Odhner tandem
Another machine that was bought at auction by the Arithmeum, and needed a bit of TLC to both look and work properly.
The issues at the very start were clear - there was a big dent at the top, the right machine did not properly show its input control register, the right input did not clear with the wingnut, the machine was blocked, pulling out the crank would not free up the input control registers, the result registers were also blocked, didn't clear, and as it appeared later, the revolution counter did not have functioning tens' carry.
All in all, a good day's work for the aspiring office machine mechanic!
All of these problems could be related to hardened oil and lack of lubrication (apart from the dent, that is), and this is indeed mostly how it panned out. I took the covers off the machine and repaired the input registers not freeing up by lubricating everything in the chain of events that leads to their clearing. The text reads "During transport, the locking screw must be screwed in". If you go looking for this locking screw, you'll find a more extensive set of instructions right there - see below.
It turned out that the decisive issue which would block the input control register clearing time and again was dried oil between the input register numeral wheels and the axle passing through them - while the numeral wheel is normally carried slightly too far in normal operation, it is then supposed to spring back to a position determined by the pinwheel cylinder, but in this case, the movement of the numeral wheels was so sluggish that they just stayed stuck at their furthest position. A bit of clock oil fixed that.
Once that was done, also all the other gear trains in the top part of the machine were lubricated (including the direction switching and +/- for the counter register), which also made some of the safeties properly operational again, and now the machine could be reliably rotated in either direction.
The next issue was the carriage. The tens' carry in the revolution counter didn't operate, and the clearing of the middle register was blocked completely.
I had started taking the peripherals of the carriage apart, but since the gear for tens' carry that rotates together with the rest of the machine and slides from left to right on an axle at the back of the carriage as it is moved is held captive on said axle, the only way to disassemble the carriage on one of these Odhners would be to undo all the screws fixing the machine to the bottom plate, and take all the rest of the machine off the base. The instructions are quite thorough, and read (in Swedish) : "During transport, the locking screw shall be screwed in, while the carriage must stand so that the yellow arrow points between 6 and 7. Before use of the machine, unscrew this locking screw so far that it sits against the stop plate." Just in case you would be a bit blind, or stupid (after all this machine was mainly used in the army, and would be packed away by perhaps what we would call "not the sharpest knife in the drawer"...) the Odhner factory added a large arrow next to the text, pointing to the locking screw.
So I started doing exactly that, but halfway into it I came to my senses. What, again, was there that couldn't be remedied by some cleaning and oiling with the carriage in place? As soon as some cleaner was applied and some fresh oil, and everything started moving freely in the revolution counter, the effort for clearing it went down about tenfold, and the carries started working again. The carriage of these machines, by the way, rolls very lightly, on four roller bearings.
I also oiled the quick clearing gear trains of the result registers. The right clearing started working reliably and lightly, so then it was just down to the middle register in the carriage. It functions with a tilting tooth that grips into a three-toothed ratchet wheel. The problem seemed to be that on release of the lever, it would spring back, but the tooth would be a few tenths of a mm short from being able to fall into the next cutout in the ratchet wheel. There was no slop in the mechanism, the lever was not loose on the axle, nothing seemed to be out of place, so I didn't understand why this didn't work. It turned out that there was just an accumulation of dirt on the bottom side of the lever, so that it would spring back, but not completely. Adding a bit of oil and pushing it down completely for a few times cleared the blockage, and now also the last register clears reliably.
Now we were down to three issues - one was that there was a big dent at the top of the machine, where something must have fallen on top of it. I carefully massaged that out as well as I could - it's still visible, but by far not as bad as when I started.
The second issue were the numeral wheels in the result registers and the counter. Where they had been sitting under the windows in the carriage top plate, apparently sunlight and/or cigarette smoke caused the top layer of the result wheels to crack/peel off, completely obfuscating the numbers.
I very carefully filed this cracked layer away, and scratched all remaining white paste out of the result register numeral wheels with a pick.
I then filled the zeroes (and part of the ones where the paste was missing) in again with off-white latex paint, so they would be well visible again.
All of the result wheels were scrubbed to try and get the nicotine stains out, but this was not completely succesful - some numbers, particularly those that spent a lot of time at the bottom of the register, are still quite brown.
The final issue, I thought, was the paint job on the machine. Odhner used a kind of varnish on their paint jobs, and on this machine clearly the varnish perished. The entire machine was a dull black, and also spent a good part of its life in a very nicotine-rich environment, because what I started to scrub off was very dark brown to outright black, and smelling of cigarettes. It's a long and tiring process to bring a paint job like this one back from the dead. The extra annoying thing is that this varnish, if there are any remnants on the machine when it is given a coat of wax, will turn light brown, becoming very clearly visible. I have a few Odhners in the collection which display this problem, so I was determined for this machine to try and take the varnish and nicotine off completely, and make it shine again. You can already see in the picture above that the base came up quite nicely, so I then started on the left side of the machine - before:
during:
and after:
Much better!
In places the paint has worn very thin on the machine, which was sort of hidden by the layer of brown grime and decomposed varnish. More scrubbing - before:
After:
before (left) and after (right) in the same picture for the top plates
So now the machine was pretty much finished, but a baffling problem started to occur - as soon as the machine was rotated, suddenly the result registers would no longer clear, because the security lock to prevent the clearing of the carriage kicked in when it should not - namely with the main crank handle in the zero position, where it should be for clearing the carriage. Taking the back off the machine again, I investigated, and what I found was not pretty.
The brass gear which operates the disk with the cutout that proved the interlock had a tooth which was sheared off at a weak point created by drilling for one of the three pins that fixes the gear to the locking disc, and in addition the gear was not engaging over its entire width with the gear that drives it. Luckily this is the one single gear that is the easiest of all gears in the entire machine to remove - just one single screw and it comes right out!
Sure enough, guess what I found on the left cover of the bottom plate when I took this off earlier?
Clearly someone had taken this gear off the machine before - I found the washer which is supposed to shift the gear a bit to the left to provide full engagement - the lack of it explains how the damage could occur - as well as the missing tooth itself. This damage means buying a new gear - it is a 20 tooth gear, with modulus 0.75. After checking with many suppliers, all of which sell gears with modulus 0.5, 0.6, 0.7 and 0.8, but not 0.75, it was everyone's favorite auction site to the rescue - sure enough, I found a seller there with 6 of these brass gears, 20T, mod 0.75 in stock - so I ordered two, to have a contingency plan if the drilling and turning to width of the first one would somehow go wrong. These are the new and old gears next to each other.
Here I knocked out the brass pins that pin the old gear to the steel plate with the cutout:
Then I drilled out the central hole to 5mm to fit over the axle.
I then reversed the gear in the chuck, because the new gears were too long, and needed to be cut down. And this is where my foresight really paid off, because of course, trying to be careful not to damage the teeth of the gear, I didn't tighten the lathe chuck enough, and the gear was, eh, destroyed.
Not to worry, I had ordered a second one ;) Then the gear had to be drilled to be locked to the steel plate. I superglued the gear to the steel plate in the correct position.
Then I transferred the holes to the gear and drilled through with 1.7mm. It then became clear why the gear that had been in the machine had had issues (apart from insufficient engagement). One of the holes in the gear is so close to one of the teeth that the tooth in question is barely attached. I decided not to use this hole for locking the gear, but instead hammered two brass pins into the remaining holes, and glued in a piece of wire with some clearance and cyanoacryalte glue to strengthen the tooth. I think the Arithmeum should apply for a warranty repair with the Original Odhner company ... this is no way to conduct a business, with machines just breaking after barely 80 years of use!
This picture illustrates how much thicker the gear now is.
The backside, pins in place:
And finally, the gear in place in the machine:
Now for some pretty pictures:
This is a video of my favourite example coordinate calculation, rotating two coordinates in the same grid over 25° clockwise
...and finally, some pictures of the machine with its "simple" siblings, the Odhner 18 and 20.
