# 159 – Analog Computers

Rate Episode |

This episode is about analog computers, which are computers that compute using a physical model of a real system, often using analog electronic devices. Our guest is Bernd Ulmann who runs the Analog Computer Museum near Wiesbaden, Germany. In the episode we talk about what analog computers are (mechanical, electronic, digital), how they are programmed, what they are used for and why they should (and will?) be used in modern computing as well. We close the episode with a short discussion of the VAX and the AN/FSQ-7, both computer systems near and dear to Bernd.

- http://vaxman.de
- Bernd Ulmann
- The Analog Computer Museum
- http://analogmuseum.org (WP)
- AN/FSQ-7: the computer that shaped the Cold War (Book)
- Analog Computing
- Dynamical system (WP)
- Field-programmable gate array (WP)
- Digital differential analyzer (WP)
- Analog computer (WP)
- Operational amplifier (WP)
- General Purpose Computer (WP)
- Telefunken RA 770 (WP)
- The Mathworks Simulink (WP)
- Data flow diagram (WP)
- Potentiometer (WP)
- Voltage divider (WP)
- Damping (WP)
- Harmonic oscillator (WP)
- Differential Equation of Second Degree (WP)
- Kelvin water dropper (WP)
- Pullman Company (WP)
- George A. Philbrick
- Helmut Hölzer (WP)
- V-weapons (WP)
- Kuhn’s Paradigm shift (WP)
- Green 500 List (WP)
- Runge–Kutta methods (WP)
- Euler Method (WP)
- Integration (WP)
- Cellular differentiation (WP)
- Differential (mechanical device) (WP)
- Historic Videos (7 parts) that explains mechanical analog computing elements (thanks @Dr4k3_LE)
- Disc integration (WP)
- ISER – Informatik Sammlung Erlangen
- Stochastic differential equation (WP)
- Quantum computer (WP)
- Oscilloscope (WP)
- Differential Equations
- DESY Accelerator (WP)
- Quadrupole magnet (WP)
- Octopole Magnets (WP)
- Magnetic bearing (WP)
- Ultracentrifuge (WP)
- SAGE/AN/FSQ-7 (WP)
- NORAD (WP)
- Lockheed SR-71 Blackbird (WP)
- North American X-15 (WP)
- Function generator (WP)
- Nondimensionalization (WP)
- Telefunken RA 1 (WP)
- Telefunken RA 463/2
- Vacuum tube (WP)
- Field-effect transistor (WP)
- EarlyComputers: The Donner Analog Computer Model 3500.
- EAI 580
- EAI 680
- EAI Pacer 700
- Diode–transistor logic (DTL) (WP)
- German Aerospace Center (DLR)
- Ariane (rocket family) (WP)
- Link Trainer (WP)
- Black Magic and Gremlins (Book)
- Prof. Dr. Meier-Brötz
- Telefunken Transistorized Table-Top Computer
- Punched card (WP)
- Prof. Dr. Flenn Cowan
- The End of Digital Tyranny: Why the Future of Computing Is Analog (Article)
- General-Purpose Code Acceleration with Limited-Precision Analog Computing (Article)
- Analog Computation: Everything Old Is New Again (Article)
- Darpa Has Seen the Future of Computing … And It’s Analog (Article)
- Analog Computation and Dynamical Systems
- Hybrid Computers (WP)
- Dornier Do 31 (WP)
- Dornier DO-960 (WP)
- PDP-11 (WP)
- Hexapod (WP)
- Telefunken RA 742 (WP)
- DEX 100
- EAI 1000
- CMOS (WP)
- EAI 180 (WP)
- Digital Equipment Corporation (WP)
- Computer History Museum
- Whirlwind I (WP)
- Light gun (WP)
- JOVIAL (WP)
- Ones\u0026#39; complement (WP)
- Boeing E-6 Mercury (WP)
- Ken Olsen (WP)
- PDP-11 (WP)
- VAX (WP)
- Dr. Margarete Hartmann

Pingback: Die letzten und nächsten 24h, Sonntag, 16.11.2014 | die Hörsuppe

I found this episode rather nostalgic:

– My introduction to computers was at a summer session for school students around 1975, where staff at a nuclear research reactor introduced us to both analog and digital computers.

– Like Bernd, I visited the Computer Museum when it was in Boston. The air-defence computer was designed to be a high-availability system. As I recall it, with around 20,000 vacuum tubes, each with an MTBF around 20,000 hours, this was a real challenge. They solved this problem by having two of these computers in adjacent rooms. They would plan a switchover every 15 minutes, with the time in-between spent running diagnostics on the offline machine while stressing it by running on high and low voltages. They also had an Intel 8080 microprocessor as part of the display, commenting that the microprocessor speed was comparable, but the reliability was far higher.

You’re right, we didn’t talk much about reliability at all. I guess we should have.

Fascinating! And I was stunned to learn that the Analogmuseum is located more or less in my backyard – I will definetely contact Mr Ulmann to visit his place.

Oh, and Marcus, let me know when you’re in Wiesbaden next time and have some spare time, I’d like to buy you a drink. Or two. ;-)

Will do :-)

Most enjoyable episode.

Having started on CDC, PDP 8, 11 then the Eclipse myself, I could almost visualize the ongoings in the house.

Well done Markus and many thanks Bernd – very generous of you to share all of that with us.

I had completely forgot about analog computers. We used them in my sophomore level systems lab. I searched and found the one we had, http://ferretronix.com/march/analog_computers/eai_analog_computer_2.jpg

Thanks for the memories.

Great episode. I actually knew analog computers were more about physical models of systems and less about analog components, but only because I had an electrical engineer room mate. I really should have looked into them more.

I’ll be visiting a friend in Berlin early next year and may fly in through Frankfurt, so I’d love to get in touch with Mr. Ulmann when I have definite plans to see if he can give us a tour.

Have fun there :-)

Fascinating episode – one of the best! Thanks for producing it. Also, many thanks to your most knowledgable guest, Mr Ulmann.

Awesome suff. Watched some of Mr. Ullmann’s clips on youtube. Brilliant. Didn’t understand anything but it’s just so cool.

Fantastic episode. I wonder how much of a role maths education (or the lack thereof) played in the downfall of analog computers — seems like computer engineers rarely study much in the way of differential equations, we mostly focus on discrete math.

Also, I was someone disappointed to learn that the only computations available are physical representations of differentials and integrals, and their sums. Seems like the advantages analog computers have over their von neumann brothers here are somewhat small. Are there analog techniques that can take advantage of the parallel nature of physics to solve problems that are computationally unfeasible on traditional computers? For example, is there an analog way to solve the travelling salesman problem in effectively constant time?

To some extent, quantum computers are what you are looking for: they use physical phenomena and work differently. Parallelism and “exploring all states at once” is one of their properties, which is envisioned to be exploitable in the context of security (cracking cyphers). Early days of course.

Pingback: 1p – Analog Computers | Profit Goals

Great episode!

I don’t understand very much of it (due to missing mathematical background), but it’s great anyway! ;)

Great. Fascinating to see (or better, hear) how someone lives his passion.