How I decide a platen is too hard to keep using

Part 01

The thumbnail test and why I stopped trusting it

The usual advice is to press a thumbnail into the platen and see whether it leaves a mark. Fresh rubber takes the mark and lets it go; old rubber does not take it at all. It is a real test and it works, in the sense that a platen which feels like a wooden dowel is definitely finished. The trouble is the middle. Between obviously soft and obviously dead there is a wide band where my thumb reported whatever I expected to find, and I only noticed because I started recording my guesses next to what the typing looked like afterwards.

Out of thirty-one machines through the room in 2022 I called eleven platens hard. When I went back through the sheets in January the following year, seven of those eleven had typed perfectly acceptably once the rest of the machine was sorted out, and two that I had passed as fine were cutting through the ribbon on capital letters. That is a poor record for a test I was charging my own time against. The thumb also varies with the weather. A roller at fourteen degrees on a February morning feels considerably deader than the same roller after an afternoon in a room at twenty-one.

What finally decided it was a wide-carriage standard from about 1954 that came in with a complaint about faint letters. I told the owner the platen was fine and spent a full evening on the ribbon drive and the type alignment. The letters were still faint, and the platen was the problem all along. A gauge costs less than that evening did.

Part 02

What a hard platen actually does to the type

A typebar arrives at the paper at speed and has to stop dead. The rubber is what stops it, and a small part of the rubber deforms for a few thousandths of a second to spread that stop over an area rather than a line. When the rubber has gone glassy it cannot deform, so the energy has to go somewhere else. Some of it goes back up the typebar as a bounce, which shows as a light or doubled impression. Some of it goes into the ribbon, which is a few hundredths of a millimetre of woven nylon or cotton being crushed between two hard surfaces, and that is why a dead platen chews ribbons.

The letters that suffer first are the ones with the least printing surface. A full stop, a comma, the apostrophe, the numeral one: all of them concentrate the whole blow onto a tiny area and punch through. On a machine with a really dead roller you can often read the last page typed on it as a pattern of pinholes held up to the light. The wide letters, m and w and the capitals with long bars, go the other way and print faint at the edges because the type can no longer settle flat against an unyielding surface.

There is also the noise. A machine with good rubber has a sound like a firm knock. A machine with a dead platen rings, and the ring carries down the frame into the table. I have started listening for it deliberately when a machine first comes out of its case, before I touch anything, because the ear gets it right more often than my thumb ever did.

Part 03

Measuring instead of guessing

The gauge is a Shore A durometer, the pocket sort with a dial, and it cost about what two evenings of my time is worth. The procedure I settled on takes ninety seconds. I take four readings along the length of the roller, at roughly a quarter, a half and three quarters of the width plus one at the very end where the platen knob sits, then a fifth on the underside where nothing has ever typed. The underside reading matters because it shows what the rubber was before the machine was used, and the difference between it and the top surface is the wear rather than the recipe.

New platen rubber for these machines sits somewhere around 85 to 88 on the Shore A scale, which is already quite firm; a car tyre tread is around 65. Anything I measure below about 92 types well. From 92 to 95 the machine works but the light characters start to bite, and I will mention it on the sheet. Above 95 the gauge needle simply runs to the stop and the roller is finished for typing, whatever it looks like. The hardest I have measured was a machine stored in a loft in the north of the city since the 1980s, and the needle sat against the pin at 100 on all five points, including the underside, which told me the rubber had oxidised right through rather than just glazed on the surface.

Part 04

Ring and cylinder: why a new roller changes the alignment

A recovered platen almost never comes back at exactly the diameter it left at. The shop that does the work grinds the new rubber to a nominal size, and half a millimetre either way is normal and not a complaint. The problem is that the type alignment was set against the old diameter. The adjustment is the one usually called ring and cylinder, and it governs how far the typebar travels before the ring around the typeface stops it against the platen surface.

Set with the platen too far back, the top of each letter prints and the bottom fades, because the bar is still rising when the ring catches it. Too far forward and the reverse happens: heavy feet, light heads, and an o that looks like a c with the gap at the top. On most machines the adjustment is at the typebar itself, bending the bar a fraction at the heel with a proper forked tool, and it has to be done letter by letter for the worst offenders rather than as a global change. This is slow work and I have never enjoyed it.

The relevant detail for a platen job is that the key ring alignment has to be checked afterwards, not before, and that I should have written that on my own wall in 2020 instead of learning it three times. I now type a full alphabet on a sheet of ordinary paper, hold it against the window and look at the heads and feet of each character before I call a platen job finished. Two out of three machines need a handful of bars nudged. One in three needs nothing at all, and I have never been able to predict which.

Part 05

What happens to the rollers that are past it

Recovering a platen means sending the metal core away to be stripped and re-rubbered, and it takes some weeks and costs more than most people expect for what looks like a stick with a skin on it. For a machine that somebody types on every day it is worth it and I say so. For a machine that sits on a shelf and gets used twice a year, it is not, and I say that too, which is an odd position for somebody who works on these to take.

The interim fix is a backing sheet. One sheet of ordinary paper wrapped around the platen under the work takes the worst of the harshness out and stops the ribbon being punched through. Two sheets are better for the impression and worse for the crispness of the letter, because the type starts to emboss into the padding rather than print flat. I keep a strip of thin card cut to carriage width by the bench for testing, and if a customer's complaint disappears with the strip in place, the platen is the fault and the rest of the machine is fine.

I have also stopped trying to soften old rubber. There are several methods in circulation involving solvents, oils and various household liquids, and I tried three of them on a spare roller in 2021. Two did nothing that lasted more than a fortnight. The third made the surface tacky for about a month, after which the roller was both hard and sticky, which is worse in every way than it had been. The rubber has cross-linked and there is no undoing it at a bench like mine.