Remove two screws and nut and bolt.


Figure 2 : Remove two screws.


Figure 3 : Remove one screw.


Figure 4 : Remove one screw. Note washer.

Once the lighting unit is out of the way you can get access to all the twelve round-headed screws that hold the sheet metal cover around three sides of the clockwork mechanism. Remove the screws and cover (Figure 5).

Figure 5 : Remove twelve screws and cover.

This partially exposes the mechanism. Concentrate on getting as much excess oil from the low power end of the gear train. The stars in Figure 6 show the important areas. There is plenty of power further up the gear train, but it will still pay patiently to remove as much oil as you can see and get at.

Figure 6 : Important areas from which to remove excess oil.

Re-assembly is the reverse of dis-assembly. Good luck!


Sealing A10 vapour pressure bubble chambers

13 10 2010

This post is preceded by “Bubble illumination of Mk V and AN 5851 bubble sextants” ,  “Refilling Mark V/AN5851 bubble  chambers” ,  “Overhaul of MkV/An5851 bubble chamber” ,  “AN5851-1 : jammed shades carrousel” ,  “A Byrd sextant restored” ,  “Update on Byrd Aircraft Sextant” and “A nautical sextant bubble horizon”

The popular WW II A10 bubble sextant, in common with nearly every bubble sextant of the period, was prone to leakage of fluid from the bubble units. The fluid was xylene and the seals between the bubble chamber and the top and bottom glasses were lead washers sitting in  grooves of semi-circular section

Early units were filled through a tapered hole below the Lucite light pipe, the hole being closed with a brass taper pin, another potential source of leakage, that was cut off flush with the outer surface of the chamber, making it not only nearly invisible, but also very troublesome (though not impossible) to remove and replace.

Recently, I was asked to refill three A10 bubble units for a new friend and when I received them I found that they had a modified filling system that makes refilling them somewhat easier (Figure 1). A tapped hole has been made nearly all the way through the front wall, terminating in a smaller hole at the entry into the chamber. The hole is closed by a screw, the end of which squeezes a lead ball into the smaller hole to seal it. There is a secondary seal in the form of a lead washer that sits in a shallow counterbore underneath the screw head.

Figure 1 : Sketch of chamber closure

Providing that the glasses are clean and their seals are still intact, something that becomes obvious as soon as the unit has been re-filled with xylene and sealed, refilling is much simplified. Using a glass syringe (plastic ones are rapidly attacked by xylene) and needle, the bubble chamber is filled with xylene and the control worked back and forth until no more bubbles issue from the control unit  into the bubble chamber, refilling as necessary. It is then sealed temporarily with the screw and holding the unit with the control unit down, it is flicked briskly towards the floor several times. This will usually force  more air out of the control chamber into the bubble chamber, which is then refilled to the brim of the filling hole, repeating the flicking process until no more air is present.

I found it rather fiddly to remove the lead ball from the filling hole and impossible in one case, so I omitted it, so as not to cause problems for the next person to refill the unit. In any case, lead shot is now frowned upon in New Zealand and I could not obtain any. However, if the seal beneath the screw head looked suspect, I replaced it with a new one made from thin sheet lead. This washer does not of course have to be round, though the hole in it does, and I made this using a punch. As a back up, I painted on several coats of shellac, which is not soluble in xylene, over the screw head and surrounding metal.

If you cannot remove the lead ball, you will have to remove at least the bottom glass, and it is probably better to remove the top one as well so you can give it a through cleaning, as my experience is that the old seals nearly always leak.  017  O rings made of Viton make very satisfactory replacements. They are resistant to attack by xylene, unlike common O rings made of Neoprene rubber. Neoprene rings can of course be used in the A10-A air-chamber bubble units to reseal the glasses, the air chamber and the diaphragm, as, according to the handbook, these were filled with alcohol.

If on operating the control of the A10 unit a stream of bubbles issues from the side of a glass, there is a leak and it will have to be re-sealed. A well-sealed unit gives a faint click as the control is turned to maximum. A bubble may not appear if it forms inside the control unit and you will have to give it a quick flick to force xylene into it and the bubble out into the bubble chamber. You can in some units avoid having to do this by forming the bubble with the control unit downwards, when a stream of bubbles flows up out of the control unit. In a well-functioning unit, only a tiny bubble is left if the control is turned fully anti-clockwise and this disappears after a little while as the vapour is reabsorbed into the fluid. If this does not happen, air may have been dissolved in the xylene (or there may be a leak) and you will have to add a little more xylene through the filling hole. Tighten the screw down hard and paint over with several coats of shellac (Figure 2).

Figure 2 : Later A10 bubble unit

 Needless to say, older units can be converted to this way of filling by carefully drilling and tapping an M3 hole through the front wall. The lead ball is optional and if you cannot find some lead shot 2.5 mm in diameter (the core size of an M3 screw), a steel one may do just as well, as in the A12 unit. However, the availability of Viton O rings means that the unit can now be refilled easily, though somewhat messily, by removing the bottom glass and then filling and refilling until all bubbles have been removed.

A Nautical Sextant Bubble Horizon

2 09 2010

This post is preceded by “Bubble illumination of Mk V and AN 5851 bubble sextants” ,  “Refilling Mark V/AN5851 bubble  chambers” ,  “Overhaul of MkV/An5851 bubble chamber” ,  “AN5851-1 : jammed shades carrousel” ,  “A Byrd sextant restored” and “Update on Byrd Aircraft Sextant”

A little while ago on e-bay I saw an adaptation of an A10-A bubble unit to a nautical sextant fail to reach its reserve at over $300, even though it was offered with a copy of my overhaul manual for the A10 series aircraft sextant. I recalled that a couple of months previously, I had made a very similar adaptation for a friend who lives in Paris, where natural horizons are not easily to be found. Since my means are relatively limited, I am always looking for ways of paying for my addiction to nautical sextants, so I decided to make another and this time to offer it for sale on the internet.

Most aircraft bubble units are of Second World War vintage, and after sixty five years, the fluid has leaked out of nearly all of them. The exceptions in my experience are the British Mark IX series, which were sealed with shellac and solder. US instruments sometimes sealed the glasses with shellac, but closed the filling hole with a taper pin or, as in the case of the A12, a ball bearing forced down upon its seat with a grub screw. Others used seals of lead or plastic and almost without exception, they leaked sooner or later. In the case of the A10A bubble unit, there were no fewer than six places where it could leak: two holes sealed with taper pins, one for filling and the other to allow a passage to be drilled btween the bubble and reservoir chambers, the top and bottom glasses, the joint between the diaphragm and the body of the unit and the joint between the reservoir and the body of the unit.

It is not possible to re-seal the A10-A units with shellac without damaging or destroying the Lucite illuminating ring. O rings had been patented by Niels Christensen in 1937 and during WWII the patent was taken over by the government in the national interest, but, curiously, did not find their way into sextant bubble units. It may be that, as most of them were filled with xylene, the elastomeres of the day were not equal to the task, but the A10-A units were, according to the official overhaul handbook, filled with relatively benign alcohol, just like the units in the German SOLD sextants and the later Russian copies of the SOLD. Although I have resealed units using home-made lead washers, it is much easier to remove the old seals and replace them with standard O rings if re-filling with alcohol or with Viton (fluorocarbon) O rings if using xylene.

So, having cleaned a bubble chamber and  resealed it with O rings I addressed the matter of attaching it and its optical attachments to a nautical sextant. Figure 1 shows the light path.

Figure 1 Light path through unit

The bubble lies at the focus of a spherical mirror, so that the rays that make up the image of the bubble reflected from the mirror are parallel and the bubble appears to be at infinity. These rays are intercepted by a partially reflecting surface or beam splitter and diverted into the eye. The eye also sees the image of the heavenly body, whose light rays, also apparently at infinity, pass straight through the beam splitter, so the images of the bubble and the object can be superimposed by adjusting the sextant. In daylight, the bubble is illuminated by the light from the sky and at night by a lamp that conducts the light through a Lucite strip that surrounds the top glass. Providing that the reflected rays from the spherical mirror are at right angles to the plane that contains the bubble, a line of sight through the centre of the bubble will always be horizontal. The mirror mounting allows it to be adjusted to this condition, and I give full details in my restoration manual. Providing it is collimated in this way (from the Latin collimare, which would have meant “to put in line” if a medieval scribe had not mis-copied collinare) it can be mounted on the nautical sextant without further adjustment.  A small index error may remain and have to be determined by observations from a known position.

The unit is attached to the sextant by a rising piece that I make using a shaping machine, the machine tool par excellence for cutting one-off vee ways. Rather than drill more holes into the unit, I removed the shouldered screw that held the shades and the top of the two  screws that limited their movement. I cut off the bottom screw short and used it to blank off the hole. I drilled out the holes and tapped them 4 BA. It is as well to dismantle the unit completely to avoid damage to internal parts when doing this. Instructions for dismantling are again given in my manual.

In day time the bubble is illuminated from above via a ground-glass diffuser screen that can be moved aside to view the bubble when adjusting its size. At night, a tiny bulb throws light onto the ends of a Lucite (UK : Perspex) strip that surrounds the top glass and the light is conducted around by total internal reflections. These bulbs are becoming hard to find nowadays, so I have experimented with using  a high-intensity red light emitting diode instead and it works quite well. The main difficulty with the adaptation is in reducing the diameter of the LED to fit the existing fitting. It is relatively simple to solder the LED to the base of a defunct bulb. The brightness of the lamps, incandescent or LED, is controlled by a potentiometer in the battery box. Incidentally, the Lucite strip does not seem to make a lot of difference to the quality of the lighting if for some reason it disintegrates or has to be dispensed with.

Here is another view of a bubble unit, from the rear of the sextant:

Figure 2 Rear view of unit



Update on Byrd Aircraft Sextant

11 08 2009

This post is preceded by “Bubble illumination of Mk V and AN 5851 bubble sextants” ,  “Refilling Mark V/AN5851 bubble  chambers” ,  “Overhaul of MkV/An5851 bubble chamber” ,  “AN5851-1 : jammed shades carrousel” and “A Byrd sextant restored”

Since writing the previous post about the Byrd sextant (or should it be the de Florez sextant?) I became disatistfied with the restoration and, for those interested in air navigation intruments I decided to add to the photographs. You will need to read the previous post first.

As noone followed my hint to donate me a set of index shades, I have had to make them myself. For the outline, it was simple enough to scribe around a genuine shade on to a sheet of 2.5 mm brass four times, drill and ream the mounting holes and saw out the blanks using a piercing saw. This tool is a little like a small fret saw and takes very fine blades having as many as 40 teeth per inch. It cuts on the down stroke and is used with the blade held vertically. Here is a picture of me cutting out a clock wheel blank. Note that it helps to have lots of light and vision…

Copy of 100_0066

I could then bolt the four blanks together with a close-fitting screw through the mounting holes and files the outlines to shape. It is actually easier and quicker to do the straight bits using a vertical milling machine – if it is already set up – and just to file the rounded corners by hand. The block of four could then be mounted and centred in the four jaw chuck of a lathe and all four drilled through and bored to 22 mm diameter. The outermost blank was then counterbored to 24 mm to a depth of 2 mm and removed, then the next counterbored and so on for all four, taking care to loosen and tighten the same pair of chuck jaws each time. This is not good practice, but it saves time when accurate centring is not vital.

It is hard to discover sources of neutral density glass, so I made a trephine out of mild steel and cut out discs of plastic Cokin filter material. This is used in photography and seems to be flat and parallel enough for a sextant likely to have observational errors of the order of minutes. I made the fit so the discs just popped into the brass frames and saved myself the trouble of having to swage them into place, though they might have looked more “genuine” if I had. Making the shouldered screw that holds them all in place was staightforward turning and I made the Belleville spring washer by making a thin brass washer, sitting it on the end grain of a block of hard wood and hitting it hard with a ball bearing. The finished set of shades shows well in the next photograph, as does the semi-circular lens that allows the level to be seen in focus.

Copy of 100_2948

I wasn’t happy with the fiducial line in my original restoration. It was simply a piece of fine thread wrapped around the level vial and secured in place with clear varnish. Since it is wrapped around a curved surface, it can be viewed only from one angle and still be seen as a straight line. In any case, it was rather too thick, so I scribed a thin line on some perspex and then cut and filed and drilled a tiny piece to size, securing it to the vial carrier with two 12 BA screws. These are only 1.3 mm in diameter and I was greatly relieved to have tapped the two blind holes without breaking my only tap of this size. The next photograph shows this small but important part. The flash has made it look dustier than it was in reality. I have no idea what was used in the original models. The Smithsonian Museum has two examples, but both are incomplete and I have only web photographs to look at.


Making the case from African mahogany needed only normally careful woodworking. Dovetail joints for the corners had by the 1920s given way to comb (finger) joints, but as some later American aircraft  cases used corner rebates, which are much easier to make without special machinery, this is what I used, with brass pins across the joints to prevent disaster if the very strong glue should fail:

Corner rebate

I copied the hook latches from a Hughes and Son sextant case and the handle is a very close copy of the handle used for a Brandis Aeronautical Sextant Mark 1 Mod 4 of 1931. I am not good at sheet metal work, so will gloss over the battery box, with its belt loop. The pick for the two capstan headed screws was simple to make and the mirror-adjusting wrench required only the ability to convert a small round hole to a small square hole using a file. It remained  to dismantle the instrument to its component parts and spray-paint them using a satin finish paint that, while not perfectly imitating the original finish, at least has the merit of pleasing its owner.

The final photograph shows the sextant in its case with its furniture and fittings. It is certainly not an easy sextant to use on land, but since these latest retoration efforts, there has been little clear sky around for me to make a serious assessment. If there is some particular aspect of this instrument that you would like discuss or to see illustrated, do contact me.


A Byrd Sextant Restored

30 05 2009

This post is preceded by “Bubble illumination of Mk V and AN 5851 bubble sextants” ,  “Refilling Mark V/AN5851 bubble  chambers” ,  “Overhaul of MkV/An5851 bubble chamber” and “AN5851-1 : jammed shades carrousel”

I recently acquired  a Brandis nautical vernier sextant without case, telescope, or any shades. It appeared to have an extra mirror in front of the horizon mirror and I recognised it as an early bubble sextant of the type used by the then Commander Richard Byrd on his claimed flight to the North Pole in 1926. There are several magazine photographs extant that show Byrd in posed pictures, using a similar sextant, this one, for example:


K Hilding Beij, writing in the Bureau of Standards Report 198, Astronomical Methods in Aerial Navigation in about 1926, refers to the sextant as a “Byrd sextant”, though Luis de Florez, a prolific  inventor, claimed priority. He had filed for a patent for exactly this type of bubble sextant in March 1919 and he was granted US Patent number 1,536,286 in May 1925. My sextant looked like this when I received it:


The one Byrd is using is a full-size Brandis vernier quintant with an arc of 180 mm radius reading to 30 seconds, whereas my example is unusually small for a vernier quintant, having an arc radius of only 140 mm, also reading to 30 seconds, so my hopes of possessing an historic instrument were disappointed. Nevertheless, it is a rare and early instrument dating from about 1920 and I felt it was worthwhile  to restore it to working order. As I have no access to original instruments, I did not set out to make exact copies of the attachments that make a nautical quintant into an aeronautical one, but I did follow the same principles, while retaining all the original parts. Needless to say, I dismantled it completely to begin with, and cleaned all the individual parts. A photograph of the restored instrument will perhaps best help to explain its workings.


The optical path for the heavenly body is as usual, via the index mirror and silvered half of the horizon mirror. An ordinary spirit level vial is held in a carrier and viewed via an auxiliary mirror set at 45 degrees above it, through the plain half of the horizon mirror. The image of the bubble would be out of focus viewed directly through the x 2 Galilean telescope and so an extra, semicircular, lens is interposed in the light path to bring it into sharp focus. The auxiliary mirror may be swung downwards to allow direct view of the natural horizon by pressing a spring-loaded catch.

The sensitivity of the vial has to be carefully chosen. If too sensitive, it is never at rest when the instrument is held in the hand and if not sensitive enough it is not possible to get meaningful results. I settled for one where the bubble moves 2 mm for 6 minutes change in level. This is of the same order of sensitivity as most other bubble sextants.  The bubble is illuminated from one end and, to try to get even illumination, the vial is painted white over most of its surface, including the end distant from the lamp. This has parallels in the lighting of some circular bubble cells, where an attempt is often made to conduct the light around the periphery of the cell with some sort of light guide. The next photograph shows a view of the lighted vial through the telescope (the view is somewhat more extensive than this in reality).


In sextants with circular bubble cells, one usually aims either to centre the body in the bubble or to align it with its equator, but this cannot be done with a linear cell, so a more-or-less central datum line is used and the sextant adjusted so that when the bubble is centred, the datum line lies on the horizontal. As is usual for a Galilean telescope, each half of the objective lens “sees”  its own half of the field of view (compare this with the inverting or astronomical telescope, where obscuring half the objective simply cuts out half the total light). Thus, the auxiliary, semicircular lens in the telescope attachment sees the left half-field and  is used to bring the bubble into focus by sliding back or forth.

The heavenly body is seen on the other half-field and, as is usual, there is a narrow band of overlap where both the sky and the horizon or bubble may be seen together. I have not made any significant trials as yet, but it is relatively easy on dry land to align a star, the moon or the sun with the datum line while trying to keep the bubble centred. However, the instrument gives no indication of lateral tilt and in an aircraft the results must have been very uncertain.

The scale lighting is particularly good. A shaded lamp shines via a standard diffusing screen on to the scales, which are viewed through a simple magnifier. Unlike many such systems it gives a very even illumination so that the main and vernier scales are seen with equal contrast, making reading relatively easy and rapid.


The handle seems to be a modified early Brandis battery handle. Power is now from an external source. Scale lighting is via the original press switch, while the bubble unit is switched by a rather crude rotary switch on the front of the handle, seen in the general arrangement photo above. The following photo shows the wiring layout, pretty well as found except for the decayed silk covered wire.

Byrd handle

I have coated the parts that I have made – telescope and attachment, and vial carrier – with modern paint. The original paintwork on the rest of the instrument is careworn and I cannot help but feel that it would look better for a fresh coat of paint. I would repaint a more modern instrument in the same condition and wonder how readers might feel about that. Is it sufficiently “historic” to preserve it as found? Would repainting it devalue it in some way.”? After all, noone is likely to mistake it for a modern fake, repainted or not.

The index shades, by the way, I borrowed from another Brandis sextant, so that if anyone has a spare set of Brandis or US Navy Mark II index shades that I can beg or buy, I should be glad to hear from them.


Get every new post delivered to your Inbox.

Join 41 other followers