Halden

Halden Calculex 1

Halden's Calculex is a serious calculator with an air of luxury - just the right size for a gentlemen engineer to pocket in his waistcoat, says the 2" square gilt-edged manual of 95 tissue-paper pages. The case is of tooled leather. The instrument is of silver and glass.

Each of the glass faces bears a cursor line, and can be independently turned with one's thumbs. The outer scales of both sides are fixed. The inner scales are linked, and are rotated together by pinching both sides of the center knob between thumb and finger and twisting.

Note that the outer A scales of both sides are clockwise, while the Bs of both sides run in opposite directions. Because the Bs move together, one side of the Calculex is always performing the inverse operation of the other. That is, whenever the A-index of one side is positioned opposite a value on that side's B, the A-index on the other side also points to the same B value on its side. However, on one side higher A values point to higher B values (multiplication), and on the other higher A values point to lower B values (division).

This Calculex needed adjustment as can be seen in the large pictures below. As you can see, though the indexes of the A and B scales are aligned in the right picture, they are not in the left picture. So I set out to fix it.

It turns out that the inner scales of each side are actually independent brass disks that turn in tandem only because they're clamped together by the two halves of the hub. [In other (later?) models I've got (below) the disks are riveted together.] The 3 screws (2 slightly offset from center on one side, and a larger one in the center of the other side) pull the two halves together, clamping a spacer between the two inner dials. Though the 2 offset screws help to ensure alignment, if the screws loosen just a little, and if there's just a little play in the alignment holes, the two inner disks can wander out of synch.

I found the repair difficult as there's no way to work on just one side. All 3 screws have to be removed to take the hub apart, and as soon as you do that, both sides come off. Getting it all aligned and back together requires steady hands, a light touch, and patience. I managed, but don't recommend it.

Sold for $200 in January 2021 to Capacity Engineering Ltd, in Canada.

Calculexes 2 and 3

One of the pleasures of collecting is finding things you can rescue. In 2013, Melvyn, in England, contacted me about a broken Calculex he wanted to fix and/or sell. One of the glass faces was gone, and so also one of the two small screws that fasten one side of the hub. I suggested he should try to get a new glass made, and might find a new screw at a jeweler. Melvyn tried to find a place to make a new glass, but ultimately had to settle for a plexi replacement. And he wasn't able to find a matching screw.

Then in late 2014 I found somebody selling another partial Calculex on eBay. It was also missing one glass, had a bent scale, and was missing the top ring. I figured if I could get them both I could swap parts to make one perfectly good instrument with leftover parts to sell/trade. Melvyn's also came with a leather case and booklet like above, but the other came with a metal case lined in purple velvet.

I first got the partial from eBay. I was surprised to find that it was constructed differently than the one I already had, above. Then Melvyn's arrived and I found that while it was of similar construction to the other partial, it was also different enough that few parts could be exchanged between the two. Here are the details.

Below is a schematic of the cross section of my original Calculex, which I'll call model 1. The key outward feature of this model is that the two glass faces are inset within the surrounding metal ring. The ring houses a 3 layer "sandwich" of the scales, that forms an inner disk that rotates within an outer ring. The outer ring is riveted together and attached to the inside wall of the surrounding ring so that they're a solid assembly. The inner disk is not part of that assembly. It's just 3 pieces that get compressed together when the screws clamp the two halves of the hub.

This is why it's so easy for the two inner disk scales to wander apart: though there are two uncentered holes though the 3 parts of the inner disk, there's unavoidably a tiny bit of play in them. Thus, you've got to get the two inner disk scales perfectly aligned and keep them aligned, as you gingerly put the screws in and tighten. If you jog things even a little, the two inner scale disks can dis-align easily and there's little way to realign them short of taking at least one glass off, nudging things back into alignment and trying again.

In contrast, in the construction of the two partials, the scales are all an assembly separate from the surrounding ring, and the glass is not inset. Both the 3 parts of the inner disk and the 3 parts of the outer ring and riveted through so it's a freestanding piece, and clamping the halves of the hub together just binds it to the inner disk and does nothing to bind the parts of the inner disk to one another. This approach makes it much easier to assemble since the alignment of the inner disks was ensured when the inner disks rivets were put in.

One may wonder, however, how in this construction the outer ring is held still while the inner disk is rotated. The answers are that (a), the scale assembly fits very closely in the surrounding ring, close enough to be just shy of tight, and (b) the ring and ball at the apex of the Calculex screw through the surrounding ring and engage a notch in the outer ring spacer. As a result, the outer ring is secured to the surrounding ring and won't turn with the inner disk.

So much for the construction differences between model 1 and the partials. The partials, which I'll call models 2a (from eBay) and 2b (from Melvyn), are also different than one another, both cosmetically and in the exact sizes of their parts.

The most noticeable different is in the surrounding rings. 2a's is smooth while 2b's is ridged:

Also obvious, 2a has squared-off hubs, while 2b's are beveled:

Note, both hubs have larger diameters than model 1.

Finally, model 2a has scales that are smooth surfaced and white while 2b's scales have a grainy, silvery surface. It this picture, 2b's scale assembly lies on top of 2a's.

Other differences between 2a and 2b are that

• their screws are of entirely different diameters and thread sizes

• the diameter of the shaft of of 2b's hub is larger than that of 2a

  • the 2 side holes in each scale assembly only fit the screw positions of its own hub, so you can't cross hubs between scale assemblies

• 2b's surrounding ring and ball (mine is missing the ring) seem to be made of a baser metal than 2a

• 2a's ball is smaller and slightly flatter than 2b's

• 2b sports a "cowling" between the ball and the surrounding ring that overlaps the glass faces. (It's possible that my 2a is just missing that cowling, but I don't think so. The shank of the screw on the ball is too short - I tried putting 2b's cowling on 2a, and then, due to the cowling's thickness, the ball-screw doesn't engage the slot in the outer ring's spacer very well.)

Here's a picture of the parts and pieces of both partials - 2b is at left.

The hole in the 2b glass was filled with a metal (maybe solder) "spacer" to conform to the stem of the hub. The 2a had resin spacers that were actually off center. Perhaps they were replacements for lost metal ones. In both cases, the fit between the spacers and the walls of the inexact holes in the glass are perfect, as if the metal/resin had been poured between into the hole in the glass and around the hub shaft. You can see the spacers in these two pictures - the 2a is on the left and the 2b on the right.

Because of all the differences, basically the only parts that can be exchanced are the glasses. I liked the appearance of the 2a better than the 2b, and also I had all the screws for the 2a, so I decided to use 2b's glass on 2a. This involved carefully filing the spacer for 2b's missing glass to fit the hole in 2a's remaining glass. Then, because the hub shaft of 2a was of a slightly smaller diameter than 2b's, I had to pad the inside of the spacers with an additional layer of thin brass so their fit around the 2a hub shaft wouldn't be loose.

I also had to fix 2a's bent outer disk scale. One of the rivets had come loose and I was able to reinsert it, but even so the scale was bowed. So I wound up using a little cyanoacrylate to glue it down to the outer disk spacer.

Here's what I wound up with for the 2a.

Then, in order to be able to put the 2b together as well, I cut my own plexi replacement to add to the one Melvyn had made, and I fashioned a new ring for the ball out of brass rod. Here's what I wound up with.

I'm calling it a success and a half. I've got one complete Calculex with two glass faces, and one missing a screw with a replacement ball ring and two plexi replacement "glasses". All told, I'm into the project for only $100.

Calculex 4

I love these little wonders, and so did somebody else. The owner of this next one fashioned a super hardcase for it from Bakelite!

This version is similar to my first one (above) but not quite. No two of my Calculexes have the same dimensions or construction! I get the sense that they were all handmade.