The Thacher Calculating Instrument

A long while back, I heard that K&E used to replace the scales on their Thacher Calculating instruments. One could send that treasured calculator back to the shop for a repair job. Now, of course, there's no repair shop for them, and no source for replacement scales. So I had the notion that maybe I could develop facsimile scales that could be used to replace the scales on a Thacher. Additionally, the same scales might be usable to create a new home-made version of the calculator.

PostScript Thacher drum top left

At the bottom this page you will find attached replica scales for the Thacher Calculating instrument. I suggest you go check out Bob Wolfson's site first: Bob's Calculators and Slide Rules. He worked with me on the latest stage of the project, has built several replicas as well as refurbished a wreck, and deserves full billing for the work he put in on it.

It turns out that you can build your own Thacher, if your carpentry skills are - let's say - well above average (see Bob's Calculators and Slide Rules, above). Likewise, you can replace scales on a worn out Thacher.

Both of these options look to me to be beyond the average home builder - which is sad because what got me started was a desire to build my own. The scales have been a work in progress for a decade, or rather, took two versions to get right.

This is the chronology of the project:

Pre-2001

I first saw a Thacher Calculating instrument in my office in the late 90s. The company I worked for at the time, had a mint Thacher on display that had seen service in the Mechanical Engineering department. Seeing it on a daily basis seriously twigged my interest in owning one, as I was just starting collecting slide rules as a hobby. In those days the going price of a Thacher was so high that I expected I'd never own one.

So, I naively tried making one. Since I had no notion of how to tackle the mechanical side of the build, I started with the scales. I used MSExcel for calculating the tick locations, and laid them out manually using a drawing tool, I forget which one. That effort lasted a couple of months. The original draftsmen were much sturdier than I; I made it to scale 6 or something like that, before I caved in.

Not too much after that, I first saw a PostScript slide rule scale on the Web, the Otis King, designed by Peter Monta. Peter wrote a program in C to generate inline Postscript commands that rendered the scales. His scales were being used as replacement scales for damaged slide rules, as I understand it. Peter's Postscript files are in the Slide Rule Museum.

Peter's method was the technique I started to use for my own scales.

2001 - first version

Early in 2001 Cyril Catt asked for the International Slide Rule group for help with the creation of slide rule scales to repair his old scales on his Thacher. I signed up. Michael O'Leary kindly sent me a complete set of high resolution photos of his own Thacher, because at the time none of sufficient quality were available.

Overall, the generation was based on Peter Monta's technique: Postscript scales were generated by Pascal programs and were largely inline code, thus avoiding most of the benefits of the programming capabilities of Postscript. This approach makes for large PS files, for slide rules of this complexity. On the other hand, it's simple and obviates the difficulties I have with mastering Postscript.

The problem with scale replacement as opposed to a simple home-brew, is that we want to replicate the scales as authentically as possible, and the people who originally lettered the scales were not generating scales with the same consistency as a modern computer might. The effect was that the draftsmen showed a bit of personality. And that made it very difficult to replicate by straightforward computer generated scales.

The tick lines were pretty much in the right places and followed a pattern that could be easily automated, save for the bar centre. It was lettering styles that weren't quite the most difficult to follow.

To support the hand lettered scales, I developed fonts for the scales and script. It wasn't possible to get an exact match, but I think it's pretty close. It looks a lot more credible when it's printed on the page. Nonetheless, I thought differences are enough that it would be unworkable to replace parts of scales, as opposed to a complete scale replacement.

I was never sure that the slide rule scales were sufficiently authentic, but sent them back to Michael and Cyril, and retired from the project exhausted. The scales have been available on the International Slide Rule Museum site.

2011 Version 2 (Bob Wolfson)

Bob Wolfson was working on his own home made Thacher, and needed some help with problems in the scales. It was Bob's patience and attention to details that actually brought the scales to I what I think is a credible replica. Bob unluckily became my tester in order to build his home made Thacher.

Bob found numerous issues with the bars in particular, because the bars have some decidedly odd behaviour at the ends and the centre.

He had a terrible time getting the fonts installed too. That's an issue I find with portability of replica scales. If you need to change the size of the images, you need to have and edit the PostScript files. But if you have the PostScript, you also need the fonts. If I deliver as PDF, it's usually hard to re-scale them locally.

2018 Revision (Peter Fox)

When "reasonable facsimile" isn't good enough:

In July 2018, Peter Fox contacted me with an improvement request. He was trying to replace a portion of each scale on the bars of his Thacher, as the bars had become unreadable on the ends. He found the 2011 design didn't match his Thacher, when the new scales were laid side by side on the bars.

Peter, because he was trying to replace a portion of a bar scale needed to line up the original and facsimile scales out side by side - which draws attention to any inaccuracies in the copy. He observed: My bar font did not look exactly like the original when overlain in situ; font sizes and tick thicknesses were not exactly like the original, and there was a slight misalignment of the horizontal lines.

The remediation project completed in September 2018. The exercise revealed a few false assumptions I'd been living with:

Wrong assumption 1: When I first built the scales, the notion I had as that the entire scale would be replaced. This made my 'reasonable facsimile' notion more easily attainable. If you are replacing the entire scale, it needs to be accurate and close enough to original to look consistent when used to patch an area. It now can support such patching.

Wrong assumption 2: I was having a lot of difficulty believing Peter when he told me the scales were not sized correctly when he printed them. There was a fair bit of trial and error while we tried debug that problem. It turned out that with Peter's Thacher, the full scale length was just a bit off. On my own Thacher the scales are "on spec", that is to say exactly 18 inches long. I had not realized that such may not be the case for all of them, even though I'm very familiar with how paper stretches with heat and humidity. Of course, I should have asked up front about the dimensions of his scales.

The attachments are below.