back

We replaced the metallic left side panel of our restored Burroughs machine with a transparent Plexiglas window giving a view on the impressive system of levers. After waiting without success for the cutout to be made by a German company specializing in Plexiglas cutting, Claude Baumann made the work on his refurbished CNC machine, using his special written software (link). The fit was perfect! I added a LED strip for illumination.

LillyPython code v. 1.34 (zipped folder, with executable, a library of examples and full source code)

Baumann Claude
Massen Francis
(Mar 2024)

This is a huge 54 pages report on the creation of an original Python compiler for the 50 years old HP-65, the first programmable hand-held scientific calculator of the world introduced in 1974. Recursive routines using the Labview programming environment are created to produce a virtual instrument (VI) allowing to input simple Python programs to create the HP-65 specific routines. The report is divided into 2 parts:
Part I : Recursive expression evaluator
Part II: Recursive cross-compiler

The LilliPython package contains all the code needed to run the software, using the free Labview Run-Time package (not included).The application to start is LilliPYTHON1.34.exe.
To install the free Labview Run-Time, simple google or try this link; I installed the 2021 SP1 32 bit version on an older Win10 (64bit) laptop, and choose the online install. You may need to register to allow the download.

Here the first screen you get after running LilliPYTHON1.34.exe:

Claude Baumann has created here a practically world-unique software package, using his extensive skills of Labview, recursive programming and compiler building.


The Computarium has 3 HP-65 (all made in Singapore, 1975), of which only one had a working magnetic card reader. The others had to be repaired and also needed a replacement battery, and one even a new CPU board from Tony Nixon (yellow board is HP original with some faulty registers, green board is the new CPU board of T. Nixon, with a Bluetooth module on the upper left corner)

see video

I auctioned a Mercedes-Euklud M30 (the Stopdivision variant) on Ebay. It did not work, the power cable was brutally cut off, the collector of the electrical motor was a gooey mess etc..., but the external aspect is excellent, and also the inside in very good shape. Spending about 6 hours cured most problems. The "Stopdivision" feature is quite strange, and you may look at the video for a demonstration. This is one of the many machines I am preparing for our upcoming Science Day in May.

see album

Update: the machine is unrepairable, due too too many faults from bent parts by a former fall. The machine is now exposed "as-is" in our large magenta cabinet.

The restoration is nearly done.
What's working:
- use the 2 counters (duplex feature)
- use the automatical carriage with change between non-add, total in counter1 and total in counter2
 - automatic CR/LF
- all totals ok (grand total and clear all)
- subtotals (= key R) ok
- programmed as used by the BCEE: 1st column: non-add, 2nd col. counter1, 3rd column: counter2
- total key pulls up ok, no manual help needed anymore
- all numbers print correctly, as well as symbols
What's not working:
- transfer of total1 to counter2; we will leave this and have the TT key immobilized (*)
- Non-Add key, stub on stem broken (but automatic Non-Add is ok); key immobilized
What remains to be done:
- we plan to replace the metal left side by a transparent acrylic glass
- we still do not understand the working of the electrical gear box, but the engine works fine and smoothly

(*) An  interview (link) with O.M.Smith, a former worker on Burroughs bookkeeping machines suggests that the TT key transfers only from counter1 to counter2

20-Dec: A trolley was made with blue felt cover (8 wheels in total), the left moving tablet was restored (new axis) and the electrical plumping changed, keeping the vintage bakelite switch (at the rear border of the table) and the distribution box underneath the table. Cover is replaced, but we wait for the plexiglas to be inserted at the left side.

29Nov2023: see small no-frills  video of the machine working.

We have two very old Burroughs printing adders, they are more of a booking-machine type. The Class 1 (from 1911, serial 9-174439) is a 9 column manually driven machine, the Class 2 a heavy 60kg beast with an electrical motor (17 columns, dual totals, probably 1926) which was used by the Banque et Caisse d'Epargne de l'Etat (BCEE) in the 1920/30's. Both are shown in the Computarium, but neither works.
I started to look seriously at these machines: the smaller (seen at the back of the picture) has many problems, the main being a fault in the printing engine which causes an incomplete movement of the crank, what hinders an important pawl from taken the correct position. The bigger (at the front, serial 2-1010374) has a blocked motor and a pressed down Total key that did not move. Today, after many hours spent in the preceding days, I was able for the first time to manually operate that machine, using the compatible crank of the Class 1 model. There remains much work to be done, but today was despite the outdoor pouring rain a moment of pure joy when I could make the first Addition, Total and Reset on that heavy 60kg thing (electrical motor and printing engine not yet operational, keys capture temperamental for some columns...).

This electrical FACIT ESA0 calculator is exhibited in a transparent Plexiglas enclosure, and regularly shown in operation. This morning I found that the machine was completely blocked. I brought it on the workbench, and tried to free the pin wheel drum that was blocked at the left side of the machine. Fiddling around and switching the motor on off finally started the machine unexpectedly, with the drum sinking a sharp metal nose deep in my left finger. Blood flowed like hell, but our housekeeper Guy Weber (which knows how to make first aid) very professionally put a band-aid that stopped the blood... running down from 3rd floor attic to ground level nevertheless left a bloody trace! As you may see, restoration work is not only an intellectual pleasure, but may be challenging for your physical health!
You may watch this very interesting video on the Rapid Collapse of the Swedish Mechanical Calculator Industry., which tells the history of FACIT.

Anne BAUMANN (daughter of our member Claude) donated a rare Burroughs Class 5 keypunch calculator, for which we are very grateful and say many thx!. The serial number suggests the machine being from 1925. It is in an excellent visual state (exterior and inside), but totally blocked, crank and keys and dial wheels are stuck.. The machine stood probably 50-60 years somewhere without anything moving, so the complex planetary system to make the carries is blocked. Loosen these parts is a job for a Benedictine monk: it took me at least 40 hours to make finally 6 of the 9 columns working. I can not find any technical manual, just two videos on Youtube where the guys also complain on the work...
The moving columns make an overshoot sometimes (pressing 1 is ok, but keying in  3 increments by 5 for instance), but I can not quite see where to adjust for this. So this restoration works really is not finished at all... For the above mentioned Youtube videos click here and here.

The Computarium acquired a brand-new THAT analog computer built by the new Anabrid company founded by Prof. Dr. Bernd Ulmann. Here it computes an Euler (or Cornu) spiral and displays it on a Fluke oscilloscope. The THAT is fiddly about clean power supply, I use a USB Powerbank which works just fine.

I restored two Burroughs class 9 printing adders: one crank driven from 1953 (at the left) and one from 1954 driven by an electrical motor. The first (donated by Marcel Baumann) was in a bad shape, with awful lint and dust inside, and also much rusty corrosion. But after a thorough cleaning it worked flawlessly (I really was surprised!). The electrical model (which some years ago I auctioned on eBay) is in a very nice state, no rust, no dust, and it purrs like a cat.

One of our two vintage A3 inkjet printers was unable to print black. The K print head is definitively dead, and was replaced by the Yellow one, so that we now have a K C M only printer. Click on image for a short album.
30 Jun 2023: Both printers have only 4 MB memory (on a 80-pin card), so they abort when an image is too large. I added today a 72pin 32MB card (from MemoryStock) to each printer, which goes into a well hidden special connector on the motherboard. I first goofed thinking I should replace to 80pin card and got upset seeing that I bought 72pin cards; finally all is well, and both machines now correctly print large images.

A protection of the RGB LED stripe was missing, so I added transparent plastic water hose pieces (cut in half or simply with a longitudinal slit) over the LED's, and some decoration on the side panels. The Jukebox is now a playable showpiece in our first room in the main exhibition.

The Computarium participates to the late night organisation (29th April) of our lycée, with a Retro Juke-Box playing 8-bit games music and several very vintage (1970's) games, which will be mounted near our big cabinet, with a lot of bells and whistles. The upper picture shows the JukeBox which really is a vintage Epson POS; the lower shows an old rolling cabinet with a very heavy Philips TV set and a PONG and ATARI 2600 video game. The Atari runs the FERRARI play. More at the exhibitions page.

Many people were involved in this repair. This very rare machine is now working 100%, and exposed at the Computarium. See album!

See the short album!

The tube is mounted above our Millionaire calculator, which is also a device from the first decades of the 20th century.

More on Gundelach tubes and other Roentgen tubes here.

PIC based signal generator (lab report on this PIC generator here.)

We wanted to show the LA function without many clutter; so Claude built an 8 channel signal generator using one of his PIC's, all this in a small case holding the PIC and a 9V battery:

CH.0 is an UART serial stream, sending the text "COMPUTARIUM" in regular bursts. Ch.1 is the pulse wave triggering the send of transmission; the next 6 channels are pulses with smaller and smaller frequencies (mostly dividing the previous by two). Here a screenshot of the signals captured by a "modern" Kingst LA1010 USB LA (highly recommended!): See the lab report on this PIC generator here.

Recall: version 1.1 of the lab report on creating the TTY/USB converter is here! 3 minute video here!

The 1963 ASR-32 Telex restoration is mostly done; the whole equipment is mounted on a wheeled dolly, together with a laptop running Teraterm; Teletype documentation (in yellow cradle) is stored below. The TTY/USB interface is fixed by Velcro so that the Receive/Transmit toggle switch is in easy reach, and that the box can quickly be removed when needed. A "Figs" problem caused by dirty signals has been solved by adding a 1 MF capacitor on the receive line (i.e. the line receiving the data from the Telex). All functions including tape punching and reading now work like a charm. The front panel has been salvaged from an ASR-33, as that of the ASR-32 was missing.

2. Yesterday (4-Nov) I saw that a large cabinet which holds a part of our software collection has crashed: the highest board became loose, and fell down on that below, starting an avalanche until all boards were down. Here a picture with the first 2 boards cleared away:

I made new sturdy supports which are now held by metal screws, and not by the former plastic stubs which were unable to support the weight:

BTW these cabinets are disused equipment of the physics department of the Lycée Technique Hôtelier Alexis Heck (now EHTL), donated by my late wife Colette; the Computarium recycles whenever possible!
 

Read version 1.1 of the lab report on creating this converter here! (update!)

(Oct.2022)

Claude has built a TTY/USB converter for the ASR-32 Telex, which allows communication between a PC (using a serial terminal software, 9600-N-8-1) and the 5-bit ASR-32, that runs at 74 baud. The converter was heavy work, with Claude pulling all strings to program a PIC using the PICLAB software, Assembler and Labview.

The device is a small box, which many LEDS showing Rx/Tx and FIGS/LTRS settings, and having a FT232R USB/serial converter on the PC side, and a 3 wire connection on the ASR-32 side:

Claude has managed to set the different timings (using complicated interrupts) so good, that the interface manages to correctly interpret the terrible signals sent out by the ASR-32. The left picture shows what the Telex sends to the PC, and the right the smooth signal in the other direction PC to Telex:

There remains one annoying mechanical problem with the carriage return, which is not executed, neither in local mode by pushing the RETURN key, nor in ONLINE mode when receiving it from the PC. But the electrical communication works without any fault on this device, which probably is a world-première!
 

HP 16500A
"mainframe" logic analyzer

(Oct.2022)

Claude who is a really expert in PIC microcontroller programming and Labview usage had built a small ASCII to BAUD serial  converter, allowing to send text from a PC with 9600 bit/s and usual parameters to the ASR-32 which receives it at a much slower speed and different start/stop parameters. His interface has a DIP switch which allows to try out the many different parameters that exist. Our first surprise was that the ASR-32 (which has a 3 wire output, and not the standard 4 wire dual loop) curiously is an active device (all ASR-33 teletypes are passive), working in half-duplex. We rather quickly found out the working combination and could send text from a buffer to the telex. The other way round still has some hiccups, but changes in the programming should resolve this. There is a common GND problem between the Rx and Tx lines that we are unable to solve, except by adding a toggle switch to change between sending and receiving. Claude will build a small electronic module to handle the transmission and code changes between a PC using its serial RS232 port and the ASR-32.
We spent quite a lot of time (smoothed by a bottle of good French red wine), also checking the different pulses and decoding. The ASR-32 has two mechanical problems that still exist: the keys FIGS and LTRS send the correct binary codes, but do not switch correctly the Telex between Numbers/Symbols and Letters. Another problem (which we also see in an ASR-33 8 bit Teletype of very similar but not identical CCU) is that the RETURN key does not engage properly to execute the return of the carriage to the left (but the bit pattern sent = keyboard decoding, is ok) This problem waits to be solved, it may be difficult to find where a mechanical part does not operate as it should.

The Computarium acquired an HP16500A logic analyzer (probably from 1992), which they called a "mainframe". It is very heavy, uses a Motorola 68000 uP and has two 3.5" 720kB (DD) floppies as storage medium: one at the rear which holds the system, and one on the front where configuration files etc. can be stored. They use the HP LIF file structure which is non-compatible with DOS. It is a beautiful beast, with a single module for the logic analysis. With a bit of luck we may find the DSO module (a digital oscilloscope board) which would be an interesting add-on. We were able to make an image of the single system disk (OS v. 5.01) present, using IMAGEDISK. But LIF formatting and disk duplication also works fine on the 16500A itself.  I found DOS versions of SYS 6.0 at archive.org, and was able to create the first system disk using the  LIFUTIL utility software of HP on our DOS CopyStation.

Again, many new things to learn, as both Claude and myself are neophytes in this domain of logic analyzers.!

I restarted work on the ASR-33 donated by Carlo Mullesch (a 115VAC model, where I added a 220/115 transformer into the stand). Everything works fine, with the exception of the automatic tape reader, which refuses to start either manually or through software commands (but it works by manually pushing the trip lever). The problem probably is caused by one of the 2 power supply/relay-logic modules in the stand.

_____________
04 Jul 2022:

I finally made something of a kludge to bring the tape reader to work: I routed a 117VAC line via an On/Off switch directly to the solenoid of the trip-magnet; this switch is mounted on the frontpanel of the stand: switch it on, and the reader starts, switch it off and it halts. This bypasses all "automatic" logic and relays, and also the manual (and broken) start lever, the tape-out and tape-tight switches. It works very well!

The call unit of this model is somewhat particular, with 5 square switches each sitting on a relay. I did not find any information for this wiring, so this part also is switched off with no regrets!

I also programmed the answer-back drum to send a "COMPUTARIUM" message when the HERE IS key is pressed; breaking off the correct tins of the drum is a messy work. Look at the picture with my preparation paper and the broken off tins...  and the printed result (the exclamation mark was not intended, I goofed with my pliers...

So this TTY now works fine, as do the puncher and reader.

end 04 Jul 2022
---------------------------

I also switched on a very old TTY (donated by Jean Mootz+ in 2012), which resembles an ASR-33 but in fact is the 5 bit version ( 32ASR or ASR-32, typename on plate is 32 TAJ). It works with some faults of wrong decoding, no CR, but the tape puncher and manual tape reader seem to be fine. This is a 220VAC model with a big transformer and a curious keyboard with FIGS (figures) and LTRS (letters) keys to switch on the upper rows of numerals or the lower 3 rows of characters; no SHIFT and CTRL keys. This keyboard has 4 rows of keys, whereas the ASR-32 normally should have only 3 (see here). This machine was used by WESTREX. (a Western Electric subsidiary) and probably bought by Jean Mootz at a "stock militaire" ( = military used equipment store) in Belgium.

Look at the tape reader which has a "normal" 8 hole plate, but only 5 pins to sense the holes in the paper tape and stubs spaced for the narrow 5bit tape...

"HELLO COMPUTARIUM" punched on the 5-bit Telex paper tape.

Today I got all keys/codes and the tape puncher and reader working properly, with the exception of a non-responsive RETURN key, and also no automatic CR/LF after the 72th character printed. This and the non printing of numerals remain the last 2 problems. Connection to a computer via the Horter V24/20mA converter has not yet been checked.
I programmed the drum so that the HERE IS key prints COMPUTARIUM.

We have a working demo exhibit with an TTY ASR-33 (115VAC model) connected to an OSI single board computer with 3kB BASIC in ROM. The TTY started making very disturbing, ratchy noises after a few minutes, or even seconds of running, and needed an emergency shutdown. I did not see anything mechanical becoming loose, so oiled most rotating axis and bearings with silicon oil. This helped somewhat, but the problem did not disappear. Finally I put white lithium grease on the plastic sprocket wheels, because the noise could be caused by one wheel ratching over the second which demanded too much torque. This helped definitively (tongue in cheek!). As there was than another problem with the print head not printing, I first build a second V24/TTY active converter from a Horter kit (easy to build, inexpensive!, see second photo), and operation with Teraterm on a XP laptop (110, 7bit, no parity, no flow control, 2 stop bits) was ok after some supplementary cleaning. The paper tape puncher of the TTY works fine, but the reader does only return gibberish...I do not find the fault. But this is not needed for the demo...
24June22: tape reader now ok. 3 sensing pins were out of their guiding holes...silly fault, I should have noticed without delay...

The DEC VT180 (anno 1982), a VT100 terminal modified to become a computer by adding a Z80 board, was ill: only a single line was shown on the CRT, and the second floppy drive was missing. The first problem was easy to solve, a contact problem. The second was more difficult: in my collection of big 5.25" drives I found none that worked flawlessly. Finally I found a narrow Matsushita drive, that when used as drive A: and configuring the original DEC drive (a Shugart 400, SSDD) as B: made a successful CP/M 2.2 (v.1.0) system. This system runs MUSIMP, the famous system for doing symbolic calculation developed by the Software Warehouse in Hawaii from LISP, starting 1976. The VT180 was donated by Jean Mootz (+), who was a fan of symbolic algebra.

PDP-11/34 non repair ongoing album! (last update: 25Jun2022, album finished!)

The (non-working) DEC PDP-11/34 system has been definitively brought to its resting place near the flashy magenta Computarium cabinet suite. The upper panel has been replaced by a plexiglass window allowing to peek inside and see the numerous boards.

A lock enables to open the upper RL02 drive drawer for demonstration. The bottom panels have received new (yellow) fixations which have been printed by our Makerspace crew (Scheer Tom). The picture shows the old fixation with broken off holders and the new yellow models made on the 3D printer.

Posters at both sides of the rack give further explanations and QR links to videos about DEC and the PDP-11 minicomputers.

An album showing the several months long repair effort will follow asap... stay tuned!

1. Our contributor Nico Beckerich donated among other items an Atari Portfolio in mint condition, with memory card (which is seen as drive A:) and parallel and serial interfaces (the 2 modules left and right on the picture). I tried with the these to make transfers from PC to Portfolio and vice-versa. Running the FT.com on a DOS PC allows transfers via the parallel port using the inbuilt features of the Portfolio. Making transfers through the serial port was a bit more complicated. I made a new Nullmodem cable (with CTS-RTS jumpered together, a must for the Portfolio) and tried AMCOM (same software on Portfolio and PC); a subset of the features worked. But best results were obtained by installing XTERM2 on the Portfolio, and using TERATERM on a Windows PC (albeit a WIN2K machine I use for this serial stuff) and transfering with XMODEM. Sure, getting XTERM2 into the Portfolio is a bit a chicken and egg problem...

2. PDP-11/34
Despite spending an awful lot of time, I am unable to bring this computer to work; the most I realized was that the RL02 lights go on (and show a FAULT) when trying to boot. But I did not get any signal from the serial cards (I tried several of these) which are used for the console. I also changed a lot of the grant continuity cards for the later G7273 model, checked parts of the Unibus, all this to no avail.
So sadly I gave up, and the machine will be exposed  in a non-working state close to our beautiful magenta cabinet. I installed a plexiglass window and some LED stripes so that the cards of the computer can be seen; the side-panels will receive some pictures of the wire-rap UNIBUS backplane and the interior of the RL02 drive. I added a lock to allow opening the RL02 drive only for demonstration. Assembling everything was heavy work, and I was helped by two of our students and my cleaner in lifting up the drives and the computer in its rails.
Many plastic pieces to hold the front-panel are partially broken; a team of our students under the guidance of Prof. Tom Scheer will print new ones in their 3D printer lab. So you see that finally quite a lot of people contributed to this exhibit, which will be put in place at the end of the month.

and tried to achieve a communication via the DL11 serial board: no success for the moment.

Jean Wagner donated a Burroughs B22 system (actually built by Convergent Technologies in 1981) consisting of a client station and a master computer (a tower), The client is a nice assembly of monitor and i8086 computer on a common base. The monitor was crippled with white fungus, which had grown in the explosion protection plastic sheet (ca. 2mm thick) covering the glass CRT. The glue between glass and plastic had partially liquefied, and removal and cleaning was quite a job. The computer boots with an SOS-beep, possibly because I do not yet have the keyboard (which will be delivered asap). The monitor remains off, possibly because of the non-booting state. I am unable to find any technical details on the 20 pin square connector from the video card to the monitor, which seems to be a CGA or variant digital type (not a BAS) monitor). Any help will be appreciated!

The client station has no floppy drive neither HD, so relies on the server which has a 8" Shugart 8" floppy drive and an enormous HD. Again, no documentation, no software! Interestingly the material seems to come from Piedboeuf SA, which is a famous Belgian brewery in Liège!

The 35 kg H650 power supply was separated from the main cabinet (see upper picture), and all power modules (H744 = 5V, H745 = -15V and 5411086 = +15V)   cleaned and checked under load. It was very difficult to separate the Mate-N-Lok connectors who were very tight. All power modules are ok, with no big ripple on the regulated voltages. Reconnection will be done next month, this is a 4 hands job as the wires from the backplane to the connectors are very short, and the H650 so heavy!

Two terminals came with our PDP11/34, donated by Guy Schintgen. The Decscope VT52 is in bad shape (this is the big thing at the right showing its bare underbelly!), the monitor screen does not display any cursor or characters, and on one of the big motherboards the IC voltage is 4.75V, a bit too low. I am searching for the cause.

The Decwriter III (LA-120) which is on the left of the picture had many unresponsive keys, among them the most important SETUP key. But luckily, the culprit were bad contacts. As the key mechanism is a wonder of simplicity and easy to access, that repair was gratifying! The serial terminal needed a good deal of cleaning, and I built a platform with rollers to make moving it more easy. On the picture the Decwriter is wired by a null-modem cable to a laptop running Teraterm (at 9600 bit/s), and everything works very smoothly!

PS, Feb.2022: : After 3 days of good work, the Decwriter has become seriously ill: on power up, the print head does not move (to its resting position left), and it shows a constant "Paper Out" fault (the paper switch is ok). Not sure were the problem is (encoder? bad motor driver?).... Such is life for a restorator, do never become too confidant!

DECwriter III: restoration done.

Alas, the IBM Selectric is beyond repair, at least for me. Several steel traction ribbons have become loose, and I am unable to put them again in proper position.

The CT-1024 (also called TV-Typewriter) was the original alphanumerical terminal sold as a kit with the SWTPc6800. It has a RF modulator to be connected to the antenna port of an ordinary B/W TV, which so became the display of a cheap computer terminal. Jean Mootz (+) and I built each one of these in 1977, and this is Jean's. He made a modification with the external connectors and added a RS232 DB25 socket. The "normal" baud-rate was 1200, switchable to 110 (= the baud rate of the famous ASR33 TTY). Here 1200 works unreliably, so the default is 110 (N81), with an internal switch for 1200. Notice the many "flying" wires on the different circuit boards. The CT-1024 kit was introduced in 1975. Look also here.

I tried to bring back to life one of our old SWTPc 6800 computers that my colleague Jean Mootz (+) and I assembled in 1976/77 (look here for the catalog). The tin contacts of the SS50 motherboard were badly corroded, but a MP-A2 CPU board with the SWTBUG Rom worked "out of the box". I fixed an MP-S ACIA serial board with some new solder, and used a 300 bd speed for serial communication, first with a laptop, than later with a real serial terminal. A GIMIX 32kb memory board looked like new, with it's silk shining intact after more than 42 years. The two 5.25" floppy drives are the original Wangco SS, SD 35 track drives. One had a bad bearing, but some silicon oil corrected the problem later on. I found only one last TSC Flex 1.0 system disk (Flex 1.0 manual here), all other were Flex9 for the SWTPc 6809. As I had only a one working drive at the beginning, relying on one single diskette was very strenuous. I tried Imagedisk, Anadisk, CopyIIPC, Teledisk etc. for imaging, without great success: impossible to make a bootable copy of the image of the SD 128bytes/sector system disk, that has a track 0 different from the remaining ones. My best result was a copy that was partially readable, but could not boot; visually comparing the dumps of the original and the copy did not show up any difference, so this remains a problem; I am grateful for every serious suggestion!

After a treatment with silicon oil the second disk became usable, and I could make backup's of the Flex 1.0 system disk. The original SWTPc CT-64 terminal (assembled by J. Mootz) works, but loses characters even at the slow 300 bd.. The "newer" CT-82 has a bad video monitor. So I finally attached a much newer (1997) Televideo 990 terminal donated by Carlo Mullesch, which gives a splendid readable display. BASIC, TSC Editor and Assembler are on the system disk (Flex 1.0 was also called "miniFlex" or even "FDOS" at the beginning). I am still flabbergasted how much software the old developers could put on a single diskette!

So the last serious problem remaining is the making of a correct bootable image of the system disks. I know that Vetusware has this on top of its priority list!

PS:I exchanged many emails with Mike Douglas, who is an absolute expert in vintage computer restoration (look at his breathtaking website deramp.com). He wrote two assembly programs to make Flex 1.0 images using a SWTPc 6800, which does not need to run an OS (only the monitor SWTbug is needed). His programs allow to convert a floppy disk to an image (to be sent via the serial line to a PC for instance) or to make a new floppy on the SWTPc by sending the image file from the serial terminal to the SWTPc. He also wrote a new much faster disk backup software executable on the SWTPc.
I also found a version of FDOS 1.0, an OS yet prior to Flex 1.0. This is easy to image with Imagedisk, and I uploaded a copy to Vetusware.
What remains is a kludge for making FLEX 1.0 images on a PC. Any suggestions will be appreciated.

We received a Motorola Exorset30 microcomputer (1981) from Nic. Malget in 2010. The computer was in use at Circuit Foils, a company in the northern part of Luxembourg that produces the copper foils which you find on the electric circuit and motherboards. The computer has two 5.25" BASF 6106 drives, which are single-sided only, and a FDC (floppy disk controller) which is pre-standard (w.r. to the IBM PC). One of the drives was defunct, but the famous EXORbug monitor prompt showed up. It took me a long time to replace the 2 drives by old Tandon 100-2A models (and to find drives that still worked!), the same that were used by IBM (and these here have IBM embossed on the front!). I had to make small changes on the motherboard of the drives, as the controller waits for a READY signal which is not delivered by the Tandon and other IBM compatible floppy drives. I also had to rebuild my copy station with a "faster" 486 board, and a SD capable controller to make .IMD files and recreate 5.25" disks. We now have 3 versions of the XDOS OS: XDOS 3.01, 4.0 and 4.13. The latter seems to have practically vanished from the universe. I found a xdos413.dsk file on the Net, which was corrupt, and had to reassemble the full set of 4 original disks (they are now available on Vetusware).
This "repair" was very time consuming, I spent more than 50 hours on it... but the system now works like new!

I wanted to use this beautiful machine we received in 2005 for making 720kB floppies from TD0 images. The first hours in operation were fine, but than the problems were creeping up, with a hard disk becoming more and more irritable. See the album for more details.
 

The "restoration" of this beautiful HP computer (sold for scientists and engineers) owns much to other people. Let me mention first the Vintage HP computers systems group in groups.io, for giving much needed information very fast, including images of HP software on diskettes, and bitsavers.org for the manuals.
It was at this group that I found that Dominique BERGET (dvp800@gmail.com), a former French engineer specializing among others in refurbishing/repairing  HP/Agilent instruments, had developed a BASIC 5.1 + 1MB Ram combo board. As our model was one of the first without the Basic in Rom, I contacted Dominique and was able to buy his board... It works like a charm, and allowed also to check that the 5.25" floppy drive is in good working order. Dominique is extremely helpful concerning HP (mainly the HP 9000 series) stuff, and if you are a collector, feel free to contact him... he might have that rare item you are desperately seeking!

Two key stubs were broken, and in first try I managed to glue them together with cyanide glue. Time will tell if a more solid fix is needed.

I restarted working on this beautiful machine that we acquired many years ago from an "Ingenieur-Büro" in Essen, Germany. The two 8" drives work perfectly, but we had no documentation at all, and it is not easy to find. M. Heiko Schumann , an Olivetti-only collector (see his beautiful web-site olivrea.de) gave me access to the most important manuals. The computer boots into Basic, and from there executes programs and system routines. I was finally able to find several system diskettes among those we got from Essen, and to make copies on the L1 M40.
Mr. Thomas Spangemacher (look for his edv-technik-ts shop at ebay.de where he has a lot of restored vintage computers to sell) also sent me a 8" disk with a newer version of the Olivetti MDOS system and the PDF's of many manuals.
Alas, our BASF DSDD 8" drive we use for disk imaging broke fatally down (probably a rupture inside the stepper motor), so for the moment we have no 8" capable of making images of 8" floppies. If you read this and find one in your attic, please write me a word at francis.massen@education.lu.

Carlo Mullesch donated last year also three original front-panels from DEC PDP11 8/I, /55 and /70 minicomputers. They are now on display in the Computarium .

Gilbert Maurer donated a nice HP85 machine with its large carrying case; it is a 16KB Ram computer (the 32 KB sticker is wrong), with Basic in Rom. a thermal printer and a data cartridge drive. Here a picture of the open machine (which works flawlessly):

The tape drive has originally a capstan covered by plastic, which drives the brown front wheel seen in the second-next picture. Here the plastic has crumbled away; I tried replaced the plastic by a shrink tube.

An insurmountable problem is the unusual mechanism inside the data cartridge (made by 3M for HP): the tape wheels are rotated by a plastic loop band, which itself is moved by the capstan pressing against the brown wheel seen at the front. This loop is broken and split on all cartridges we have; here in the picture I replaced it by a yellow rubber band to show the working:

Here a blow-up view showing the yellow loop band which rotates the tape reels by friction against the data tape. This clever mechanism with a constant length loop is compatible with the increasing/decreasing diameters when the data tape is moving from one reel to the other.

Our TSM ("Talents, Skills and More") workshop was held today; 4 very young students of our 1st year participated. The aim was to assemble a working XT-clone (with a 8088 uP), launch DOS 3.30  from 5.25" floppies and learn some rudiments of BASIC (MSBASIC).
The preparation took quite some time, as many of our items (as power supplies, display cards, floppy controllers and big floppy drives) which worked well 9 years ago have become defunct or unreliable. I also found that the Diskcopy version of IBM DOS 3.30 did not copy a bootable system disk correctly (the copy was non bootable), so I switched to the old faithful COPYIIPC (version 6.0), which, albeit slow, does the work without complaining.

The Computarium's miniscule Soviet vintage uP's collection was greatly enhanced by a lot of similar material from Carlo Mullesch. I introduced a new 6th section in our virtual museum for the microprocessor and support chips category. The first section finished is that of the Soviet and Warsaw Block origin. I assembled all these chips in a "beauty-case", so that they may easily be stored away and shown in one of our 3rd floor cabinets. The yellow labels are the database indices, and the white page seen behind the glass cover is more explicite.

From time to time you see extraordinary things when opening a computer. Here is the motherboard of a vintage Macintosh Classic (1990) whose lithium battery has erupted like a volcano, and the ejected goo has even eaten away the legs of several IC's and caps. Look here for a photo of the full motherboard, which is small and nicely laid-out in this last of the original Macintosh series. This item will be exposed as-is in a cabinet.

Claude Baumann restored back to life an older Colinbus CNC profiler, an heritage of our late member Jean Mootz.

The full procedure, research, investigations, electronics development and other work are described in a 62-page report titled " Bringing back to life a Colinbus Profiler CNC-Router" (46 figures); the report is located in the Literature chapter of the Computarium web-site,
The profiler will work in Claude's workshop.

Carlo Mullesch donated 2 vintage Heathkit ET-3400A microprocessor trainers of the early 1980's, and also an ETA-3400 expansion box. One of the trainers was ready to accept the expansion box, but the latter was very ill. The small  transformer became burning hot, and the serial communication with a terminal (here an old laptop) remained silent. Both of these items were kits, assembled by a guy in the USA, and so run on 120VAC. The ETA was not fully populated, but the culprit was a blown vanadium capacitor close to a 7912 -12VDC regulator. We replaced both, and after some fiddling with RS232 cables, break-out box etc. finally the monitor prompt MON> showed up. The box has a ROM with a version of TINY BASIC (15 functions and/or commands only, no FOR...NEXT loop, see manual here), which can now be started and sends a greeting:

We will mount this combo on a platform having its own 240/120 VAC transformer.

Here a selfie of my 2 helpers and myself and the empty cabinet (this last picture was taken a few years ago).

The Insite floptical drive (I 325VM) is repaired; the problem was purely mechanical: the electromechanically eject system was wrongly positioned, so that the inserted floppy was not lowered enough down into its correct position, and the computer assumed that no floppy was inserted ("floppy unloaded" message).

The SGI Indy is now fully functional (but I do not have an original 21MB MO floptical disk, so the drive can only be tested with normal 3.5" HD floppies).

This is the last machine of the contributions by Eric Dondelinger I worked on. Its a very nice SGI Indy workstation with a MIPS R4600 uP, 64 MB Ram, a SCSI HD + floptical, and a digital camera. The machine booted into its outstanding PROM, but the installed IRIX 5.3 was very sick. I looked and found an IBM SCSI HD with the 50pin connector, and installed IRIX 6.2 (found on the truly amazing repository of  FSCK Technology). The single IRIX 5.3 CDROM joined was not enough to make a full installation, one needs at leasr 4 CD's. The machine now runs like a charm; there is one remaining problem with the capricious floptical drive (a drive which can read/write standard 3.5 floppies as well as a 21MB magneto-optical disc; I do not have any of these. If you do, please contact me); it has problems with loading and ejecting the disk. This INDY is one of our nicest workstations in the Computarium collection.

This professional workstation from 1994 was donated by Eric Dondelinger (see below); it was a powerful computer based on the DEC Alpha microprocessor, and is also called Alphastation M300. It did not work, hang during the boot process and changed keyboard keys in a strange manner. The main culprit was the infamous DS1287 Dallas chip with an exhausted cell. I ordered the DS12887A as a replacement, and it worked! Boot was now ok, but the main HD (an 426MB RZ25-E SCSI disk) was terminal ill. I installed Digital Unix 3.2C (which is the first follower of the OSF/1 series) from an external SCSI CDROM drive on the second RZ25-E. The system is now working at perfection, after some strange downs of the power-supply that made my heart stop, but that I was unable to reproduce.

Many years ago (2010 !) our contributor Eric Dondelinger donated among several DEC Alphastations also 2 X-Terminals from Tektronix (TEKxp217C). These were used to connect to a big host-server and use its desktop in a X-Windows environment. Both machines were in good condition, and I tried to use a small Raspberry Pi as a host. It was a bit tricky, but now everything runs ok: the X-terminal (based on a Risc processor) fetches its code through bootp from the server (= the RPi), starts an X-session which allows either a XDMCP graphical connection or a console-type Telnet. There remains an obscure font-related problem that I am unable solve until now: the RPi desktop quality is not good, but a Scratch application for instance works in A+ quality. If you have any suggestion, please drop a word at francis.massen@education.lu.
This package allows now a simple demonstration of what  was once the standard method to work from a large server.

We have enlarged our storage rooms, and are installing many racks to tidy up the huge storage section in the attic of the LCD. Today (27-Oct) our member Claude Boes and myself started assembling the first 5 of 16 racks. This is heavy work, and we hope that the final result will reduce a bit the clutter of the storage.
The 31-Oct work continued with Waldemar Lichy (my brother in law), and we finished the installation in the first room.

The DECmate II (a 115VAC model from 1982) received from C. Mullesch without monitor and keyboard is now starting to work. I was able to find a LK201 keyboard on eBAY and built a small adapter-box splitting the DB15 cable from the computer (which has both video and keyboard signals) into an Cinch socket (for the monochrome video) and RJ9 socket for the keyboard cable.

The 5.25" introduction disk (on upper picture) and the system test disk (lower picture) are bootable and work fine; the Seagate ST-412 hard-drive seems to stuck . None of the WPS disks joined is readable. Both RX50 disk drives are ok, contrary to the test complaining.

Restoration is mostly done: I was able to bring back the stuck HD to spinning (don't ask me how!) without ruining the disk, and without loosing any software installed. This is the main MASTER MENU (first screen), a very powerful entry into all the installed four OS's and maintenance/utility programs. Communication via the serial port works good with KERMIT under DOS.

The preformatted RX50 disks can be made on an AT using 22disk from Sybex (and putting a 360 KB 5.25" disk in a MF drive). DEC forced its clients to buy preformatted disks, and only later included a full formatting software in its Rainbow microcomputer series when angry clients became louder!

The DECmate II has been mounted with all the documentation, monitor, keyboard (in a drawer), transformer etc. on a sturdy rolling table, to ease the demonstration of the working machine. Everything now works at perfection!

see also lower down!

21 Oct-2020: done!

06-Nov-2020: definitively done!

MONROE 1920 scientific calculator

Carlo Mullesch donated (among many other items) a DEC PDP-8/L mini-computer, 115 VAC power, 1968. The backplane has its connectors wire-wrapped, not soldered to a circuit board. As both upper and lower covers were missing, somebody in the past has badly handled the wire-rap pins (about 2600 in total), so that many are bent and make short circuits or unwanted contacts. I started today painfully straightening these up...

The last picture shows the many boards inserted in the backplane connectors; the bigger boards located near the center are the two core memory boards (4 KB memory, 12 bit addressing). This mini-computer was sold from 1968 to 1971.

I probably will be unable to restore it to working condition, but will try to get help from more knowledgeable people. There are only 5 museums worldwide exhibiting this mini-computer (see for instance here)

__________________________________

MONROE 1920 scientific calculator

This rare scientific desktop calculator from 1975 (115 VAC) received from C. Mullesch was working, except that it refused to make addition and multiplication. Both keys had bad contacts. Look at the nice Rockwell chips inside!

_____________

work done!

This is the definitive form of our Exorciser-Copy-Station. A single AT286 is used to interact with the Exorciser (serial, 9600 N 8 1) using TELIX; the copying is done with the same machine using Dave Dunfield's Imagedisk (or other software). Writing an image to a 8" floppy works good, but sometimes floppies that are read into a file show errors when booting on the Exorciser. I guess the time of head-scratching is not yet quite over... if it will anytime!
The BASF 6104 drive used in writing/reading images is enclosed in plexiglass, so that it's working can easily by observed. The picture below shows the read/write head, the stepper motor driving it with screwed axis and the big 220 VAC spindle motor on the upper right.

The TFT Miracle monitor is a rare MDA/CGA/EGA device (still sold for an horrendous price) ; it would be nicer to use a vintage CRT-terminal, but these beasts take too much of the available space.

finished Oct-2020

Mr. Georges Eicher donated a rare Sanyo MBC-555 computer from 1982. This is a 8088 8 bit machine which was sold below 1000 US$, a world-first sensation. It is not quite "IBM compatible", as a part of the BIOS is on the floppy disk (there are 2 drives, no HD). We received the machine without any software, and booting with a standard MSDOS or IBM DOS disk is futile. So I looked around to find a master image of the boot disk, and found one in .TD0 format. Using the excellent LIBDSK software suite, much tinkering and head scratching, I finally was able to make a real boot disk (also using CopyIIPC). The Sanyo works like a charm, and is in excellent condition. BTW the 2 floppy drives also are not quite standard!

This would not have been possible without the outstanding suite IMAGEDISK of Dave Dunfield; his software by-passes the OS (here DOS 6.20) and sends commands directly to the floppy disk controller (FDC), so that the whole bunch of parameters can be individually set. A BIG problem is finding an FDC capable of SD (single density). None of the "newer" integrated controller chips can do this. I have a couple (not many) in stock and finally choose after many tests a combi ISA FDC&HDC controller board from Adaptec (ST506 ATA two cable type, not an SCSI board!), whose FDC chip is known and was tested being capable of SD. As a final test I tried to make an image from an MDOS 2.20 Exorciser disk (SSSD) and write it back on another disk...and ho...Exorciser boots fine from this!

One problem is that I do not have many SD 8" floppy disks in the collection, so if you have still some of these dinosaurs lying around, please send me a word (francis.massen@education.lu).

This is one of the big joyful moments in retro-computing life, which makes one forget the long hours spent, and also the many adventures which ended unsuccessfully!

In 2005 we received from the LTAM (Lycée Technique des Arts et Métiers) a complete Motorola Exorciser system (6800 based) with MPU, Exorciser II (dual 8" floppy system), EXORterm150 (terminal) and a printer. I never had the time to correctly look into this jewel, and started now. The MPU boots into EXBUG 1.2, executes the MAID command etc... The EXORterm150 has problems, possibly with a half-dead keyboard encoder. Many keys are not working, but data sent is shown on the screen. I replaced it for the moment by a vintage laptop running TELIX, and communication with E-7-1 1200 runs fine.
I did not find any software, until this morning: the drive :1 (the second) had a diskette that was, o miracle!, booting into MDOS 2.2. I am unable to make a backup or generate a MDOS diskette on drive :1 (for reasons not yet found), so life is dangerous with only one single 8" boot disk at hand. I am in touch with Mr. Thomas Spangemacher (www.edv-technik-ts.de) from whom we bought several vintage machines; he tries to make a 8 disk from an .IMD image file.

I really must built a 8" copying station asap; I have most material to do, just lacking time, as always...

By an uncredible luck I found one 8" diskette in drive :1 (the second drive), and it was bootable after probably more than 35 years containing MDOS 2.2. After some tinkering, I was able to make backups and generate new MDOS disks (DOSGEN command).

MDOS is not MSDOS, but a different beast (Motorola DOS).

Thank you "Jacob Schneider" (= name on the label), the anonymous benefactor who forgot this diskette and saved the world!

We received quite a lot of vintage Compaq workstations, and all need some tinkering. Most gruesome was the AP500,  a sturdy and very heavy workstation from 1999. It's BIOS battery was flat, and horribile est dictu, the guys at Compaq had soldered the coin-cell straight onto the motherboard (metallic contacts were cold-welded to the cell and than soldered to the mainboard). I cut the lower fixation and soldered two wires going to an AA battery-case to the pads. How could such an expensive machine have a such terrible design-fault? (or was this planned obsolescence?). I installed W2k+SP4 on the SCSI disk, and the beast runs very smoothly.

I built a second stand-alone (external) player trying to cram everything in a much smaller case (14cm*8cm) as the first one. This model uses a Raspberry Pi Zero (see upper border), a small open-frame 220VAC/5VDC power supply (right side) and a USB/serial converter (TTL levels) with the Prolific chip. The prototyping board holds both this chip and the 1-transistor level adapter. The white Y-connector is a kludge because I could not find a small enough non-DIN power socket.

I auctioned at eBAY a rather rare MAGIS Club Minitel, probably built in 2002 (agreement date is 2000). This was one of the last models, with a very stylish case and a hinged keyboard. The vendor alas made a very poor inside packaging, so that the plastic case and many items inside the Minitel were broken (more than 35 different pieces! see upper picture.); the keyboard with the front plate was ripped off. By extraordinary luck, the electronics remained in working condition. It took some hours to repair the case and many fixations inside, as for instance the magnetic compensation ring on the collar of the CRT tube. The "repaired" beautiful Magis Club works fine and remains quite sexy.

Instead adding an outside box, I tried to cram everything inside this older Minitel1 case (which has more inside room than the Telic  models). I used a Raspberry Pi Zero, an open-frame 220VAC/5VDC power supply and the usual stuff. Switching the Minitel on starts the slideshow in a loop.
Look at the album.

Replacing the Ram chips did not help, as the replacement of the few other chips on sockets. I was lucky to stumble on an excellent article from Marc Brumlik, a former repair technician for TANDY: "How to diagnose and repair a TRS-80 Model 1" (link). In this article the exact misbehavior of the computer was described (pressing the key B shows J), and  the cause is a faulty video Ram (2102 chip, static, 1kbit). These video chips are soldered in, so it took some work to remove the chip, solder in a socket and put in a 2102 chip from a defunct 8K memory board of one of our SWTPC computers. Click for very short video of repaired computer.


 

Look here for a first video of the running player...

2. Minitel as a terminal for the Raspberry Pi

2. I finally found the time to make this happen, following an excellent text by jcquentin (see doc). Works like a charm! Pictures show the interface for the USB serial (TTL) connection and the login screen of Raspbian Buster.

Done.

The Computarium was always very "resource-conscious": our main lighting were design lamps from the 1960's originally used in one of the LCD cantines and removed in 2009 for being sent to the scrap yard. Luckily I was able to salvage them for our Computarium rooms, where they are in action since 2009. The fluorescent tubes (FT) lamps were very inefficient, and the coils became burning hot. So we decided to replace the  FT tubes by LED tubes, and our lycée takes over the cost. Replacement is simple in theory, but not in practice: many of the plastic tube sockets are broken by age and heat, and can not be replaced as these models do not exist anymore. So when needed the light tubes are fitted with screw joints and the broken sockets removed.

I auctioned a vintage Tim Unitas II calculator (the model with sliders, serial 05753) to complement our collection of Unitas II machines (we have the older wooden case model and the newer keyboard model). The machine was sold as defective, and it was in a bad state. Somebody has dropped it, so the upper carriage was bent and did not slide through, the right indicator on the upper carriage blocked and the crank is broken away. Some prudent work with a hammer and a Dremel grating disk restored the machine to full function. We were not able to fix a new crank, as the remaining hardened steel stub is (for the moment) impossible to remove. The machine can be operated with a screw-driver; not aesthetic but it works. It is a great machine to show the stepped drum and the engaging wheel, as the front cover is very easy to remove .

Mr. Felix Hansen donated a rack containing a A350 Modicon PLC (=Programmable Logic Controller = industrial computer) and two separate A120 micro models, with DOS based software (Modsoft_AKF) on a Compaq Presario and a lot of documentation. I transferred the software to a vintage Maxdata laptop, changed the rack to contain the laptop and one A120 Modicon Micro (with the 612CPU), seen above the laptop. The A120 runs fine, the AKF software is similar to the Simatic S5 one (both follow the IEC 61131 standard).
The upper panels will be modified in the future to contain several output devices and power connections.
Modicon was the US company which created in 1968 the world-first PLC controller (in German a "SPS" = "Speicher Programmierbare Steuerung"). Our systems are from 1993/95 and use an Intel TN80C188 microprocessor (see here).

Finished mounting the Braille Printing Station, which allows to demonstrate Braille printing on a Index Braille Everest Printer (model V2) from the early 2000's (this model was manufactured from 1995 to 2003). The sound-proof cabinet has been mounted on wheels and risen up, so that a vintage Maxdata laptop running Win2k and Word2000 can be comfortably used to write and print a text.

This is a printout of a well known two-liner much in use for checking telecommunications (hint: it speaks of a fox and a dog!). Look if you can decode the characters limited by the blue lines (one character is represented by a dot structure in a 3x2 matrix (3 lines, two columns).

I finished building the PiDP8i kit that I ordered in 2016 during the Vintage Computer Festival in Berlin from Oscar Vermeulen and didn't find time to assemble. The kit represents the original console of the DEC PDP 8/I in miniature (300mm x 150mm) with all its switches and lamps. Inside works a Raspberry Pi ZeroW with Raspbian Jessie Lite and the SIMH emulator  modified by Oscar Vermeulen and Warren Young (see here and here) emulating OS/8. The kit will be driven by the original ASR-33 TTY, which creates some problems as this old machine is CAPITALS-ONLY and 100 baud. Everything works fine with an old RS232 laptop and Teraterm, the last test with the real TTY will be done asap.

I salvaged a Superdrive floppy from a dead MAC IIsx to replace a non-functional floppy drive in a MAC LC. The drive was dirty beyond imagination. Tedious cleaning, removing tussle, dust and hair and a tiny drop of oil cured the beast, which is now again capable to Read/Write/Format 400/800/1440 kB diskettes (using System 7.1)

Repaired a Mercedes-Euklid M30 electro-mechanical calculator auctioned at eBAY, which constantly ran into a blocked state. The main cause was dirt gathered around the keys which did not always engage properly, so that a misaligned toothed wheel blocked the proportional levers. I had to dismount all key-columns, which is relatively easy.The M30 has no automatic division.

As usual the RS232 transfer was a bit tricky to set up, as the usual settings of the COM port through the inbuilt utility was ignored by the HCOM software, which insisted on a 38400 bit/s default speed. The serial cable is extremely simple: 3 wires are all what's needed!
On the PC side, I ran the Husky software on a reliable old MAXDATA Win2k laptop, which still has very good DOS compatibility.

A vintage (probably ~1937) printing calculator from the famous Wanderer Werke (Chemnitz-Schönau in Saxony, Germany) was donated by C. Baumann. The looks of the machine are very good, but several number rods were stuck and did not move. Repair was easy, the most difficult part being two very tenacious screws resisting to loosen.

In 2010 we received an old IBM 29 card puncher; a quick test at that moment showed that the machine was not working. After seeing a similar system running at the Computer History Museum in Mountainview (California) I decided to try a "repair". After two days, the first result is that the feeding/transporting mechanism works again (see very short video). There is a serious problem with the keyboard, so I am not sure that a complete restoration will be feasible.

I found another portable electrotherapy device in the legacy of Jean Mootz; he had repaired it, but then cut off the power-wires so that the device was not working. I "repaired" it by adding a new battery holder and replacing Jean's rather ugly screws by more stylish ones. After some cleaning the device works beautifully, and the electrical shocks it delivers can be quite invigorating! The instrument will be used as a "vintage-atmosphere" enhancement for the Computarium.
See album !

I transformed an existing rolling presentation stand to show the complete Simatic S5 system, together with the PG675 computer, the handheld programmer and the S5-100U  PLC. A board with several switches and buttons is available for some quick demonstrations. The raw wooden boards were beautified with black felt and black foil.

I repaired a vintage Mercedes Euklid M21 electro-mechanical calculator, auctioned at eBAY. The electrical cable was completely rotten inside the machine, and had to be replaced by removing the shut-off board. After a few trials, the machine blocked, probably due to a missing screw on the motor block. Removing the carriage and some fiddling cured this. Inside the machine looks like new, without any traces of rust. The main surprise was a live little spider crawling out of the machine where it had built a tiny comfortable nest.

I finished today the creation of 8 RetroGameMachines, based on HP T610 thin clients running Win7e (Windows 7 starter embedded) with 4GB Ram and a 16 GB SSD Sata disk.. All these second hand computers were bought at eBAY. Eight emulators allow to play a selection of vintage games, using  joypad and/or keyboard and mouse. The computers are very robust (no moving parts) and tamper proof, frozen by a write filter. Cloning is easy with a USB stick. Besides the gaming user  there is a user "LOGO" to learn Logo (I installed the excellent FMS logo). The machines will first be used at the next LCD Science Déeg (10 May 2017).

We are working on an vintage Siemens SPS S5 system donated by the LTPEM. The PG675 computer (8088, 5.25" floppy SSDD drives) runs its original software under CP/M86-1.2.
We build several new disks from .TD0 images. Knowing how to use the AG S5 SPS system and the handheld PG675 programmer will take a bit more time...

Communication between the PG675 and the AG (S5_100U) is now ok (03-Feb-2017).

Look here for a picture of the assembled calculator and the video which shows the strange multiplication.

We auctioned a rare Everest Multarapid-S calculator (from Officine Serio, Italy, 1958). This is an electromechanical three function calculator using a phone dial to enter the multiplicator and sort of a handle bar for the usual sub-total, total and negative actions. It was in a bad state, very dirty, with neither multiplication nor zero-keys working. Some metal work with a Dremel was needed.

There remains an unsolved problem with the multiplicant, which must be entered with trailing 000 to give a proper print-out. A special bar probably defining the grouping of decimals is missing; this would have been activated by the right level marked with dots (see picture of assembled calculator). Also missing is the key-cap to loosen the paper-friction.

I auctioned a Monroe CSA10 fully automatical 4 function calculator on eBAY; the vendor specified that the machine was broken (motor spins idle). After quite some hours of work and a bit of luck the machine is back to 100% working condition. This is a very nice, compact calculator, in mint condition (no rust, no scratches); it came with an original German Monroe instruction manual.

I repaired an old Madas 12e, which was donated by Guy Bollendorf in 2006 in an absolute blocked state. A first tentative repair 10 years ago did not succeed, but experience helping, I was able to bring back to working conditions this nice machine today. Several elements were severely blocked by (invisible) hardened grease or oil, and a weakly spring had to be replaced by a twin pair to give the necessary pull to bring back the main black action key.

Album.

This is a rare original IBM 5150, the first "real" IBM computer after the IBM Junior. Has a DIN connector for an audio cassette player used as storage. Runs now PCDOS 1.1.

A non working HP25 has been disassembled and the different elements put into a frame, so that the 5 IC's can clearly be seen.

I finally build two sets of Genaille-Lucas rulers in a form that is similar to the Napier-Bones. I found wooden prismatic rods in a market, and adapted the template using Photoshop.  Kudos to our member Claude Baumann who found an error in one of the division templates (ruler 7, look at 6) from the GERMEA group of the University of Pau. The beautiful boxes were bought from Amazon.

See two videos on YouTube: the first shows the robot following a very large and simple track, In the second video the track detection has been optimized and the robot must follow a somewhat nasty track.

Video 1, video 2, album

See other other pictures here and here (the machine has two graphics cards: a CGA and a MDA/parallel combo).

Only one channel (of 9) was working ok. Many wires broken at the points of solder. All valves ok (surprise!). Some balance potentiometers a bit "scratchy". Linearity of ramp integrators is good (check done with a vintage Heathkit IR-5207 XY plotter).

January 2016: Capacitors replaced, a first set of modules made.

Printing calculator from the1930's. Was in a very bad state, non working

More see here.

Screen shows graphical desktop of Rasbian-Wheezy.

2. Monroe CSA-8
 

3. Win286

2. I auctioned (very cheaply) a Monroe CSA-8 fully automatic electromechanically calculator from the 1950's. It is in a bad shape, and after having spent many hours (and removing the carriage), I am still unable to get the multiplication and the division working properly.

3. I am putting a very old Windows286 on our BULL Micral45 which is on display. It works (still a mouse problem; solved!), but the EGA monitor is showing its age. This becomes a major problem for the Computarium: we have only very few CGA and EGA monitors, and my efforts to make a Jamma board working are without success until now. If you have an old EGA or CGA monitor, please do not throw it away, the Computarium would be glad to take it into its collection.

An Edmund Scientific analog computer bought at eBay USA came in a non working state This primitive device has only 3 potentiometers, a push button switch and a galvanometer. The cardboard case badly needed a wooden frame to give some stability. One of the 50 Ohm wire potentiometers had broken contacts, and one of the contacts of galvanometer broke away. The plastic hands on the dials also were broken off.
Read here an article in Popular Electronics (1961) how this "computer" works.

15Dec2014:
The analog computer is now fully functional. Click on the pictures left to see the patching and display of a solution of the differential equation ax"+bx'+cx=0 (exponential damped oscillation). The timing signal of the GP-6 triggers the oscilloscope.

See video.

Watch video!

...and fitted with a European power cord. Most blocks worked fine, but one 741 opamp was dead and has been replaced. This computer has 3 modules for addition/substraction and 2 modules for integration. The pictures show the switch board and the electronic boards.

See album of Hamann Manus R

Vintage bubble jet printers from 1991 (donated by the LTHAH) will be used, as these dot-matrix printers are reasonable fast. Making them work was quite a job, as the ink channels and print heads were thouroughly blocked.
A short video showing an open printer is here.

The son of our late member Jean-Claude Asselborn donated some BBC material to the Computarium; among these items was a Valiant mechanical, remote controlled Turtle. It is probably one of the first models from 1982. The 10 individual AA NiCad batteries were dead, there was no software and no manual. We had some electrical problems with very bad contacts, but finally brought the turtle back to life with new NiMH batteries, some tinkering, an old diskimage of the Valiant Mover software found on the internet and a working BBC. mod.B.
 

This is one of the first "tablet-like" computers from 1993. The whole computer (main-board, HD, 3.5" floppy) is contained in the touch-sensitive screen, which reacts to a wireless pen. System was "MS Windows for Pen Computers 1.0". On this machine DOS 6.2 is installed, and the stylus is missing. One hinge is broken off; LCD screen is very dim (screen shows a Mandelbrot picture being drawn by a GWBASIC program... very slowly!)

Restauration of a vintage Siemens & Halske telegraph, which was used at a railway station. Thorough cleaning, some slight changes in wiring to operate device with one single battery, searching and finding a paper tape as replacement of the original.

Three of us were unable to assemble the electrical motor correctly but Claude found the trick. The reassembled motor ran for a very short time, but than blocked as the bearings seem very used.  A second problem is the flaky insulation, which makes using this engine during a demonstration an impossible risky adventure.
So I looked for replacement AC motor (~1360 rpm, 60 W, single phase) and found one at an eBAY shop. There was some work needed to make a new support, but today (27Oct12) the task is mostly finished.

This booking machine from 1930 (serial number 4708) was in a bad shape, extremely dirty, and the electrical motor removed and disassembled. The beautiful photographs at eBAY surely were made a long time ago, when the machine was still in better condition.

The auction price will be paid by the association of former LCD students (AALCD). Many thanks!

Machine after heavy cleaning, oiling an minor repair (original motor replaced by a temporary one):, upper part switched back:

Look at the beautiful keyboard:

See short album and YouTube video

See also here at the website of EarlyComputers

I continued the 14/15th June with chart recorders, audio generator, frequency scale/counter and digital multimeter.
See YouTube videos here and here.

See You Tube video here (from another web-site)

2. Finally brought to working state a FACIT ESA-0 fully automatic electromechanical calculator (from the fifties, see here)

and much more....

Work on the first TTY will resume asap.

The big A0 plotter donated by C. Bechoux did its demo flawlessly, but we were unable to drive it from a computer

Also repaired a non working MONROE LA7160 from another bid.