Spacing of 9.5 characters per Inch

Being easily distracted, started taking apart a recently acquired little 3-bank typewriter for a proper cleaning. When scrubbing the escapement rack, noticed that it had a number stamped in it: 2,67

A quick check with the vernier confirmed that 2.67 indeed is the pitch in mm. Somehow had not expected that. For a 'normal' Pica machine assumed 10 characters per Inch as common standard.

One Inch at 25.4 mm by 2.67 mm suggest 9.51 times - with a design goal of 9 and a half characters per Inch, the pitch would be 25.4 divided by 9.5 equals about 2.673; effectively the 2.67 as marked. Very likely then that these machines were not 10 cpi, but 9½ cpi.

As extra; the mostly emptied frame of this little machine - no prizes :-)

To fix the Hammond linespacing

To fix the wobbly linespacing of the Hammond Multiplex typewriter, the rear-roller was covered with a single layer of heat-shrink tubing. The rubber of the rollers on most Hammond machines has aged badly, usually cracked from shrinkage. When acquired 2 years ago, the roller looked pretty bad.

It had already been sanded to make its surface even. Now also one length of 25mm (shrunk) diameter polyolefin tube was fitted over the old rubber. This worked to increase the diameter of the rear feed-roller and hide all the cracks.

With a clean, new roller-surface. The excess bits were cut off with a scalpel.

Whilst it is indeed possible to remove the rear-roller without taking the whole carriage apart, it is not recommended. Mr Hammond could certainly have made this a lot less difficult. It may be just this particular machine or it is the effect of accumulated design-changes that obscured the original design-intent. The tolerances are almost-but-not-quite-enough for cleanly taking out the roller. If only three parts would've been made about a mm shorter, it would've worked without issues. The parts being as they are, means that the erasing-plate even needs to be loosened to give the side-frames a way to flex (!) a little bit. These side-frames are delicate castings, that makes removal of the roller a risky undertaking. Fortunately it all went well and no broken casting - but removing the rear-roller still not a recommended thing to do.

The adjustments after re-fitting are also an interesting look into 1880s engineering - not before on a typewriter had I seen the need to adjust the linespacing detent to match the position of the linespacing ratchet. And there's so much more to adjust!

Result of it all is a Multiplex with a very firm grip on the paper and spacing lines as it should :-)

Parenthetical Corona in a bit of a state

An early Corona folding typewriter. It is in a bit of a state, rusty with a cracked platen and some missing bits.

Perhaps overpaid a bit for this one, but these early parenthetical Corona's are getting hard to find and it will be a nice project to completely re-build and discover how these early machines differ from the later solid-segment Corona's.

This machine still has the 'Standard Folding Typewriter' line in parentheses under the Corona branding - hence 'parenthetical' :)

It already has the pigeon-decal, wide spacebar and the fan-shaped ribbon-selector; so it is not of the very first 1912 variant of Corona, but a bit later. The serial number probably is 28806, that would make this the 8806th Corona (Standard Folding Typewriter) 3 that was made. That would have been in early 1913.

Normally the typebar-pivots are best kept clean and free from oil, but these first need to be made freely-moving again. All screws were also given a drop of sewing-machine oil and the little typewriter will be kept on the shelf for the oil to creep and help loosen all screws. 

A parenthetical Corona 'new' on the shelf - awaiting restoration and a full re-build.

New typeshuttles for the Hammond - almost there

After printing and testing a lot of shuttles, it's almost there; the ability to create new typeshuttles for the Hammond typewriter :-)

The latest improvement was to finally get a new 10 mm wide ribbon, moderately inked with a nicely dense-weave - looked better than most new regular half-inch ribbons (usually over-inked and coarse weave). Wound on some new spools, tried this on the machine and the improved evenness of printing is immediately obvious.

This new ribbon does very well with the crisp resin-printed shuttles; this Hammond can now credibly type in Attic typeface. (But it does really need work on its feed-rollers.)

Especially the resin-printed shuttles are amazingly crisp, allowing for different typefaces to be clearly visible - any limitations on resolution coming from the ribbon and paper, not from the type. How well some typefaces come out in FDM remains to be tested, to see what the limits are of FDM for making shuttles.

The other major step forward was in FDM or filament-printed shuttles. The 3D models were modified to compensate for some of the limitations of the printing-process. In the below image the effect can be seen most clearly in the capital-letter 'O'. In the regular FDM-printed shuttle (1) the face of the 'O' is slightly curved inward, the letter will print top and bottom heavy. The compensated print (3) has this curving mostly removed - it is compensated for in the 3D model. The resin-printed shuttle (2) of course is pretty-much perfect without any artefacts - but harder to manufacture.

Having come up with a method to compensate for some of the limitations of FDM printing, also did a test with a standard (rough) 0.4 mm nozzle on the printer. Using a 0.4 mm nozzle would make printing of new shuttles much easier and faster. This experiment gave a relatively rough looking shuttle (4). Placed on the Hammond Multiplex typewriter - this shuttle actually does type!

That typing-sample above was done with the home-made ribbon with its un-even inking.

All 'flavours' of manufacturing technique now print legibly, but of course differences can be seen. The 'cheap-n-cheerful' 0.4 shuttle does print less crisp and the most 'costly' resin-printed shuttle is best.

The status of the slow project to create new Hammond typeshuttles is thus looking positive - most of the issues seem to have been solved and it should be possible to streamline the creation of 3D models for different layouts and typefaces - and printing processes for different qualities as desired. 

The angles and line-spacings may need tweaking; with teamed-up efforts on Hammond shuttles this will for sure be jointly solved. With more shuttles, we'll get more insight into the effect of different materials and additives. And of course the durability of the various materials is to be explored.

Almost there :-)

New spools for a Dalton Adding Calculating Machine

And a few fixes.

The most glaring issue with this dusty and rusty Dalton was that it had no spools and no ribbon.

The spool for a Dalton of the "Super Model" series that launched in 1921 is a proprietary format, but not too complex in shape. Quickly modelled in CAD and two form-fit-functional replacements created on an FDM-printer. 

A small nail to catch the ribbon, the top-flange glued to the hub and thus two new, replacement spools.

The left spool-capstan was much lower than the right one. It also did not advance at a machine-cycle. A bit more peeking revealed that the left-capstan had slipped its mounting-clip and been pushed down into the machine. Its ratchet-wheel thus free from the driving ratchets/dog.

Re-forming the clip (bending) and fiddling the ratchet back into engagement, the capstan is again at correct height and whole mechanism again works. Including the automatic ribbon-reverse - this does require a lot of pull on the ribbon though! (Best to firmly nail that ribbon to the hub.)

New spools and ribbon on the machine - this specimen turns out to be one of the 'Special' 100-dollar model 181-4 machines that do not have two-color printing. The only model in the Dalton range to be monochrome; black ribbon, very austere.

Having the front panel off was a good opportunity to clean the position-indicator. The dots are again visible through yellowed celluloid.

Another small improvement; the handle. The rough, rusty crank-rod had been hollowing-out the wooden handle. Taken off the crank the wood was cleaned, oiled and given a copper-tube insert to have a bit less free play on the de-rusted rod of the handle. 

With ribbon, the machine adds just fine; 8 times 12345679 indeed is 98765432. (Machine still looks terrible, though :)

Dalton 10-key-swap (and a surplus Dalton Adding Listing Calculator)

The number-keys of the Dalton 10-key adding-listing machine were a bit damaged. Some keys were 'mangled' as if by being burned (cigarette?) and the legends were worn off.  From a local pick-up wreck-machine, a surprisingly clean set of original keys could be taken:

Replacing with clean keys, the cleaned-up woodgrain Dalton's keyboard looks much better - readable numbers:

The worn keys were then put on the 'wreck' machine. Swap.

This dusty-brown donor-machine originally was lacquered black. It's probably the same basic Dalton model 181-4 as my woodgrain specimen, but is a few (~3) years older. (The woodgrain finish probably came in around 1929 or 1930, superseded by a plain dark-green probably around 1932 or 1933. By then, these large 'little giants' of the 'super model' line were obsoleted by the new 'Portable' model derived from the Brennan.) The finish has not aged well; apart from large rust-areas either the paint or a varnish-layer is 'cracking' and drawn-up in spots.

The machine obviously lacks spools and ribbon - otherwise it actually still functions. After a few cycles to overcome some sluggish spot in the input-register, it calculates fine. The only thing noticeable now is that the crank-return is not dampened - internal damper must have lost all its oil. Still impressive.

The serial number places it in late 1926 or early 1927 - so this machine was made right around the time of the merger of Dalton with Remington, Rand Kardex and Baker-Vawter Ledger into the new Remington-Rand company.

Now what to do with a surplus Dalton new "Super Model" machine? - cosmetically terrible, but looks to be functionally perfect. Perhaps first some new spools to be made for it, then see what to do :-)