DIGI-COMP I - Show Case Project

Parent Category: Articles Category: DIGI-COMP I Written by Administrator

DIGI-COMP I is designed to be a hand-held: hold it in one hand, and operate the clock slider with the other one.

In manual operation it is difficult to show on conventions like these Vintage Computer Festivals ... there you wish to run it unattended in "Kiosk-Mode".

Get Ready for the Show

To operate automagically, I added a motor drive to push the clock slider for and back endlessly. Speed can be set by a "slow/fast/manual" selector.

And the whole setup should be protected by a show case, against dust (and inquisitive VCF-visitors).

digicomp1 vitrine total small

 

Internals

I made heavy use of 3D-printed parts. The motor is a regular NEMA 14 stepper motor, controlled by an Arduino UNO. The stepper motor is driven by an A3988 module, found in many 3D printers.

A3988 supports "micro stepping" and current control. Both is important: micro stepping makes the motor less noisy, and current control prevents the strong stepper from breaking DIGI-COMP in case of any mechanical jam.

The Arduino software also manages two end switches and the speed dial. All in all its not more than a "Hello world" application.

digicomp1 driver total

 

digicomp1 driver gear

The whole assembly is powered by a 12V Notebook power supply.  I did not mounted a fixed power jack, instead let it hang flexible out of the bottom plate center. This is because you want to look at the working DIGI-COMP from all sides, a fixed cable would be annoying.

 

Development took weeks until I understood DIGI-COMPs mechanical properties, resulting forces, friction, geometric problems and heat management.

And all the stuff had to fit into a slim show case socket.

 

digicomp1 misprints

Endless trial-and-error iterations were needed to make it run: I'm not exactly a natural born maker genius!

Kinematics are hard!

However, with these problems I'm well in line with more prominent mechanical devices: Leibniz' calculator from 1700 did not work fully due to kinematic problems even after 4 prototypes, it was finally running in 1990. Also Zuse's Z1 was unreliable, the 1937 original as well as the 1989 reconstruction.

Apparently its much easier to imagine a mechnical calculator on paper than to actually build one.

 

Attachments:
Download this file (BOM.ZIP)BOM.ZIP[Bill of Material with 3D files]14702 kB
Download this file (firmware.zip)firmware.zip[Arduino UNO firmware, cabling info documented in sources]3 kB