tor 2017-04-06 klockan 19:59 +0200 skrev Mattis Lind via cctalk: > A late follow up on this topic. > > I got my IBM 1627 running. Actually it is not mine, it belong to my father > that got it used for his home built computer somewhere around 1977. It > became a project that never was finished... There are still some simple > drawings he did on how to interface it. > > http://i.imgur.com/8JK8wwO.jpg > http://i.imgur.com/8YJA0Qb.jpg > > First thing I recognized was that the Mallory FP type capacitors in this > one both measured below 1nF capacitance. A quite serious problem. It turned > out that when I pulled them apart the positive electrode was subject to > severe corrosion. It was simply cut off.. > > http://i.imgur.com/XUaQtPn.jpg > http://i.imgur.com/nMLoFbN.jpg > > One of the stud mount 1N2547 rectifiers diodes was broken at off from its > studmount in the middle of the insulating ceramic material. Since this type > of rectifier was hard to find I replaced it with a modern part. > > Then on I replaced other electrolytics that all measured marginal (after > all this machine is from 1967, 50 years old!). The two Mallory AC > capacitors which are the run capacitors for the paper feed system were > replaced as well. Not that they measure bad but I simply don't like oil > filled caps from the sixties. They might contain PCB oil. And I didn't find > a list that certified that Mallory RP-3301 is PCB free. > > Then the PSU was fine and all voltages measured well. > > On to the mechanics. The drum was working fine in both directions, but the > carriage was not moving in either direction. Checking the electronics > revealed that it tried to drive the stepper motor but nothing happened. It > was stuck. > > The motor was removed and disassembled. There was a small gearbox on top of > the stepper motor and it turned out that the motor axis was completely > stuck. > > http://i.imgur.com/O043J9J.jpg > > By removing the last six screws it was possible to get the axis out by > knocking on the axis with a screwdriver handle. It turned out that there > was quite some rust in there. Cleaning, re-lubricating and new grease and > it worked quite well. > > One note for others who repair these. Do not remove the big wheel that is > driven by the stepper motor and drives the carriage. It is not necessary > unless you plan to put some new grease in the gearbox. It is a small hell > to get the carriage cable back into position again.. > > After this the plotter moved nicely in all directions and then it was time > for serious testing with pens which bring us back to the topic of the > thread. > > This plotter was supplied with non-pressurized pens. As far as I understand > there are pen holders for both varieties of pens as well as felt pens. Mine > are definitely non-pressurized. > > http://i.imgur.com/ZquhQMT.jpg > > As per Randys suggestion I put them into an ultrasonic cleaner and they > become much cleaner. Then onto filling them which turned out to be a > complex and a messy business.. > > From what I understand there are different types of ink. The type of ink > that are used for reservoir pens or Harry Potter feathers which is more > liquid and then a thicker ink which is used in ballpoint pens. The former > is sold in small bottles the latter is not sold except for in ballpoint > refills. > > The next step is to try to extract ink from the ballpoint refill and use a > ink refill syringe to push it into one of the cleaned pens. We'll see how > that works out. >
Ballograf in Mölndal is nowadays the only roller-ball pen producer in sweden (Scandinavia too?) Maybe they would be amenable to sending a bottle of ink (they are world-famous for producings roller-ball pens with archive-longevity inks.) > The other option is to find new pens with the same diameter which can be > cut into the correct length and tip dimensions. Maybe even plastic ones > will work if they are stiff enough. > > In parallell I am working on the interfacing part. As far as the > description in the manual interfacing is quite simple. Either negative > going or positive going pulses, at least 10V amplitude and more the 4us > long with less than 10us rise time. Source impedance less than 500 ohms. > > A ULN2004 or similar together with some simple AVR microcontroller will > probably do it. > > /Mattis
